Avery

AVERY’S DISEASES OF THE NEWBORN Ninth Edition Christine A. Gleason, MD W. Alan Hodson Endowed Chair in Pediatrics Professor of Pediatrics Head, Division of Neonatology Department of Pediatrics University of Washington Seattle Children’s Hospital Seattle, Washington Sherin U. Devaskar, MD Mattel Endowed Executive Chair and Distinguished Professor Department of Pediatrics David Geffen School of Medicine Assistant Vice Chancellor for Children’s Health University of California, Los Angeles Health System Physician in Chief Mattel Children’s Hospital Los Angeles, California 1600 John F. Kennedy Blvd. Ste 1800 Philadelphia, PA 19103-2899 AVERY’S DISEASES OF THE NEWBORN  ISBN: 978-1-4377-0134-0 Copyright © 2012, 2005, 1998, 1991, 1984, 1977, 1971, 1965, 1960 by Saunders, an imprint of Elsevier Inc. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies, and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions. This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein). Notice Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary. Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility. With respect to any drug or pharmaceutical products identified, readers are advised to check the most current information provided (i) on procedures featured or (ii) by the manufacturer of each product to be administered, to verify the recommended dose or formula, the method and duration of administration, and contraindications. It is the responsibility of practitioners, relying on their own experience and knowledge of their patients, to make diagnoses, to determine dosages and the best treatment for each individual patient, and to take all appropriate safety precautions. To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. Library of Congress Cataloging-in-Publication Data Avery’s diseases of the newborn. -- 9th ed. / [edited by] Christine A. Gleason, Sherin U. Devaskar. p. ; cm. Diseases of the newborn Includes bibliographical references and index. ISBN 978-1-4377-0134-0 (pbk. : alk. paper) 1. Newborn infants--Diseases. I. Gleason, Christine A. II. Devaskar, Sherin U. III. Avery, Mary Ellen, 1927- IV. Title: Diseases of the newborn. [DNLM: 1. Infant, Newborn, Diseases. WS 421] RJ254.S3 2012 618.92’01--dc23 2011020667 Acquisitions Editor: Judith Fletcher Developmental Editor: Dee Simpson Publishing Services Manager: Anne Altepeter Associate Project Manager: Jessica L. Becher Design Direction: Steve Stave Printed in the United States of America Last digit is the print number: 9 8 7 6 5 4 3 2 1 CONTRIBUTORS Steven H. Abman, MD Professor Department of Pediatrics University of Colorado School of Medicine Director Pediatric Heart Lung Center Co-Director Pulmonary Hypertension Program The Children’s Hospital Aurora, Colorado Amina Ahmed, MD Pediatric Infectious Disease Department of Pediatrics Carolinas Medical Center Levine Children’s Hospital Charlotte, North Carolina Adjunct Clinical Associate Professor Department of Pediatrics University of North Carolina Chapel Hill, North Carolina Marilee C. Allen, MD Professor of Pediatrics Johns Hopkins University School of Medicine Co-Director Neonatal Intensive Care Unit Developmental Clinic Kennedy Krieger Institute Baltimore, Maryland David Askenazi, MD, MSPH Assistant Professor Division of Nephrology and Transplantation Department of Pediatrics University of Alabama at Birmingham Birmingham, Alabama Stephen A. Back, MD, PhD Associate Professor of Pediatrics and ­Neurology Oregon Health and Science University Clyde and Elda Munson Professor of ­Pediatric Research Director Neuroscience Section Papé Family Pediatric Research Institute Portland, Oregon H. Scott Baldwin, MD Professor of Pediatrics and Cell and ­Developmental Biology Vanderbilt University Medical Center Chief Division of Pediatric Cardiology Co-Director Pediatric Heart Institute Monroe Carell Jr. Children’s Hospital at Vanderbilt Nashville, Tennessee Roberta A. Ballard, MD Professor Department of Pediatrics and Neonatology University of California, San Francisco School of Medicine San Francisco, California Eduardo Bancalari, MD Professor of Pediatrics Director Division of Neonatology University of Miami Miller School of ­Medicine Chief Newborn Service Jackson Memorial Hospital Miami, Florida Carlton M. Bates, MD Associate Professor Department of Pediatrics University of Pittsburgh School of Medicine Chief and Program Director Pediatric Nephrology Children’s Hospital of Pittsburgh Rangos Research Building Pittsburgh, Pennsylvania Diana W. Bianchi, MD Natalie V. Zucker Professor of Pediatrics, Obstetrics and Gynecology Tufts University School of Medicine Vice Chair for Research Department of Pediatrics Floating Hospital for Children Boston, Massachusetts Gil Binenbaum, MD, MSCE Assistant Professor Department of Ophthalmology University of Pennsylvania School of ­Medicine Attending Surgeon Department of Ophthalmology The Children’s Hospital of Philadelphia Philadelphia, Pennsylvania Sureka Bollepalli, MD Assistant Professor Department of Pediatrics University of South Florida Diabetes Center Tampa, Florida Sonia L. Bonifacio, MD Associate Professor of Pediatrics Division of Pediatric Nephrology Emory University School of Medicine Director Pediatric Hypertension Program Children’s Healthcare of Atlanta Atlanta, Georgia Assistant Adjunct Professor Department of Pediatrics and Neonatology University of California, San Francisco School of Medicine Co-Director Neurological Intensive Care Nursery University of California, San Francisco ­Medical Center Benioff Children’s Hospital San Francisco, California Stephen Baumgart, MD Mitchell S. Cairo, MD Donald L. Batisky, MD Professor of Pediatrics Department of Neonatology George Washington University School of Medicine and Health Sciences Children’s National Medical Center Washington, DC Thomas J. Benedetti, MD, MHA Professor Department of Obstetrics and Gynecology University of Washington School of Medicine Seattle, Washington Gerard T. Berry, MD Professor Department of Pediatrics Harvard Medical School Director Metabolism Program Division of Genetics Children’s Hospital Boston Boston, Massachusetts Professor of Pediatrics and Medicine and Pathology Chief Division of Blood and Marrow ­Transplantation Department of Pediatrics New York-Presbyterian Morgan Stanley ­Children’s Hospital Columbia University Medical Center New York, New York Katherine H. Campbell, MD, MPH Fellow Department of Obstetrics, Gynecology, and Reproductive Sciences Yale School of Medicine New Haven, Connecticut v vi CONTRIBUTORS Michael Caplan, MD Chairman Department of Pediatrics NorthShore University HealthSystem Evanston, Illinois Professor Department of Pediatrics University of Chicago Pritzker School of Medicine Chicago, Illinois Stephen Cederbaum, MD Professor Emeritus Departments of Psychiatry and Pediatrics and Human Genetics University of California, Los Angeles Attending Physician Department of Pediatrics Ronald Reagan UCLA Medical Center Los Angeles, California Consulting Physician Department of Pediatrics Santa Monica UCLA Medical Center Santa Monica, California Sudhish Chandra, MD, FAAP Medical Director Department of Neonatology Neonatal Intensive Care Unit St. Anthony Medical Center Crown Point, Indiana Ming Chen, MD, PhD Assistant Professor Department of Pediatrics University of Michigan Ann Arbor, Michigan Nelson Claure, MSc, PhD Research Associate Professor of Pediatrics Director Neonatal Pulmonary Research Laboratory Department of Pediatrics Division of Neonatology University of Miami Miller School of ­Medicine Miami, Florida Ronald I. Clyman, MD F. Sessions Cole, MD Park J. White, MD, Professor of Pediatrics Assistant Vice Chancellor for Children’s Health Director Division of Newborn Medicine Washington University School of Medicine Chief Medical Officer St. Louis Children’s Hospital St. Louis, Missouri Lawrence Copelovitch, MD Assistant Professor University of Pennsylvania School of ­Medicine Attending in Nephrology Department of Pediatrics The Children’s Hospital of Philadelphia Philadelphia, Pennsylvania Michael Cunningham, MD, PhD Professor and Chief Division of Craniofacial Medicine Department of Pediatrics University of Washington Medical Director Craniofacial Center Seattle Children’s Hospital Seattle, Washington Alejandra G. de Alba Campomanes, MD, MPH Assistant Professor of Ophthalmology Division of Pediatric Ophthalmology and Strabismus University of California, San Francisco Director Department of Pediatric Ophthalmology and Strabismus San Francisco General Hospital San Francisco, California Ellen Dees, MD Assistant Professor of Pediatrics Division of Pediatric Cardiology Monroe Carell Jr. Children’s Hospital at Vanderbilt Nashville, Tennessee Professor Department of Pediatrics Senior Staff Cardiovascular Research Institute University of California, San Francisco San Francisco, California Scott C. Denne, MD Bernard A. Cohen, MD Mattel Endowed Executive Chair and Distinguished Professor Department of Pediatrics David Geffen School of Medicine Assistant Vice Chancellor for Children’s Health University of California, Los Angeles Health System Physician in Chief Mattel Children’s Hospital Los Angeles, California Professor of Pediatrics and Dermatology Johns Hopkins University School of Medicine Director Pediatric Dermatology Johns Hopkins Children’s Center Baltimore, Maryland Professor of Pediatrics Indiana University School of Medicine Riley Hospital for Children Indianapolis, Indiana Sherin U. Devaskar, MD Robert M. DiBlasi, RRT-NPS, FAARC Respiratory Research Coordinator Center for Developmental Therapeutics Seattle Children’s Research Institute Seattle, Washington Reed A. Dimmitt, MD, MSPH Associate Professor of Pediatrics and Surgery Director Division of Neonatology and Pediatric ­Gastroenterology and Nutrition University of Alabama at Birmingham Birmingham, Alabama Eric C. Eichenwald, MD Associate Professor Vice Chair and Division Director Neonatology Department of Pediatrics University of Texas Health Science Center Texas Children’s Hospital Houston, Texas Eli M. Eisenstein, MD Senior Pediatrician Department of Pediatrics Hadassah-Hebrew University Medical Center Mount Scopus, Jerusalem, Israel Jacquelyn R. Evans, MD Medical Director Newborn/Infant Intensive Care Unit The Children’s Hospital of Philadelphia Associate Division Chief Department of Neonatology University of Pennsylvania School of ­Medicine Philadelphia, Pennsylvania Kelly Evans, MD Fellow Division of Craniofacial Medicine Department of Pediatrics University of Washington Craniofacial Center Seattle Children’s Hospital Seattle, Washington Diana L. Farmer, MD, FAAP, FACS, FRCS Professor of Surgery, Pediatrics, and ­Obstetrics, Gynecology, and Reproductive Sciences Chief Division of Pediatric Surgery Vice Chair Department of Surgery University of California, San Francisco School of Medicine Surgeon-in-Chief University of California, San Francisco ­Medical Center Benioff Children’s Hospital San Francisco, California CONTRIBUTORS Patricia Ferrieri, MD Christine A. Gleason, MD Donna M. Ferriero, MS, MD Michael J. Goldberg, MD Professor Chairman’s Fund Endowed Chair in Lab Medicine and Pathology Division of Infectious Diseases Department of Pediatrics University of Minnesota Medical School Minneapolis, Minnesota Professor and Interim Chair of Pediatrics Professor of Neurology Co-Director Newborn Brain Research Institute University of California, San Francisco School of Medicine Physician-in-Chief University of California, San Francisco ­Medical Center Benioff Children’s Hospital San Francisco, California Neil N. Finer, MD Division of Neonatal-Perinatal Medicine Department of Pediatrics University of California, San Diego San Diego, California Maria Victoria Fraga, MD Fellow Division of Neonatology and Pediatrics The Children’s Hospital of Philadelphia Philadelphia, Pennsylvania Lydia Furman, MD Associate Professor of Pediatrics Case Western Reserve University School of Medicine Rainbow Babies and Children’s Hospital Cleveland, Ohio Susan Furth, MD, PhD Associate Professor Department of Pediatrics Johns Hopkins University School of Medicine Associate Professor Department of Epidemiology Johns Hopkins Bloomberg School of Public Health Baltimore, Maryland Estelle B. Gauda, MD Professor Department of Pediatrics Division of Neonatology Johns Hopkins University School of Medicine Baltimore, Maryland Bertil Glader, MD, PhD Professor of Pediatrics (Hematology/­ Oncology) and Pathology Stanford University School of Medicine Stanford, California W. Alan Hodson Endowed Chair in Pediatrics Professor of Pediatrics Head, Division of Neonatology Department of Pediatrics University of Washington Seattle Children’s Hospital Seattle, Washington Clinical Professor Department of Orthopedics and Sports Medicine University of Washington Director Skeletal Health Program Department of Orthopedics Seattle Children’s Hospital Seattle, Washington Fernando Gonzalez, MD Assistant Professor Department of Pediatrics Division of Neonatology University of California, San Francisco ­Medical Center Benioff Children’s Hospital San Francisco, California Sameer Gopalani, MD Clinical Assistant Professor Department of Obstetrics and Gynecology University of Washington School of Medicine Division of Perinatal Medicine Swedish Medical Center Seattle, Washington P. Ellen Grant, MD Associate Professor Department of Radiology Harvard Medical School Founding Director Center for Fetal-Neonatal Neuroimaging and Developmental Science Center Chair Department of Neonatology Children’s Hospital Boston Boston, Massachusetts Carol L. Greene, MD Professor Departments of Pediatrics and Obstetrics, Gynecology, and Reproductive Sciences Division of Genetics University of Maryland School of Medicine Baltimore, Maryland Susan Guttentag, MD Associate Professor Department of Pediatrics University of Pennsylvania School of Medicine The Children’s Hospital of Philadelphia Philadelphia, Pennsylvania Chad R. Haldeman-Englert, MD Assistant Professor Department of Pediatrics Section on Medical Genetics Wake Forest University School of Medicine Winston-Salem, North Carolina Thomas Hansen, MD Professor Department of Pediatrics University of Washington School of Medicine Chief Executive Officer Seattle Children’s Hospital Seattle, Washington Anne V. Hing, MD Associate Professor Division of Craniofacial Medicine Department of Pediatrics University of Washington Craniofacial Center Seattle Children’s Hospital Seattle, Washington A. Roger Hohimer, PhD Associate Professor Department of Obstetrics Division of Perinatology Oregon Health and Science University Portland, Oregon Margaret K. Hostetter, MD Professor and Chair Department of Pediatrics Yale School of Medicine New Haven, Connecticut Andrew D. Hull, MD, FRCOG, FACOG Professor of Clinical Reproductive Medicine University of California, San Diego Director Maternal Fetal Medicine Fellowship University of California, San Diego Medical Center La Jolla, California J. Craig Jackson, MD, MHA Surgeon University of California, San Francisco ­Medical Center Benioff Children’s Hospital San Francisco, California Professor Department of Pediatrics Division of Neonatology University of Washington Neonatal Intensive Care Unit Medical Director Seattle Children’s Hospital Seattle, Washington Jean-Pierre Guignard, MD Lucky Jain, MD, MBA Salvador Guevara-Gallardo, MD Honorary Professor of Pediatrics Lausanne University Medical School Lausanne, Switzerland vii Executive Vice Chairman Department of Pediatrics Emory University Medical Director Emory Children’s Center Atlanta, Georgia viii CONTRIBUTORS Vandana Jain, MD Nanda Kerkar, MD Halima Saadia Janjua, MD John P. Kinsella, MD Associate Professor Division of Pediatric Endocrinology Department of Pediatrics All India Institute of Medical Sciences New Delhi, India Pediatric Nephrology Fellow Department of Pediatrics Ohio State University Nationwide Children’s Hospital Columbus, Ohio Sandra E. Juul, MD, PhD Professor Department of Pediatrics University of Washington Seattle, Washington Satyan Kalkunte, MPharm, PhD Research Associate Superfund Basic Research Program Department of Pediatrics Women and Infants’ Hospital of Rhode Island Warren Alpert Medical School of Brown University Providence, Rhode Island Bernard S. Kaplan, MB, BCh Professor of Pediatrics University of Pennsylvania School of ­Medicine Attending in Nephrology Department of Pediatrics The Children’s Hospital of Philadelphia Philadelphia, Pennsylvania Roberta L. Keller, MD Assistant Professor of Clinical Pediatrics University of California, San Francisco Director Neonatal Extracorporeal Membrane ­Oxygenation Program University of California, San Francisco ­Medical Center Benioff Children’s Hospital San Francisco, California Thomas F. Kelly, MD Clinical Professor and Chief Division of Perinatal Medicine Department of Reproductive Medicine University of California, San Diego School of Medicine La Jolla, California Director Maternity Services University of California, San Diego Medical Center San Diego, California Steven E. Kern, PhD Associate Professor of Pharmaceutics, ­Anesthesiology, and Bioengineering University of Utah Salt Lake City, Utah Department of Pediatrics Division of Pediatric Hepatology Recanati-Miller Transplantation Institute The Mount Sinai Medical Center New York, New York Professor of Pediatrics Section of Neonatology University of Colorado School of Medicine Medical Director Newborn/Young Child Transport Service Co-Director Newborn Extracorporeal Membrane ­Oxygenation Service The Children’s Hospital Aurora, Colorado Roxanne Kirsch, MD, FRCPC, FAAP Cardiac Intensivist Departments of Anesthesia and Critical Care The Children’s Hospital of Philadelphia Philadelphia, Pennsylvania Monica E. Kleinman, MD Associate Professor of Anesthesia Department of Pediatrics Harvard Medical School Clinical Director Medical-Surgical Intensive Care Unit Department of Anesthesia Division of Critical Care Medicine Department of Anesthesia, Perioperative, and Pain Medicine Medical Director Critical Care Transport Program Children’s Hospital Boston Boston, Massachusetts Thomas S. Klitzner, MD, PhD Jack H. Skirball Professor and Chief Pediatric Cardiology David Geffen School of Medicine University of California, Los Angeles Los Angeles, California Sarah M. Lambert, MD Assistant Professor of Surgery in Urology University of Pennsylvania School of ­Medicine The Children’s Hospital of Philadelphia Philadelphia, Pennsylvania John D. Lantos, MD Professor of Pediatrics University of Missouri at Kansas City Director Children’s Mercy Bioethics Center Children’s Mercy Hospital Kansas City, Missouri Visiting Professor of Pediatrics University of Chicago Chicago, Illinois Tina A. Leone, MD Assistant Professor of Pediatrics University of California, San Diego San Diego, California Mary Leppert, MB, BCh Assistant Professor of Pediatrics Johns Hopkins University School of Medicine Attending Physician Neurodevelopmental Medicine Kennedy Krieger Institute Baltimore, Maryland Harvey L. Levy, MD Professor Department of Pediatrics Harvard Medical School Senior Physician in Medicine and Genetics Department of Medicine Children’s Hospital Boston Boston, Massachusetts Mark Lewin, MD Professor and Chief Pediatric Cardiology University of Washington School of Medicine Co-Director Heart Center Seattle Children’s Hospital Seattle, Washington Karen Lin-Su, MD Clinical Associate Professor of Pediatrics Pediatric Endocrinology Weill Cornell Medical College New York, New York Mignon L. Loh, MD Professor of Clinical Pediatrics Department of Pediatrics University of California, San Francisco Pediatric Hematological Oncologist University of California, San Francisco ­Medical Center Benioff Children’s Hospital San Francisco, California Scott A. Lorch, MD, MSCE Assistant Professor Department of Pediatrics University of Pennsylvania School of ­Medicine Attending Neonatologist Division of Neonatology and Center for Outcomes Research The Children’s Hospital of Philadelphia Philadelphia, Pennsylvania Ralph A. Lugo, PharmD Professor and Chair Department of Pharmacy Practice Bill Gatton College of Pharmacy East Tennessee State University Johnson City, Tennessee Volker Mai, PhD University of Florida Microbiology and Cell Sciences Emerging Pathogens Institute Gainesville, Florida CONTRIBUTORS Bradley S. Marino, MD, MPP, MSCE Associate Professor of Pediatrics University of Cincinnati College of Medicine Director Heart Institute Research Core Attending Physician Cardiac Intensive Care Unit Divisions of Cardiology and Critical Care Medicine Cincinnati Children’s Hospital Medical Center Cincinnati, Ohio Barry Markovitz, MD, MPH Dennis E. Mayock, MD Jeffrey C. Murray, MD William L. Meadow, MD, PhD Josef Neu, MD Ram K. Menon, MD Maria I. New, MD Professor Division of Neonatology Department of Pediatrics University of Washington School of Medicine Seattle, Washington Professor Department of Pediatrics University of Chicago Chicago, Illinois Professor of Clinical Anesthesiology and Pediatrics University of Southern California Keck School of Medicine Director Critical Care Medicine Children’s Hospital Los Angeles Los Angeles, California Professor of Pediatrics and Molecular and Integrative Physiology Director Division of Endocrinology Department of Pediatrics University of Michigan Medical School Ann Arbor, Michigan Kerri Marquard, MD Professor of Pediatric Neurology Department of Pediatrics Catholic University Rome, Italy Clinical Fellow Reproductive Endocrinology and Infertility Department of Obstetrics and Gynecology Division of Reproductive Endocrinology and Infertility Washington University School of Medicine St. Louis, Missouri Camilia R. Martin, MD, MS Assistant Professor of Pediatrics Harvard Medical School Associate Director Neonatal Intensive Care Unit Director Cross-Disciplinary Research Partnerships Division of Translational Research Beth Israel Deaconess Medical Center Boston, Massachusetts Richard J. Martin, MD Drusinsky-Fanaroff Chair in Neonatology Professor Rainbow Babies and Children’s Hospital Professor of Pediatrics Case Western Reserve University Cleveland, Ohio Katherine K. Matthay, MD Mildred V. Strouss Professor of Translational Research Chief of Pediatric Hematology-Oncology University of California, San Francisco ­Medical Center Benioff Children’s Hospital San Francisco, California Dana C. Matthews, MD Associate Professor University of Washington School of Medicine Director Clinical Hematology Pediatric Hematology and Oncology Seattle Children’s Hospital Seattle, Washington Eugenio Mercuri, MD, PhD Sowmya S. Mohan, MD Professor Departments of Pediatrics, Biology, Nursing, and Epidemiology University of Iowa Iowa City, Iowa Professor of Pediatrics Division of Neonatology University of Florida College of Medicine Gainesville, Florida Professor of Pediatrics Director Adrenal Steroid Disorders Program Mount Sinai School of Medicine New York, New York Annie Nguyen-Vermillion, MD, FAAP Department of Neonatology Northwest Permanente, PC Providence St. Vincent Medical Center Neonatal Intensive Care Unit Portland, Oregon Victoria Niklas, MD Neonatal-Perinatal Medicine Fellow Department of Pediatrics Division of Neonatology Emory University Atlanta, Georgia Associate Professor of Pediatrics Division of Neonatal Medicine University of Southern California Keck School of Medicine Children’s Hospital Los Angeles Los Angeles, California Kelle Moley, MD Saroj Nimkarn, MD James P. Crane Professor of Obstetrics and Gynecology Division of Reproductive Endocrinology and Infertility Vice Chair Basic Science Research Washington University School of Medicine St. Louis, Missouri Thomas J. Mollen, MD Associate Medical Director Infant Breathing Disorder Center The Children’s Hospital of Philadelphia Philadelphia, Pennsylvania Jeremy P. Moore, MD Assistant Professor of Pediatrics Division of Pediatric Cardiology Mattel Children’s Hospital David Geffen School of Medicine University of California, Los Angeles Los Angeles, California ix Assistant Professor of Pediatrics Associate Director Pediatric Endocrinology Weill Cornell Medical College New York, New York James F. Padbury, MD Oh-Zopfi Professor of Pediatrics and ­Perinatal Biology Vice Chair for Research Department of Pediatrics Warren Alpert Medical School of Brown University Women and Infants’ Hospital of Rhode Island Providence, Rhode Island Marika Pane, MD, PhD Institute of Neurology Catholic University Rome, Italy Nigel Paneth, MD, MPH Professor and Chairman Department of Reproductive Medicine University of California, San Diego San Diego, California University Distinguished Professor Departments of Epidemiology and Pediatrics and Human Development College of Human Medicine Michigan State University East Lansing, Michigan David A. Munson, MD Thomas A. Parker, MD Thomas R. Moore, MD Assistant Professor of Clinical Pediatrics University of Pennsylvania School of Medicine Associate Medical Director Newborn and Infant Intensive Care Unit The Children’s Hospital of Philadelphia Philadelphia, Pennsylvania Associate Professor of Pediatrics Director Training Program in Neonatal-Perinatal Medicine University of Colorado School of Medicine Aurora, Colorado x CONTRIBUTORS Janna C. Patterson, MD, MPH Assistant Instructor Department of Pediatrics Division of Neonatology University of Washington Seattle, Washington Christian M. Pettker, MD Assistant Professor Department of Obstetrics, Gynecology, and Reproductive Sciences Yale School of Medicine New Haven, Connecticut Lauren L. Plawner, MD Acting Assistant Professor of Neurology University of Washington Pediatric Neurologist Seattle Children’s Hospital Seattle, Washington Dan Poenaru, BSc, MD, MHPE Adjunct Professor Department of Surgery Queen’s University Kingston, Ontario, Canada Medical Director BethanyKids at Kijabe Hospital Kijabe, Kenya Brenda B. Poindexter, MD, MS Associate Professor of Pediatrics Section of Neonatal-Perinatal Medicine Indiana University School of Medicine Indianapolis, Indiana Michael A. Posencheg, MD Assistant Professor of Clinical Pediatrics University of Pennsylvania School of ­Medicine Associate Medical Director Intensive Care Nursery Medical Director Newborn Nursery Division of Neonatology and Newborn Services Hospital of the University of Pennsylvania Attending Neonatologist The Children’s Hospital of Philadelphia Philadelphia, Pennsylvania Sanjay P. Prabhu, MBBS, DCH, MRCPCH, FRCR Instructor Department of Radiology Harvard Medical School Director Advanced Image Analysis Lab Department of Radiology Children’s Hospital Boston Boston, Massachusetts Katherine B. Püttgen, MD Assistant Professor Departments of Dermatology and Pediatrics Johns Hopkins University School of Medicine Baltimore, Maryland Graham E. Quinn, MD, MSCE Professor of Ophthalmology Division of Pediatric Ophthalmology The Children’s Hospital of Philadelphia University of Pennsylvania School of ­Medicine Philadelphia, Pennsylvania Tonse N.K. Raju, MD, DCH Medical Officer Eunice Kennedy Shriver National Institute of Child Health and Human Development National Institutes of Health Bethesda, Maryland Gladys A. Ramos, MD Associate Physician Department of Reproductive Medicine Division of Perinatology University of California, San Diego San Diego, California Benjamin E. Reinking, MD Clinical Assistant Professor Department of Pediatrics University of Iowa Iowa City, Iowa C. Peter Richardson, PhD Associate Research Professor Department of Pediatrics University of Washington Associate Research Professor Department of Pulmonary and Newborn Care Seattle Children’s Hospital Principal Investigator Center for Developmental Therapy Seattle Children’s Research Institute Seattle, Washington David L. Rimoin, MD, PhD Professor of Pediatrics, Medicine, and ­Medical Genetics David Geffen School of Medicine University of California, Los Angeles Director Medical Genetics Institute Steven Spielberg Chair Cedars-Sinai Medical Center Los Angeles, California Elizabeth Robbins, MD Susan R. Rose, MEd, MD Professor of Pediatrics and Endocrinology University of Cincinnati Pediatric Endocrinologist Cincinnati Children’s Hospital Medical Center Cincinnati, Ohio Mark A. Rosen, MD Professor Departments of Anesthesia and Perioperative Care and Obstetrics, Gynecology, and Reproductive Sciences University of California, San Francisco Director Obstetric Anesthesia University of California, San Francisco San Francisco, California Lewis P. Rubin, MPhil, MD Pamela and Leslie Muma Endowed Chair in Neonatology Professor of Pediatrics, Obstetrics and ­Gynecology, Pathology and Cell Biology, and Community and Family Health University of South Florida Medical Director Newborn Service Line Tampa General Hospital Tampa, Florida Inderneel Sahai, MD, FACMG Assistant Professor Department of Pediatrics University of Massachusetts Chief Medical Officer New England Newborn Screening Program Division of Genetics and Metabolism Massachusetts General Hospital Department of Pediatrics Harvard Medical School Boston, Massachusetts Sulagna C. Saitta, MD, PhD Assistant Professor of Pediatrics Division of Genetics The Children’s Hospital of Philadelphia University of Pennsylvania School of ­Medicine Philadelphia, Pennsylvania Pablo J. Sánchez, MD Clinical Professor Department of Pediatrics University of California, San Francisco San Francisco, California Professor of Pediatrics University of Texas Southwestern Medical Center Children’s Medical Center Dallas Dallas, Texas Richard L. Robertson, MD Gary M. Satou, MD Associate Professor of Radiology Harvard Medical School Radiologist-in-Chief Children’s Hospital Boston Boston, Massachusetts Mark D. Rollins, MD, PhD Associate Professor Department of Anesthesia and Perioperative Care University of California, San Francisco San Francisco, California Associate Clinical Professor David Geffen School of Medicine University of California, Los Angeles Director Pediatric Echocardiography Co-Director Fetal Cardiology Program Mattel Children’s Hospital Los Angeles, California CONTRIBUTORS Richard J. Schanler, MD Professor Department of Pediatrics Hofstra North Shore-LIJ School of Medicine Hempstead, New York Associate Chairman Department of Pediatrics Chief Neonatal-Perinatal Medicine Steven and Alexandra Cohen Children’s Medical Center of New York New Hyde Park, New York Mark S. Scher, MD Professor of Pediatrics and Neurology Case Western Reserve University School of Medicine Chief of Pediatric Neurology Rainbow Babies and Children’s Hospital University Hospitals of Cleveland Cleveland, Ohio Mark R. Schleiss, MD Professor of Pediatrics Director Division of Pediatric Infectious Diseases and Immunology Associate Chair for Research Department of Pediatrics University of Minnesota Medical School American Legion Endowed Chair in Pediatric Infectious Diseases Co-Director Center for Infectious Diseases and ­Microbiology Translational Research Minneapolis, Minnesota Thomas D. Scholz, MD Children’s Miracle Network Professor of Pediatrics Director Division of Pediatric Cardiology University of Iowa Carver College of Medicine Iowa City, Iowa Andrew L. Schwaderer, MD Assistant Professor Department of Pediatrics Ohio State University Columbus, Ohio Istvan Seri, MD, PhD, HonD Professor and Chief Division of Neonatal Medicine Department of Pediatrics University of Southern California Keck School of Medicine Director Center for Fetal and Neonatal Medicine Children’s Hospital Los Angeles Los Angeles, California Surendra Sharma, MD, PhD Professor Department of Pediatrics Warren Alpert Medical School of Brown University Women and Infants’ Hospital of Rhode Island Providence, Rhode Island Evan B. Shereck, MD Endre Sulyok, MD, PhD, DSc Eric Sibley, MD, PhD Peter Tarczy-Hornoch, MD, FACMI Assistant Professor of Pediatrics Division of Pediatric Hematology and ­Oncology Oregon Health and Science University Doernbecher Children’s Hospital Portland, Oregon Associate Professor of Pediatrics Stanford University School of Medicine Stanford, California Caroline Signore, MD, MPH Medical Officer Pregnancy and Perinatology Branch Eunice Kennedy Shriver National Institute of Child Health and Human Development Bethesda, Maryland Rebecca Simmons, MD Associate Professor of Pediatrics Children’s Hospital Philadelphia University of Pennsylvania School of ­Medicine Philadelphia, Pennsylvania Jeffrey B. Smith, MD, PhD Professor Department of Pediatrics David Geffen School of Medicine University of California, Los Angeles Medical Director Newborn Nursery Mattel Children’s Hospital Los Angeles, California Lorie B. Smith, MD, MHS Staff Pediatric Nephrologist Walter Reed National Military Medical Center Bethesda, Maryland Clara Song, MD, FAAP Professor of Pediatrics Faculty Health Sciences University of Pecs Institute of Public Health and Health ­Promotion Pecs, Vorosmarty, Hungary Head and Professor Division of Biomedical and Health ­Informatics Department of Medical Education and ­Biomedical Informatics Professor Division of Neonatology Department of Pediatrics Adjunct Professor Computer Science and Engineering University of Washington Seattle, Washington George A. Taylor, MD John A. Kirkpatrick Professor of Radiology Department of Pediatrics Harvard Medical School Radiologist-in-Chief Emeritus Children’s Hospital Boston Boston, Massachusetts James A. Taylor, MD Professor Department of Pediatrics University of Washington Seattle, Washington Janet A. Thomas, MD Associate Professor Department of Pediatrics Section of Clinical Genetics and Metabolism University of Colorado School of Medicine The Children’s Hospital Aurora, Colorado Assistant Professor of Pediatrics Division of Neonatal-Perinatal Medicine The University of Oklahoma Health Sciences Center Children’s Hospital at Oklahoma University Medical Center Oklahoma City, Oklahoma George E. Tiller, MD, PhD Robin H. Steinhorn, MD Resident Physician Department of Dermatology Johns Hopkins Hospital Baltimore, Maryland Professor and Division Head Department of Pediatrics Children’s Memorial Hospital Northwestern University Feinberg School of Medicine Chicago, Illinois Frederick J. Suchy, MD Professor of Pediatrics Associate Dean for Child Health Research University of Colorado School of Medicine Chief Research Officer and Director The Children’s Hospital Research Institute The Children’s Hospital Aurora, Colorado xi Regional Chief Department of Genetics Southern California Permanente Medical Group Los Angeles, California Mark M. Tran, MD Michael Stone Trautman, MD Clinical Professor of Pediatrics Section of Neonatal-Perinatal Medicine Indiana University School of Medicine Riley Hospital for Children Indianapolis, Indiana Jeffrey S. Upperman, MD Associate Professor of Surgery Department of Pediatric Surgery Program Director Pediatric Surgery Fellowship Children’s Hospital Los Angeles Los Angeles, California xii CONTRIBUTORS Carmella van de Ven, MA Department of Pediatrics Columbia University Senior Research Staff Associate Pediatric Blood and Marrow Transplantation New York-Presbyterian Morgan Stanley ­Children’s Hospital Columbia University Medical Center New York, New York Margaret M. Vernon, MD Assistant Professor Department of Pediatrics Division of Cardiology University of Washington School of Medicine Children’s Heart Center Seattle Children’s Hospital Seattle, Washington Jon F. Watchko, MD Professor of Pediatrics, Obstetrics, ­Gynecology, and Reproductive Sciences Division of Newborn Medicine Department of Pediatrics University of Pittsburgh School of Medicine Senior Scientist Magee-Women’s Research Institute Pittsburgh, Pennsylvania Gil Wernovsky, MD Conrad Taff Professor of Pediatrics and Nutrition Harvard Medical School Mucosal Immunology Laboratory Boston, Massachusetts Professor of Pediatrics Department of Pediatric Cardiology University of Pennsylvania School of ­Medicine Medical Director Neurocardiac Care Program Associate Chief Department of Pediatric Cardiology Director Program Development and Staff Cardiac Intensivist The Cardiac Center The Children’s Hospital of Philadelphia Philadelphia, Pennsylvania Linda D. Wallen, MD Klane K. White, MD, MSc W. Allan Walker, MD Clinical Professor of Pediatrics Associate Division Head Neonatal Clinical Operations University of Washington Associate Medical Director Neonatal Intensive Care Unit Seattle Children’s Hospital Seattle, Washington Sarah A. Waller, MD Maternal Fetal Medicine Fellow Department of Obstetrics and Gynecology University of Washington Seattle, Washington Bradley A. Warady, MD Professor of Pediatrics University of Missouri at Kansas City School of Medicine Senior Associate Chairman Department of Pediatrics Chief Section of Pediatric Nephrology Director Dialysis and Transplantation Pediatric Nephrology Children’s Mercy Hospitals and Clinics Kansas City, Missouri Robert M. Ward, MD Professor Department of Pediatrics Attending Neonatologist Adjunct Professor Pharmacology and Toxicology Director Pediatric Pharmacology Program University of Utah Salt Lake City, Utah Assistant Professor Department of Orthopedics and Sports Medicine University of Washington Seattle, Washington Pediatric Orthopedic Surgeon Department of Orthopedics and Sports Medicine Seattle Children’s Hospital Seattle, Washington Calvin B. Williams, MD, PhD Professor of Pediatrics Chief Section of Pediatric Rheumatology Medical College of Wisconsin D.B. and Marjorie Reinhart Chair in ­Rheumatology Children’s Hospital of Wisconsin Milwaukee, Wisconsin David Woodrum, MD Professor of Pediatrics Division of Neonatology University of Washington School of Medicine Seattle, Washington George A. Woodward, MD, MBA Professor Department of Pediatrics University of Washington School of Medicine Chief Emergency Medicine Medical Director Transport Services Seattle Children’s Hospital Seattle, Washington Dakara Rucker Wright, MD Pediatric Dermatologist Johns Hopkins University School of Medicine Johns Hopkins Children’s Center Baltimore, Maryland Jeffrey A. Wright, MD Associate Professor Department of Pediatrics University of Washington Seattle, Washington Linda L. Wright, MD Deputy Director Center for Research for Mothers and Children Director Global Network for Women’s and Children’s Health Research Eunice Kennedy Shriver National Institute of Child Health and Human Development National Institutes of Health Bethesda, Maryland Christopher M. Young, MD Fellow Neonatal-Perinatal Medicine Department of Pediatrics Division of Neonatology University of Florida Gainesville, Florida Guy Young, MD Associate Professor of Pediatrics University of Southern California Keck School of Medicine Director Hemostasis and Thrombosis Center Center for Cancer and Blood Disorders Children’s Hospital Los Angeles Los Angeles, California Elaine H. Zackai, MD Division of Genetics Department of Pediatrics The Children’s Hospital of Philadelphia Philadelphia, Pennsylvania Stephen A. Zderic, MD Professor of Surgery in Urology University of Pennsylvania School of ­Medicine John W. Duckett Endowed Chair The Children’s Hospital of Philadelphia Philadelphia, Pennsylvania Preface “The neonatal period … represents the last frontier of medicine, territory which has just begun to be cleared of its forests and underbrush in preparation for its eagerly anticipated crops of saved lives.” Introduction from the first edition of Schaffer’s Diseases of the Newborn The first edition of Diseases of the Newborn was published in 1960 by Dr. Alexander J. Schaffer, a well-known Baltimore pediatrician who coined the term neonatology to describe this emerging pediatric subspecialty that concentrated on “the art and science of diagnosis and treatment of disorders of the newborn infant.” Schaffer’s first edition was used mainly for diagnosis, but also included reference to neonatal care practices (i.e., the use of antibiotics, temperature regulation, and attention to feeding techniques)—practices that had led to a remarkable decrease in the infant mortality rate in the United States, from 47 deaths per 1000 live births in 1940 to 26 per 1000 in 1960. But a pivotal year for the new field of neonatology came 3 years later in 1963, with the birth of President John F. Kennedy’s son, Patrick Bouvier Kennedy, at 36 weeks’ gestation (i.e., late preterm). His death at 3 days of age, from complications of hyaline membrane disease, accelerated the development of infant ventilators that, coupled with micro-blood gas analysis and expertise in the use of umbilical artery catheterization, led to the development of intensive care for newborns in the 1960s on both sides of the Atlantic. Advances in neonatal surgery and cardiology, along with further technological innovations, stimulated the development of neonatal intensive care units and regionalization of care for sick newborn infants over the next several decades. These developments were accompanied by an explosion of neonatal research activity that led to improved understanding of the pathophysiology and genetic basis of diseases of the newborn, which in turn has led to spectacular advances in neonatal diagnosis and therapeutics—particularly for preterm infants. These efforts led to continued improvements in the infant mortality rate in the United States, from 26 deaths per 1000 livebirths in 1960 to 6.5 per 1000 in 2004. Current research efforts are focused on decreasing the striking global disparities in infant mortality rates, decreasing neonatal morbidities, advancing neonatal therapeutics, and preventing prematurity and newborn diseases. We neonatologists would like to be put out of business one day! Dr. Mary Ellen Avery joined Dr. Schaffer for the third edition of Diseases of the Newborn in 1971. For the fourth edition in 1977, Drs. Avery and Schaffer recognized that their book now needed multiple contributors with subspecialty expertise and they became co-editors, rather than sole co-authors, of the book. In the preface to that fourth edition, Dr. Schaffer wrote, “We have also seen the application of some fundamental advances in molecular biology to the management of our fetal and newborn patients”—referring to the new knowledge of hemoglobinopathies. Dr. Schaffer died in 1981, at the age of 79, and Dr. H. William Taeusch joined Dr. Avery as co-editor for the fifth edition in 1984. Dr. Roberta Ballard joined Drs. Taeusch and Avery for the sixth edition in 1991, with the addition of Dr. Christine Gleason for the eighth edition in 2004. Drs. Avery, Taeusch, and Ballard retired from editing the book in 2009, and became “editors emeriti.” Dr. Gleason was joined by Dr. Sherin Devaskar as co-editor for this, the ninth edition. What’s new and different about this edition? The book has been completely (and often painfully) revised and updated by some of the best clinicians and investigators in their field. Some chapters required more extensive revision than others, particularly those that deal with areas in which we have benefitted from new knowledge and/or its application to new diagnostic and therapeutic practices. This is particularly true in areas such as the genetic basis of disease, neonatal pain management, information technology, and the fetal origins of adult disease—an area that is now embedded within many of the chapters of this book. Some of the book’s sections were reorganized to reflect our field’s continued evolution. For example, Chapter 52, Persistent Pulmonary Hypertension, previously in Part X, Respiratory System, has found a new home in Part XI, Cardiovascular System. Finally, we’ve added new chapters that reflect the continued growth and development of our subspecialty. These include Chapter 4, Global Neonatal Health; Chapter 29, Stabilization and Transport of the High-Risk Infant; Chapter 33, Care of the Late Preterm Infant; Chapter 74, Disorders of the Liver; and Chapter 95, Craniofacial Malformations. With the incredible breadth and depth of information immediately available to neonatal caregivers and educators on multiple internet sites, what’s the value of a textbook? We, the co-editors of this ninth edition, believe that textbooks such as Diseases of the Newborn and all forms of integrative scholarship, will always be needed—by clinicians striving to provide state-of-the-art neonatal care, by educators striving to train the next generation of caregivers, and by investigators striving to advance neonatal scholarship. A textbook’s content is only as good as its contributors and this textbook, like the previous editions, has awesome contributors. They were chosen for their expertise and ability to integrate their knowledge into a comprehensive, readable, and useful chapter. They did this despite the demands of their day jobs in the hopes that their syntheses could, as Ethel Dunham wrote in the foreword to the first edition, “spread more widely what is already known … and make it possible to apply these facts.” Textbooks of the future will undoubtedly take advantage of online, xiii xiv PREFACE interactive publishing technologies, making their content readily accessible and more real-time, with continued revision and updating. However, in 2011—a full 50 years after the publication of the first edition of this book—we continue to find copies of this and other textbooks important to our subspecialty lying dog-eared, coffee-stained, annotated, and broken-spined in all of the places where neonatal caregivers congregate. These places, these congregations of neonatal caregivers, are now present in every country around the world. The tentacles of neonatal practice and education are spreading—ever deeper, ever wider—to improve the outcome of pregnancy worldwide. Textbooks connect us to the past, bring us up to date with the present, and prepare and excite us for the future. We will always need them, in one form or another, at our sites of practice. To that end, we have challenged ourselves to meet, and hopefully exceed, that need—for our field, for our colleagues, and for the babies. We wish to thank key staff at Elsevier—Deidre Simpson, senior developmental editor, and Judith Fletcher, publishing director, both of whom demonstrated patience, guidance, and persistence, and Jessica Becher, associate project manager, for our book. We also wish to thank our academic institutions and our administrative assistants, Mildred Hill at the University of Washington and Kristie Smiley at the University of California, Los Angeles. They kept us grounded, on track, and basically saved our lives! We are indebted to our contributors, who actually wrote the book and did so willingly, enthusiastically, and (for the most part) in a timely fashion—despite the myriad of other responsibilities in their lives. And we are deeply grateful for the support of our families throughout the long, and often challenging, editorial process. Finally, we thank the editors emeriti of this book, Drs. Mary Ellen Avery, Bill Taeusch, and Roberta Ballard, for their enormous contributions to the field of neonatology and to the lives of babies throughout the world, and their wise influence on us, the editors of the ninth edition of this text. Christine A. Gleason Sherin U. Devaskar To our parents, Peter and Vera and Sitaram and Santha, who have inspired us To our husbands, Erik and Uday, who are the wind behind our sails To our children, Kristen, Lauren, and Erin, and Chirag and Chetan, who are our future Christine A. Gleason Sherin U. Devaskar Editors Emeritis Mary Ellen Avery, MD H. William Taeusch, MD Roberta A. Ballard, MD P A R T I Overview C H A P T E R 1 Neonatal and Perinatal Epidemiology Nigel Paneth EPIDEMIOLOGIC APPROACHES TO THE PERINATAL AND NEONATAL PERIODS The period surrounding the time of birth (the perinatal period) is a critical episode in human development, rivaling only the period surrounding conception in its significance. During this period, the infant makes the critical transition from its dependence on maternal and placental support—oxidative, nutritional, and endocrinologic—and establishes independent life. The difficulty of this transition is indicated by mortality risks that are higher than any occuring until old age (Kung et al, 2008) and by risks for damage to organ systems, most notably the brain, that can be lifelong. Providers of care in the perinatal period recognize that the developing human organism cannot always demonstrate the immediate effects of even profound insults. Years must pass before the effects on higher cortical functions of insults and injuries occurring during the perinatal period can be detected reliably. Epidemiologic approaches to the perinatal period must therefore be bidirectional—looking backward to examine the causes of adverse health conditions that arise or complicate the perinatal period, and looking forward to see how these conditions shape disorders of health found later in life. Traditionally the perinatal period was described as from 28 weeks’ gestation until 1 week of life, but the World Health Organization has more recently antedated the onset of the perinatal period to 22 weeks’ gestation (World Health Organization, 2004). For this discussion we will view the term perinatal more expansively, as including the second half of gestation—by which time most organogenesis has occurred, but growth and maturation of many systems have yet to occur—and the first month of life. The neonatal period, usually considered as the first month of life, is thus included in the term perinatal, reflecting the view that addressing the problems of the neonate requires an understanding of intrauterine phenomena. HEALTH DISORDERS OF PREGNANCY AND THE PERINATAL PERIOD KEY POPULATION MORTALITY RATES Maternal and child health in the population have traditionally been assessed by monitoring two key rates— maternal mortality and infant mortality. Maternal mortality is defined by the World Health Organization as the death of a woman from pregnancy-related causes during pregnancy or within 42 days of pregnancy, expressed as a ratio to 100,000 live births in the population being studied (World Health Organization, 2004). Because pregnancy can contribute to deaths beyond 42 days, some have argued for examining all deaths within 1 year of a pregnancy (Hoyert, 2007). When the cause of death is attributed to pregnancy-related causes, it is described as direct. When pregnancy has aggravated an underlying health disorder, the death is termed an indirect maternal death. Deaths unrelated to pregnancy that occur within 42 days of pregnancy are termed incidental maternal deaths or sometimes pregnancy-related deaths (Khlat and Guillaume, 2006). These distinctions are not always easy to make. Homicide and suicide, for example, are sometimes found to be more common in pregnancy, and thus might not be entirely incidental (Samandari et al, 2010; Shadigian and Bauer, 2005). Since 2003, the U.S. Standard Certificate of Death has included a special requirement for identifying whether the decedent, if female, was pregnant or had been pregnant in the previous 42 days, or from 43 days to 1 year before the death, thus enhancing the monitoring of all forms of maternal death (Centers for Disease Control and Prevention, 2003). This addition has added to the number of recorded maternal deaths. In most geographic entities, infant mortality (IM) is defined as all deaths occurring from birth to 365 days of age in a calendar year divided by all live births in the same year. This approach is imprecise, because some deaths in the examined year occurred to the previous year’s birth cohort, and some births in the examined year may die as infants in the following year. In recent years, birth-death linkage has permitted vital registration areas in the United States to provide IM rates that avoid this imprecision. The standard IM rate reported by the National Center for Health Statistics links deaths for the index year to all births, including those taking place the previous year. This form of IM is termed period infant mortality. An alternative procedure is to take births for the index year and link them to infant deaths, including those taking place the following year; this is referred to as birth cohort infant mortality, and it is not used for regular annual comparisons because it cannot be completed in as timely a fashion as period IM (Mathews and MacDorman, 2008). Infant deaths are often divided into deaths in the first 28 days of life (neonatal death) and deaths later in the first year (postneonatal death). Neonatal deaths, which are largely related to preterm birth and birth defects, tend to reflect the circumstances of pregnancy. Postneonatal 1 PART I Overview deaths, when frequent, commonly result from infection, often in the setting of poor nutrition. Thus in underdeveloped countries, postneonatal deaths predominate; in industrialized countries the reverse is true. In the United States, neonatal deaths have been more frequent than postneonatal deaths since 1921. In recent years, the ratio of neonatal to postneonatal deaths in the United States has consistently been approximately 2:1. Perinatal mortality is a term used for a rate that combines stillbirths and neonatal deaths in some fashion (World Health Organization, 2004). Stillbirth reporting before 28 weeks’ gestation is probably incomplete, even in the United States, where such stillbirths are required to be reported in every state. Nonetheless, stillbirths continue to be reported at a levels not much lower those of neonatal deaths, and our understanding of the causes of stillbirth remains very limited. TIME TRENDS IN MORTALITY RATES IN THE UNITED STATES Maternal mortality and IM declined steadily through the twentieth century. By 2000, neonatal mortality was 10% of its value in 1915, and postneonatal mortality less than 7%. Maternal mortality in this interval declined 74-fold; the rate in 2000 was less than 2% of the rate recorded in 1915. The contribution to these changes of a variety of complex social factors including improvements in income, housing, birth spacing, and nutrition have been documented widely, as has the role of ecologic-level public health interventions that have produced cleaner food and water (Division of Reproductive Health, 1999). Public health action at the individual level, including targeted maternal and infant nutrition programs and immunization programs, have made a lesser but still notable contribution. Medical care was, until recently, less critically involved, except for the decline in maternal mortality, which was highly sensitive to the developments in blood banking and antibiotics that began in the 1930s. To this day, hemorrhage and infection account for a large fraction of the world’s maternal deaths (Khan et al, 2006). A notable feature of the past 50 years is the sharp decline in all three mortality rates, beginning in the 1960s after a period of stagnation in the 1950s (Figure 1-1). The decline began with maternal mortality, followed by postneonatal and then by neonatal mortality. The contribution of medical care of the neonate was most clearly seen in national statistics in the 1970s, a decade that witnessed a larger decline in neonatal mortality than in any previous decade of the century (Division of Reproductive Health, 1999). All of the change in neonatal mortality between 1950 and 1975 was in mortality for a given birthweight; no improvement was seen in the birthweight distribution (Lee et al, 1980). The effect of newborn intensive care on mortality in extremely small babies has been striking. In 1960, 142 white singletons weighing less than 1000 g in the United States survived to age 1, less than 1% of births of that weight. In 2005, the survival rate for infants weighing 501 to 999 g was 70%, and the number of survivors at age 1 was almost 18,000 (Mathews and MacDorman, 2008). In retrospect, three factors seem to have played critical roles in the rapid development of newborn intensive MORTALITY RATES IN THE U.S. 1955–2005 50 45 40 35 30 Neonatal mortality rate per 1000 Postneonatal mortality rate per 1000 Maternal mortality rate per 100,000 25 20 15 10 5 0 1955 1957 1959 1961 1963 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2 FIGURE 1-1 Neonatal, postneonatal, and maternal mortality, 1955 to 2005. care programs. These programs that largely accounted for the rapid decline in birthweight-specific neonatal mortality that characterized national trends in the last third of the twentieth century. The first factor was the willingness of medicine to provide more than nursing care to marginal populations such as premature infants. It has often been noted that the death of the mildly premature son of President John F. Kennedy in 1963 provided a stimulus to the development of newborn intensive care, but it should be noted that the decline in IM that began in the 1970s was paralleled by a similar decline in mortality for the extremely old (Rosenwaike et al, 1980). A second factor was the availability of government funds, provided by the Medicaid program adopted in 1965, to pay for the care of premature newborns, among whom the poor are overrepresented. Whereas these two factors were necessary, they would have been insufficient to improve neonatal mortality had not new medical technologies, especially those supporting ventilation of the immature newborn lung, been developed at approximately the same time (Gregory et al, 1971). Advances in newborn care have ameliorated the effects of premature birth and birth defects on mortality. Unfortunately, the underlying disorders that drive perinatal mortality and the long-term developmental disorders that are sometimes their sequelae have shown little tendency to abate. With the important exception of neural tube defects, whose prevalence has declined with folate fortification of flour in the United States and programs to encourage intake of folate in women of child-bearing age (Mathews et al, 2002), the major causes of death— preterm birth and birth defects—have not declined, nor has the major neurodevelopmental disorder that can be of perinatal origin, cerebral palsy (Paneth et al, 2006). Progress has come from improved medical care of the high-risk pregnancy and the sick infant, rather than through understanding and preventing the disorders themselves. The period since 1990 has witnessed much less impressive mortality improvement. While infant, neonatal, and postneonatal mortality have declined since 1990, in the 1995-2005 decade these rates were at a near standstill 3 CHAPTER 1 Neonatal and Perinatal Epidemiology TABLE 1-1 U.S. Perinatal Mortality, Morbidity, Interventions, Health Conditions, and Behaviors, 1990 to 2005 1990 1995 2000 8.2 9.2 5.8 3.4 7.5 7.1 7.6 4.9 2.6 6.9 9.8 6.9 4.6 2.3 6.6 10.6% 11.0% 1.9% 2005 Net Change 1990-2005 (%) Death Rates and Ratios Maternal mortality ratio (per 100,000 live births) Infant mortality rate (per 1000 live births) Neonatal mortality rate (per 1000 live births) Postneonatal mortality rate (per 1000 live births) Fetal mortality ratio (per 1000 live births)* 15.1 6.9 4.5 2.3 6.2 +84% –25% –22% –32% –17% 11.6% 12.7% +20% 1.9% 1.9% 2.0% +5% 7.1 7.0 7.2 7.6 +7% 7.0% 1.3%. 6.3 7.3% 1.4% 6.6 7.6% 1.4% 7.1 8.2% 1.5% 7.3 +12% +15% +16% 22.6% 20.8% 22.9% 30.3% +34% 17.7% 3.2% 6.1% 7.9% 13.7% 1.5% 4.2% 9.3% 12.2% 0.9% 3.9% 11.6% 10.7% 0.7% 3.5% 11.4% –40% –78% –42% +44% 28.0% 32.1% 33.2% 36.9% +32% 23.3 26.1 31.1 33.8 +45% 70.9 64.6 65.3 66.7 –5.9% Morbidity Rates Preterm birth (<37 weeks’ gestation) per 100 live births Very preterm birth (<32 weeks’ gestation) per 100 live births Extremely preterm birth (<28 weeks’ gestation) per 1000 live births Low birthweight (<2500 g) per 100 live births Very low birthweight (<1500 g) per 100 live births Extremely low birthweight (<1000 g) per 1000 live births Cesarean Section Per 100 live births Health Behaviors (per 100 live births) Cigarette smoking† Alcohol intake‡ Late or no prenatal care Inadequate weight gain (<16 lb) at 40 weeks’ ­ gestation§ Unmarried Multiple Births Per 1000 live births Fertility Rate Per 1000 women aged 15-44 years *Fetal deaths with stated or presumed period of 20 weeks’ gestation or more. †1990-2000 smoking data based on 45 states and Washington, D.C.; 2005 data based on 36 states, New York City, and Washington, D.C., which used pre-2003 smoking definitions. ‡1990-2000 alcohol data based on 46 states and Washington, D.C.; 2005 data based on 36 states, New York City, and Washington, D.C., which used pre-2003 drinking definitions. §1990 weight gain data based on 48 states and Washington, D.C. Remaining data based on 49 states and Washington, D.C. (Table 1-1). Infant mortality actually rose in 2002, the first 1 year since 1958. Data from the Vermont Oxford Neonatal Network encompassing hundreds of neonatal units show stability in weight-specific mortality rates for all categories of infants weighing less than 1500 g, beginning in approximately 1995 (Horbar et al, 2002). It appears that the pace of advancement in newborn medicine and the expansion of newborn intensive care to previously underserved populations, factors that have exerted a constant downward pressure on infant mortality since the 1960s, have lessened greatly in the past decade. Maternal mortality has actually climbed substantially, but this is almost certainly the effect of improved reporting (Hoyert, 2007). Two key changes were the implementation of the International Classification of Diseases, Tenth Revision, in 1999, which was less restrictive in including indirect maternal deaths, and the introduction of the separate pregnancy question on the U.S. Standard Certificate of Death in 2003. The risk of preterm birth has been increasing in recent years. This increase is mostly noted in moderately preterm babies, and is likely to reflect increased willingness on the part of obstetricians to deliver fetuses earlier in gestation who are not doing well in utero, in addition to the increased prevalence of twins and triplets, who are generally born preterm, resulting from in vitro fertilization. The small rise in extremely small and preterm babies is likely to reflect trends toward increased reporting of marginally viable immature babies as live births rather than stillbirths (MacDorman et al, 2005). The cesarean section rate continues its long-term increase, from 5% in 1970 to 23% in 1990 to 32% in 2005. The reasons for this unprecedented trend are multifactorial and include pressures from patients, physicians, and the medical malpractice system. The steady reduction in smoking during pregnancy is likely to be real, whereas trends in the self-reporting of alcohol use in pregnancy may be influenced by societal norms and expectations. Fewer women seem to have late or no prenatal care in recent years, but more women have been found to have inadequate pregnancy weight gain at term. A slight increase in the fertility rate follows a long-term (since approximately 4 PART I Overview 1960) decline in fertility in the United States. One third of mothers in the United States are now unmarried when they give birth. International Comparisons The United States lags in IM compared with other developed nations; the United States ranked thirtieth in the world in IM in 2005 (MacDorman and Mathews, 2009). This surprising finding, in light of more favorable socioeconomic and medical care circumstances in the United States than in many nations with lower IM, cannot be attributed to inferior neonatal care. Mortality rates for low-birthweight infants are generally lower in the United States than in European and Asian nations, although mortality at term may be higher. The key difference, however, is that the United States suffers from a striking excess of premature births. Whereas the U.S. African American population is especially vulnerable to premature birth, and especially severe prematurity, premature birth rates are also considerably higher in white Americans than in most European populations. It is likely that the recording of marginally viable small infants as live births rather than stillbirths is more pronounced in the United States than in Europe (Kramer et al, 2002). Although this practice makes a contribution to our higher prematurity and IM rates, it cannot fully explain them. Premature birth, fetal growth retardation, and IM are tightly linked, in every setting in which they have been studied, to most measures of social class and especially to maternal education; however, uncovering precisely why lower social class drives these important biologic differences has been elusive. Factors such as smoking have at times been implicated, but can only explain a small fraction of the social class effect. It is unlikely that this situation will change until a better understanding of the complex social, environmental, and biologic roots of preterm birth is achieved. Health Disparities in the Perinatal Period In 2005, 55.1% of all U.S. births were to non-Hispanic white mothers, 23.8% were to Hispanic mothers, 14.1% were to African American mothers, and 7% were to mothers of other ethnic groups. Health disparities are especially prominent in the perinatal period, with African American IM stubbornly remaining about double that of white IM in the United States, even as rates decline in both populations (Table 1-2). Preterm birth is the central contributor to this racial disparity in IM, and the more severe the degree of prematurity, the higher the excess risk for African American infants. The risk of birth before 37 weeks’ gestation was 76% higher in African American mothers than among non-Hispanic whites in 2005, but the risk of birth before 32 weeks’ gestation was 310% higher. Reduction in IM disparities in the United States thus requires a better understanding of the causes and mechanisms of preterm birth. Birth defect mortality shows a less pronounced gradient by ethnic group, and it does not contribute in a major way to overall IM disparities (Yang et al, 2006). The Hispanic paradox is a term often used to describe the observation that IM is the same or lower in U.S. citizens classified as Hispanic than in non-Hispanic whites, despite the generally lower income and education levels of U.S. Hispanics (Hessol and Fuentes-Afflick, 2005). The IM experience of Hispanic mothers in the United States reflects the principle that premature birth and low birthweight are key determinants of IM, because these parameters are also favorable in Hispanics. Smoking is much less common among Hispanics in the United States, but this factor alone does not fully explain the paradox. MAJOR CAUSES OF DEATH Analyzing the cause of death, a staple of epidemiologic investigation, has limitations when applied to the perinatal period. Birth defect mortality is probably reasonably accurate, but causes of deaths among premature infants are divided among categories such as respiratory distress syndrome, immaturity, and a variety of complications of prematurity. Choosing which particular epiphenomenon of preterm birth to label as the primary cause of death is arbitrary to some extent. Some maternal complications, such as preeclampsia, are also occasionally listed as causes of newborn death. However categorized, prematurity accounts for at least one third of infant deaths (Callaghan et al, 2006). Before prenatal ultrasound examination could be used to estimate gestational age with reasonable accuracy, a high fraction of neonatal deaths were attributed to low birthweight, but most of these deaths occurred in premature infants, because premature birth is much more important as a cause of death than is fetal growth restriction. Extreme prematurity makes a contribution to IM well beyond its frequency in the population; 1.9% of births, that occurred before 32 weeks’ gestation, accounted for 54% of all infant deaths in 2005. Following premature birth, another important group of causes of death is congenital anomalies (Mathews and MacDorman, 2008). With the exception of folate supplementation to prevent neural tube defects, there is no clearly effective primary prevention program for any birth defect. Pregnancy screening and termination of severe defects, however, is an option for many mothers, and there is evidence that this practice contributes to a reduced prevalence of chromosomal anomalies at birth (Khoshnood et al, 2004). The major postneonatal cause of death since the 1970s in the United States is sudden infant death syndrome. This cause of death has declined substantially in the United States in parallel with successful public health efforts to discourage sleeping in the prone position during infancy (Ponsonby et al, 2002). MAJOR MORBIDITIES RELATED TO THE PERINATAL PERIOD The principal complications of preterm birth involve five organs: the lung, heart, gut, eye, and brain. Management of respiratory distress syndrome and its short- and long-term complications is the centerpiece of neonatal 5 CHAPTER 1 Neonatal and Perinatal Epidemiology TABLE 1-2 Ethnic Disparities in Key Perinatal Outcomes and Exposures in 2005 Non-Hispanic White Prevalence rate African American Prevalence rate Hispanic RR* Prevalence rate RR* Deaths Rates and Ratios Maternal mortality ratio (per 100,000 live births) Infant mortality rate (per 1000 live births) Neonatal mortality rate (per 1000 live births) Postneonatal mortality rate (per 1000 live births) Fetal mortality ratio (per 1000 live births)† 11.7 39.2 3.4 9.6 0.8 5.7 13.7 2.4 5.8 1.0 3.8 9.1 2.4 3.9 1.0 1.9 4.6 2.4 1.8 1.0 4.8 11.1 2.3 5.4 1.1 11.6% 1.6% 5.6 7.3% 1.2% 5.5 4.5% 3.6% 18.3% 4.1% 18.9 14.0% 3.3% 1.8% 4.6% 3.5% 1.6 2.6 3.4 1.4 2.8 3.3 1.0 1.0 11.6% 1.8% 6.2 6.9% 1.2% 5.9 2.8% 3.8% 1.0 30.4% 26.6% 33.6% 19.7% 1.1 0.7 29.0% 15.5% 1.0 0.6 13.9% 8.5% 0.7 2.9% 0.2 2.2% 9.4% 5.6% 16.3% 2.5 1.7 5.1% 14.2% 2.3 1.5 25.3% 58.3 69.9% 67.2 2.8 1.2 48.0% 99.4 1.9 1.7 38.2 37.5 1.0 22.8 0.6 Morbidity (percent of live births) Preterm birth Very preterm birth Extremely preterm birth Low birthweight Very low birthweight Extremely low birthweight Pregnancy-associated hypertension Diabetes in pregnancy 1.2 0.9 1.0 1.1 0.6 1.1 Interventions (percent of live births) Cesarean section Induction of labor Health Behaviors Smoking‡ Alcohol intake Late or no prenatal care§ Inadequate weight gain (<16 lb) at 40 weeks’ gestation¶ Unmarried Fertility rate (women 15-44 years old) Multiple births *Relative risk compared with non-Hispanic white (rounded to nearest decimal point). †Fetal deaths with stated or presumed period of 20 weeks’ gestation or more ‡Based on 36 states, New York City, and Washington, D.C., which used pre-2003 smoking definitions §Based on 37 states, Washington, D.C. and New York City ¶Based on 49 states, New York City, and Washington, D.C. medicine. Surgical or medical management of symptomatic patent ductus arteriosus is the major cardiac challenge in premature infants, and there is limited understanding of the striking variations, by time and place, of necrotizing enterocolitis—a disorder that in its most extreme forms can cause death or substantial loss of bowel function. Retinopathy of prematurity is closely related to arterial oxygen levels. The epidemic level of this disorder encountered in the 1950s, when oxygen was freely administered without monitoring, was a major setback for neonatal medicine (Silverman, 1980). However, even with more careful management of oxygen, retinopathy of prematurity continues to occur. The largest unsolved problem in neonatal medicine remains the high frequency of brain damage in survivors of premature birth. The extraordinary decline in mortality rates has not been paralleled by similar declines in rates of neurodevelopmental disabilities in survivors. The key epidemiologic feature of cerebral palsy rates in population registries toward the end of the twentieth century was a modest overall increase in the prevalence of that disorder, which was attributable entirely to the increasing number of survivors of very low birthweight. There are suggestions that this rise may now be leveling (Paneth et al, 2006). FACTORS AFFECTING PERINATAL HEALTH HEALTH STATES IN PREGNANCY The major causes of neonatal morbidity, prematurity and birth defects, generally occur in pregnancies free of antecedent complications. Having a previous birth with an anomaly or a previous preterm birth raises the risk for recurrence of the condition. For preterm birth, no other known risk factor carries as much risk for the mother as having previously delivered preterm. 6 PART I Overview More than a quarter of preterm birth is iatrogenic, the result of induced labor in pregnancies in which the fetus is severely compromised (Morken et al, 2008). Generally the reason is preeclampsia with attendant impairments in uterine blood flow and poor fetal growth, but poor uterine blood flow and impaired fetal growth can also occur independently of diagnosed preeclampsia; the other major complication of pregnancy is diabetes, most often gestational but sometimes preexisting. Insulin resistance in the mother promotes the movement of nutrients towards the fetus, and typically the infant of the diabetic mother is large for gestational age. Severe diabetes, however, can be accompanied by fetal growth retardation. HEALTH BEHAVIORS The most carefully studied and well-established health behavior affecting newborns is maternal cigarette smoking, which has a consistent effect in impairing fetal growth (Cnattingius, 2004). Infants with growth retardation from maternal cigarette smoking paradoxically survive slightly better than do infants of the same weight whose mothers did not smoke, but the net effect of smoking, which also shortens gestation slightly, is to increase perinatal mortality. Although the subject is much debated, it has not been conclusively shown that prenatal maternal smoking has independent long-term effects on children’s cognitive capacity (Breslau et al, 2005). Alcohol is less clearly a growth retardant, but mothers who drink heavily during pregnancy are at risk of having infants with the cluster of defects known as fetal alcohol syndrome. Cocaine use in pregnancy is almost surely a severe growth retardant, and it may affect neonatal behavior, but the long-term effects of this exposure on infant cognition and behavior are not as grave as initially feared (Bandstra et al, 2010). PERINATAL MEDICAL CARE In light of the potent effects of medical care on the neonate, it has been important to develop systems of care that ensure, or at least facilitate, provision of care to neonates in need. This concept was first promoted by the March of Dimes Foundation, which in its committee report of 1976 recommended that all hospitals caring for babies be classified as level 1 (care for healthy and mildly sick newborns), level 2 (care for most sick infants born in the hospital, but not accepting transfers), or level 3 (regional centers caring for complex surgical disease and receiving transfers) (The Committee on Perinatal Health, 1976). This concept of a regional approach to neonatal care, with different hospitals playing distinct roles in providing care, was endorsed by organizations such as American College of Obstetricians and Gynecologists and the American Academy of Pediatrics and by many state health departments. Whereas it is important to transfer sick babies to level 3 centers when needed, it is preferable to transfer mothers at risk of delivering prematurely or of having a sick neonate, because transport of the fetus in utero is superior to any form of postnatal transport. Birth at a level 3 center has been shown consistently to produce lower mortality rates than birth in other levels of care (Paneth, 1992). The overall system of care, which includes selecting mothers and babies for transfer to other hospitals, is highly dependent on cooperative physicians and health systems. Concern has been raised that economic considerations may threaten ideal systems of regionalization of newborn intensive care (Bode et al, 2001). EPIDEMIOLOGIC STUDY DESIGNS IN THE PERINATAL PERIOD Epidemiologic studies have contributed substantially to better understanding the patterns of risk and prognosis in the perinatal period, to tracking patterns of mortality and morbidity, to assessing regional medical care, and to assisting physicians and other providers in evaluating the efficacy of treatments. A variety of study designs have been used in this research. VITAL DATA ANALYSES Routinely collected vital data serve as the nation’s key resource for monitoring progress in caring for mothers and children. All data presented in this chapter’s figures and tables are derived from the annual counts of births and deaths collected by the 52 vital registration areas of the United States (50 states, Washington, D.C., and New York City) and then assembled into national data sets by the National Center for Health Statistics. Unlike data collected in hospitals, in clinics, or from nationally representative surveys, birth and death certificates are required by law to be completed for each birth and death. Birth and death registration has been virtually 100% complete for the entire United States since the 1950s. The universality of this process renders the findings from vital data analyses stable and generalizable. Figure 1-2 illustrates the most recent nationally recommended standard for birth certificate data collection, which has been adopted by most states. Birth certificates contain valuable information for neonatologists. Especially of note are variables such as to whether the mother or infant was transferred for care (not shown), breastfeeding plans, alcohol and cigarette smoking histories, and patterns of maternal weight gain and prenatal care. The limitations of vital data are well known. Causes of death are subject to certifier variability and, perhaps more importantly, to professional trends in diagnostic categorization. The accuracy of recording of conditions and measures on birth certificates is often uncertain and variable from state to state and hospital to hospital; however, the frequencies of births and deaths in subgroups defined objectively, such as birthweight, are likely to be valid. COHORT STUDIES IN PREGNANCY AND BIRTH Studies that follow populations of infants over time, beginning at birth or even before birth and continuing to hospital discharge, early childhood, and adult life, are the leading sources of information about perinatal risk factors for disease and adverse outcomes. As with all observational studies, cohort studies produce associations of exposures and CHAPTER 1 Neonatal and Perinatal Epidemiology FIGURE 1-2 U.S. national standard birth certificate, 2003 revision. 7 8 PART I Overview outcomes whose strength and consistency must be carefully judged in the light of other biologic evidence, and with attention to confounding and bias. Collaborations across centers in assembling such data are highly valuable. One noteable collaboration is the Vermont-Oxford Network, which provides continuous information on the frequency of conditions observed and diagnoses made in hundreds of U.S. and overseas hospitals, with a particular emphasis on using these data for improving care (­Horbar et al, 2010). The neonatal network supported by the National Institute of Child Health and Human Development (NICHD) has been a rich source of randomized trials and has produced observations about prognosis based on large samples of low-birthweight babies (Fanaroff, 2004). These collaborations focus mainly on the period until hospital discharge. Multicenter cohort studies focusing on diagnosis and follow-up of brain injury in premature infants—such as the Developmental Epidemiology Network (Kuban et al, 1999), Neonatal Brain Hemorrhage (Pinto-Martin et al, 1992), and Extremely Low Gestational Age Newborns (O’Shea et al, 2009) studies—have contributed to our understanding of the prognostic value of brain injury imaged by ultrasound in the neonatal period, because they include follow-up to the age of 2 years or later. Of particular value have been regional or population-wide studies of low-birthweight infants with follow-up to at least school age, among which are included the Neonatal Brain Hemorrhage study from the United States and important studies from Germany (Wolke and Meyer, 1999), the Netherlands (Veen et al, 1991), the United Kingdom (Wood et al, 2000), and Canada (Saigal et al, 1990). Newborn intensive care has been in place long enough that the first reports of adult outcomes in small infants are now emerging (Saigal and Doyle, 2008). These reports paint a picture that is perhaps less dire than anticipated. From 1959 to 1966, the National Collaborative Perinatal Project assembled data on approximately 50,000 pregnancies in 12 major medical centers and followed them to age 7 (Niswander and Gordon, 1972). This highly productive exercise, one of whose major contributions was to show that birth asphyxia is a rare cause of cerebral palsy, has now been followed by the development of even larger birth cohort studies starting in pregnancy. For reasons that are not entirely clear, a sample size of 100,000 has been adopted in studies in Norway (Magnus et al, 2006), Denmark (Olsen et al, 2001), and the United States (Lyerly et al, 2009). A major difference from the National Collaborative Perinatal Project is that each of the studies aims to obtain some degree of national representativeness. The U.S. National Children’s Study, currently underway, plans to enroll women in early pregnancy and possibly before conception and to follow the offspring to age 21 in 105 locations in the United States, selected by a stratified random sampling of all 3141 U.S. counties. RANDOMIZED CONTROLLED TRIALS Few areas of medicine have adopted the randomized trial as wholeheartedly as has newborn medicine. The number of trials mounted has been large and have created a strong influence on practice. A notable influence on this field has been the National Perinatal Epidemiology Unit at Oxford University, established in 1978, which prioritized randomized trials among their several investigations of perinatal care practices and other circumstances affecting maternal and newborn outcomes. The NICHD neonatal research network was established in 1986, principally to support trials. Hundreds of trials have been mounted by these two organizations, but many other centers have contributed to the trial literature. Trials in pregnancy or in labor have also been supported by the National Perinatal Epidemiology Unit and by NICHD, who support a network of obstetric centers to conduct trials in pregnancy and labor, the NICHD maternal-fetal research network. These trials have often had important implications for newborns and mothers, most notably for showing that the risk of preterm birth can be reduced by administering 17-OH hydroxyprogesterone caproate in midgestation to high-risk women (Meis et al, 2003). Most newborn trials have focused on outcomes evident in the newborn period, such as mortality, chronic lung disease, brain damage visualized on ultrasound exam, and duration of mechanical ventilation or hospital stay. Recently however, trials extending into infancy or early childhood that incorporate measures of cognition or neurologic function have been a welcome addition to the trial arena. In the past few years, such trials have shown that moderate hypothermia can reduce mortality and brain damage in asphyxiated term infants (Shankaran et al, 2005), and that both caffeine for apnea treatment (Schmidt et al, 2007) and magnesium sulfate administered during labor can reduce the risk of cerebral palsy (Rouse et al, 2008). Trials in which both mortality and later outcome are combined raise complex methodologic issues. Imbalance in the frequency of the two outcomes being combined can result in a random variation in the more common outcome, overwhelming a significant finding in the other. Precisely how best to conduct such dual- or multiple-outcome trials is the subject of discussion and debate in the neonatal and epidemiology communities. As the number of trials increases, not all of them sufficiently powered, the methodology for summarizing them and drawing effective conclusions has become increasingly important to neonatologists. The terms systematic review and metaanalysis have firmly entered the research lexicon, especially the randomized trial literature. The Cochrane Collaboration is an international organization that uses volunteers to systematically review trial results in all fields of medicine. The collaboration, established in 1993, began in the field of perinatal medical trials. Systematic reviews of neonatal trials reviewed by the Cochrane Collaboration are hosted on the website of the NICHD (http://www. nichd.nih.gov/cochrane). SUMMARY The patterns of disease, mortality, and later outcome in the perinatal period are complex. Some factors are reasonably stable (e.g., long-term trends in preterm birth and birthweight), whereas others can undergo rapid change (e.g., the rates of cesarean section and twinning). The success of CHAPTER 1 Neonatal and Perinatal Epidemiology newborn intensive care is well established. No other organized medical care program, targeted at a broad patient population, has had such remarkable success in lowering mortality rates in such a short period of time. Much of that success is due to the evidence-based nature of neonatal practice. Nonetheless, this success has opened the door to new problems as survivors of intensive care face the challenges of the information age. Resource allocations similar to those that permitted the development of newborn intensive care are now needed to address the educational and rehabilitative needs of survivors. A hopeful sign is the success of some recently studied interventions in reducing the burden of brain damage. On the nontechnologic front, targeted epidemiologic efforts to address perinatal disorders have yielded progress. Careful study of the circumstances surrounding infant sleep patterns led to active discouragement of sleeping in the prone position, which has reduced mortality from sudden infant death syndrome by half (The Committee on Perinatal Health, 1976). Observational research, followed by two important randomized trials in Europe, led to interventions that increased folate intake in women of child-bearing age and a substantial reduction in the birth prevalence of neural tube defects (Czeizel and Dudás, 1992; Medical Research Council, 1991). The population-level study of health events occurring in pregnancy and infancy, their antecedents, and their long-term consequences have been an important component to the success of newborn care. Careful selfevaluation through monitoring of vital data and collaborative clinical data, rigorous assessment of new treatments 9 through randomized trials, and alertness to opportunities to implement prevention activities after discovering important risk factors should continue to guide the care of the newborn. ACKNOWLEDGMENTS The author thanks Ariel Brovont and Kimberly Harris for assisting in the preparation of the figures and tables. SUGGESTED READINGS Division of Reproductive Health: National Center for Chronic Disease Prevention and Health Promotion: achievements in public health, 1900-1999: healthier mothers and babies, MMWR 48:849-858, 1999. Fanaroff A: The NICHD neonatal research network: changes in practice and outcomes during the first 15 years, Semin in Perinatol 27:281-287, 2004. Horbar JD, Soll RF, Edwards WH: The Vermont Oxford Network: a community of practice, Clin Perinatol 37:29-47, 2010. Hoyert DL: National Center for Health Statistics: maternal mortality and related concepts, Vital Health Stat 33:1-13, 2007. Khan KS, Wojdyla D, Say L, et al: WHO analysis of causes of maternal death: a systematic review, Lancet 367:1066-1074, 2006. MacDorman MF, Mathews TJ: Centers for Disease Control and Prevention National Center for Health Statistics: behind international rankings of infant mortality: how the United States compares with Europe, NCHS Data Brief 23:1-8, 2009. Niswander KR, Gordon M, editors: The Collaborative Perinatal Study of the National Institute of Neurological Diseases and Stroke: the women and their pregnancies, Philadelphia, 1972, WB Saunders. Saigal S, Szatmari P, Rosenbaum P, et al: Intellectual and functional status at school entry of children who weighed 1000 grams or less at birth: a regional perspective of births in the 1980s, J Pediatr 116:409-416, 1990. Silverman WA: Retrolental fibroplasias: a modern parable, New York, 1980, Grune & Stratton. World Health Organization: International statistical classification of diseases and related health problems, Tenth Revision, vol 2, ed 2, Geneva, 2004. Complete references and supplemental color images used in this text can be found online at www.expertconsult.com C H A P T E R 2 Evaluation of Therapeutic Recommendations, Database Management, and Information Retrieval Peter Tarczy-Hornoch BACKGROUND At a fundamental level, the practice of neonatology can be considered an information management problem. The care provider is combining patient-specific information (history, findings of physical examination, and results of physiologic monitoring, laboratory tests, and radiologic evaluation) with generalized information (medical knowledge, practice guidelines, clinical trials, and personal experience) to make medical decisions (diagnostic, therapeutic, and management). The Internet has made possible a revolution in the sharing and disseminating of knowledge in all fields, including medicine, with continued growth and maturation of online clinical information resources and tools. Although medicine remains a quintessentially human endeavor, computers are playing a growing role in information management, particularly in neonatology. Patient-specific and generalized information (medical knowledge) are becoming increasingly available in electronic form. In the United States, a growing number of hospitals are adopting electronic medical record systems to manage patient-specific information, with the approaches ranging from electronic flow sheets in the intensive care unit to entirely paperless hospitals. The American Recovery and Reinvestment Act of 2009 (ARRA; i.e., the federal economic stimulus plan) has a provision for the investment of $19 billion in health information technology to motivate physicians to adopt electronic health records (in the Health Information Technology for Economic and Clinical Health [HITECH] act, a part of AARA) and $1.1 billion to research the effectiveness of certain health care treatments. These ARRA provisions are predicated on the belief that quality, safety, and efficiency of clinical care can be improved through electronic medical records and through evidence-based practice. Parallel with and related to the adoption of information technology is the growth of societal pressures to improve the quality of medical care while controlling costs. These changes are beginning to affect the way in which medicine and neonatology are practiced. In turn, it is becoming important for neonatologists to understand basic principles related to biomedical and health informatics, databases and electronic medical record systems, evaluation of therapeutic recommendations, and online information retrieval. This expansion of information technology in clinical practice and the concurrent growth of medical knowledge have great promise in addition to potential pitfalls. One pitfall that must not be underestimated, and which is as great a danger today as when Blois (1984) first cautioned 10 against it, is the unquestioning adoption of information technology: “And, since the thing that computers do is frequently done by them more rapidly than it is by brains, there has been an irresistible urge to apply computers to medicine, but considerably less of an urge to attempt to understand where and how they can best be used.” A present and real challenge is information overload. Bero and Rennie (1995) observed, “Although well over 1 million clinical trials have been conducted, hundreds of thousands remain unpublished or are hard to find and may be in various languages. In the unlikely event that the physician finds all the relevant trials of a treatment, these are rarely accompanied by any comprehensive systematic review attempting to assess and make sense of the evidence.” The potential of just-in-time information at the point of care is thus particularly appealing, especially considering that the growth in published literature continues at an accelerating rate, with a flood of new knowledge coming from the latest research in genomics, proteomics, metabolomics, and systems biology. A vision to address this was articulated by one of the editors of the British Medical Journal: “New information tools are needed: they are likely to be electronic, portable, fast, easy to use, connected to both a large valid database of medical knowledge and the patient record” (Smith, 1996). Although these goals are close to being achieved, there is still progress to be made before this vision is a reality. This chapter aims to provide an overview of the current progress in this direction. BIOMEDICAL AND HEALTH INFORMATICS In the 1970s, clinicians with expertise in computers became intrigued by the potential of these tools to improve the practice of medicine, and thus the field of medical informatics was born. The importance of this field addressing the issues of information management in health care is growing rapidly, as seen in the activities of the American Medical Informatics Association (AMIA; www.amia.org). Medical informatics can be concisely defined as “the rapidly developing scientific field that deals with storage, retrieval, and optimal use of biomedical information, data, and knowledge for problem solving and decision making” (Shortliffe and Blois, 2006). A more extensive definition can be found at the AMIA Web site under About AMIA, including professional and training opportunities. The University of Washington CHAPTER 2 Evaluation of Therapeutic Recommendations, Database Management, and Information Retrieval Web site (www.bhi.washington.edu) contains a review of the discipline (found under History, About Us, Vision). The field includes both applied and basic research, with the focus in this chapter being on the applied aspects. Examples of basic research are artificial intelligence in medicine, genome data analysis, and data mining (sorting through data to identify patterns and establish relationships). As our knowledge of the genetic mechanisms of disease expands and more data about patients and outcomes are available electronically, the role of informatics in medicine will expand, particularly in the field of neonatology. The applied focus of the field in the 1960s and 1970s was data oriented, focusing on signal processing and statistical data analysis. In neonatology, the earliest applications of computers were for physiologic data monitoring in the neonatal intensive care unit (NICU). As the field matured in the 1980s, applied work focused on systems to manage patient information and medical knowledge on a limited basis. Examples are laboratory systems, radiology systems, centralized transcription systems, and, probably the best-known medical knowledge management system, the database of published medical articles maintained by the National Library of Medicine known first as MEDLARS, then as MEDLINE and currently as PubMed (www.ncbi.nlm.nih.gov/pubmed). For example, neonatologists began to develop tools to aid in the management of patients in the NICU, such as computerassisted algorithms to help manage ventilators, although the algorithms have not been successfully deployed on a large scale in the clinical setting. As computers and networking became mainstream in the workplace and home in the 1990s, informatics researchers began to develop integrated and networked systems (Fuller, 1992, 1997). With the explosion of information from the Human Genome Project, the intersection between bioinformatics and medical informatics began to blur, leading to the adoption of the term biomedical informatics. The 1990s saw the development of a number of important systems. In terms of patient-specific information retrieval, these systems included integrated electronic medical record systems that in their full implementation can encompass—in a single piece of easy-to-use software—interfaces to physiologic monitors; electronic flow sheets; access to laboratory and radiology data; tools for electronic documentation (charting), electronic order entry, and integrated billing; and modules to help reduce medical errors. The Internet has permitted ready access and sharing of this information within health care organizations and limited secured remote access to this information from home. In terms of patient population information retrieval, a number of tools were developed to help clinicians and researchers examine aggregate data in these electronic medical records to document outcomes and help to improve quality of care. The Internet, particularly the World Wide Web, has transformed access to medical knowledge (Fuller et al, 1999). Health sciences libraries are becoming digital and paper repositories. Journals are available online. Knowledge is now available at the point of care in ways that were not previously possible (Tarczy-Hornoch et al, 1997). In 2004, in recognition of the unfulfilled potential of health care information technology, the Office of the National Coordinator for Health 11 Information Technology (ONC) was established (www. healthit.hhs.gov) to achieve the following vision: Health information technology (HIT) allows comprehensive management of medical information and its secure exchange between health care consumers and providers. Broad use of HIT has the potential to improve health care quality, prevent medical errors, increase the efficiency of care provision and reduce unnecessary health care costs, increase administrative efficiencies, decrease paperwork, expand access to affordable care, and improve population health. In the upcoming decade, the focus will be shifting from demonstrating the potential of electronic medical record and information systems toward implementing them more broadly to realize their benefit (e.g., the ARRA legislation). Evidence-based medicine is considered by many as a part of informatics as an approach to the evaluation of therapeutic recommendations and their implementation (see later discussion), and it is part of the ARRA and some approaches to health care reform being proposed in 2009. Neonatologists have been involved in informatics for a long time. Duncan (2010) maintains an excellent continually updated bibliographic database on the literature about computer applications in neonatology. In 1988, as one of the earlier groups to develop national databases of clinical care, neonatologists established and expanded the Vermont Oxford Network (www.vtoxford.org) to improve the quality and safety of medical care for newborn infants and their families. As part of the activities, the Vermont established and maintained Oxford Network a nationwide database about the care and outcome of high-risk newborn infants. In 1992, Sinclair et al (1992) published one of the earlier ­evidence-based textbooks, Effective Care of the Newborn Infant. DATABASES In broad terms, a database is an organized, structured collection of data designed for a particular purpose. Thus, a stack of 3 × 5 cards with patient information is a database, as is the typical paper prenatal record. Most frequently, the term database is used to refer to a structured electronic collection of information, such as a database of clinical trial data for a group of patients in a study. Databases come in a variety of fundamental types, such as single-table, relational, and object-oriented. A simple database can be built using a single table by means of a spreadsheet program such as Microsoft Excel, or a database program such as Microsoft Access (Microsoft, Redmond, Washington). The advantage of such a database is that it is easy to build and maintain. For an outcomes database in a neonatology unit, each row can represent a patient and each column represents information about the patients (e.g., name, medical record number, gestational age, birth date, length of stay, patent ductus arteriosus [yes/no], necrotizing enterocolitis [NEC; yes/no]). The major limitation of such a database is that a column must be added to store the information each time the researcher wants to track another outcome (e.g., maple syrup urine disease [MSUD]). This limitation can result in tables with dozens to hundreds of columns, which then become difficult to maintain. The challenges can be illustrated with a few examples. The first 12 PART I Overview example of a challenge is that which results from adding a new column (e.g., MSUD); one must either review all records (rows) already in the spreadsheet for the presence or absence of MSUD or flag all existing records (rows) in the spreadsheet as unknown for MSUD status. The second example of a challenge results from the logistics of managing an extremely wide spreadsheet—imagine not adding the tenth column but the 1000th column. The majority of databases and electronic medical records in neonatology are built using relational database software. To build a simple outcomes relational database that permits easy adding of new outcome measures, one could use a three-table database design (Figure 2-1). The first table contains all the information for each patient (e.g., name, medical record number, gestation in weeks, birth date, admit date, discharge date). The second table is a dictionary that assigns a code number to each diagnosis or outcome being tracked (e.g., patent ductus arteriosus = 1; NEC = 2; MSUD = 10234). The third table is the diagnosis-tracking table; it links a patient number to a particular code and assigns a value to that code. Adding a new diagnosis to track would require adding an entry to the diagnosis dictionary table. To add a diagnosis to a patient, one would add an entry to the tracking table. For example, Girl Smith (medical record number 00-00-01) has a diagnosis of NEC. To add the diagnosis, add to the diagnosis table an entry that has a value of 00-00-01 in the medical record column, a value of 2 (the code for NEC) in the code column, and a value of 2 in the value column (the code for surgical). Although relational databases are harder to build, they provide greater flexibility for expansion and maintenance and thus are the preferred implementation for clinical databases. They address the challenges in a simple spreadsheet by tracking dates that new diagnostic codes were added and by user interfaces that allow one to easily view only diagnoses present for a given patient rather than all potential diagnoses for a patient. The distinction between an NICU quality assessment– quality improvement (i.e., outcomes) database and an electronic medical record is largely a matter of degree. Some characteristics typical of a neonatal outcomes database are data collection and data entry after the fact, limited amount of data collected (a small subset of the information needed for daily care), lack of narrative text, lack of interfaces to laboratory and other information systems, and the episodic (e.g., quarterly) use of the system for report generation. Some characteristics typical of an electronic medical record are real-time (daily or more frequent) data entry, a large amount of data collected (approximating all the information needed for daily care in a fully electronic care environment), narrative text (e.g., progress notes, radiology reports, pathology reports), interfaces to laboratory and other information systems, and, most important, the use of the system for daily patient care, including features such as results review, messaging or alerting for critical results, decision support systems (drug dosage calculators, drug-drug interaction alerts, among others), and computerized electronic order entry. In the past, the majority of neonatal databases and firstgeneration NICU electronic medical record systems were FIGURE 2-1 Example of a relational database. CHAPTER 2 Evaluation of Therapeutic Recommendations, Database Management, and Information Retrieval developed locally by and for neonatologists. The literature describing these efforts is available online (Duncan, 2010). Unfortunately the majority of these systems were never published or publicly documented, and thus a number of important and useful innovations are lost or must be repeatedly rediscovered. Anybody thinking about building their own neonatology database would be well advised to review the existing literature and existing commercial products before embarking on this path. That said, there is room for improvement of the existing products, and neonatologists continue to develop their own databases today. With the national push toward interoperable electronic medical records through U.S. Department of Health and Human Services Office of the National Coordinator for Health Information Technology (ONC; since 2004) and ARRA HITECH (since 2009), it is likely that the market will consolidate into a smaller number of neonatology practice tailored systems that are certified and that interoperate with the National Health Information Network (2010). The largest neonatal outcomes database is the centralized database maintained by the Vermont Oxford Network with the mission of improving the quality and safety of medical care for infants and their families. One of the key activities of the network is their outcomes database, which involves more than 800 participating intensive care nurseries both in the United States and internationally collecting data on approximately 55,000 low-birthweight infants each year. Other activities of the network are clinical trials, follow-up of extremely low–birthweight infants, and NICU quality and safety studies. The focus of the database initially was very low–birthweight infants (401 to 1500 g), but this focus has expanded to include data on infants weighing more than 1500 g. Presently the network collects data on more than two thirds of the very low–birthweight infants born in the United States. Participants in the network submit data and in return receive outcome data for their own institution and comparative data from other nurseries nationwide, including custom reports, comparison groups, and quality management reports. Members also have the ability to participate in collaborative research projects and collaborative multicenter quality-improvement collaborations. All data except their own are anonymous for all participants. The network does have access to both the individual and aggregate data. The network database is maintained centrally, and data quality monitoring and data entry are centralized. Initially the process involved paper submission of data by participating nurseries. Currently a number of the commercial and custom NICU databases and electronic medical record systems can export their data in the format required by the Vermont Oxford Network as well as the option to submit data directly using the custom eNICQ software developed by Vermont Oxford. Submissions of data are thus a combination of paper forms and reports generated by commercial and custom software packages. The database focuses on tracking outcomes. With the passage of the Health Insurance Portability and Accountability Act of 1996 (HIPAA) and federal regulations governing the confidentiality of electronic patient data, some of the anonymous demographic data that were collected by the network in the past have decreased. This change has grown from concerns that, in combination with identity of the referring center, these data could be used to uniquely identify patients, which is a violation of the HIPAA. 13 ELECTRONIC HEALTH RECORD An electronic health record (EHR; also known as an electronic medical record) is much more complex than an outcomes database, because the system is intended to be used continually on a daily basis to replace electronically some, if not all, of the record keeping, laboratory result review, and order writing that occur in a neonatal intensive care nursery (or more generically in any inpatient or outpatient clinical setting). The complexity of this task becomes evident if one imagines that, for a paperless medical record environment, every paper form in a nursery would need to be replaced with an electronic equivalent, which is also true for every paper-based workflow and process. Organizations are moving in this direction because of a combination of forces, such as the desire to reduce error and to control the spiraling costs of health care. These reasons are addressed at great length in two reports from the Institute of Medicine (Institute of Medicine Committee on Improving the Patient Record, 1997; Kohn et al, 2000). These benefits are typically achieved when information is available electronically (e.g., results of laboratory tests, radiology procedures, transcription) and input into the system (e.g., problem lists, allergies), and when both sets of information are combined and checked against electronic orders. Only with electronic orders has it been shown that errors can be reduced and care provider behavior clearly changed. Combining just electronic laboratory results (e.g., creatinine level) and electronic order entry (e.g., a drug order), for example, enables one to verify that drug dosages have been correctly adjusted for renal failure. This approach would work well in adults, but in neonates, whose renal function is more difficult to assess and for whom drug dosage norms depend on gestational age and post-delivery age, additional information must be entered into the system (e.g., urine output, gestational age), requiring a more sophisticated EHR. Despite more than one decade of work deploying EHRs, the proportion of acute care hospitals that are members of the American Hospital Association and have a comprehensive EHR is remarkably low (1.5%), and computerized order entry has been implemented in only 17% of hospitals (Jha et al, 2009). Results review systems include basic demographic data, such as name, age, and address from the hospital registration system. These systems require a moderate amount of work to tie them to the various laboratory, radiology, and other systems and to train users. The benefits are hard to quantify, but users typically prefer them to the paper alternative because of the more rapid access to information. The challenge in moving beyond the results review level to the integrated system level is that the documentation level and order entry level are essentially prerequisites for the integrated system level, but they have marginal benefit, particularly given the human and financial costs. Integrated systems require significant work to implement, including the presence of computers at each bedside, as well as significant work to train users. The benefits accrue mainly to the organization, in the form of reduced costs of filing, printing, and maintaining paper records and, if providers are forced to enter notes instead of dictate them, significant savings in transcription costs. 14 PART I Overview The challenge is that the end users often find that it takes much longer to do their daily work with electronic documentation. Without moving to electronic order entry, if not an integrated system, the users do not realize major day-to-day benefits. The ARRA HITECH provisions noted earlier essentially are designed to create incentives for provider adoption of EHRs to overcome this activation barrier. The benefits start to accrue more clearly at the next level—electronic order entry. The complexity of implementing and deploying an electronic order entry system cannot be overstated. Interfaces need to be built with all the systems that are part of results review in addition to other systems. Furthermore, a huge database of possible orders must be created to allow users to pick the right orders. This database and the menu of choices are needed because computers are poor at recognizing and interpreting a narrative text typed by a human. Finally and most important, there is a huge training challenge, because writing orders electronically is more complex and time consuming than writing them by hand. The change management issues become apparent when one considers that typically these systems take the unit assistant out of the loop; therefore much of the oversight that can occur at the unit assistant level does not, or the burden of oversight is borne by the person entering the orders. After overcoming the barriers to electronic order entry, organizations can start to benefit from integrated systems. For this reason, the trend today is not a stepwise move from results review to documentation of integrated systems. Instead, organizations are moving from results review directly to integrated systems. Interestingly, the technical complexities and the training and usage complexities of integrated systems are not much higher than those for order entry. Integrated systems add tools to make life easier for care providers using all the data in the system. As an analogy, an integrated EHR system is like an office software suite that encompasses a word processor, a spreadsheet, a slide presentation tool, a graphic drawing tool, and a database, all of which can communicate with one another, making it easy to put a picture from the drawing tool or a graph from a spreadsheet into a slide show. Integrated systems include (1) checking orders for errors, (2) alerts and reminders triggered by orders or by problems on the problem list or other data in the system, (3) care plans tied to patient-specific information, (4) charting modules customized to the problem list, (5) charting and progress notes that automatically import information (e.g., from laboratory tests, flow sheets) and that help generate orders for the day as the documentation occurs, (6) modules to facilitate hyperalimentation ordering, and (7) modules to assist in management. For example it is possible to imagine a system in which reminders for screening studies (e.g., for retinopathy of prematurity, intraventricular hemorrhage, and brainstem auditory evoked response) were triggered by gestational age, a problem list, and previous results of screening studies. Similarly, admitting a neonate at a particular gestational age with a particular set of problems could trigger pathways, orders, and reminders specific to that clinical scenario. An important caveat is that all such systems are only as good as the data and rules put into them. The issues raised in the section on evaluation of therapeutic recommendations are important to consider in the context of electronic order entry and integrated systems. The EHR market is still relatively young and continually evolving; this is true of products designed specifically for the NICU and more generic products designed to be used throughout a hospital or health care system. The ONC was established in part to address this young marketplace by creating standards and certification bodies. In particular, the ONC created the Certification Commission for Healthcare Information Technology (CCHIT) and the Health Information Technology Standards Panel to begin to bring more standardization to the marketplace. The CCHCIT examines criteria for certifying systems, including functionality, security, and interoperability. Order entry and documentation systems are beginning to be more widely adopted, however truly integrated systems are much less broadly implemented. The major reason for this situation is that the needs of different health care systems vary significantly, and the existing products are not flexible enough to meet all these needs in one system. Furthermore, there is a trend among health care organizations, EHR developers, and vendors to move away from niche systems tailored to particular subsets of care providers, such as neonatology, and toward a focus on systems that are generically useful. There are two important drivers behind this trend. The first and most important reason for adopting a single integrated system is that the benefits of an EHR system begin to accrue only when an entire organization uses the same one. Consider the following scenario: a woman receives prenatal care in the clinic of an institution and is then admitted to the emergency department in preterm labor. Her infant is delivered in the labor and delivery department, hospitalized in the neonatal intensive care nursery, and discharged to an affiliated pediatric follow-up clinic. In the current era of paper medical records, paper is used to convey information from one site to the other. In an integrated EHR system, all the information for both mother and infant is in one place for all providers to see. Interoperability is important as well if the care described crosses organizational boundaries, such as an outpatientfocused health maintenance organization contracting inpatient obstetric care to one hospital system and neonatal or pediatric care to a children’s hospital in a different health care system. If a single unified institution were to adopt niche software tailored to the needs of each site, a provider caring for the infant might need to access an emergency department system, an obstetric system, an NICU system, and an outpatient pediatric system to gather all the pertinent information. Each system would require the user to learn a separate piece of software. Learning a site-specific piece of software is a considerably greater burden on care providers than learning to use a site-specific paper form. If care crosses organizations and electronic systems are not interoperable, then care transitions most often remain on paper. The second factor driving adoption of integrated systems is economies of scale. The ideal EHR system contains electronic interfaces that automatically import the system data from laboratory, pharmacy, radiology, transcription, integrated electronic orders, error checking, and electronic CHAPTER 2 Evaluation of Therapeutic Recommendations, Database Management, and Information Retrieval documentation by care providers. Given that development of these interfaces, training, and maintenance cost more than the purchase of the system itself, it is far more cost effective to install one system with one set of interfaces and one set of training and maintenance issues than to replicate the process multiple times. The neonatal intensive care environment poses some unique challenges for EHRs. As a result, it is important to ensure that when health care systems are making decisions about the purchase of an EHR, neonatologists and other neonatal health care providers are involved in the process. An excellent source of information about NICU medical record systems and databases is an article by Stavis (1999). Neonatologists in the position of helping to select an EHR system must acquire the necessary background through reading some basic introductory texts on medical informatics, focusing on EHRs. It is then critical that they survey other organizations similar to their own to discover which systems have worked and which ones have not. For example, the needs of a level III academic nursery that performs extracorporeal membrane oxygenation are different from those of a community level II hospital that does not perform mechanical ventilation. Systems that work well in teaching hospitals with layers of trainees may not work well in private practice settings and vice versa. Most important, when using a medical informatics framework, the neonatologists must develop a list of prioritized criteria specific to their institution and compare available products in the marketplace with this list, while also considering the recommendations of the CCHIT as described earlier. All end user needs must also be considered. If residents, nurse practitioners, nutritionists, pharmacists, and respiratory therapists are expected to use the system, their input must be solicited. Ensuring broad-based input is especially relevant if the goal is an EHR system into which a lot of data will be entered by health care providers (e.g., electronic charting, note writing, medication administration records, order entry). The reason to ensure acceptance by all users of systems that require data entry is that a significant percentage of systems requiring data entry has ultimately been unsuccessful because of lack of user acceptance. Unfortunately there is little literature on this issue, because institutions rarely publicize and publish their failures in this arena, although the situation is beginning to change. A review of some of these challenges and a theoretical framework for looking at them is provided by Pratt et al (2004). The final step in evaluating a potential system is to develop a series of scenarios and to have potential users test the scenarios. Evaluating usage scenarios typically involves visits to sites that have installed the EHR system under consideration. An example of a scenario might be for a nurse, a respiratory therapist, a resident, and an attending physician to try to electronically replicate, on a given system under consideration, the bedside charting, progress note charting, and order writing for a critically ill patient who undergoes extracorporeal membrane oxygenation and then decannulation. The reason for developing and testing such scenarios is that this approach is the best way to ensure that aspects of charting, note writing, and documentation unique to the NICU are supported by the system. 15 EVALUATING THERAPEUTIC RECOMMENDATIONS Once all the data about a patient, whether in electronic or paper form, are in hand, the clinician is faced with the challenge of medical decision making and applying all that he or she knows to the problem. It is vital that the clinician understand what is known and what is still uncertain in terms of the validity of therapeutic recommendations. The evaluation of new recommendations arising from a variety of sources, including journal articles, metaanalyses, and systematic reviews, is a critical skill that all neonatologists must master. Broadly speaking, this approach has been termed evidence-based medicine. A full discussion of this approach to evaluation of new approaches in clinical medicine is beyond the scope of this chapter. Two outstanding sources of information are the works by Guyatt and Rennie (2002) and Straus et al (2005). An excellent overview of the progress in evidence-based medicine over the last 15 years is provided by Montori and Guyatt (2008). An important caveat is that evidence-based medicine is not a panacea. It is not helpful when the primary literature does not address a particular clinical situation, such as one that is rare or complex. This approach also does not necessarily address broader concerns, such as clinical importance or cost effectiveness, although it sometimes does. There are potential challenges when combining evidence-based medicine with the emerging humanistic approach to health care, which can heavily weight patient preferences potentially over the evidence base. In the early days of medicine, the standard practice was observation of individual patients and subjective description of aggregate experiences from similar patients. As the science of medicine evolved, formal scientific methods were applied to help to assess possible therapeutic and management interventions. Important tools in this effort are epidemiology, statistics, and clinical trial design. Currently medicine in general and neonatology in particular are faced with an interesting paradox. For some areas, there is a wealth of information in the form of randomized controlled clinical trials, whereas there is scant information to guide clinical practice for others. A wealth of well-designed clinical trials on the use of surfactant has been published, for example, but there are essentially no trials addressing the management of chylothorax. One might assume that the practice of medicine reflects the available evidence, but this is not the case. McDonald (1996) summarized the problem as follows: “Although we assume that medical decisions are driven by established scientific fact, even a cursory review of practice patterns shows that they are not.” A study of 2500 treatments in the British Medical Journals Clinical Evidence Database (http://clinical­evidence.bmj.com), shows that, as of the summer of 2009, 49% of treatments are of unknown effectiveness, 12% are clearly beneficial, 23% are likely to be beneficial, 8% are a tradeoff between beneficial and harmful, 5% are unlikely to be beneficial, and 3% are likely to be ineffective or harmful. As a result, neonatologists have a responsibility to identify what knowledge is available in the literature and elsewhere and to critically evaluate this information before applying it to practice. Furthermore, because this information is constantly evolving, practitioners must continually revisit the underlying literature as it expands (e.g., the recommendations regarding the use of steroids for chronic lung disease). 16 PART I Overview The evidence-based practice of medicine is an approach that addresses these issues. It is helpful to consider the process as involving two steps—the critical review of the primary literature and the synthesis of the information offered in the primary literature. Critical review of the primary literature is an area in which most neonatologists have significant experience, with journal clubs and other similar forums. The approach involves systematically reviewing each section of an article (i.e., background, methods, results, discussion) and asking critical questions for each section (e.g., for the methods section: Is the statistical methodology valid? Were power calculations made? Was a hypothesis clearly stated? Do the methods address the hypothesis? Do the methods address alternative hypotheses? Do the methods address confounding variables?). The formal evaluation of each section must then be synthesized into conclusions. A helpful question to ask is, “Does this paper change my clinical practice, and if so, then how?” Additional resources for systematic review of the primary literature are listed in the Suggested Readings. It is important to note that guidelines for systematic review of a single article differ according to whether it describes a preventive or therapeutic trial (e.g., use of nitric oxide for chronic lung disease), evaluation of a diagnostic study (e.g., use of C-reactive protein level for prediction of infection), or prognosis (e.g., prediction of outcome from a Score for Neonatal Acute Physiology score). The second, and arguably more important, step is to determine not the effect of one article on one’s practice, but the overall effect of the body of relevant literature on one’s practice. For example, if the preponderance of the literature favors one therapeutic recommendation, then a single article opposing the recommendation must be weighed against the other articles that favor it. This task is complex, and the most complete and formal statistical approach to combining the results of multiple studies (i.e., metaanalysis) requires significant investment of time and effort. Part of the evidence-based practice of medicine approach therefore involves the collaborative development of evidence-based systematic reviews and metaanalyses by communities of care providers. Within the field of neonatology, Sinclair et al (1992) laid the seminal groundwork for this approach; their textbook Effective Care of the Newborn remains an important milestone, but it illustrates the problem of information currency. Because the book was published in 1992, none of the clinical trials in neonatology in the last decade and a half are included. The Internet has permitted creation and continual maintenance of up-to-date information by a distributed group of collaborators, lending itself well to the maintenance of a database of evidence-based medicine reviews of the literature. This international effort is the Cochrane Collaboration, and the Cochrane Neonatal Review is devoted to neonatology (Cochrane Neonatal Collaborative Review Group). A limitation of the Cochrane approach is illustrated by the relatively restricted scope of topics covered at the National Institute of Child Health and Human Development Web site (www.nichd.nih.gov/cochrane/cochrane.htm). The existence of a review requires adequate literature on a topic and a dedicated and committed clinician to create and update the review. It is important to distinguish between these formal approaches to reviewing the literature (i.e., systematic literature reviews and metaanalyses) that have specific methodologies and more ad hoc reviews of the literature. Evidence-based medicine aggregate resources such as the Cochrane Collaboration take a more systematic approach, but review articles published in the literature vary in their approach. Metaanalyses are easy to distinguish, but systematic reviews versus ad hoc reviews are harder to distinguish. Systematic reviews focus on quality primary literature (e.g., controlled studies rather than case series or case reports) and must include (1) a methods section for the review article that explicitly specifies how articles were identified for possible inclusion and (2) what criteria were used to assess the validity of each study and to include or exclude primary literature articles in the systematic review. Systematic reviews also tend to present the literature in aggregate tabular form, even when metaanalyses of statistics of all the articles cannot be done. One commonly used source of overview information in neonatology—the Clinics in Perinatology series—is a mix of opinion (written in the style of a book chapter), ad hoc literature review, systematic literature review, and metaanalysis. Guidelines (e.g., screening recommendations for group B streptococcal infection), although based on primary literature review, are typically neither metaanalyses nor systematic reviews of the literature. Whereas formal methods are used to derive conclusions with metaanalyses and systematic reviews, guidelines are developed frequently instead by consensus among committee members; this is true of both national and local practice guidelines. General textbooks of neonatology are typically based on ad hoc literature review that includes both primary literature and systematic literature review. When reading overviews of the aggregate state of current knowledge on a given topic in neonatology, it is important to keep these distinctions in mind. Anyone interested in developing evidence-based reviews on a particular topic should review some of the textbooks on evidence-based practice listed at the end of this chapter. Initially, it is a good idea to collaborate with someone who has experience in systematic review and metaanalysis. The process consists of the following steps: (1) identifying the relevant clinical question (e.g., management of bronchopulmonary dysplasia); (2) narrowing the question to a focus that enables one to determine whether a given article in the primary literature answers it (e.g., does prophylactic highfrequency ventilation have positive or negative effects on acute and chronic morbidity—pulmonary and otherwise?); (3) extensively searching the primary literature (frequently in collaboration with a librarian with expertise searching the biomedical literature) and retrieving the articles; (4) critically, formally, and systematically reviewing each article for inclusion, validity, utility, and applicability; and (5) formally summarizing the results of the preceding process, including conclusions valid throughout the body of included primary literature. ONLINE INFORMATION RETRIEVAL Because of the rapid growth of biomedical information and because it changes over time, investigators in informatics and publishers believe that print media will soon CHAPTER 2 Evaluation of Therapeutic Recommendations, Database Management, and Information Retrieval no longer be used for sharing and distributing biomedical information (Smith, 1996; Weatherdall, 1995). Economic realities will dictate, however, that quality information will generally come at a price. The medical digital library at the University of Washington (www.healthlinks.washington.e du, under Care Provider) serves to illustrate the current state of the art. The information available is a combination of locally developed material (e.g., University of Washington faculty writing practice guidelines), material developed by institutions elsewhere (faculty elsewhere writing such guidelines), material developed by organizations (e.g., the neonatal Cochrane Collaboration), and electronic forms of journals, textbooks, and evidence-based medicine databases. The last (journals, textbooks, databases) generally are not free, and the University of Washington pays a subscription fee (termed a site license) to provide access to these resources for their faculty, staff, and students. Libraries will likely remain the primary source of information, but will shift their attention from paper to electronic records. Health sciences libraries can provide invaluable training in the efficient use of online medical resources, and most offer training sessions and consultation. A number of online resources are valuable for neonatologists. For accessing the primary literature, the most valuable resource is the National Library of Medicine’s database of the published medical literature accessible (PubMed; www.ncbi.nlm.nih.gov/entrez) along with many other databases accessible from the Health Information Web site (www.nlm.nih.gov/hinfo.html). The PubMed system is continually being enhanced; therefore it is useful to review the help documentation online and regularly check the New/ Noteworthy section to see what has changed. One of the most powerful yet underused tools is the Find Related Link that appears next to each article listed on PubMed. This link locates articles that are related to the one selected (Liue and Altman, 1998). The PubMed system applies a powerful statistical algorithm with complex weightings to the article selected to each word in the title, to each author, to each major and minor keyword (Medical Subject Heading terms), and to each word in the abstract and then finds statistically similar articles in the database. In general, this system outperforms novice to advanced health care providers performing a complex search, and it begins to approach the accuracy of an experienced medical librarian. Another powerful search tool within PubMed is the Clinical Query (www.ncbi.nlm. nih.gov/corehtml/query/static/clinical.shtml). This tool facilitates searches for papers by clinical study category (e.g., etiology, diagnosis, therapy, prognosis), focuses on systematic reviews, and performs medical genetics searches—the last of which is useful in the context of neonatology. All the major pediatric journals are available online either through a package at local hospital libraries or by subscription, instead of or in addition to a print subscription. The value of having an electronic subscription to a journal is that it provides access to current and past issues. The best free online source of information on evidence-based practice is the Cochrane Neonatal Review. Subscriptions to the full Cochrane database can be purchased online as well. Given the growing role of genetics in health care, and in particular the importance of genetic diseases in infants, there are two notable genetics databases that are available 17 free of charge online (in addition to the Medical Genetics Searches mentioned previously). The first is Online Mendelian Inheritance in Man (www.ncbi.nlm.nih.gov/omim). This database, which is a catalog of human genes and genetic disorders, is an online version of the textbook by the same name. It is a diachronic collection of information on genetic disorders, meaning that each disease entry chronologically cites and summarizes key papers in the field. The second is the GeneTests database (www.genetests.org), which is a directory of genetic testing (what testing is available on a clinical and research basis, where, and how one sends a specimen) and a user’s manual (how to apply genetic testing). The user’s manual section consists of entries for a growing number of diseases or clinical phenotypes of particular importance. Entries are written by experts, peer reviewed both internally and externally, subjected to a formal process similar to a systematic review, and updated regularly online. As of the winter of 2011, the GeneTests database includes GeneReviews (user’s manual entries) for 527 diseases, and the directory includes 1189 genetics clinical, 595 laboratories, and information on testing for 2270 diseases (2005 clinically and 265 on a research basis). An excellent site that maintains links to the majority of locally developed, neonatology-specific content around the country is the site maintained by Duncan (2010). In addition to a database of links to clinical resources around the country, the site also has a job listing and a database of the literature on computer applications in medicine. The clinician must realize that, unlike journals, textbooks, and guidelines, the material on the World Wide Web (whether accessed from Duncan’s site or based on a search) is not necessarily subject to any editorial or other oversight as to what is published; therefore, as stated by Silberg et al (1997), “caveat lector” (reader beware). A number of articles and Web sites address criteria for assessing the validity and reliability of material on a Web site. (Health on the Net Foundation, www.hon.ch Mitretek Systems). With caution in mind, a search on the entire World Wide Web using a sophisticated search engine (e.g., Google, Google Scholar) can yield valuable information, though search results typically include a lower proportion of quality resources compared with curated resources. Google also has a sophisticated statistical algorithm that allows a user to find similar Web pages after identifying a particular one of interest. SUGGESTED READINGS Cochrane Neonatal Collaborative Review Group: Cochrane neonatal review. Available from www.nichd.nih.gov/cochrane. Jha AK, DeRoches CM, Campbell EG, et al: Use of electronic health records in U.S. hospitals, NEJM 360:1628-1638, 2009. Montori VM, Guyatt GH: Progress in evidence-based medicine, JAMA 300:1814-1816, 2008. Norris T, Fuller SL, Goldberg HI, et al: Informatics in primary care: strategies in information management for the healthcare provider, New York, 2002, SpringerVerlag. Pratt W, Reddy MC, McDonald DW, et al: Incorporating ideas from computer supported cooperative work, J Biomed Inform 37:128-137, 2004. Shortliffe EH, Cimino JJ, editors: Biomedical informatics: computer applications in health care and biomedicine, ed 3, New York, 2006, Springer Science+Business Media. Straus SE, Richardson WS, Glasziou P, et al: Evidence-based medicine: how to practice and teach EBM, ed 3, London, 2005, Churchill Livingstone. Complete references used in this text can be found online at www.expertconsult.com C H A P T E R 3 Ethics, Data, and Policy in Newborn Intensive Care William L. Meadow and John D. Lantos PHILOSOPHY Ethics in the neonatal intensive care unit (NICU), as in all clinical contexts, starts with the traditional triangular framework of autonomy (do what the patient, or in this case the parent, thinks is right), paternalism (do what the doctor thinks is right), and beneficence and nonmaleficence (do the right thing). These concepts, independent of context or data, are timeless. The problem with timeless concepts, of course, comes in knowing what to do in real time. What exactly is the right thing? What facts should be brought to bear in the decision? What weight should each fact be given? And whose opinion should count in the end? Nonetheless, some applications of traditional ethical concepts in the NICU are already universally adopted (e.g., avoid futility, do not torture, and intervene when the data provide compelling evidence to do so). Unfortunately, much of NICU care falls between not resuscitating 21-week births, obligatory support of 28-week births, and not performing cardiopulmonary resuscitation on infants with lethal anomalies. The traditional ethical solution to medical dilemmas is to ground concerns in context, take data into account, and be sympathetic to patient preferences when the balance of benefits and burdens is not clear. The precise problem in the NICU is that the burdens are real, immediate, long term, and significant—months of painful procedures such as intubation, ventilation, intravenous catheterization, and any permanent sequelae that might ensue—whereas the benefits of NICU interventions are distant, statistical, and unpredictable. Moreover, NICU success is often viewed as “all or none”—that is, in most of the NICU follow-up literature a Bayley mental developmental index MDI or psychomotor developmental index PDI greater than 70 is classified as normal, whereas an MDI or PDI less than 70 is classified as an adverse outcome. Faced with a difficult case, it is rare that simply applying principles will help to devise a solution. Difficult cases are usually ones in which the principles themselves are in conflict, or their application to the case is ambiguous. DATA What kind of information would parents, physicians, or judges want to know about the NICU? The essential truth at the intersection of NICU epidemiology and ethics is that survival depends sharply on gestational age (GA), within relatively precise boundaries. In the United States, as in virtually all the industrialized world, infants born after 27 weeks’ gestation have, from any ethical perspective, no increased mortality over infants born at term. Consequently, for these infants, the ethical principle of best interests requires their resuscitation, in the same way that sick children born at term deserve resuscitation. 18 Conversely, for infants born before 22 weeks’ gestation, survival is essentially zero. Consequently, these infants and their parents deserve our compassion, but not our interventions, on the ethical grounds of strict futility. In between, spanning roughly one gestational month, from 23 to 26 weeks, we will require not just data, but interpretation. First, there is the intriguing finding that GA-specific mortality for infants resuscitated in this gestational range has more or less reached a plateau. Within 10 to 20 years ago, the “border of viability” was steadily decreasing, thus improving outcomes for infants in the gray zone. Not so much, recently, and there is little reason to think that things will change for infants born in this gestational range in the foreseeable future. Nonetheless, epidemiologic progress has been made. Tyson et al (2008), using the vast data base of the National Institute of Child Health and Human Development network, attempted to go “beyond gestational age” and quantify additional risk factors for both mortality and neurologic morbidity in infants born on the cusp of viability. The analysis revealed that singleton status, appropriate in utero growth, antenatal steroids, and female gender all improve the likelihood of survival and intact neurologic outcome, independent of GA. Tyson’s algorithm, using information available at the time of birth, significantly improved predictive value and sensitivity of prognostication over GA alone. However, two problems remain. First, for many infants the predictive value of the Tyson algorithm is still not very good—that is, many of the lower-risk patients will still die, and many of the higher-risk patients will survive. Second, the Tyson algorithm, like GA, ignores a potentially important feature of NICU care—time. The algorithm stops accumulating data at the time of birth, and it does not account for prognostic features that might become available as the infant’s course unfolds in the NICU. In theory, making decisions over the course of time offers two advantages over prognostication in the delivery room. First, parents often appreciate the opportunity to get to know their baby as an individual, as opposed to evaluating anonymous population-based prognostications at the time of birth. Second, there is valuable information to learn while the baby is in the NICU. Two time-sensitive prognostic features have been evaluated in the context of infants born at the border of viability—serial illness severity algorithms (Score for Neonatal Acute Physiology [SNAP] scores) and serial that intuitions that the patient would “die before NICU discharge” (Meadow et al, 2008). Unfortunately, although SNAP scores on the first day of life have good prognostic power for death or survival, their power diminishes over time. Intriguingly, serial intuitions that an individual baby will die before discharge—offered by medical caretakers for CHAPTER 3 Ethics, Data, and Policy in Newborn Intensive Care patients who require mechanical ventilation and for whom there is an ethical alternative to continued ventilation, namely extubation and palliative care—are remarkably accurate in predicting a combined outcome of either death or survival with neurologic impairment (MDI or PDI <70). Children with abnormal results from a cranial ultrasound examination and corroborated predictions of death have a less than 5% chance of surviving with both MDI and PDI greater than 70 at 2 years, independent of their gestational age. The predictive power of these data, acquired over time during an individual infant’s NICU course, though not perfect, is greater than any algorithm available at the time of birth. What do prospective parents or medical caretakers consider when they are asked to decide whether or not to resuscitate their micro-premie? A fascinating insight has been offered by Janvier et al, (2008), who have done extensive surveys comparing responses to requests for resuscitation of sick micro-premies with resuscitation of comparably sick patients at other ages (from term infants to 80-year-olds). Consistently, it appears that micropremies are devalued—that is, for comparable likelihood of survival and comparable likelihood of neurologic morbidity in survivors, more people would let a micro-premie die first, or at least offer to resuscitate them last. There is no theory to account for these findings. Indeed, for male infants born at 23 weeks’ gestation, the likelihood of intact survival (i.e., neither death in the NICU nor permanent neurologic morbidity) is less than 10%. Given that the likelihood of burdensome therapy is 100%, how can resuscitation be justified? Intriguingly, combining the outcomes of death in the NICU with abnormal neurologic function in later life may not reflect the emotional reality of many NICU parents. For these parents, death in the NICU is not necessarily the worst outcome, because acknowledging that the baby was too small might be better consolation than not ever giving the baby a chance at life. If trying and failing is seen as a positive process, then the long-term burden of NICU care for parents should perhaps be calculated not as a function of all live births (the current practice), but as a function of infants who survive to discharge. Numerous studies analyzing various populations in several countries have converged on the same surprising observation: the incidence of neurologic morbidity in NICU survivors is not very different when comparing infants at 23, 24, 25, and 26 to 27 weeks’ gestation. The essential epidemiologic difference for infants born in this gestational range appears to be whether the baby will survive at all. Once the baby leaves the NICU, the risk of severe morbidity is largely the same; this is true in singlecenter and multicenter studies, in the United Kingdom, Canada, Europe, and the United States (Johnson et al, 2009; Tyson et al, 2008). Paradoxically, if avoiding survival with permanent crippling neurologic injury is the driving force behind resuscitation decisions, it appears that we should not be worrying about 23- and 24-­weekers–rather we should not be resuscitating 26- or 27-weekers, since many more of them will survive, and survive with disability. But that seems very odd. Finally, there is epidemiologic difficulty in assigning value to morbidity in surviving infants in the NICU. Verrips et al (2008) have attempted to assess the effects of permanent 19 residual disability for NICU survivors and their immediate families; they have demonstrated consistently that children with disabilities and their parents place a much higher value on their lives, and the quality of those lives, than do either physicians or NICU nurses. The vast majority of infants who survive the NICU, even those with significant permanent neurologic compromise, have “lives worth living,” as judged by the people most affected by those lives. GA-specific mortality seems to preclude resuscitation for infants born before 22 weeks’ gestation and require resuscitation for infants born after 27 weeks’ gestation. In between, the outcomes are murky, prognostic indices are imperfect, and sociologic analyses of human behaviors (of parents and physicians) appear inadequate to develop any uniform approach that is satisfactory. POLICY Neonatologists in the United States are sued for medical negligence approximately once every 10 years, and the average U.S. neonatologist will be sued more than once during his or her career (Meadow et al, 1997). Other than case-specific failures or oversights, are there any overarching themes that have arisen from medical malpractice cases in neonatology? In the United States most malpractice allegations are state based, as opposed to federal cases. Consequently, the decisions of lower state courts, or even state supreme courts, are not binding in any other state, but they can be informative. The first important legal case in neonatology in the United States was Miller v HCA (2003). In 1990, Mrs. Miller came to the Hospital Corporation of America (HCA) in Texas in labor at 23 weeks’ gestation. The fetus was estimated to weigh 500 to 600 g. No baby born that size had ever survived at that hospital. Mrs. Miller, her husband, and the attending physicians agreed that the baby was previable and that no intervention was indicated. The baby was born, but a different physician performed resuscitation, and the infant survived with brain damage. As a result, the Millers sued the hospital for a breach of informed consent, and they were awarded $50 million by a trial jury. The case wound its way to the Texas Supreme Court, which dismissed the verdict and articulated an “emergency exception” for physicians—that is, if a Texas physician finds himself or herself in the emergency position of needing to resuscitate a patient to prevent immediate death, the physician can try to perform resuscitation without being obligated to obtain consent from anyone. Whether it would be acceptable for a physician not to perform resuscitation in an emergency was left unarticulated by the Texas court. In Wisconsin, the case of Montalvo v Borkovec (2002) took the legal obligations of neonatologists and parents to a different place. The case derived from the resuscitation of a male infant born between 234⁄7 and 242⁄7 weeks’ gestation, weighing 679 g. The parents claimed a violation of informed consent, arguing that the decision to use “extraordinary measures” should have been relegated to the parents. The Wisconsin Appellate Court disagreed, holding that “in the absence of a persistent vegetative state, the right of a parent to withhold life-sustaining treatment from a child does not exist.” Because virtually no infant is 20 PART I Overview born in a persistent vegetative state, this decision would apparently eliminate the ethical possibility in Wisconsin of a “gray zone” of parental discretion. No other jurisdiction in the United States has adopted this position. The Wisconsin Appellate Court, like the Texas Supreme Court, was silent on whether physicians have discretion not to resuscitate. However, in Texas and Wisconsin, physicians are apparently not liable if they choose to do so. A number of other state courts have addressed issues of treatment or nontreatment. In general, the courts are permissive of physicians who resuscitate infants. If courts are asked to sanction decisions to allow infants to die, most will do so only if there is consensus among physicians and parents, and occasionally ethics committees. Courts are not eager to punish physicians who treat infants over parental objections or to empower physicians to stop treatment when parents want it to continue. BABY DOE REGULATIONS Of course, state-by-state civil malpractice cases are not the only administrative means for redrawing the interface between neonatal medicine and the law. The federal government has contributed as well. The Baby Doe Regulations were one of the first attempts to codify and impose a federal vision of appropriate ethical behavior in the NICU. In 1982, a baby with Down syndrome and esophageal atresia was born in Bloomington, Indiana. At that time, the standard of care for babies with Down syndrome was shifting. A decade or two earlier, most babies with Down syndrome who survived infancy were institutionalized. If babies needed surgery for an anomaly like Baby Doe’s atresia, the parents were given the option. Approximately half chose surgery and half chose palliative care (Shaw, 1973). Similarly, half of pediatricians thought that palliative care, rather than surgery, was the better option (Todres et al, 1977). By the early 1980s, this practice had started to change. More parents opted for active intervention to save their babies (Shepperdson, 1983), and more physicians believed that withholding treatment was medical neglect (Todres et al, 1988). Baby Doe was born into this shifting cultural milieu. Baby Doe’s parents refused to consent to surgery and chose palliative care instead. The pediatrician alerted the state child protection agency, which investigated the case. At a court hearing, the parents claimed that they were following the advice of their obstetrician and not their pediatrician. The court found that parents only had to follow the advice of a licensed physician and that, since they were doing so, they were not neglectful. The court did not take protective custody. The doctor and hospital appealed. The Indiana Supreme Court refused to hear the appeal, and the baby died after 8 days (Lantos, 1987). These facts became publicly known and led to a national controversy that eventually reached the Oval Office. President Reagan demanded federal action. It was difficult to decide which action should be taken, because the federal government has no jurisdiction over child abuse and neglect; it is the domain of the states. The federal government oversees civil rights enforcement, however, and the Reagan administration devised a legal strategy that defined not treating babies with Down syndrome or other congenital anomalies as discrimination against people with disabilities, rather than medical neglect. This strategy gave the federal government a justification for oversight. The new regulations were implemented, and bright red signs containing federal hotline telephone numbers were posted in NICUs across the country. These signs proclaimed that withholding treatment on the basis of disability was a federal civil rights violation (Annas, 1984) and that federal investigative squads could review medical records to determine whether discrimination had taken place. These regulations were challenged and eventually struck down by the U.S. Supreme Court (Annas, 1984). A diluted version of the original Baby Doe guidelines was eventually incorporated into the federal Child Abuse and Treatment Act (Annas, 1986; Kopelman, 1988). That law, however, is primarily a funding mechanism to channel federal funds to state child protection agencies; it is not a regulation that can be enforced for physicians or hospitals. The Baby Doe regulations do not exist today and have not existed since 1984; however, they still hold symbolic power. The mention of Baby Doe strikes fear in the hearts of pediatricians who lived through the events, in part because pediatricians had made pediatricians into villains in the societal battle over child protection. There is a certain irony in the controversy over Baby Doe regulations. The original goal—to decrease the range of cases in which withholding treatment of newborns is permissible—did not need federal input. Progress was already being made. Many diseases that used to be considered incompatible with life or that were seen as leading to an unacceptable quality of life were being treated routinely. Metabolic therapies for genetic diseases (phenylketonuria [PKU], hypothyroidism, Gaucher disease) and surgical therapies for organ dysgenesis (congenital heart corrections, congenital diaphragmatic hernia (CDH) repair, extra corporeal membrane oxygenation (ECMO), dialysis) have been forcefully advanced by medical subspecialists in their corresponding fields (e.g., genetics, cardiac surgery, general surgery). With few exceptions (hypoplastic left ventricle being one of the best recognized), these advances have been relatively uncontroversial. Once sufficient data are gathered to demonstrate moderate efficacy, the innovations are widely and rapidly adopted. There is still controversy when treatments enable survival but have a high likelihood, or certainty, that survival will be accompanied by severe neurologic impairment. As a result, two questions must be asked. First, how severe will the neurologic impairment be? Second, what is the likelihood that the child will have the most severe possible impairment? The prognostic spectrum of Down syndrome is broad, but few infants with trisomy 21 are severely impaired. In contrast, almost all infants with trisomy 13 or 18 either die in infancy or are left with profound neurologic impairment. The outcomes for these chromosomal anomalies can be used to define the spectrum within which clinical decisions are made. For babies whose outcomes are likely to be similar to those seen in trisomy 21, it is no longer permissible to withhold life-sustaining treatment. For babies whose outcomes are likely to be similar to babies with trisomy 13, it is permissible to withhold or withdraw life-sustaining treatment and offer palliative care instead. The calculus becomes more complex in conditions associated with a wider range of outcomes, such as extreme prematurity or high myelomenigocele with hydrocephalus. CHAPTER 3 Ethics, Data, and Policy in Newborn Intensive Care The shift in moral standards regarding babies with Down syndrome was not related to technology, but rather sociology. The capacity to repair Arnold–Chiari malformation and duodenal atresia existed long before it was applied to children with myelomeningocele and Down syndrome. What has changed the mood of the country is a growing recognition that disability is as much a social construct as a medical construct, although it is always both and not one or the other. BORN ALIVE INFANT PROTECTION ACT In 2002, the U.S. Congress passed a law called the Born Alive Infant Protection Act (BAIPA). BAIPA, like the discredited Baby Doe regulations, was an attempt to insert federal values into medical deliberations. There are some interesting similarities and distinctions between Baby Doe and BAIPA. The Baby Doe regulations addressed disability, whereas BAIPA applies to abortion. BAIPA declares that “for the purposes of Federal law, the words ‘person’, ‘human being’, ‘child’, and ‘individual’ shall include every infant who is born alive at any stage of development.” As Sayeed (2005) noted, “The agency arguably substitutes a nonprofessional’s presumed sagacious assessment of survivability for reasonable medical judgment.” It is unclear what the implications of this law have been. On the one hand, all infants born alive before BAIPA were also treated as human beings; however, that did not necessarily mean that they received all available life support and resuscitation. After all, patients can have do-not-resuscitate orders or receive palliative care rather than intensive care. Still, the purpose of BAIPA seemed to be less about the treatment of babies and more about restrictions on abortion. Although some authors have expressed concern that the influence of BAIPA may transform neonatal care of infants born at, or even below, the threshold of viability, it appears to have had little measurable effect to date. Partridge et al (2009) surveyed neonatologists in California and found that they were concerned about the implications of BAIPA. They write, “If this legislation were enforced, respondents predicted more aggressive resuscitation potentially increasing risks of disability or delayed death.” There have been no cases to date in which BAIPA has been invoked or in which physicians and hospitals have been found to be in violation of its requirements. FUTURE DIRECTIONS There is no new technology in development that appears likely to affect outcomes significantly. Consequently, for infants who receive resuscitation in the delivery room, birthweight-specific mortality and morbidity are unlikely to change much in the near future. Nonetheless, three developments may change the way we think about newborns, and consequently shift the terrain of neonatal bioethics. HIGH RISK MATERNAL-FETAL MEDICINE CENTERS Many children’s hospitals are now developing high-risk, maternal-fetal medicine centers. The goal of these centers is to identify fetuses at risk—particularly those with 21 congenital anomalies—and to care for those fetuses and their mothers in centers where there is expertise in fetal diagnosis, therapy, and neonatal care. The hope is that such centers will allow more timely, and therefore more effective, intervention for babies with congenital heart disease, congenital diaphragmatic hernia, or other anomalies. The medical effectiveness of fetal centers will depend on two distinct developments. First, on a population basis, these centers will only be as effective as fetal screening and diagnosis. The existence of these centers will almost certainly create an expectation and a demand for better fetal screening. Such screening is likely to include both better imaging and better screening tests that can be performed on maternal blood; both will lead to earlier diagnosis of fetal anomalies. These diagnoses will create more complex dilemmas for perinatologists and parents who will need to decide, in any particular case, whether to terminate the pregnancy, offer fetal therapy, or offer either palliative care or interventions after birth. Ironically, better fetal diagnosis may increase the likelihood of pregnancy termination, even when postnatal treatment is possible, such as in hypoplastic left heart syndrome. Second, the effectiveness of fetal centers will depend on the effectiveness of fetal interventions. Perhaps surprisingly, other than in utero transfusion for Rhesus disease or vascular ablation for twin-twin transfusion syndromes— neither of which are particularly new and neither of which is performed by pediatric surgeons or pediatricians—there is little evidence that any fetal intervention has had any effect on any neonatal outcome. This lack of demonstrated effectiveness has, thus far, not suppressed the proliferation of fetal intervention centers. There may be other factors, including institutional prestige, finances, and recruitment of “desirable” patients. EXPANDED NEWBORN SCREENING In recent years, the number of diseases and conditions that can be diagnosed through newborn screening has expanded dramatically. Such screening is under the purview of states, rather than the federal government, and there is wide variation in the number of tests that are performed. In 1995 the average number of tests per state was five (range: zero to eight disorders). Between 1995 and 2005 most states added tests, so that the average number of screening tests done by 2005 was 24 (Tarini et al, 2006). The expansion of newborn screening raises three problems. First, even the most accurate test has false positives. For rare conditions, the percentage of positive tests that are false positives is increased. Thus, the more rare conditions that are added to a newborn screening panel, the more false positives there will be. False positives are associated with considerable parental anxiety and can lead to potentially dangerous and unnecessary diagnostic procedures or treatments. Second, expanded newborn screening costs money. Interestingly, the tests themselves are astoundingly inexpensive, which is why policy makers are tempted to add more to the panels. However, the follow-up counseling and testing after positive tests are expensive, and without such followup the screening programs will not work. The Centers for Disease Control and Prevention has recently expressed concern about these costs (Centers for Disease Control 22 PART I Overview and Prevention, 2008). Finally, there is the potential for discrimination against patients for whom documented heterozygous carrier status conveys no recognized medical infirmity, but social or psychological stigma may be real. There is little funding available to assist or counsel these patients. FINANCIAL CONSTRAINTS The American Academy of Pediatrics guidelines on neonatal resuscitation suggest that resuscitation should be obligatory at 25 weeks’ gestation or greater, optional at 24 weeks’ gestation, and unusual at 23 weeks’ gestation. These recommendations are thought to reflect the best understanding of both the ethical discussion and the epidemiologic facts surrounding NICU outcomes. The recommendations purport to reflect the traditional paradigm that data drive policy. However, particularly in the context of NICU survival for infants at 23 to 25 weeks’ gestation, there is good evidence that causation can work in the reverse direction—that is, policy drives data. In the Netherlands, Canada, and some parts of Oregon, survival at 25 weeks’ gestation is comparable and comparably good; more than 50% of infants born at 25 weeks’ gestation will survive to discharge. However, in the Netherlands, virtually no infant survives birth at 24 weeks’ gestation, whereas in Canada and Oregon the survival rate is 40%. In addition, in Oregon and some parts of Canada, survival at 23 weeks’ gestation is close to zero, whereas in other parts of Canada and the United States the survival rate is 20% or greater. Why? Certainly not because the Dutch, Canadians, or Oregonians have forgotten how to resuscitate small infants. Rather, they have chosen not to. Why not? Perhaps it is non-maleficence–the fear that survival with permanent neurologic morbidity may be cruel to the child, the family, or society at large. However, the incidence of neurologic morbidity in survivors is not different when comparing infants at 23, 24, 25, and 26 to 27 weeks’ gestation. If the fear of a permanent crippling neurologic injury is the driving force, we should not be resuscitating 26 or 27 weekers, since many more of them will survive, and survive with disability. But that seems odd. The approach in the Netherlands is consistent; there is a limited budget and a communitarian ethic. There is a certain rationale behind spending money on all pregnant women, instead of 1% of micro-premies. The United States appears ambivalent–we value individuals over community, are fascinated with high-technology, and claim to prize our children. On the other hand, we will not spend money to prevent unwanted teen pregnancy or to provide visiting nurses for new mothers. Finally the concept of generational conflict must be considered. We appear quite comfortable calling delivery-room resuscitation of 24 weekers “optional,” based on ­gestational age alone. It is difficult to imagine the AMA recommending that resuscitation for 85-year-olds who come to the emergency department is “optional,” based on age alone. NICU care is often criticized as “too expensive.” We ask, “Compared to what?” Fewer than 5% of infants will be in the NICU for more than a short stay. The vast majority of these patients will survive. Only 24,000 infants of 4 million births die each year in the United States, and half of these will die within fewer than 7 days. In contrast, nearly 3 million adults will die in the United States each year and one third of these will have been admitted to an medical intensive care unit (MICU) in the 6 months before dying. MICU costs outpace NICU costs by at least 100 to 1. SUMMARY Ethical philosophy is a place to start, not a place to ­finish. Data are relatively easy to acquire and agree on. Policy is intriguingly insensitive to data, but that may reflect social and political realities that exist beyond the NICU— perceptions of disability, abortion politics, individual versus communitarian emphasis, fascination with technology, discrimination, publicity, financial constraints—so that an ethical course of action in one country, one city, or one family might seem perverse elsewhere. SUGGESTED READINGS Annas GJ: The Baby Doe regulations: governmental intervention in neonatal rescue medicine, Am J Public Health 74:618-620, 1984. Born-Alive Infants Protection Act of 2001: Report together with additional and dissenting views of the House Committee on the Judiciary, 107th Congress, 1st Session, August 2, 2001. 1-38, 3. (Purpose and Summary). Centers for Disease Control and Prevention: Impact of expanded newborn screening: United States, 2006, MMWR Morb Mortal Wkly Rep 57:1012-1015, 2008. Janvier A, Leblanc I, Barrington KJ: The best-interest standard is not applied for neonatal resuscitation decisions, Pediatrics 121:963-969, 2008. Johnson S, Fawke J, Hennessy E, et al: Neurodevelopmental disability through 11 years of age in children born before 26 weeks of gestation, Pediatrics 124:E249-E257. Lantos J: Baby Doe five years later: implications for child health, N Engl J Med 317:444-447, 1987. Meadow W, Lagatta J, Andrews B, et al: Just in time: ethical implications of serial predictions of death and morbidity for ventilated premature infants, Pediatrics 121:732-740, 2008. Miller v HCA, Inc, 118 S.W. 3d 758, 771 (Texas 2003). Montalvo v Borkovec, 647 N.W. 2d 413 (Wis. App. 2002). Shepperdson B: Abortion and euthanasia of Down’s syndrome children: the parents’ view, J Med Ethics 9:152-157, 1983. Todres ID, Guillemin J, Grodin MA, et al: Life-saving therapy for newborns: a questionnaire survey in the state of Massachusetts, Pediatrics 81:643-649, 1988. Tyson JE, Parikh NA, Langer J, et al: National Institute of Child Health and Human Development Neonatal Research Network. Intensive care for extreme prematurity: moving beyond gestational age, N Engl J Med 358:1672-1681, 2008. Complete references used in this text can be found online at www.expertconsult.com C H A P T E R 4 Global Neonatal Health Linda L. Wright CHILD MORTALITY In 2000, the 192 United Nations member states and many international partners adopted the Millennium Development Goals (MDGs) in response to morbidity and mortality rates in the developing world that were alarmingly high. These eight international development goals were to eradicate extreme hunger and poverty; achieve universal primary education; promote gender equality and empower women; reduce child mortality; improve maternal health; combat HIV/AIDS, malaria, and other diseases; ensure environmental sustainability; and develop a global partnership for development. The agreement included 18 specific targets and 48 technical indicators to measure progress toward the MDGs between 1990 and 2015 (United Nations General Assembly, 2001). Although significant progress has been made toward achieving many of the MDGs, progress has been uneven, with huge disparities across the goals and among countries. This disparity is especially true for MDG 5 (i.e., improving indicators of maternal health, including maternal mortality during pregnancy or the 42 days following the end of pregnancy, per 100,000 deliveries) and MDG 4 (i.e., reducing child mortality, expressed as deaths before 5 years per 1000 live births). Regions and countries that have the highest maternal mortality rates also have the highest child mortality rates (Table 4-1). Twenty percent of all deaths in the world are child deaths (Save the Children, 2007; United Nations Children’s Fund, 2007), and greater than 99% of deaths occurring in children aged 5 years or younger are in the developing world. Of the 136 million babies born in 2007, an estimated 9.7 million died before the age of 5 years (approximately 26,000 per day) (Save the Children, 2007; United Nations Children’s Fund, 2007). This staggering number equals approximately half of all U.S. children younger than 5 years (Save the Children, 2007; U.S. Census Bureau’s American Community Survey, 2007). The latest estimates suggest that 8.8 million children died in 2008 (United Nations Children’s Fund, 2008b). The rate of decline in mortality in children younger than 5 years is grossly insufficient to meet the MDG 4 goal by 2015, particularly in Sub-Saharan Africa and South Asia (United Nations Children’s Fund, 2009). At the current rate, the target of fewer than 5 million annual child deaths will not be met until 2045. To meet the goal of fewer than 5 million child deaths in 2015, deaths in children younger than 5 years must be cut in half between 2008 and 2015 (Table 4-2) (United Nations Children’s Fund, 2007, 2009a). The regions with the highest numbers of child deaths are Sub-Saharan Africa (which has high fertility rates and the highest child mortality rates [144 deaths per 1000 live births], and 4.8 million children [1 in 7] dies before the age of 5 years) and South Asia (3.1 million deaths [1 in 13] before the age of 5 years) (United Nations Children’s Fund, 2007). Sub-Saharan Africa accounts for 51% of all deaths among children younger than 5 years, followed by Asia with 42% (You et al, 2009). In 2008, 75% of deaths in children younger than 5 years occurred in only 18 countries, and 40% occurred in only three countries: India, Nigeria, and the Democratic Republic of the Congo. Of the 34 countries with mortality rates exceeding 100 per 1000 live births in 2008, all were in Sub-Saharan Africa, except for Afghanistan (You et al, 2009). Equally troubling, only 10 of the 67 countries with high mortality rates (at least 40 per 1000 live births) were on track to meet MDG 4 before the 2009 economic crisis (You et al, 2009). However, a number of relatively poor countries with low gross national incomes have made considerable progress in improving survival, including Malawi, Tanzania, Madagascar, Nepal, ­Bangladesh (Save the Children, 2007), Eretria, the Lao People’s Democratic Republic, Mongolia, and Bolivia (Figure 4-1) (You et al, 2009). More than 70% of deaths in children younger than 5 years are caused by newborn problems, pneumonia, and diarrhea. Pneumonia kills more children than any other illness— more than HIV, malaria, and measles combined (Figure 4-2). Pneumonia results in death for more than 2 million children younger than 5 years each year, or approximately 20% of child deaths worldwide. More than 95% of all new pneumonia cases, representing an estimated 150 million episodes of pneumonia annually, occur in children younger than 5 years in developing countries. Sub-­Saharan Africa and South Asia together have more than half the total number of pneumonia cases. Effective prevention strategies include immunization against measles, whooping cough, Haemophilus influenzae type b, and Streptococcus pneumoniae; exclusive breastfeeding and improved nutrition or low birthweight; zinc supplementation; reducing indoor air pollution; and prevention and management of HIV infection (Qazi et al, 2008); however, the protection afforded by immunizations will not prevent neonatal pneumonia. Universal use of simple standardized management guidelines for identifying and treating pneumonia in communities and primary health care centers, with the World Health Organization (WHO) Integrated Management of Childhood Illness (Niessen et al, 2009), may reduce the child pneumonia fatality rate (Niessen et al, 2009). Sazawal and Black (1992) suggested that community-based acute respiratory infection case management might reduce mortality by more than 20% in children younger than 4 years. Failing prevention, prompt diagnosis and treatment are necessary to improve pneumonia mortality and morbidity; however, prompt diagnosis and effective treatment of pneumonia and hypoxemia are often not available. Radiology, laboratory tests, and pulse oximetry, which can predict response to antibiotic therapy in cases of severe pneumonia (Fu et al, 2006), are not available in most first-level (i.e., rural) 23 24 PART I Overview TABLE 4-1 Countries With the Highest Numbers of Child Deaths Also Have High Rates of Maternal Death Country Ranking for Number of Child Deaths Number of Child Deaths Ranking for Number of Maternal Deaths Number of Maternal Deaths India 1 1,919,000 1 136,000 Nigeria 2 1,043,000 2 37,000 DR Congo 3 589,000 4 24,000 Ethiopia 4 509,000 4 24,000 Pakistan 5 473,000 3 26,000 China 6 467,000 9 11,000 Afghanistan 7 370,000 7 20,000 Bangladesh 8 274,000 8 16,000 Uganda 9 200,000 12 10,000 Angola 10 199,000 9 11,000 6,043,000 child deaths Approximately 60 percent of global total 315,000 maternal deaths Approximately 60 percent of global total Sources: Child deaths UNICEF. State of the World’s Children 2007, Table 1; Maternal deaths: World Health Organization, United Nations Children’s Fund and United Nations Population Fund, Maternal Mortality in 2000: Estimates Developed by WHO, UNICEF and UNFPA. From Save the Children: State of the world’s mothers, 2007: saving the lives of children under 5. TABLE 4-2 Global Progress in Reducing Child Mortality is Insufficient to Reach Millennium Development Goal 4* Average annual rate of reduction (AARR) in the under 5 mortality rate (U5MR) observed for 1990–2006 and required during 2007–2015 in order to reach MDG 4 U5MR AARR No. of Deaths per 1,000 Live Births 1990 2006 Observed % Required % 1990–2006 2007–2015 Progress Toward the MDG Target Sub-Saharan Africa 187 160 1.0 10.5 Insufficient progress Eastern and Southern Africa 165 131 1.4 9.6 Insufficient progress West and Central Africa 208 186 0.7 11.0 79 46 3.4 6.2 Middle East and North Africa South Asia No progress Insufficient progress 123 83 2.5 7.8 Insufficient progress East Asia and Pacific 55 29 4.0 5.1 On track Latin America and Caribbean 55 27 4.4 4.3 On track CEE/CIS 53 27 4.2 4.7 On track Industrialized countries/­ territories 10 6 3.2 6.6 On track 103 79 1.7 9.3 Insufficient progress 93 72 1.6 9.4 Insufficient progress Developing countries/­ territorries World United Nations Children’s Fund: The state of the world’s children 2008: child survival, 2007, New York. *Progress towards MDG 4, with countries classified according to the following thresholds: On track: U5MR is less than 40, or U5MR is 40 or more and the average annual rate of reduction (AARR) in under 5 mortality rate observed from 1990 to 2006 is 4.0% or more. Insufficient progress: U5MR is 40 or more and AARR observed for the 1990-2006 period is between 1.0% and 3.9%. No progress: U5MR is 40 or more and AARR observed for 1990-2006 is less than 1.0%. Source: UNICEF estimates based on the work of the interagency Child Mortality Estimation Group. hospitals, despite high mortality rates in children who have hypoxemia or HIV. Randomized controlled trials of parenteral antibiotic treatment in hospitals compared with home-based treatment have demonstrated the safety and efficacy of treating pneumonia with oral antibiotics outside of a hospital setting in older children. New evidence regarding home treatment of severe pneumonia is changing concepts about the need for hospitalization. The first randomized trial to compare outcomes of hospital treatment of severe pneumonia, without underlying complications, with home-based oral antibiotics in Pakistan demonstrated that home-based antibiotics are safe and effective. Of 2037 children with severe pneumonia aged 3 to 59 months, randomized to either parenteral ampicillin for 48 hours followed by 3 days of oral ampicillin or home-based oral amoxicillin for 5 days, there were equal numbers of failures in the hospitalized group (8.6%) and in the ambulatory group (7.5%) by day 6. Just 0.2% children died within 14 days of enrollment and none of the deaths were considered to be associated with treatment CHAPTER 4 Global Neonatal Health 25 POVERTY DOES NOT HAVE TO BE A DEATH SENTENCE FOR CHILDREN UNDER 5 Above-average GNI per capita Above-average reduction in under-5 mortality Percent reduction in under-5 mortality (1990–2005) Below-average GNI per capita Above-average reduction in under-5 mortality Bangladesh Nepal Malawi Madagascar Tanzania Cameroon Cambodia Côte d’Ivoire Below-average GNI per capita Below-average reduction in under-5 mortality South Africa Equatorial Guinea Above-average GNI per capita Below-average reduction in under-5 mortality GNI per capita (purchasing power parity) Priority countries for under-5 child survival A for effort Failing to save children FIGURE 4-1 Poverty does not have to be a death sentence for children under 5. (From Save the Children: State of the world’s mothers 2007: saving the lives of children under 5, 2007.) WHY CHILDREN DIE Newborn disorders, pneumonia, and diarrhea account for 73% of all deaths among children under 5 Pneumonia 19% Newborn causes 37% Diarrhea 17% Preterm 28% Sepsis/pneumonia 26% Asphyxia 23% Congenital 8% Tetanus 7% Diarrhea 3% Other 7% Malaria 8% Measles 4% AIDS 3% Injuries 3% Other 10% Undernutrition is an underlying cause of 53% of deaths among children under age 5 FIGURE 4-2 Why children die. (From Save the Children: State of the world’s mothers 2007: saving the lives of children under 5, 2007.) allocation (Hazir et al, 2008). A small percentage (approximately 2% to 3%) of severely ill children will still require early community detection and transport to a hospital for evaluation of hypoxia, infection, pneumonia, malaria, and parenteral antibiotics with or without oxygen (Mwaniki et al, 2009). The WHO has spearheaded efforts to reduce the global burden of child pneumonia with the Integrated Management of Childhood Illness, a Global Action Plan for Control and Prevention of Pneumonia (GAPP) (Qazi et al, 2008), publication of new comprehensive global, regional, and national disease burden statistics for pneumococcal and hepatitis b disease, the Global Coalition against Childhood Pneumonia, and the first World Pneumonia Day on November 2, 2009 (United Nations Children’s Fund, 2009b). Diarrheal diseases are the second most common cause of child deaths globally. Worldwide, diarrhea accounts for 18% of deaths among children younger than 5 years, or an estimated 1.7 million child deaths every year, and it accounts for 10% to 80% of growth retardation in the first few years of life (Baqui and Ahmed, 2006). Exclusive breastfeeding provided significant protection from diarrheal diseases before the WHO-recommended introduction of complementary feeds at 6 months. Children in poverty are especially prone to diarrheal diseases after the introduction of complementary feeding, because diarrhea is spread by poor hygiene and sanitation facilities; contaminated water, formula, food, or utensils; low rates of vitamin A supplementation; low zinc intake; and limited access to rotavirus immunization. Most diarrhea-related deaths in children are due to dehydration. The WHO recommendations for oral rehydration therapy (ORT) for childhood diarrheal diseases have changed. Research has demonstrated that homemade fluids that contain lower concentrations of sodium and glucose, sucrose, or other carbohydrates (e.g., cereal-based solution) can be as effective as ORT. Current recommendations include increased fluid intake and continued feeding, as well as the use of zinc and low-osmolarity ORT to prevent and treat diarrheal episodes (Fischer-Walker et al, 2009; World Health Organization and United Nations Children’s Fund, 2004). Lower-osmolarity ORT reduces stool output, vomiting, and unscheduled intravenous therapy (Baqui and Ahmed, 2006). In Sub-Saharan Africa there has been little progress in diarrhea prevention and treatment in the last decade— the percentage of children younger than 5 years who received the recommended treatment increased from 32% in 2000 to only 38% in 2008 (United Nations Children’s Fund, 2009). Progress on case management of childhood 26 PART I Overview pneumonia, diarrhea, and malaria will depend on strengthening the integrated community-based prevention and treatment of these pervasive childhood diseases within the health system (United Nations Children’s Fund, 2009). Undernutrition is another major factor in mortality of children younger than 5 years; it is responsible for 7% of the total disease burden in any age group, making it the highest of any risk factor for overall global burden of disease (Black et al, 2008b). Undernutrition is a contributing factor in more than half of infectious disease deaths in children (Save the Children, 2007) and is the underlying cause of more than one third of all deaths among children younger than 5 years (United Nations Children’s Fund, 2008a). The effects of undernutrition reach beyond the individual child. Maternal or child undernutrition is a complex intergenerational problem that includes intrauterine growth restriction, severe wasting, and stunting. Virtually no progress has been made toward overcoming undernutrition in the last 25 years (United Nations Children’s Fund, 2008a). Wasting (weight-for-height Z score less than –2) is associated with acute weight loss. Of the estimated 55 million children younger than 5 years with wasting (10% of all children), more than half are in south central Asia or Sub-Saharan Africa; 19 million children have severe acute malnutrition or wasting (weight-forheight Z score less than –3) and are in need of urgent therapeutic feeding (United Nations Children’s Fund, 2008a). Stunting (height-for-age Z score less than –2), which is more common, indicates chronic restriction of a child’s potential growth. An estimated 178 million children younger than 5 years have stunting—almost one third of children in low- to middle-resource settings. Ninety percent of them (160 million) live in just 36 countries and represent almost half of the children in those countries. India’s 61 million children with stunting represent more than half of all Indian children younger than 5 years and 34% of all children with stunting worldwide (Bhutta et al, 2008a). This urgent problem must be solved, because the period between birth and 24 months old is critical. If children do not grow appropriately before 2 years old, they are more likely to be short as adults, have lower educational achievement and economic productivity, and give birth to smaller infants who repeat the cycle in the next generation. Although the problem of undernutrition has been overshadowed by concerns over obesity, there is no evidence that rapid weight or linear growth in the first 2 years of life increases the risk of chronic disease in adults (fetal origin of adult disease), even in children with poor fetal growth (Victora et al, 2008). A recent review of interventions that affect maternal and child undernutrition suggested that counseling about breastfeeding and supplementation with vitamin A and increased zinc intake have the greatest potential to reduce the burden of child morbidity and mortality (Bhutta et al, 2008a). The promotion of breastfeeding has had an effect on the improved survival of infants and young children, but its effect on stunting has been negligible. Among populations with inadequate food, food supplements are beneficial with or without educational interventions (increased height-for-age Z scores by 0.41; 95% confidence interval [CI], 0.05-0.76) and can reduce stunting and the related burden of disease. In populations with sufficient food, complementary feeding education increased height-forage Z scores by 0.25 (95% CI, 0.01-0.49) compared with controls. Facility management of severe acute malnutrition, according to WHO guidelines, reduced the case fatality rate by 55% (relative risk [RR], 0.45; 95% CI, 0.32-0.62), and observational data suggest that ready-to-use prepared foods in treatment of severe malnutrition may be effective in community settings as well. Recommended micronutrient interventions for children include strategies for vitamin A supplementation, zinc supplements to prevent and treat diarrhea and lower respiratory tract infections, iron supplements in areas where malaria is not endemic, and universal promotion of iodized salt (Bhutta et al, 2008a). A subsequent metaanalysis of neonatal vitamin A supplementation concluded, on the basis of six trials in the developing world, that there was no evidence for a reduced risk of mortality and morbidity during infancy and thus no justification for neonatal vitamin A supplementation as a public health measure to reduce infant mortality and morbidity in developing countries (Gogia and Sachdev, 2009). The efficacy of zinc supplementation in reducing overall mortality in neonates has been questioned as well (Sazawal et al, 2007). Existing interventions designed to improve nutrition and prevent related disease may reduce stunting at 36 months old by as much as 36% and mortality between birth and 36 months old by approximately 25%. However, elimination of stunting will require longterm investment in improved nutrition and interventions to improve early childhood education, maternal education, and women’s economic status (Black et al, 2008a). Growth (length, height, and weight) should be monitored routinely and regularly in view of its importance as a marker for undernutrition and stunting (Victora et al, 2009). The nutrition of the children of India and Sub-Saharan Africa are of greatest concern. India is a concern because it represents 34% of the world’s children with stunting, who often start life with intrauterine growth retardation and as adults are members of families with intergenerational stunting that live on less than $2 per day. In Africa, neonates are born with normal birthweight, but develop stunting and wasting because of poverty and civil unrest. Although exclusive breast­feeding protects the neonate, providing adequate amounts of nourishing food in early infancy is critical to the future of children from these two continents. Among the areas needing further research are assessments of the effectiveness and cost-effectiveness of a national health system’s nutritional interventions on stunting rates and weight gain; long-term effects of maternal nutritional interventions on maternal and child health and cognitive outcome; research on the reversibility of stunting and cognitive impairment in children aged 36 to 60 months; studies of community-based prevention and treatment strategies for severe acute malnutrition; and studies of the effectiveness of various zinc delivery strategies. EPIDEMIOLOGY OF NEONATAL MORTALITY (BEFORE 28 DAYS) Developed countries have experienced significant declines in child mortality in the last 30 years, but only 1% of the world’s neonatal deaths occur in 39 high-income countries whose average neonatal mortality rates are 4 to 6 per 1000 27 CHAPTER 4 Global Neonatal Health TABLE 4-3 Regional or Country Variations in Neonatal Mortality Rates and Numbers of Neonatal Deaths, Showing the Proportion of Deaths in Children Younger than Age 5 Years NMR per 1000 Livebirths (range across countries) Number (%) of Neonatal Deaths (1000s) Percentage of Deaths in ­ hildren Aged Younger than C 5 Years in the Neonatal Period Percentage Change in NMR between 1996 and 2005 Estimates* 4 (1–11) 33 (2–70) 42 (1%) 3956 (99%) 63% 38% –29% –8% Africa 44 (9–70) 1128 (28%) 24% 5% Americas 12 (4–34) 195 (5%) 48% –40% Eastern Mediterranean 40 (4–63) 603 (15%) 40% –9% Europe 11 (2–38) 116 (3%) 49% –18% Southeast Asia 38 (11–43) 1443 (36%) 50% –21% Western Pacific 19 (1–40) 512 (13%) 56% –39% 30 (1–70) 3998 (100%) 38% –16% Income groups High-income countries† Low-income and middleincome ­countries WHO regions Overall From Lawn JE, Cousens S, Zupan J: 4 Million neonatal deaths: when? where? why? Lancet 365:891-900, 2005. *The data inputs cover at least a 5-year period before each set of estimates. Period of change may be assumed to be up to 15 years. †Thirty-nine countries with NMR data of 54 countries with gross national income per person of US$9386.10. live births. In contrast, little progress has been made in reducing maternal and neonatal deaths in the developing world, where the disparity is large and growing (Lawn et al, 2005) and the average neonatal mortality rate (deaths in the first 28 days of life per 1000 live births) in 2000 was 33, with a range of 2 to 70. Together Africa and Southeast Asia account for approximately two thirds of neonatal deaths (Table 4-3) (Lawn et al, 2005). More than half of all maternal and newborn deaths occur at birth or in the first few days after birth, when health coverage is the lowest. Cultural norms, financial constraints, and small fetuses that are considered nonviable also limit the reporting of neonatal deaths. As a consequence, the births of an estimated 51 million children per year go unrecorded in any formal registration system (United Nations Children’s Fund, 2007). These children are often born in slums or rural poverty to very young or older mothers who lack access to education and basic health and reproductive services, and who also live in countries that have experienced recent political unrest (United Nations Children’s Fund, 2007). As a result, fewer than 3% of neonatal deaths occur in countries that have high-coverage vital registration data or recent, reliable data on causes of neonatal death; therefore global analysis is based on estimates (Lawn et al, 2005) derived from statistical modeling. Until the middle to late 1990s, estimates of neonatal deaths were drawn from historical data. More rigorous estimates using demographic and health surveys of newborn deaths at a national level were available in 1995 and 2000 (Lawn et al, 2005; United Nations Children’s Fund, 2008a), which produced more reliable neonatal mortality rates. The lack of reliable data from most high-risk countries, and the complex methods for developing estimates and for estimating uncertainty, makes even these improved neonatal mortality data inherently uncertain (United Nations Children’s Fund, 2008a). As a result, data from different sources are difficult to compare and interpret; this will be especially true when comparing data from before and after 2008, when the Inter-agency Group for Child Mortality Estimation (IGME) incorporated a substantial amount of new data and developed a new method to adjust mortality related to HIV/AIDS (You et al, 2009). The latest available data suggest that 3.7 million deaths (40% of deaths in children younger than 5 years) occur in the first month and that there are almost as many stillbirths per year worldwide (3.3 million) (Stanton et al, 2006; United Nations Children’s Fund, 2008a; World Health Organization, 2006). The first hours and days of a baby’s life are the most critical (Figure 4-3). Every year 2 million babies die on the day they are born (Save the Children, 2007), representing almost 50% of all neonatal deaths, and an astounding 75% of neonatal deaths occur within the first 7 days of life (Murray et al, 2007; Save the Children, 2007). As a result, a neonate is approximately 500-fold more likely to die in the first day of life than at 1 month of age (United Nations Children’s Fund, 2007). Without a major reduction in early (7 days) deaths in high-mortality countries, it will be impossible to meet the MDG 4 (Figure 4-4) (World Health Organization, 2006). MATERNAL RISK FACTORS FOR NEONATAL DEATH Because 40% to 90% of women in low-resource settings deliver their baby in the home without a skilled birth attendant or access to facility care for themselves or their newborn, intrapartum complications put the fetus or neonate at increased risk for death, especially maternal bleeding after the eighth month, hypertensive disorders, obstructed labor, prolonged second-stage of labor or malpresentation, maternal fever or rupture of membranes for longer than 24 hours, multiparity, malaria or syphilis, meconium staining, and maternal HIV (Lawn et al, 2005). 28 PART I Overview Up to 50% of newborn deaths are in the first 24 hours – 2 million deaths per year Daily risk of death (per 1,000 survivors) IN THE FIRST MONTH OF LIFE,THE FIRST DAY AND WEEK ARE MOST RISKY 9 8 7 6 5 4 3 2 1 0 0 7 14 21 Days of life A child’s risk of dying on the first day of life is about 500 times greater than their risk of dying when they are 1 month old. The first hours and days of a baby’s life are critical. 75% of newborn deaths are in the first week – 3 million deaths per year Mortality rate per 1000 live births FIGURE 4-3 In the first month of life, the first day and week pose the highest risk. (Modified from ORC Macro: Measure Demographic Health Survey STAT compiler, 2006. Available at www.measuredhs.com. Accessed April 6, 2006. Based on the analysis by J. Lawn of 38 DHS datasets [2000 to 2004] with 9022 neonatal deaths.) 250 200 150 Late neonatal mortality Early neonatal mortality Global mean under-5 mortality rate Global MDG 4 target 100 50 0 68 28 32 1960 1970 1980 1990 2000 2010 FIGURE 4-4 Progress toward Millennium Development Goal 4 for child survival showing the increasing proportion of deaths in children younger than 5 years. (Modified from Lawn JE, Kerber K, EnweronuLaryea C, et al: Newborn survival in low resource settings: are we delivering? BJOG 116[Suppl 1]:49-59, 2009.) STILLBIRTHS Current estimates attribute at least 3.2 million deaths worldwide to stillbirth—defined as having no signs of life at delivery in a fetus at 28 weeks’ gestation or greater— ranging from 5 per 1000 in wealthy countries to 32 per 1000 in South Asia and sub-Saharan Africa (Lawn et al, 2009c; Stanton et al, 2006); however, stillbirths are largely unrecorded and uncounted. Although differentiating a macerated stillbirth from a recent stillbirth seems simple, differentiating a stillbirth from an early neonatal death is challenging, especially in community settings with weak health systems and limited access to facility care. The task requires training in monitoring fetal heart rate and early signs of life, prompt and appropriate resuscitation, and emergency caesarian section, as required (McClure et al, 2007). Lack of training of birth attendants in resuscitation and careful assessment for signs of life, variability in what is considered the lower limit of viability, gender bias, and the influence of financial or other burdens with an assignment can result in the misclassification of early deaths as stillbirths. Cultural norms that discourage the weighing of dead infants and dictate prompt burial may serve to perpetuate misclassification. Stillbirth is commonly defined as fetal death within the last 12 weeks of pregnancy (i.e., at least 28 weeks’ gestation or weighing 1000 g); however, some community and national standards define stillbirth as fetal death after 22 weeks’ gestation or weighing 500 g. Because most of these deaths occur in settings where women have limited access to skilled birth attendants, it is likely that the current estimates of stillbirth numbers are too low and that a cause of death may not be established in the majority of stillbirths that occur in the developing world. Better data on the number and causes of stillbirth are urgently needed to prioritize action to reduce avoidable stillbirths in high-mortality settings, where rates are at least 10-fold higher than in wealthy countries (Stanton et al, 2006). Stillbirths are included in the mortality tables of the 2009 Global Burden of Disease for the first time (Lawn et al, 2009c). DIRECT CAUSES OF NEONATAL DEATH The proportion of neonatal deaths among children younger than 5 years varies, but the absolute number of neonatal deaths is determined by the size of the population and the neonatal mortality rate (NMR). Africa and Southeast Asia represent approximately two thirds of neonatal deaths, with the largest number of newborn deaths in South Asia and the highest rates of neonatal mortality in Sub-Saharan Africa, similar to child deaths. India contributes 25% of the world’s neonatal deaths (United Nations Children’s Fund, 2007). Three causes of neonatal death are responsible for approximately 75% of neonatal deaths: prematurity (28%), sepsis and pneumonia (26%), and asphyxia (23%) (Save the Children, 2007). Prematurity (birth before 37 weeks’ gestation) and asphyxia (failure to initiate and sustain spontaneous respiration) can result in long-term neurologic injury and cognitive impairment among survivors CHAPTER 4 Global Neonatal Health in countries and families that are least able to provide appropriate care. For every newborn that dies, another 20 suffer birth injury, preterm birth complications, or other neonatal conditions (United Nations Children’s Fund, 2007). Determining cause-specific perinatal and neonatal mortality would allow the development of focused interventions and evaluation of their effects on perinatal and neonatal survival. However, establishing a primary cause of stillbirth and neonatal death in the 50% (2 million) of neonates that die in the first day—often in homes without access to health care systems, skilled birth attendants, or diagnostic techniques—is extremely challenging. To overcome these problems, verbal autopsy techniques have been developed to assign a cause of death in such cases, usually based on an interview with the infant’s mother or caretaker within 6 months of the child’s death. The cause of death is assigned by a panel of physicians or compared to a reference standard from prospective hospital-assigned causes of death. The diagnostic accuracy of verbal autopsy techniques to establish neonatal cause of death is limited because of the lack of standardization of case definitions, cause of death classifications, methods of assigning cause of death, and the limited generalizability of hospital reference standards to infants who die in the community (Edmond et al, 2008; World Health Organization, Development of Verbal Autopsy Standards). In the absence of critical vital registry data (Setel et al, 2007), global estimates of the causes of neonatal deaths are possible only through statistical modeling. The cause-specific distribution of neonatal deaths correlates with the NMR. In high-mortality settings (>45 per 1000 live births), the risk of neonatal death because of severe sepsis and pneumonia is approximately 11-fold higher, and the risk of dying because of birth asphyxia is approximately eightfold higher than the risk in low-mortality countries (<15 per 1000 live births). The proportion of deaths as a result of prematurity drops in countries with a high NMR because of the deaths due to infection; however, the risk of death attributable to the complications of prematurity is still threefold higher than in low-mortality countries (Lawn et al, 2005). In addition to significant differences in cause-of-death distribution between countries, there is often substantial variation within countries (Lawn et al, 2005), especially between urban and rural areas. CONSEQUENCES OF PRETERM DELIVERY AND LOW BIRTHWEIGHT Prematurity and low birthweight are often not distinguished because of the lack of gestational age dating and failure to weigh all babies at birth. As many as 18 million babies worldwide may be born annually at a low birthweight (<2500 g). Although low birthweight affects approximately 15% of births, preterm delivery and low birthweight may account for up to 60% to 80% of neonatal deaths (Awasthi et al, 2006). South Asia contributes 50% of the world’s babies with low birthweight, because 29% of babies have a low birthweight. In contrast, only 14% of Sub-Saharan African babies have a low birthweight (United Nations Children’s Fund, 2007). In developed countries, preterm birth is the leading cause of morbidity and mortality, and the percentage of 29 deaths attributed to preterm delivery is more than twofold the percentage of deaths due to asphyxia, sepsis and pneumonia, and congenital anomalies (death from diarrhea does not occur) (Lawn et al, 2005). The current rate of 12.5% in the United States represents an increase of more than 30% since 1981 and is almost double the 5% to 9% frequency in other developed countries. The highest rates of preterm birth in the United States occur among racial and ethnic minorities and in older women who conceive by artificial reproductive technology (Preterm birth: crisis and opportunity, 2006). Whether the number of preterm deliveries is rising in the developing world is unknown. Despite 30 years of research, little is known about the etiology and prevention of preterm delivery, but improved management in neonatal intensive care units has increased the survival of extremely immature fetuses in the developed world. Complex technology is not required to prevent deaths attributable to prematurity and birth asphyxia in the developing world. Low-cost interventions, including resuscitation training (Deorari et al, 2001), nutritional and thermal support through kangaroo-mother care (Charpak et al, 2005), and exclusive breastfeeding may reduce deaths attributable to asphyxia and preterm delivery after 34 weeks’ gestation. The WHO has developed an Essential Newborn Care package, which includes clean delivery practices, neonatal delivery care (including prompt stimulation and bag-mask resuscitation as required), thermoregulation with skin-to-skin care, early initiation of exclusive breastfeeding, care of the moderately small baby at home, and recognition of common illnesses (World Health Organization and Department of Reproductive Health and Research, 1996). The addition of neonates to the WHO– United Nations Children’s Fund Integrated Management of Childhood Illnesses package, which has been adopted by India, represents a significant resource to improved neonatal survival. The package provides skill-based training with elements similar to the Essential Newborn Care package, but it adds immunization and several postpartum home visits by health workers to help mothers recognize and manage minor conditions and refer severe cases in a timely manner (World Health Organization, 2003a). The Global Alliance to Prevent Prematurity and Stillbirth (GAPPS) has recently been launched to address stillbirth and prematurity with a comprehensive review of published and unpublished data on preterm birth and stillbirth research and interventions. In collaboration with diverse global partners in science, public health, and policy, GAPPS plans to advance research on the etiology of preterm delivery and stillbirth, accelerate delivery of low-cost effective interventions, and raise awareness of the effects of prematurity and stillbirth (www.gapps.org). NEONATAL SEPSIS AND PNEUMONIA Neonatal infections are responsible for more than 1 million of the 3.7 million annual deaths in the developing world. More than 95% of all deaths from birth to 2 months of age occur in developing countries. Risk factors include chorioamnionitis, low birthweight, unhygienic delivery, skin care, cord care, and environments. Ideally, simple preventive strategies such as clean delivery kits, hand washing, and cord care would be effective, but such data are not 30 PART I Overview available, and the effectiveness of using chlorhexidine to prevent community-acquired neonatal sepsis and mortality is still unsettled (Cutland et al, 2009; Mullany et al, 2006). The quality and quantity of data on neonatal deaths caused by infections in the developing world are extremely limited. The estimated total is at least 1.6 million annual deaths (26% for sepsis and pneumonia, excluding tetanus and diarrhea) (Lawn et al, 2005). The majority of births and deaths are thought to occur at home without coming to medical attention (Lawn et al, 2004), because of cultural norms that prescribe seclusion for mothers and neonates; lack of trained caretakers and facilities; high out-of-pocket costs for transport, hospitalization, and medications; and loss of wages for the mother, the father, and often another family member. As a result, the current WHO recommendation of 10 to 14 days of inpatient treatment with broad-spectrum parenteral antibiotics is unavailable or unacceptable to most families (Zaidi et al, 2005). Among those babies born in hospitals, the risk of nosocomial infection is threefold to 20-fold higher (6.5 to 38 per 1000) than in industrialized countries because of poor intrapartum and postnatal infection-control practices (Zaidi et al, 2005). Many hospitals are overcrowded and understaffed and lack even basic infection control procedures, despite the guidelines of WHO–United Nations Children’s Fund Integrated Management of Pregnancy and Childbirth and the Newborn Problems Handbook: A Guide for Doctors, Nurses, and Midwives (World Health Organization, 2003a, 2003b; Zaidi et al, 2005). The major pathogens among babies born in a hospital (11,471 isolates) are Klebsiella pneumonae, other gram-negative rods (Escherichia coli, Pseudomonas spp., Acinetobacter spp.), and Staphylococcus aureus (8% to 22%) (Zaidi et al, 2009). Newborns in a hospital often receive empiric therapy with broad spectrum parenteral antibiotics (imipenem and amikacin), because of the lack of culture facilities and concerns about resistance. Data from community settings during the first week of life are almost nonexistent. A 2009 Pediatric Infectious Disease supplement (Qazi and Stoll, 2009) reviewed evidence on community-acquired neonatal sepsis in the developing world from 32 studies published since 1990. The tremendous heterogeneity in studies, suggesting that infections may be responsible for 8% to 80% of all neonatal deaths and up to 42% of deaths in the first week, make the data difficult to interpret. Among neonates from 0 to 60 days old, rates of clinically diagnosed neonatal sepsis were as high as 170 per 1000 live births versus 5.5 blood-culture confirmed sepsis cases per 1000 live births (Thaver and Zaidi, 2009). Gram-negative rods (Klebsiella spp. [25%], E. coli [15%]) and S. aureus were the major communityacquired pathogens. Group B streptococcus was relatively uncommon (7%) in the first week, but group B streptococcus, S. aureus, and nontyphoid Salmonella spp. infection rates increased to 12% to 14% after the first week. Only 170 isolates, predominantly gram negative, were reported among home-delivered babies. The authors concluded that hospital-based and community studies suggest that most infections in the first week are attributable to gram-negative pathogens that may be environmentally acquired during unhygienic deliveries, rather than maternally acquired. As with hospital-born infants, empiric therapy with broad spectrum antibiotics was the norm, based on clinical diagnosis or on algorithms, because advanced technology was neither available nor affordable. Because the signs and symptoms of neonatal sepsis and pneumonia are nonspecific and medical systems are weak, delays in recognition, referral, and treatment were common and were reflected in both the high mortality rate (22%) (Bang et al, 2005) and frequent prescription of broad-spectrum antibiotics. A recent population-based nested observational study of community and hospital-born neonates randomized to a package of neonatal and maternal interventions in Mirzapur, Bangladesh, represents the difficulty of obtaining reliable infection data in the developing world. Of the 239 neonates who died without being enrolled, 59% and 87% died within the first 2 and 7 days, respectively, and were thought to be the result of birth asphyxia, prematurity, or both. Among the 7310 neonates who were assessed at least once by community health workers, the incidence of early neonatal sepsis was only 3 in 1000 live births. The 29 positive blood cultures represent an incidence of bacteremia of only 2.9 cases per 1000 live births; 38% of these cultures were obtained in the first three postnatal days. Fifty percent of the organisms were gram negative and 50% were gram positive; 10 in 15 gram-positive organisms were S. aureus, and one was group B streptococcus. The case fatality rate was 13% (2/15) in the gram-positive and 27% in the gramnegative infections. Seventy percent of the isolates were sensitive to the combination of ampicillin plus gentamicin or ceftriaxone. The authors noted that the incidence rate was roughly comparable to reported early-onset neonatal sepsis in the United States; however, the reported rates are likely to be low because infants who died early were not enrolled, parents were not compliant with referrals, and the intervention package may have prevented some infections. For the many reasons noted, the current recommendations for hospitalization and parenteral therapy are simply not feasible in the developing world. Therefore several multicenter trials have been launched in Asia (2009) and (2010) Africa to test the safety and efficacy of simplified antibiotic regimens to treat possible serious bacterial infections in 0- to 59-day-old infants in the community or first-level facilities (S. Qazi, personal communication, 2010). Studies evaluating the effect of prenatal and postnatal home visits by community health workers to improve newborn care practices, and identification and referral of positive serious bacterial infection were completed in 2010 in Ghana and Uganda. Studies are ongoing in Tanzania (R. Bahl, personal communication, 2011). Such research is a priority to guide community management of infections and prevent unacceptably high neonatal mortality rates in developing countries. Effective and simple interventions for the prevention and treatment of neonatal infections exist, but poor coverage of health services, a shortage of health care providers, access to referral services, and lack of knowledge on how to implement existing cost-effective interventions at scale in lowresource settings prevent them from reaching community neonates in the developing world. A methodology developed by the WHO Department of Child and Adolescent Health and Development (CAH) provides a systematic method, the Child Health and Nutrition Research Initiative (CHNRI) methodology, for setting priorities in health research investments at any level (institutional to global) (Rudan et al, 2007, 2008). Applying the CHNRI methodology to CHAPTER 4 Global Neonatal Health the prevention and treatment of neonatal infection identified the need for health policy and systems research to understand the barriers to implementation, effectiveness, and optimized use of available interventions (Bahl et al, 2009). The need for point-of-care diagnostics for neonatal pneumonia, hypoxia, bacterial sepsis, and antibiotic resistance is urgent because standard laboratory and radiologic technology are not available. To clarify the contribution of vertical transmission to neonatal mortality, sepsis data during the first 3 days of life are also a high priority. Malaria and HIV infection are threats to neonatal health in Sub-Saharan Africa. The burden of malaria in pregnancy is exacerbated by HIV, which increases susceptibility in pregnancy, in addition to reducing the efficacy of antimalaria interventions and complicating their use because of potential drug interactions. Important progress has been made in preventing malaria with intermittent preventive treatment in pregnancy and insecticide-treated nets, but coverage of these treatments with funds is still unacceptably low (Menendez et al, 2007). HIV has devastated Sub-Saharan Africa, but progress is being made. The latest guidelines are available on the WHO Web site (www.who.int/hiv). Other useful, sites for current recommendations for treating pregnant women, reducing mother-to-child transmission, and infant feeding include http://AIDSinfo.nih.gov (for U.S. guidelines and access to information on trials and drugs), http://unaidstoday.org, www.accessdata.fda-gov, and www.cdc.gov/hiv/dhap.htm. ASPHYXIA The major causes of stillbirth and early neonatal death (during the first 7 days after birth) are birth asphyxia (defined by the WHO as the failure to initiate and maintain spontaneous respiration), low birthweight, and preterm delivery. Concerns about identifying stillbirth prevention strategies, the appropriate timing for such interventions, misclassification of early neonatal deaths as stillbirths, and the limitations of verbal autopsy have led to proposals to use terms that describe the timing of an insult (intrapartum deaths and intrapartum-related neonatal deaths) and specific adverse outcome (neonatal encephalopathy) rather than the term asphyxia (Lawn et al, 2009b, 2009c). Classification of the timing of death as previable versus antepartum (macerated) or intrapartum (fresh stillbirth) may be possible, even among the 60 million annual home births; however, the consequences of intrapartum fetal organ damage caused by poor oxygenation are often difficult to distinguish from those associated with infection and trauma; therefore differentiating the specific outcomes associated with each condition might not be important. Intrapartum fetal organ damage caused by poor oxygenation may be the final common pathway for many stillbirths and early neonatal deaths (Goldenberg and McClure, 2009). Several studies suggest that improved neonatal resuscitation skills reduce misclassification of stillbirths and improve neonatal survival (Cowles, 2007; Daga et al, 1992), including a before-and-after study that provided college-educated Zambian midwives equipment and training in essential newborn care and resuscitation. The training resulted in a reduction of stillbirths from 23 to 16 per 1000 births without an increase in neonatal deaths (Chomba et al, 2008), 31 suggesting that resuscitation training of providers can decrease misclassification of stillbirths and improve neonatal survival. Of the approximately 136 million babies born every year, approximately 10% (14 million) require only stimulation at birth to establish regular respiration. As many as 3% to 6% (4 million) require stimulation and basic resuscitation with room air and a self-inflating resuscitation bag and mask, and less than 1% (1.4 million) require advanced resuscitation and postresuscitation care (Wall et al, 2009). Because less than 1% of neonates require advanced resuscitation, and few of those would survive without mechanical ventilation, advanced neonatal resuscitation is not a priority unless neonatal intensive care is available. The 1997 WHO Basic Newborn Resuscitation: A Practical Guide, which will be revised and published in 2011, provides guidelines for resuscitation training that are appropriate for first-referral level facilities in low-resource settings. A new educational resuscitation training program, Helping Babies Breathe, by the American Academy of Pediatrics and others (Niermeyer, 2009), is designed to support resuscitation training in low-resources settings. It emphasizes the “golden moment,” provides clear graphics for decision making in basic resuscitation and hands-on exercises. It is being rolled out globally by USAID and partners. New low-cost resuscitation bags and infant resuscitation models will facilitate hands-on resuscitation training initiatives. A recent review of resuscitation in low-resource settings describes the available evidence for which newborns should be resuscitated, when resuscitation should not be initiated, and when it should be stopped; management of meconium-stained infants; equipment needed for ventilation during resuscitation; evidence to support resuscitation with room air; evidence of the effects of resuscitation training in facilities and communities; postresuscitation management; and considerations for improving neonatal resuscitation in low- and middle-income countries (Wall et al, 2009). Improvement will require providing essential newborn care to newborns in all settings and frequent retraining to maintain resuscitation knowledge and skills. The authors estimate that systematic implementation of personnel using standard neonatal and competency-based training could avert an estimated 192,000 intrapartumrelated neonatal deaths per year and an additional 5% to 10% of deaths as a result of complications of preterm birth (Lawn et al, 2009; Wall et al, 2009). Among critical issues neonatal resuscitation are: • How to deliver neonatal resuscitation in settings with the highest burden, but the weakest health systems • How to implement and sustain national vital registries • How to document the actual number of births • How to document the number of intrapartum stillbirths • Whether improved survival is associated with increased numbers of disabled survivors • How to improve monitoring of the proportion of infants requiring resuscitation and their outcome • How to deliver cost-effective neonatal care, resuscitation training methods, and maintenance of resuscitation skills by different levels of providers in facilities and communities • How to determine whether infants should be suctioned, including those with meconium staining 32 PART I Overview • Early infant stimulation methods to ameliorate the effects of perinatal hypoxia Equally important are methods to improve the quality of care for mothers and neonates, including maternal and perinatal death reviews, criterion-based audits, and emergency drills (van den Broek and Graham, 2009). Dissemination of the new Helping Babies Breathe curriculum represents an important opportunity for implementation research to improve newborn survival in the developing world. PROVIDING A CONTINUUM OF CARE Early efforts to improve neonatal mortality focused on high coverage levels of a few simple and cost-effective interventions in low-resource settings. Because simple interventions did not decrease neonatal mortality, the emphasis has subsequently shifted to comprehensive packages of community interventions and to a continuum of levels of care from home to hospital. Interventions that improve access to quality health care systems and can provide training, skilled birth attendants, transportation, timely emergency obstetric and neonatal care, and early postnatal care are likely to simultaneously reduce stillbirths and early neonatal deaths, as well as maternal morbidity and mortality; however, they have only been achieved in the context of research. The proposed components include: • Empowerment of women • Increased training of all levels of birth attendants in essential newborn care and resuscitation • Emphasis on increased institutional deliveries ANTENATAL CARE • Tetanus toxoid immunization • Nutrition: iodine, iron/folate (periconceptional) • Maternal infections: syphilis, malaria (endemic areas) • Breastfeeding counseling • Birth preparedness* • Danger signs* Antenatal Care Labor & Delivery Care INTRAPARTUM CARE • Clean delivery • Skilled care at delivery* • Danger signs* • Mobilization of communities to identify and transfer high-risk pregnancies and neonates • Improved strategies for community treatment of postpartum hemorrhage, eclampsia, and sepsis • Increased postpartum home visits • The Safe Childbirth Checklist (World Health Organization, Safe Childbirth Checklist) The joint statement from the WHO and United Nations Children’s Fund recommending several early postpartum visits to deliver effective elements of care to newborns and their mothers is based on studies in Bangladesh and ­Pakistan, where such visits have been associated with reductions in newborn deaths and improved care practices. However, postpartum visits in a large-scale communitybased integrated nutrition and health program in Uttar Pradesh, India, improved care practices but did not reduce the primary outcome (i.e., neonatal mortality) at the population level (Baqui et al, 2008a). A 2007 review of the effects of packaged interventions on neonatal health (Haws et al, 2007) found no true effectiveness trials among 19 randomized controlled trials. No trial targeted women before pregnancy, and antenatal interventions were largely micronutrient supplementation. Intrapartum interventions were limited principally to clean delivery, and few increased the demand for care or improved the delivery of interventions to large populations. Subsequent trials using existing human and material resources and documenting external input are limited, but there is an increased emphasis on rural community-based interventions that could be improved. The early Bang Gadchiroli trial in ­Mahrashtra, India—which achieved a greater than 60% reduction in IMMEDIATE NEWBORN CARE • Newborn reuscitation • Prevention of hypothermia: drying, warming • Prevention of hypoglycemia: immediate breastfeeding • Prophylactic eye care (areas endemic for gonorrhea) Immediate Newborn Care Postpartum Care for Mother & Newborn POSTNATAL CARE • Exclusive breastfeeding • Clean umbilical cord care • Maintenance of temperature • Pneumonia and sepsis management • Early postpartum visit* • Birth spacing* CONTINUUM OF CARE FIGURE 4-5 Summary of priority antepartum, intrapartum, and postnatal interventions. (Modified from Bhutta ZA, Darmstadt GL, Hasan BS, et al: Community-based interventions for improving perinatal and neonatal health outcomes in developing countries: a review of the evidence, Pediatrics 115[Suppl 2]:519-617, 2005.) CHAPTER 4 Global Neonatal Health neonatal mortality by intensive training of community health workers to resuscitate asphyxiated infants, manage infants with low birthweight, and treat suspected bacterial infections with oral and injectable antibiotics (Bang et al, 1999, 2005)—has not been replicated in other communities. However, community mobilization training (women’s support groups with little additional input of health system strengthening) in rural Nepal between 2001 and 2003 reduced the neonatal mortality rate in intervention clusters by 30% (adjusted odds ratio, 0.70; 95% CI, 0.53–0.94) to 26.2 per 1000 (76 deaths per 2899 live births) compared with 36.9 per 1000 (119 deaths per 3226 live births) in controls (Manandhar et al, 2004). More recently, a community-based mobilization and education trial of care of newborn babies in rural India was associated with improved household care behaviors (early initiation of breastfeeding, delayed bathing, and skin-toskin care) and a reduction of neonatal mortality (Kumar et al, 2008). This 30-month, community-based unmasked cluster randomized trial was conducted through government and nongovernmental organization infrastructures. The trial provided home care visits (two prenatal and four postnatal care home care and referral or treatment of sick neonates by a female community health worker [CHW]) or community-based promotion of care-seeking and birth or newborn care preparedness through group sessions with female and community mobilizers, in addition to a comparison arm. Neonatal mortality was reduced in the home-care arm by 34% (adjusted RR, 0.66; 95% CI, 0.47–0.93) over the last 6 months of the intervention versus the comparison arm. The community-care arm documented improved care practices, but no reduction in neonatal mortality. These favorable results were achieved despite a much lower community health worker (CHW) density and 30% of the CHW postnatal care visits, compared with the Gadcharoli trial (Bang et al, 2005). Improvement of the home care service delivery strategy for essential newborn care is underway in Bangladesh (Baqui et al, 2008b). Although it appears that communitybased preventive strategies for newborn care can improve newborn survival and care practices, it is not clear that government health care workers and CHWs can duplicate these research results (Bhutta and Soofi, 2008). Because each setting is unique, such efforts are likely to be improved after local formative research with the communities, CHW, and birth attendants (Bahl et al, 2008). Key research gaps in community management include: • How best to create the political will to prioritize community maternal and neonatal health • How to provide a continuum of care from home to hospital (effectiveness trials carefully tailored to local health needs and conducted at scale) • How to mobilize communities to identify, stabilize, and transfer at-risk pregnancies and neonates • What strategies should be used to ensure quality of care • How to manage birth asphyxia, preterm delivery, and serious neonatal bacterial infections in the community Finally, it is important to test whether the community strategies that were effective in rural Southern Asia will be equally effective in Africa and in urban slums (Bhutta and Soofi, 2008; Bhutta et al, 2008b; Kumar et al, 2008). 33 RESOURCES Some of the key challenges to global health initiatives are information, communication, and assessment. Although the data are not always available or consistent, a number of important resources are available. The United Nations Children’s Fund reports progress in maternal and child survival in the The State of the World’s Children, based on the work of the Inter-agency Group for Child Mortality Estimation (United Nations Children’s Fund, 2007). The State of the World’s Children provides summary tables of basic, health, education, demographics, economics, and progress indicators with national rankings. The detailed text discusses when, where, and why mothers and neonates die, and it documents interventions to improve outcomes. The 2009 State of the World’s Children emphasizes maternal health. The WHO Department of Child and Adolescent Health and Development is the Secretariat for the Child Epidemiology Reference Group (CHERG) that quantifies the burden of child illnesses, supports and disseminates research to understand the determinants of childhood illnesses, and develops and evaluates interventions of new delivery strategies and large-scale interventions (Bryce et al, 2005; World Health Organization, 2009). A number of other important partnerships publish current data, including The Countdown to 2015 (United Nations Children’s Fund, 2005, 2008a), which assembles and summarizes the latest published data on 68 priority countries that represent 97% of child and maternal mortality worldwide. A unique feature is data on coverage rates for interventions that are feasible for universal implementation in poor countries and have been empirically proven to reduce mortality in mothers, children, and neonates. The 2008 edition included approaches such as delivery care and reproductive health services, which can serve as platforms for delivering multiple, proven interventions to reduce maternal and neonatal mortality (United Nations Children’s Fund, 2008a); it is intended to assist policy makers, development agencies, and donors in making performance-based policy and decisions. An explicit goal is to hold governments, development partners, and the international health community accountable for the lack of progress (United Nations Children’s Fund, 2008a). In 2003 a group of technical experts published The Child Survival Series in The Lancet (United Nations Children’s Fund, 2007), which went on to become a unique series of special editions on perinatal health in the developing world. The series has played a critical role in drawing attention and resources to improved neonatal survival, which is important to the future of the developing world. The comprehensive Disease Control Priorities Project (DCPP)—a joint effort of the National Institutes of Health Fogarty International Center, the WHO, and the World Bank—was launched in 2001 to identify policy changes and intervention strategies for the health problems of countries in need. The aim of the DCPP was to generate knowledge to assist decision makers in developing countries to realize the potential of cost-effective interventions to rapidly improve the health and welfare of their populations and to detail prevalent investments that are not cost 34 PART I Overview By disease 22 21-79 Unallocable Other Health-sector support Tuberculosis Malaria HIV/AIDS 20 18 18-99 17-90 15-60 14 13-55 12-44 12 10-69 10-90 9-80 7-76 8 8-01 8-10 8-42 8-65 98 10 97 2007 US$ (billions) 16 6-61 6 5-59 6-11 5-47 4 2 07 20 06 05 20 04 20 20 03 02 20 01 20 20 00 99 20 19 19 96 19 95 19 94 19 93 19 92 19 91 19 19 19 90 0 Year FIGURE 4-6 Development assistance for health from 1990 to 2007, by disease. (Modified from Ravishankar N, Gubbins P, Cooley RJ, et al: Financing of global health: tracking development assistance for health from 1990 to 2007, Lancet 373:2113-2124, 2009.) effective. The DCPP published an expanded and updated second edition that addresses disease conditions, their burdens and risk factors, strategy and intervention effectiveness and health systems, and financing (Laxminarayan et al, 2006). FUNDING There is no comprehensive system for tracking total amounts of developmental assistance for health or how they are spent. However, recent analyses have documented a fourfold increase in developmental assistance, from $5.6 billion in 1990 to $21.8 billion in 2007. The WHO estimated that the 10-year incremental global costs for universal health coverage of maternal and child health services ranged from $39.3 billion for a moderate improvement scenario to $55l.7 billion for a rapid improvement scenario. These projections did not include the cost of health system reforms, such as recruiting, training, and retaining a sufficient number of personnel (Johns et al, 2007). Global assistance rose sharply after 2002 because of increases in public funding, especially from the United States, and from increased philanthropic donations and in-kind contributions from corporations. (The Bill and Melinda Gates Foundation is the largest single source of private developmental health assistance.) Donor funding from the United States for HIV/AIDS has increased from $300 million in 1996 to $8.9 billion in 2006 (Oomman, et al, 2007). The proportion of developmental assistance from United Nations agencies and development banks decreased between 1990 and 2007 as targeted funding increased for the Global Alliance for Vaccines and Immunization; Medicines for Malaria, the Global Fund to Fight AIDS, Tuberculosis and Malaria; and the United States President’s Emergency Plans for AIDS Relief (PEPFAR). The influx of funds has been accompanied by major changes in the institutional landscape of global health, with global health initiatives such as the Global Fund and the GAVI assuming a central role in mobilizing and channeling global health funds. Nongovernmental organizations have become a major conduit for an increasing share of developmental assistance (Ravishankar et al, 2009). The pattern is similar for research and development for drugs: Global Funds and GAVI HIV/AIDS, tuberculosis, and malaria initiatives accounted for approximately 80% of the $2.5 billion that was spent on research and drug development in 2007 for neglected diseases in developing countries (Moran et al, 2009). Drugs and vaccines—rather than diagnostics, platform technologies, or country-specific products—are also funded preferentially (Moran et al, 2009). Research and development in neglected diseases— such as pneumonia and diarrheal illness, two major causes of mortality in developing countries–are severely underfunded at less than 6% of the budgeted funding. Increasing attention has been focused on the large amount of funding being earmarked for HIV, malaria, and tuberculosis and CHAPTER 4 Global Neonatal Health 35 TABLE 4-4 The Research Pipeline of Description and Determinants, Discovery, Development, and Delivery Description and Determinants Discovery Development Delivery Research aim Descriptive epidemiology and understanding determinants, advancing definitions New science for the discovery of mechanisms and causes of neonatal disease that provides foundation for developing new interventions Developing new or adapting existing interventions to reduce the cost, increase effect, improve deliverability Delivering existing interventions in new ways or in new settings. Includes monitoring and evaluation as feedback to additional discovery and development science Types of research (research instrument) Epidemiology New drugs and vaccines. Biochemical and genetic basis for disease Refining or adapting existing technology or drugs Effectiveness trials or implementation research for scale-up in health systems Typical timeline before impact is seen Variable 5 to 15 years 5 to 10 years 2 to 5 years Investment level Variable Very high Moderate Low to moderate depending on size of trial and rigor of evaluation Probability of major impact Variable, if new epidemiology leads directly to program or intervention Very high Moderate Very high if high impact intervention and currently low coverage Risk of failure Low High Moderate Low to moderate Specific examples for global newborn health Cohort studies to better delineate preterm birth and term small for gestational age and define short-and long-term outcomes Discovering a marker for preterm birth amenable to intervention Adapting technology for head or body cooling to be effective, lower cost, feasible and safe in low-income settings Impact and cost to provide early postnatal care package using different cadres of workers in a range of varying health system contexts especially in Africa From Lawn JE, Rudan I, Rubens C: Four million newborn deaths: is the global research agenda evidence-based? Early Hum Dev 84:809-814, 2008. the missed opportunities to save more lives—especially young lives—at a lower cost by focusing on simpler interventions (Gostin, 2008). CAREER OPPORTUNITIES American universities are experiencing an unprecedented surge in interest in global health. Students and faculty have become actively engaged in operational research, project analysis, workforce training, and policy debates. A number of American universities have made long-term commitments to specific countries in the developing world, including formal opportunities for faculty and residents to work in targeted countries in low-resource settings. Opportunities range from in-depth experiences to volunteer research and service projects. The newly launched Consortium of University for Global Health 2009 survey of 37 universities found that the number of students enrolled in global health programs in universities across the United States and Canada doubled in just 3 years and that universities have established 302 training and education programs in 97 countries. Although there is currently no official “bulletin board” for international global opportunities at the faculty level, a number of Web sites offer a range of opportunities for individuals seeking additional training, career opportunities, and interaction with other global health professionals, including the American Medical Students Association (www.amsa.org); the Global Health Council career network (careers.globalhealth.org); the United States Agency for International Development (USAID) (www.usaid.gov); and the National Institutes of Health Fogarty International Center (www.nih.fogarty.org). A number of Web sites also provide research updates, including the WHO, USAID, Save the Children Newborn Research eUpdates, and Medical News Today. The Federation of Pediatric Organizations is working in the areas of international certification, cataloging international rotations, creating a checklist of requirements for international rotations, and creating global partnerships. THE WAY FORWARD: DATA, COLLABORATION, EVALUATION, AND INVOLVEMENT There is a consensus in the global research community regarding the importance of providing a continuum of care from home to hospital for mothers and neonates and evaluating the effectiveness of packages of interventions that have proved effective in smaller trials (Madon et al, 2007). Much attention has been focused on the lack of quality data; social and cultural limitations; the need for large community randomized trials and their high cost; the lack of coordination of efforts to maximize current data by prioritizing and strengthening existing programs with proven, low-cost, high-impact interventions; and the need to systematically establish research priorities (Lawn et al, 2008, 2009a). The need for a change in the design, implementation, and evaluation of programs has received less attention, to meet the needs of national governments and donors for rigorous assessment of child survival and health in general. Victora, Black, and Bryce (2009) emphasize the need for nationwide improvement and nationwide assessments of multiple programs, in collaboration with the government and other concurrent programs. They suggest three initial 36 PART I Overview steps for an ecologic evaluation platform: (1) develop and regularly update a district database from multiple sources, (2) conduct an initial survey to be repeated every 3 years to measure coverage for proven interventions and health status, and (3) establish a continuous monitoring system to document provision, use, and quality of interventions at the district level, with mechanisms for prompt and transparent reporting. Although this ambitious plan would support the analysis of combinations of interventions and delivery strategies with the ability to adjust for confounders, there is no precedent for undertaking such a massive effort. Short of their comprehensive strategy, there is increasing recognition of the need for national vital registries and recurrent surveys to provide the basis for changes in health policy. There is also clear evidence of the need for a new emphasis on implementation science and the urgency of strengthening the independent capacity for health research in the developing world (Whitworth et al, 2008), which will enable collaborators to solve their own national problems. Finally, everyone has the power to advocate for political change to support maternal and child health. Shiffman (2009) emphasizes the need to build a strong policy community to generate political attention for global maternal and neonatal health; to develop issue frames that resonate with politicians to move them to act; to cultivate strategic alliances within women’s groups, key ministers, and congressional aides; to link the health of women and neonates in the developing world with other problems; and to remember that medical professionals carry great moral authority because of their expertise and pursuit of a humanitarian cause, if they choose to exert their political power in a strategic way. SUGGESTED READINGS Baqui A, Williams EK, Rosecrans AM, et al: Impact of an integrated nutrition and health programme on neonatal mortality in rural northern India, Bull World Health Organ 86:796-804, 2008. Bhutta ZA, Ali S, Cousens S, et al: Alma-Ata. Rebirth and revision 6 interventions to address maternal, newborn, and child survival: what difference can integrated primary health care strategies make?, Lancet 372:972-989, 2008. Himawan B: State of the world’s mothers 2007: saving the lives of children under 5, 2007, New York, Save the Children. Jamison DT, Breman JG, Measham AR, et al: Disease control priorities in developing countries: a copublication of The World Bank and Oxford University Press, ed 2, New York, 2006, Oxford University Press. Madon T, Hofman KJ, Kupfer L, et al: Public health: implementation science, Science 318:1728-1729, 2007. Martines J, Paul VK, Bhutta ZA, et al: Neonatal survival: a call for action , Lancet 365:1189-1197, 2005. Morris SS, Cogill B, Uauy R: Effective international action against undernutrition: why has it proven so difficult and what can be done to accelerate progress? Lancet 371:608-621, 2008. Save the Children: Serious bacterial infections among neonates and young infants in developing countries: evaluation of etiology and therapeutic management strategies in community settings, Pediatr Infect Dis J 28:S1-S48, 2009. United Nations Children’s Fund: Countdown 2015: tracking progress in maternal, newborn & child survival, the 2008 report, vol 2, New York, 2008a. United Nations Children’s Fund: The state of the world’s children 2008: child survival, New York, 2007. United Nations Children’s Fund: The state of the world’s children 2009: maternal and newborn health, New York, 2008b. Complete references used in this text can be found online at www.expertconsult.com P A R T I I Fetal Development C H A P T E R 5 Immunologic Basis of Placental Function and Diseases: the Placenta, Fetal Membranes, and Umbilical Cord Satyan Kalkunte, James F. Padbury, and Surendra Sharma Complex yet intricate interactions between maternal and fetal systems promote fetal growth and normal pregnancy outcomes. Throughout embryonic development, organogenesis and functional maturation are taking place. This period of development coincides with a high rate of cellular proliferation and organ development, which creates critical periods of vulnerability. Adverse factors, disruption, or impairment during these critical periods of fetal development can alter developmental programming, which can lead to permanent metabolic or structural changes (Baker, 1998). For example, triggers such as undernutrition can elicit placental and fetal adaptive responses that can lead to local ischemia and metabolic, hormonal, and immune reprogramming, resulting in small for gestational age (SGA) fetuses. Maternal health, dietary status, and exposure to environmental factors, uteroplacental blood flow, placental transfer, and fetal genetic and epigenetic responses likely all contribute to in utero fetal programming (Figure 5-1). Adult diseases such as coronary heart disorders, hypertension, atherosclerosis, type 2 diabetes, insulin resistance, respiratory distress, altered cell-mediated immunity, cancer, and psychiatric disorders are now thought to be a consequence of in utero life (Sallour and Walker, 2003). It is a matter of considerable interest that, in addition to maternal predisposing factors, cytokines, hormones, growth factors, and the intrauterine immune milieu also contribute to in utero programming. Adaptations of the maternal immune system exist to modulate detrimental effects on the fetus and additional mechanisms and factors actively cross the placenta and induce regulatory T cells in the fetus to suppress fetal antimaternal immunity. These effects persist at least through adolescence (Burlingham, 2009; Mold et al, 2008). Excessive restraint of maternal immune responses could lead to a lethal infection in the newborn. On the other hand, too little modulation of maternal immune response to the fetal allograft could lead to autoimmune-mediated fetalplacental rejection. Moreover, placental growth resembles that of a tumor, evading immune surveillance and initiating its own angiogenesis. Therefore a healthy mother with healthy placentation is critical to healthy fetal outcomes. MAMMALIAN PLACENTATION The immune tolerance of the semiallograft fetus and de novo vascularization are two highly intriguing processes that involve direct interaction of maternal immune cells, invading trophoblast cells, and arterial endothelial cells. Pregnancy is considered an immunologic paradox, in which paternal antigen-expressing placental cells interact directly with and coexist with the maternal immune system (Medawar, 1953). This anatomic distinction of the immunologic interface that arises from hemochorial placentation that occurs in humans and rodents is distinct from epitheliochorial placentation as seen in marsupials, horses, and swine or the endotheliochorial placentation seen in dogs and cats. Understanding the anatomic and physiologic events that occur during placentation is the key to appreciate the uniqueness of human placentation in the phylogenetic evolution. Typically, in hemochorial placentation, maternal uterine blood vessels and decidualized endometrium are colonized by trophoblast cells, derived from trophectoderm of the implanting blastocyst. These cells come in direct contact with maternal blood and uterine tissue. A similar phenomenon is evident in murine pregnancy, except the trophoblast invasion is deeper in humans (Moffett and Loke, 2006). In epitheliochorial placentation, trophoblast cells of the placenta are in direct contact with the surface epithelial cells of the uterus, but there is no trophoblast-cell invasion beyond this layer. In endotheliochorial placentation, the trophoblast cells breach the uterine epithelium and are in direct contact with endothelial cells of maternal uterine blood vessels. EMBRYOLOGIC DEVELOPMENT OF THE PLACENTA Shortly after fertilization takes place in the ampullary portion of the fallopian tube, the fertilized ovum or zygote begins dividing into a ball of cells called a morula. As the morula enters the uterus (by the fourth day after fertilization), it forms a central cystic area and is called a blastocyst (Figure 5-2). The blastocyst implants within the endometrium by day seven (Moore, 1988). The blastocyst has two components: an inner cell mass, which becomes the developing embryo, and the outer cell layer, which becomes the placenta and fetal membranes. The cells of the developing blastocyst, which eventually become the placenta, are differentiated early in gestation (within 7 days after fertilization). The outer cell layer, the trophoblast, invades the endometrium to the level of the decidua basalis. Maternal blood vessels are also invaded. Once entered and controlled by the trophoblast, these 37 38 PART II Fetal Development Maternal Placenta Fetus Health status Uteroplacental perfusion Metabolic adaptation Diet and exposure status Ischemia Genetics and epigenetics Immune status Hormonal regulation Immune regulation Placental-fetal response Outer cell layer Inner cell mass A Cytotrophoblast Syncytiotrophoblast Fetal pole Fetal Programming FIGURE 5-1 Fetal programming. Maternal health and the placenta influence fetal adaptations. Dietary status, exposure to environmental factors, uteroplacental blood flow, placental transfer, and genetic and epigenetic changes contribute to the in utero fetal programming. maternal blood vessels form lacunae, which provide nutrition and substrates for the developing products of conception. The trophoblast differentiates into two cell types, the inner cytotrophoblast and the outer syncytiotrophoblast (Figure 5-3); the former has distinct cell walls and is thought to represent the more immature form of trophoblast. The syncytiotrophoblast, which is essentially acellular, is the site of most placental hormone and metabolic activity. Once the trophoblast has invaded the endometrium, it begins to form outpouchings called villi, which extend into the blood-filled maternal lacunae or further invade the endometrium to attach more solidly with the decidua, forming anchoring villi. B FIGURE 5-2 A, The human blastocyst contains two portions: an inner cell mass, which develops into the embryo, and an outer cell layer, which develops into the placenta and membranes. B, The outer acellular layer is the syncytiotrophoblast, and the inner cellular layer is the cytotrophoblast. (From Moore TR, Reiter RC, Rebar RW, et al, editors: Gynecology and obstetrics: a longitudinal approach, New York, 1993, Churchill Livingstone.) PLACENTAL ANATOMY AND CIRCULATION At term, the normal placenta covers approximately one third of the interior portion of the uterus and weighs approximately 500 g. The appearance is of a flat circular disc approximately 2 to 3 cm thick and 15 to 20 cm across (Benirschke and Kaufmann, 2000). Placental and fetal weights throughout gestation are presented in Table 5-1. During the first trimester and into the second, the placenta weighs more than the fetus; after that period, the fetus outweighs the placenta. With the formation of the tertiary villi (19 days after fertilization), a direct vascular connection is made between the developing embryo and the placenta (Moore, 1988). Umbilical circulation between the placenta and the embryo is evident by 51⁄2 weeks’ gestation. Figure 5-4 demonstrates aspects of the maternal and fetal circulation in the mature placenta. The umbilical arteries from the fetus reach the placenta and then divide repetitively to cover the fetal surface of the placenta. Terminal arteries then penetrate the individual cotyledons, forming capillary beds for substrate exchange within the tertiary villi. These capillaries then reform into tributaries of the umbilical venous system, which carries oxygenated blood back to the fetus. EXAMINATION OF THE PLACENTA A renaissance in placental pathology has led to a new relevance of the placenta to neonatology and early infant life, including issues of preterm birth, growth restriction, and cerebral, renal, and myocardial diseases. The placenta can give some clues to the timing and extent of important adverse prenatal or neonatal events as well as to the relative effects of sepsis and asphyxia on the causation of neonatal diseases. Placental disorders can be noted immediately in the delivery room, and others can be diagnosed through detailed gross and microscopic examinations over the ensuing 48 hours. Every placenta should be examined at the time of birth regardless of whether the newborn has any immediate problems. Most placentas invert with traction at the time of delivery, and the fetal membranes cover the maternal surface. It is important to reinvert the membranes and examine all surfaces of the placenta and membranes, looking for abnormalities. Table 5-2 lists pregnancy complications or conditions that are diagnosable at birth through examination of the placenta. 39 CHAPTER 5 Immunologic Basis of Placental Function and Diseases: the Placenta, Fetal Membranes, and Umbilical Cord Maternal lacunae Syncytiotrophoblast Cytotrophoblast A Endometrium Mesenchymal cells TABLE 5-1 Fetal and Placental Weight Throughout Gestation Gestational Age (wk) Placental Weight (mg) Fetal Weight (g) 14 45 — 16 65 59 18 90 155 20 115 250 22 150 405 24 185 560 26 217 780 28 250 1000 30 282 1270 32 315 1550 34 352 1925 36 390 2300 38 430 2850 40 470 3400 Adapted from Benirschke K, Kaufmann P: Pathology of the human placenta, ed 4, New York, 2000, Springer-Verlag. B Fetal blood vessel cases of meconium aspiration syndrome, for which legal questions may arise as to whether the aspiration occurred before or during labor. If the membranes are deeply stained, the passage of meconium by the fetus may have predated onset of labor; therefore aspiration could have occurred before labor. The umbilical cord should also be examined for the number of vessels and their insertion into the placenta. Vessels on the fetal surface of the placenta should be examined for evidence of clotting or thrombosis. FUNCTIONS OF THE PLACENTA C FIGURE 5-3 A, The cytotrophoblast indents the syncytiotrophoblast to form primary villi. B, Mesenchymal cells invade the cytotrophoblast 2 days after formation of the primary villi to form secondary villi. C, Blood vessels arise de novo and eventually connect with blood vessels from the embryo, forming tertiary villi. (From Moore TR, Reiter RC, Rebar RW, et al, editors: Gynecology and obstetrics: a longitudinal approach, New York, 1993, Churchill Livingstone.) The initial placental examination should include checking the edges for completeness. The membranes and fetal surface should be shiny and translucent. An odor may suggest infection, and cultures of the placenta may be beneficial (Benirschke and Kaufmann, 2000). Greenish discoloration may represent meconium staining or old blood; placentas with such discoloration should be sent to the pathologist for complete histologic examination. The finding of deep meconium staining of the membranes and umbilical cord suggests that the meconium was passed at least 2 hours before delivery; this fact may be helpful in To ensure normal fetal growth and development, the placenta behaves as an efficient organ of gas and nutrient exchange and as a robust endocrine and metabolic organ. Besides mediating the transplacental exchange of gases and nutrients, the placenta also synthesizes glycogen with a significant turnover of lactate. Hormones secreted by the placenta have an important role for the fetus and the mother. Placental trophoblasts are a rich source of cholesterol and peptide hormones, including human chorionic gonadotrophin (HCG), human placental lactogen, cytokines, growth hormones, insulin-like growth factors, ­corticotrophin-releasing hormones, and angiogenic factors such as vascular endothelial growth factor (VEGF) and placental growth factor (PlGF). HCG, which is detected as early as day 8 after conception, is secreted by syncytiotrophoblasts into the maternal circulation, reaches maximal levels by week 8 of pregnancy and diminishes later during gestation. HCG is essential to promote estrogen and progesterone synthesis during different stages of pregnancy. Human placental lactogen mobilizes the breakdown of maternal fatty acid stores and ensures an increased supply of glucose to the fetus. VEGF and PlGF are secreted by trophoblasts and specialized natural killer (NK) cells in the decidua, and they promote angiogenesis and vascular activity, particularly during early stages of pregnancy when spiral 40 PART II Fetal Development Umbilical vein Fetal circulation Umbilical arteries Amniochorionic membrane Decidua perietalis Smooth chorion Intervillous space Chorionic plate Mainstem villus Amnion Stump of mainstem villus Placental septum Decidua basalis Anchoring villus Myometrium Endometrial veins Endometrial arteries Cytotrophoblastic shell Maternal circulation FIGURE 5-4 Schematic drawing of a section of a mature placenta showing the relation of the villous chorion (fetal part of the placenta) to the decidua basalis (maternal part of the placenta), the fetal placental circulation, and the maternal placental circulation. Maternal blood flows into the intervillous spaces in funnel-shaped spurts, and exchanges occur with the fetal blood as the maternal blood flows around the villi. Note that the umbilical arteries carry deoxygenated fetal blood to the placenta, and the umbilical vein carries oxygenated blood to the fetus. In addition, the cotyledons are separated from each other by decidual septa of the maternal portion of the placenta. Each cotyledon consists of two or more mainstem villi and their main branches. In this drawing, only one mainstem villus is shown in each cotyledon, but the stumps of those that have been removed are shown. (From Moore KL: The developing human: clinically oriented embryology, ed 5, Philadelphia, 1993, WB Saunders.) TABLE 5-2 Pregnancy Conditions Diagnosable at Birth by Gross Placental Examination and Associated Neonatal Outcomes Pregnancy Conditions Fetal/ Neonatal Outcomes Monochorionic twinning TTT syndrome donor/recipient status, pump twin in TRAP, survivor status after fetal demise, selective termination, severe growth discordance without TTT Dichorionic twinning Less likelihood of survivor brain disease in the event of demise of one fetus Purulent acute chorioamnionitis Risk of fetal sepsis, fetal inflammatory response syndrome, cerebral palsy Chorangioma Hydrops, cardiac failure, consumptive coagulopathy Abnormal cord coiling IUGR, fetal intolerance of labor Maternal floor infarction IUGR, cerebral disease Abruption Asphyxial brain disease Velamentous cord IUGR, vasa previa Cord knot Asphyxia Chronic abruption oligohydramnios syndrome IUGR Single umbilical artery Malformation, IUGR Umbilical vein thrombosis Asphyxia Amnion nodosum Severe oligohydramnios leading to pulmonary hypoplasia Meconium staining Possible asphyxia, aspiration lung disease Amniotic bands Fetal limb reduction abnormalities Chorionic plate vascular thrombosis Asphyxia, possible thrombophilia Breus mole Asphyxia, IUGR IUGR, Intrauterine growth retardation; TRAP, twin-reversed arterial perfusion; TTT, twin-to-twin transfusion. artery transformation and trophoblast invasion occurs. In addition, the placenta is a rich source of estrogen, progesterone, and glucocorticoids. Whereas progesterone maintains a quiescent, noncontractile uterus, it also has a role in protecting the conceptus from immunologic rejection by the mother. Glucocorticoids promote organ development and maturation. Placental transport is another important function, efficiently transferring nutrients and solutes that are essential for normal fetal growth. The syncytiotrophoblast covering the maternal villous surface is a specialized epithelium that participates in the transport of gases, nutrients, and waste products and the synthesis of hormones CHAPTER 5 Immunologic Basis of Placental Function and Diseases: the Placenta, Fetal Membranes, and Umbilical Cord Anchoring Villi Placenta 41 ST CT IT Decidua ET NK ST: Syncytiotrophoblasts M T Myometrium CT: Columnar trophoblasts DC ET: Endovascular trophoblasts IT: Interstitial trophoblasts Treg GC Spiral arteries GC: Giant cells NK: Natural killer cells T: T lymphocytes Treg: Regulatory T cells M: Macrophages DC: Dendritic cells Normal Preeclampsia/IUGR FIGURE 5-5 Trophoblast differentiation and spiral artery remodeling. Progenitor trophoblast cells from villi differentiate into syncytiotrophoblasts and the extravillous cytotrophoblasts (EVTs). EVTs migrate out in columns as columnar trophoblasts and anchor the placenta to the decidua. Further differentiation takes place into invasive or proliferative EVTs. The invasive EVTs invade the decidua known as interstitial trophoblasts, and some of them fuse to form the multinucleated gaint cells. Endovascular transformation ensues as endovascular trophoblasts migrate into and colonize the spiral arteries, almost reaching the myometrium. This results in wide-bore, low-resistant capacitance blood vessels as observed in normal pregnancy. In contrast, shallow trophoblast invasion and incomplete transformation of spiral arteries is a common feature of preeclampsia and intrauterine growth restriction. that regulate placental, fetal, and maternal systems. The syncytiotrophoblast layer of the placenta is an important site of exchange between the maternal blood stream and the fetus. In addition to simple diffusion, syncytiotrophoblasts facilitate exchange by transcellular trafficking that utilizes transport proteins such as the water channels (aquaporins). Facilitated diffusion for molecules such as glucose and amino acids are performed by glucose transporters (GLUT) and amino acid transporters. In addition, adenosine triphosphate (ATP)-mediated active transport, such as the Na+, K+-ATPase or the Ca2+-ATPase, besides endocytosis and exocytosis, participates in transplacental exchange (Hahn et al, 2001; Malassiné and Cronier, 2002; Randhawa and Cohen, 2005; Siiteri, 2005). In healthy women who are not pregnant, uterine blood vessels receive approximately 1% of the cardiac output to maintain the uterus. During pregnancy, these same vessels must support the rapidly growing and demanding placenta and fetus. This evolutionary challenge is addressed by remodeling of the spiral arteries, converting them into large, thin-walled, dilated vessels with reduced vascular resistance. TROPHOBLAST DIFFERENTIATION AND REMODELING OF SPIRAL ARTERIES The placental-decidual interaction through invading trophoblasts determines whether an optimal transformation of the uterine spiral arteries is achieved. Trophoblastorchestrated artery remodeling is an essential feature of normal human pregnancy. As shown in Figure 5-5, progenitor trophoblast cells from villi differentiate along two pathways: terminally differentiated syncytiotrophoblasts and the extravillous cytotrophoblasts (EVTs) that migrate out in columns and anchor the placenta to the decidua. From these anchoring layers of EVTs further differentiation takes place into invasive or proliferative EVTs. The invasive EVTs invade the decidua and shallow parts of the myometrium and are known as interstitial EVTs. Thereafter endovascular transformation ensues as invasive EVTs migrate into and colonize the spiral arteries, almost reaching the myometrium. These trophoblasts are known as endovascular EVTs. Insufficient uteroplacental interaction characterized by shallow trophoblast invasion and 42 PART II Fetal Development incomplete transformation of spiral arteries is a common feature of preeclampsia and intrauterine growth restriction (IUGR) (Brosens et al, 1977; Meekins et al, 1994). The precise period when trophoblast invasion of decidua and spiral arteries ceases is not clear. Nevertheless it is widely believed to be completed late in the second trimester. Although our understanding of the molecular events underlying spiral artery remodeling in pregnancy remains poor, efficient trophoblast invasion is an essential feature. There are two waves of trophoblast invasion that follow implantation. The first wave is during the first trimester, when the invasion is limited to the decidual part of the spiral artery. The second wave is during the late second trimester involving deeper trophoblast invasion, reaching the inner third of myometrial segment. The initial invasion of EVTs into the endometrium initiates the decidualization process, which is characterized by replacement of extracellular matrix, loss of normal musculoelastic structure, and deposition of fibrinoid material. Displacement of the endothelial lining of spiral arteries by the invading trophoblasts further results in uncoiling and widening of the spiral artery, ensuring the free flow of blood and nutrients to meet the escalating demands of the growing fetus (Kham et al, 1999; Pijnenborg et al, 1983). A lack of spiral artery remodeling with shallow trophoblast invasion has been associated with preeclampsia. During the process of invasion in a normal pregnancy, cytotrophoblasts undergo phenotypic switching, with a loss of E-cadherin expression, and they acquire vascular endothelial-cadherin, ­platelet-endothelial adhesion molecule-1, vascular endothelial adhesion molecule-1, and α4 and αvβ3 integrins (Bulla et al, 2005; Zhau et al, 1997). Along with a repertoire of facilitators for invasion, trophoblasts express a nonclassic major histocompatibility complex (MHC) human leukocyte antigen (HLA) G, which has gained widespread interest because of providing noncytotoxic signals to uterine NK cells. It still needs to be evaluated whether intrinsic HLA-G inactivation by polymorphic changes influences the dysregulated trophoblast invasion seen in preeclampsia (Hiby et al, 1991; Le Bouteiller et al, 2007). Although the exact gestational age at which trophoblast invasion ceases is not known, recent studies have shown that late pregnancy trophoblasts loose the ability to transform the uterine arteries. Using a novel dual-cell in vitro culture system that mimics the vascular remodeling events triggered by normal pregnancy serum, we have shown that first- and third-trimester trophoblasts respond differentially to interactive signals from endothelial cells when cultured on the extracellular matrix, matrigel. Term trophoblasts not only fail to respond to signals from endothelial cells, but they inhibit endothelial cell neovascular formation. In contrast, trophoblast cells representing firsttrimester trophoblasts with invasive properties undergo spontaneous migration and promote endothelial cells to form a capillary network (Figure 5-6). This disparity in behavior was confirmed in vivo using a matrigel plug assay. Poor expression of VEGF-C and VEGF receptors coupled with high E-cadherin expression by term trophoblasts contributed to their restricted migratory and interactive properties. Furthermore, these studies showed that the kinase activity of VEGF receptor 2 is essential for proactive crosstalk by invading first-trimester trophoblast cells (Kalkunte et al, 2008b). This unique maternal and fetal cell interactive model under the pregnancy milieu offers a potential approach to study cell-cell interactions and to decipher inflammatory components in the serum samples from adverse pregnancy outcomes (Kalkunte et al, 2010). One of the inimitable contributors to trophoblast cell invasion is the specialized NK cell of the pregnant uterus. IMMUNE PROFILE AND IMMUNO VASCULAR BALANCE DURING PLACENTATION During pregnancy, trophoblast cells directly encounter maternal immune cells at least at two sites. One site is the syncytiotrophoblasts covering the placental villi that are bathed in maternal blood, and the other is by the invading trophoblasts in the decidua. Although the syncytiotrophoblasts do not express MHC antigens, the invading trophoblasts express nonclassic HLA-G and HLA-C and would elicit immune responses in the decidua. The decidua is replete with innate immune cells including T cells, regulatory T cells, macrophages, dendritic cells and NK cells (Table 5-3). Interestingly, NK cells peak and constitute the largest leukocyte population in the early pregnant uterus, accounting for 60% to 70% of total lymphocytes. These cells diminish in proportion as pregnancy proceeds. PHENOTYPIC AND FUNCTIONAL FEATURES OF UTERINE NATURAL KILLER CELLS Peripheral blood NK (pNK) cells constitute 8% to 10% of the CD45+ population in circulation. All NK cells are characterized by a lack of CD3 and expression of CD56 antigen. Based on the intensity of CD56 antigen, NK cells are further divided into CD56bright and CD56dim populations. The presence or absence of FcγRIII or CD16 further differentiates subpopulations of uterine NK (uNK) cells. Thus the majority of peripheral NK cells are of the CD56dimCD16+ phenotype (approximately 90%), and the remaining cells are CD56brightCD16– (approximately 10%). The majority of uterine NK cells (approximately 90%) are CD56brightCD16–. In the uterine decidua, uNK cell numbers cyclically increase and decrease in tandem with the menstrual cycle—low in the proliferative phase (10% to 15%), which amplifies during the early, middle and late secretory phases (25% to 30%)—falling to a basal level with menstruation (Figure 5-7) (Kalkunte et al, 2008a; Kitaya et al, 2007). With successful implantation, the uNK cell population further increases in the decidualized endometrium, reaches a peak in first-trimester pregnancy, and dwindles thereafter by the end of the second trimester. The origin of uNK cells that peak during the secretory phase of the menstrual cycle and early pregnancy is currently not well established, and the evidence indicates multiple different possibilities. These possibilities include recruitment of CD56brightCD16– pNK cells, recruitment and tissue specific terminal differentiation of CD56dimCD16+ pNK cells, development of NK cells from Lin–CD34+CD45+ progenitor cells, or proliferation of resident CD56brightCD16– NK cells. Comparative surface expression of antigens, natural cytotoxicity receptors, inhibitory receptors, and CHAPTER 5 Immunologic Basis of Placental Function and Diseases: the Placenta, Fetal Membranes, and Umbilical Cord A B C D E F G H I J K L 43 FIGURE 5-6 (Supplemental color version of this figure is available online at www.expertconsult.com.) Differential endovascular activity of first- and thirdtrimester trophoblasts in response to normal pregnancy serum. A representative micrograph of trophoblasts-endothelial cell interactions on matrigel is shown. Endothelial cells and trophoblasts are labeled with red and green cell tracker respectively, were independently cultured (A to E) or cocultured (F to I) on matrigel. Capillary-like tube structures were observed with human uterine endothelial cells (HUtECs) (A) and umbilical vein endothelial cells (HUVECs) (B), but not with first-trimester trophoblast HTR8 cells (C), third trimester trophoblast TCL1 cells (D), and primary term trophoblasts (E). However, in cocultures, HTR8 cells fingerprint the HUtECs (F) and HUVECs (G), while TCL1 cells (H) and primary term trophoblasts (I) inhibit the tube formation by endothelial cells (magnification ×4). Panels J to L show the cocultures of HTR8 with HUVECs (   J), HUtECs ( K), and term trophoblasts with HUVECs (L) at higher magnification (×10). (Reproduced with permission from Kalkunte S, Lai Z, Tewari N, et al: In vitro and in vivo evidence for lack of endovascular remodeling by third trimester trophoblasts, Placenta 29:871-878, 2008.) TABLE 5-3 Comparison of Peripheral Blood and Decidual Immune Cell Profiles Immune Cells T cells Peripheral blood (%) Decidua (%) 65-70 9-12 γδT cells 2-5 7-10 Macrophages 7-10 15-20 B cells 7-10 ND NKT cells 2-5 0.5-1.0 Tregs 2-4 6-10 NK cells 7-12 65-70 (CD56brightCD16-) ND, Not detected. chemokines and cytokines on human pNK and uNK cells are provided in Table 5-4. Furthermore, CD56bright uNK cells are different from the CD56bright minor population of pNK cells because of the expression of CD9, CD103, and killer immunoglobulin-like receptors (KIRs). Despite being replete with cytotoxic accessories of perforin, granzymes A and B and the natural cytotoxicity receptors NKp30, NKp44, NKp46, NKG2D, and 2B4 as well as LFA-1, uNK cells are tolerant cytokine-producing cells at the maternal-fetal interface (Kalkunte et al, 2008a). The temporal occurrence around the spiral arteries and timed amplification of these specialized uNK cells observed during the first trimester implicate its role in spiral artery remodeling. NK cell–deficient mice display abnormalities in decidual artery remodeling and trophoblast invasion, possibly because of a lack of uNK cell–derived interferon γ (Ashkar et al, 2000). Other studies have shown that unlike pNK cells, uNK cells are a major source of VEGF-C, Angiopoietins 1 and 2 and transforming grwoth factor (TGFβ1) within the placental bed that decrease with gestational age (Lash et al, 2006). These observations implicate uNK cells in promoting angiogenesis. Studies have provided further evidence that uNK cells, but not pNK cells, regulate trophoblast invasion both in vitro and in vivo through the production of interleukin-8 and interferon-inducible 44 PART II Fetal Development LH P4 E2 NK cell population KIR NKG2D NKp44 NK NKp46 NKp30 Inactive phase Proliferative phase ~30% ~50–60% Secretory phase First trimester Second trimester Term Menstrual cycle Pregnancy FIGURE 5-7 Biologic pattern of natural killer (NK) cells in the human endometrium and the decidua. The uterine NK cell population characterized by natural cytotoxicity receptors (Nkp30, NKp44, NKp46, NKG2D), killer immunoglobulin-like receptors, and cytolytic machinery (perforin and granzyme) cyclically increases and decreases in tandem with the hormonal changes during menstrual cycle. With successful implantation, uterine NK cells further increase in the decidua and dwindle thereafter by the end of second trimester. E2, Estradiol; LH, luteinizing hormone; P4, progesterone. protein-10, in addition to other angiogenic factors (Hanna et al, 2006). Recent studies suggest that VEGF-C, a proangiogenic factor produced by uNK cells, is responsible for the noncytotoxic activity (Kalkunte et al, 2009). As noted previously, VEGF-C–producing uNK cells support endovascular activity in a coculture model of capillary tube formation on matrigel (Figure 5-8). Peripheral blood NK cells fail to produce VEGF-C and remain cytotoxic. This function can be reversed by recombinant human VEGFC. Cytoprotection by VEGF-C is related to induction of the transporter associated with antigen processing 1 and MHC assembly in target cells. Overall, these findings suggest that expression of angiogenic factors by uNK cells keeps these cells noncytotoxic, which is critical to their pregnancy compatible immunovascular role during placentation and fetal development (Kalkunte et al, 2009). Although uNK cells seem to play a role that is compatible with pregnancy, retention of their cytolytic abilities suggests their role as sentinels at the maternal-fetal interface in situations that threaten fetal persistence. This facet of uNK cell function was elegantly demonstrated in animal models when pregnant mice were challenged with tolllike receptor (TLR) ligands that mimic bacterial and viral infections. These observations raise an important question whether uNK cells can harm the fetal placental unit and, if so, under what conditions? The antiinflammatory cytokine interleukin (IL)-10 plays a critical role in pregnancy because of its regulatory relationship with other intrauterine modulators and its wide range of immunosuppressive activities (Moore et al, 2001). IL-10 expression by the human placenta depends on gestational age, with significant expression through the second trimester followed by attenuation at term (Hanna et al, 2000). IL-10 expression is also found to be poor in decidual and placental tissues from unexplained spontaneous abortion cases (Plevyak et al, 2002) and from deliveries associated with preterm labor (Hanna et al, 2006) and preeclampsia (S. Kalkunte et al, unpublished observations). However, the precise mechanisms by which IL-10 protects the fetus remains poorly understood. IL-10–/– mice suffer no pregnancy defects when mated under pathogen-free conditions (White et al, 2004), but they exhibit exquisite susceptibility to infection or inflammatory stimuli compared with wild type animals. It is then plausible that in addition to IL-10 deficiency, a “second hit” such as an inflammatory insult resulting from genital tract infections, environmental factors, or hormonal dysregulation during gestation can lead to adverse pregnancy outcomes (Tewari et al, 2009; Thaxton et al, 2009). Our recent studies provide direct evidence that uNK cells can become adversely activated and mediate fetal demise and preterm birth in response to low doses of TLR ligands CHAPTER 5 Immunologic Basis of Placental Function and Diseases: the Placenta, Fetal Membranes, and Umbilical Cord TABLE 5-4 Phenotypic Characteristics of Surface Markers and Receptors on Natural Killer Cells Antigen CD56 Peripheral blood Decidua Dim (>90%) Bright CD16 + – CD45 + + + CD7 + CD69 – + L-Selectin – –/+ NK Receptors KIR + + NKp30 + + NKp44 –* + NKp46 + + NKG2D CD94/NKG2A + + –/+ + + Chemokine Receptors CXCR1 + CXCR2 + – CXCR3 – + + CXCR4 + CX3CR1 + – CCR7 – + Data from Kalkunte S, Chichester CO, Sentman CL, et al: Evolution of non-cytotoxic uterine natural killer cells, Am J Reprod Immunol 59:425-432, 2008. +, Present; –, absent; –/+, variable expression; KIR, killer immunoglobulin receptor; CXCR, CX-chemokine receptor; CX3CR1, CX3-C–chemokine receptor 1; CCR7, CCchemokine receptor 7. *Expression seen on activation with interleukin 2. resulting in placental pathology (Murphy et al, 2005; Murphy et al, 2009). Moreover, spontaneous abortion is associated with an increase in CD56dimCD16+ cells and a decrease in CD56brightCD16– NK cells in the preimplantation endometrium during the luteal phase (­Michimata et al, 2002; Quenby et al, 1999). Therefore a fine balance between maternal activating and inhibitory KIRs and their ligand HLA-C on fetal cells seems to be maintained in normal pregnancy. Insufficient inhibition of uNK cells can activate the cytolytic machinery, resulting in spontaneous abortion, intrauterine growth restriction, or preterm labor, depending on the timing of the insult (Varla-­Leftherioti et al, 2003). In the setting of IVF, the implantation failure has been associated with high uNK cell numbers, but direct evidence for their role in abnormal implantation is not clear (Quenby et al, 1999). Nevertheless current understanding strongly implies that uNK cells retain the ability to become foes to pregnancy under the axis of genetic stress and inflammatory trigger. REGULATORY T CELLS AND PREGNANCY The existence of regulatory mechanisms that suppress the maternal immune system was proposed in the early 1950 (Medawar, 1953). For several years, maternal tolerance toward fetal alloantigens was explored in the context of Th1/Th2 balance, with Th2 cells and cytokines proposed 45 to predominate over Th1 cellular immune response under normal pregnancy. Recently the role of specialized T lymphocytes, termed regulatory T cells (Tregs), in tolerogenic mechanisms has emerged. Tregs are potent suppressors of T cell–mediated inflammatory immune responses and prevent autoimmunity and allograft rejection. Tregs act by controlling the autoreactive T cells that have escaped negative selection from the thymus, and they restrain the intensity of responses by T cells reactive with alloantigens and other exogenous antigens. This unique functional capability to suppress responses to tissue-specific self-antigens that escape recognition by T cells during maturation is due to tissue specific expression and alloantigens, particularly in the epithelial surfaces where tolerance to nondangerous foreign antigen is essential to normal function. This capability enables Tregs to play a unique role at the maternal-fetal interface. Tregs are typically characterized by a CD4+CD25+ surface phenotype and expression of the hallmark suppressive transcription factor Foxhead Box P3 (Foxp3+). Their cell numbers increase in blood, decidual tissue, and lymph nodes draining the uterus during pregnancy. These cells are implicated in successful immune tolerance of the conceptus, mainly by producing IL-10 and TGF-β. Recent evidence suggests that fetal Tregs also play a vital role in suppressing fetal antimaternal immunity against maternal cells that cross the placenta (Mold et al, 2008). In the absence of Tregs the allogeneic fetus is rejected, suggesting their critical role in normal pregnancy. Unexplained infertility, spontaneous abortion, and preeclampsia are associated with proportional deficience, functional Treg deficiency, or both. In the context of pregnancy, the local milieu, particularly during the first trimester, that includes hCG, TGF-β, IL-10, granulocyte-macrophage colonystimulating factor, and indoleamine 2,3-dioxygenase expression has now been shown to induce CD4+CD25+ Tregs with Foxp3 expression with immunosuppressive features. This induction is thought to occur through the immature dendritic cells. In addition to immune suppressive and antiinflammatory properties, TGF-β is recognized as inducing differentiation of naïve CD4 T cells into suppressor T-cell phenotype, expressing Foxp3, and promoting the proliferation of mature Tregs. In addition to the suppressive effects of cytokines produced by these cells, contact-mediated immune suppression by Tregs results from ligation of inhibitory cytotoxic T-lymphocyte antigen (CTLA-4) and its ability to induce tolerogenic dendritic cells and influence T-cell production of IL-10 (Aluvihare et al, 2004; Schumacher et al, 2009; Shevach et al, 2009). Therefore the pregnant uterus may be a natural depot for Tregs. EPIGENETIC REGULATION IN THE PLACENTA The regulation of gene expression is a crucial process that defines phenotypic diversity. Switching off or turning on genes as well as tissue-specific variation in gene expression contributes to this diversity. Besides the genetic make-up (i.e, sequence) of the individual, the regulation of gene expression is also influenced by epigenetic factors. Epigenetic changes include the noncoding changes in DNA and chromatin, or both, that mediate the interactions between genes and their environment. Epigenetic regulation generates a wider 46 PART II Fetal Development DECIDUA Cytotoxic NK cells Non-cytotoxic NK cells CD56 CD56brightCD16 VEGF C CD16 CD56dimCD16 Intrauterine infections, Inflammation, TLR activation, Loss of IL-10 Promotes endovascular activity Upregulates TAP-1 and MHC molecules on trophoblats Efficient trophoblast invasion Poor endovascular activity Trophoblast lysis Poor trophoblast invasion Normal pregnancy Adverse pregnancy FIGURE 5-8 (See also Color Plate 1.) Angiogenic features of natural killer (NK) cells render them immune tolerant at the maternal-fetal interface. Vascular endothelial growth factor (VEGF) C–producing noncytotoxic uterine NK cell clones similar to decidual NK cells support endovascular activity in a coculture of endothelial cells and first-trimester trophoblast HTR8 cells on matrigel. By contrast, cytotoxic uterine NK cell clones similar to peripheral blood NK cells disrupted the endovascular activity because of endothelial and trophoblast cell lysis. This distinct functional feature determines whether optimal trophoblast invasion takes place and can result in normal or adverse pregnancy outcomes. diversity of cell types during mammalian development and sustains the stability and integrity of the expression profiles of different cell types and tissues. This regulation is choreographed by changes in cytosine-phosphate-guanine (CpG) islands of the DNA promoter region by methylation, histone modification, genomic imprinting, and expression of noncoding RNAs such as micro RNA (miRNA). Gene-environment interactions resulting in epigenetic changes in the placenta during the critical window of development can influence fetal programming in utero, with predisposing health consequences later in life. Using a microarray-based approach to compare chorionic villous samples from the first trimester of pregnancy with gestational age– matched maternal blood cell samples, recent studies show tissue-specific differential CpG methylation patterns that identify numerous potential biomarkers for the diagnosis of fetal aneuploidy on chromosomes 13, 18 and 21 (Chu et al, 2009). Human placentation displays many similarities with tumorigenesis, including rapid mitotic cell division, migration, angiogenesis and invasion, and escape from immune surveillance. Indeed, there are striking similarities in the DNA methylation pattern of tumor-associated genes between invasive trophoblast cell lines and first-trimester placenta and tumors (Christensen et al, 2009). This finding suggests that a distinct pattern of tumor-associated methylation can result in a series of epigenetic silencing events necessary for normal human placental invasion and function (Novakovic et al, 2008). Other studies using the placenta as a source suggest that the specific loss of imprinting because of altered methylation and subsequent gene expression can result in small for gestational age (SGA) newborns. Moreover, unbalanced expression of imprinted genes in IUGR placenta when compared with non-IUGR placenta was observed suggesting a differential expression pattern of imprinted genes as a possible biomarker for IUGR (Guo et al, 2008; McMinn et al, 2006). CHAPTER 5 Immunologic Basis of Placental Function and Diseases: the Placenta, Fetal Membranes, and Umbilical Cord The unique cytokine and hormonal milieu in utero may influence the trophoblast function and differentiation as well as immune cell regulation through histone posttranslational modification. In this regard, interferon γ produced by uNK cells and essential for spiral artery remodeling fails to induce MHC class II expression in trophoblast cells because of hypermethylation of regulatory class II MHC transactivator (CIITA) regions (Morris et al, 2002). This inability to upregulate classical MHC class II molecules by trophoblasts is essential for maintaining immune tolerance at the maternal-fetal interface. Moreover, the transcription factor regulating trophoblastic fusion protein syncytin, which is essential for the syncytialization of trophoblasts, is regulated by histone acetyl transferase and histone deacetylase activity (Chuang et al, 2006). miRNAs are small regulatory RNA molecules that can alter gene and protein expression without altering the underlying genetic code. Expression of miRNA is tissue specific, and several are expressed in the placenta. In placental pathology associated with preeclampsia, there is differential expression of miRNA (such as miR-210 and miR-182) compared with normal pregnancy placenta. This finding suggests that signature differences in placental miRNA and their detection in maternal serum may potentially be used as a biomarker for preeclampsia. Because implantation and early placentation is under the regulation of low oxygen tension, it is possible that miRNA are differentially expressed under different oxygen levels, as suggested by recent observations (Maccani and Marsit, 2009; Pineles et al, 2007). PLACENTAL DISEASES The placenta provides a wealth of retrospective information about the fetus and prospective information regarding the infant. Healthy development of the placenta requires efficient metabolic, immune, hormonal, and vascular adaptation by the maternal system as well as the fetus. Abnormal placentation and placental infections can lead to maternal or fetal anomalies as seen in preeclampsia, preterm birth, and SGA, which can have lifelong bearing on the development and health of infants. Maternal factors such as ascending infections, obesity, hypertension, genetic predisposition such as gene polymorphism of the pregnancy-compatible cytokine milieu, and environmental exposure could also contribute to the placental pathology. The following sections contain an abbreviated discussion of the pathogenesis of some of these placenta-associated disorders. PREECLAMPSIA Hypertensive disorders of pregnancy are enigmatic. They pose a major public health problem and affect 5% to 10% of human pregnancies. Preeclampsia is clinically associated with maternal symptoms of hypertension, proteinuria, and glomeruloendotheliosis. This disorder is strictly a placental condition because of its clearance after delivery. It causes morbidity and mortality in the mother, fetus, and newborn. Pregnancy-associated hypertension is defined as blood pressure greater than 140/90 mm Hg on at least two occasions and at 4 to 6 weeks apart after 20 weeks’ 47 gestation. Proteinuria is defined by excretion of 300 mg or more of protein every 24 hours or 300 mg/L or more in two random urine samples taken at least 4 to 6 hours apart (ACOG Committee on Practice Bulletins, 2002). The fetal problems most commonly associated with preeclampsia include fetal growth restriction, reduced amniotic fluid, and abnormal oxygenation (Sibai et al, 2005). However, the onset of clinical signs and symptoms can result in either near-term preeclampsia without affecting the fetus or its severe manifestation that is associated with low birthweight and preterm delivery (Vatten and Skjaerven, 2004). The heterogeneous manifestation of this disease is further confounded by preexisting maternal vascular disease, multifetal gestation, metabolic syndrome, obesity, or previous incidence of the disease. In addition, the pathophysiology of the disorder could differ from the onset before 24 weeks’ gestation and its diagnosis at later stages of pregnancy: Abnormal remodeling of spiral arteries and shallow trophoblast invasion are two hallmark features of preeclampsia. Preeclampsia is considered a two-stage disease where a poorly perfused placenta (stage I) causes the release of factors leading to maternal symptoms (stage II). However, it is also now being recognized that the maternal factors may contribute to programming of stage I of preeclampsia, suggesting that the intrinsic maternal factors stemming from genetic, behavioral, and physiologic conditions may contribute to placental pathology. Stage I initiated pathology may be particularly apparent in the oxidative stress-induced release of causative factors from the poorly perfused placenta and their effects on the maternal syndrome (Roberts and Hubel, 1999). Despite a poor mechanistic understanding of placental pathology leading to preeclampsia, several critical features are common to this disease. Multiple studies have shown that reduced vascular activity could be a major factor contributing to preeclampsia. In normal pregnancy, the circulating PlGF levels steadily increase in the first and second trimesters, peak at 29 to 32 weeks, and decline thereafter. However, free VEGF remains low and unchanged during this window. Reduced placental expression of VEGF and PlGF is consistently observed in preeclampsia. Furthermore, preeclampsia is frequently accompanied by enhanced circulation and placental expression of the antiangiogenic soluble VEGF receptor 1 (sFlt-1), which is a decoy receptor titrating out VEGFs and PlGF (Levine et al, 2004; Romero et al, 2008; Thadhani et al, 2004). A lack of available VEGF and increased sFlt-1 expression has been associated with trophoblast injury. The soluble form of endoglin (CD105), a coreceptor involved in TGF-β signaling is reported to enhance the antiangiogenic effects of sFlt-1. Soluble endoglin has been found to be elevated in the serum of preeclamptic women and is accompanied by an increased ratio of sFlt-1:PlGF and correlates with the severity of the disease. Soluble endoglin is thought to inhibit TGF-β1 signaling in endothelial cells and blocks activation of endothelial nitric oxide synthase and vasodilatation (Venkatesha et al, 2006). Several recent studies have suggested an increase in apoptosis within villous trophoblast from preeclampsia and IUGR deliveries (Allaire et al, 2000; Heazell and Crocker, 2008; Levy et al, 2002). Unlike normal pregnancy, villous placental explants 48 PART II Fetal Development from preeclamptic placenta have an increased sensitivity and susceptibility to apoptosis on exposure to proinflammatory cytokines, suggesting altered programming of apoptotic cascade pathway (Crocker et al, 2004; Levy et al, 2002). It is possible that incomplete spiral artery transformation resulting in reduced placental perfusion (stage I) in preeclampsia leads to focal regions of hypoxia with increase in apoptosis, oxidative stress, shedding of villous microparticles, and release of antiangiogenic factors such as sFlt-1.64 (Hung et al, 2002; Nevo et al, 2006; ­Redman and Sargent, 2000). Another pathway that may contribute to the etiology of preeclampsia is unscheduled and excessive activation of the complement cascade; this is highly likely as a result of the maternal immune system responding to paternal antigens and inflammation. However, in normal pregnancy the placenta expresses complement regulatory proteins such as DAF, CD55, and CD59 and may control activation of complement factors (Tedesco et al, 1993). Despite the positioning of complement inhibitory proteins for protective roles, increasing evidence supports the involvement of complement activation in the pathogenesis of preeclampsia (Lynch et al, 2008). Interestingly, recent in vitro studies suggest that hypoxia enhances placental deposition of the membrane attack complex and apoptosis in cultured trophoblasts (Rampersad et al, 2008). The upstream factors that trigger complement activation are not yet known. Recent studies also suggest increased serum levels of agonistic autoantibodies against angiotensin type 1 receptor (AT-1-AA) in preeclampsia as compared with healthy women (Zhou et al, 2008). Importantly, studies from our laboratory have shown that the full spectrum of preeclampsia-like symptoms can be reproduced in mice by injecting human preeclampsia serum containing subthreshold levels of AT-1-AA immunoglobulin G, suggesting that pregnancy serum contains some unknown causative factors. Therefore serum can be used as a blueprint to identify functional biomarkers for preeclampsia (Kalkunte et al, 2009). PRETERM BIRTH Preterm birth is the leading cause of infant morbidity and mortality in the world. Babies born before 37 weeks’ gestation are considered premature. In the United States, approximately 12.8% of births are preterm, and the rate of premature birth has increased by 36% since early 1980s (Martin et al, 2009). Babies from preterm birth face an increased risk of lasting disabilities such as mental retardation, learning and behavioral problems, autism, cerebral palsy, bronchopulmonary dysplasia, vision and hearing loss, and risk for diabetes, hypertension, and heart disease in adulthood. The majority of preterm deliveries are due to preterm labor. Other factors leading to premature birth are preterm premature rupture of membranes (PPROM), intervention for maternal or fetal problems, preeclampsia, fetal growth restriction, cervical incompetence, and antepartum bleeding. Additional risk factors for preterm birth include stress, occupational fatigue, uterine distention by polyhydramnios or multifetal gestation, systemic infection such as periodontal disease, intrauterine placental pathology such as abruption, vaginal bleeding, smoking, substance abuse, maternal age (<18 or >40 years), obesity, diabetes, thrombophilia, ethnicity, anemia, and fetal factors such as congenital anomalies and growth restriction. Activation of the hypothalamic-pituitary-adrenal (HPA) as a result of major maternal physical or psychological stress is thought to increase the release of corticotrophinreleasing hormone. In addition to the hypothalamus as a source of corticotrophin-releasing hormone, placental trophoblasts, amnion, and decidual cells also express this hormone during pregnancy. Corticotrophin hormone regulates the release of adrenocorticotropic hormone from pituitary and cortisol from adrenal glands, and it can also influence the activity of matrix metalloproteinases (MMPs). Premature activation of the HPA axis can eventually stimulate the prostaglandins, ultimately resulting in parturition via activation of proteases. In addition, activation of the HPA axis promotes the release of estrone, estradiol, and estriol that can activate the myometrium by increasing oxytocin receptors, prostaglandin activity, and enzymes such as myosin light chain kinase and calmodulin, which are responsible for muscle contraction. Concomitantly, progesterone withdrawal is expected with the raising concentration of myometrial estrogen receptors, further enhancing estrogen-induced myometrial activation and preterm birth (Dole et al, 2003; Grammatopoulos and Hillhouse, 1999; McLean et al, 1995). There is increasing evidence that approximately 50% of preterm births are associated with infection of the decidua, amnion, or chorion and amniotic fluid caused by either systemic or ascending genital tract infection. Both clinical and subclinical chorioamnionitis are implicated in preterm birth. Maternal or fetal inflammatory responses to chorioamniotic infection can trigger preterm birth. Activated neutrophils and macrophages and the release of cytokines IL-1β, IL-6, IL-8, tumor necrosis factor alpha (TNF-α) and granulocyte colony-stimulating factor can lead to an enhanced cascade of signaling activity, causing release of prostaglandins and expression of various MMPs of fetal membranes and the cervix. Furthermore, elevated levels of TNF-α and apoptosis are associated with PPROM. Non–infection-related inflammation caused by placental insufficiency and apoptosis can also cause preterm birth. In addition to augmented inflammatory responses to infections, pathogenic microbes (e.g. Staphylococcus, Streptococcus, Bacteroides, and Pseudomonas spp.) are thought to directly degrade fetal membranes by releasing proteases, collagenases, and elastases, produce phospholipase A2, and release endotoxin that stimulate uterine contractions and cause preterm birth (Goldenberg et al, 2000, 2008; Romero et al, 2006; Slattery and Morrison, 2002). The innate immune system and trophoblasts during pregnancy recognize bacterial and viral infections using TLRs. Placental transcripts for TLRs 1 to 10 have been detected in human placental tissue, and placental choriocarcinoma cell lines reportedly express TLR-2, TLR-4, and TLR-9 (Abrahams and Mor, 2005). Studies have demonstrated functionality for TLR-2, TLR-3, and TLR-4 in first- and third-trimester placental tissue (Patni et al, 2007). Decidual expression in humans has demonstrated functional receptors in term decidua of TLR-1, TLR-2, TLR-4, and TLR-6 (Canavan and Simhan, 2007). Our recent studies using mice have shown that extremely small CHAPTER 5 Immunologic Basis of Placental Function and Diseases: the Placenta, Fetal Membranes, and Umbilical Cord doses of the TLR-4 ligand lipopolysaccharide can cause preterm birth or fetal demise in pregnant IL-10–deficient mice by activating and promoting infiltration of uterine NK cells into the placenta and inducing apoptosis by secretion of TNF-α (Murphy et al, 2005, 2009). Similarly, activation of TLR-3 or TLR-9 has been shown to induce spontaneous abortion or preterm birth in IL-10–deficient pregnant mice that is attributed to immune infiltration and proinflammatory cascade in the placenta (Thaxton et al, 2009). Decidual hemorrhage leading to vaginal bleeding increases the risk for preterm birth and PPROM. Increased occult decidual hemorrhage, hemosiderin deposition, and retrochorionic hematoma formation is seen between 22 and 32 weeks’ gestation as a result of PPROM and preterm birth after preterm labor. The development of PPROM in the setting of abruption could be caused by high decidual concentration of tissue factors, which eventually generate thrombin. Thrombin activation as measured by serum thrombin-antithrombin III complex levels are elevated on preterm birth. Thrombin binds to decidual protease-­ activated receptors (PAR1 and PAR2), induces the production of IL-8 in decidua, attracts neutrophils, and promotes degradation of the fetal membrane MMPs that can result in PPROM (Lockwood et al, 2005; Salafia et al, 1995). Polyhydramnios is also a high risk factor for preterm birth. It was shown recently that exposure of IL-10– deficient pregnant mice to polychlorinated biphenyls, an environmental toxicant, can lead to preterm birth with IUGR. The IUGR was due to increased amniotic fluid volume (polyhydramnios) and placental insufficiency caused by poor spiral artery remodeling associated with reduced expression of water channel aquaporin-1 in the placenta (Tewari et al, 2009). Increasing evidence also suggests impaired vascular activity because of an increase in antiangiogenic factors such as sFlt-1 and decreased VEGF in PPROM and preterm birth (Kim et al, 2003). INTRAUTERINE GROWTH RESTRICTION IUGR is used to designate a fetus that has not reached its growth potential; it can be caused by fetal, placental, or maternal factors. Disparities between fetal nutritional or respiratory demands and placental supply can result in impaired fetal growth. Chromosomal abnormalities (aneuploidy, partial deletions, gene mutation particularly on the gene for insulin-like growth factors), congenital abnormalities, multiple gestation, and infections can also result in IUGR. Preterm birth, preeclampsia, and abruption because of placental ischemia can result in IUGR. Reduced placental weight with identifiable placental histologic abnormalities (e.g, impaired development or obstruction in uteroplacental vasculature, chronic abruption, chronic infections, maternal floor infarction, thrombosis in uteroplacental vasculature or fetoplacental vasculature) are common findings in IUGR. In addition, a single umbilical artery, velamentous umbilical cord insertion, bilobate placenta, circumvallate placenta, and placental hemangioma are some of the other structural anomalies seen in the placenta. Maternal factors such as nutritional deficiency; severe anemia; pulmonary disease leading to maternal hypoxemia; smoking; exposure to 49 toxins such as warfarin, anticonvulsants, folic acid antagonists, and caffeine; and pregnancies conceived through assisted reproductive techniques have a higher prevalence of IUGR. IUGR results in the birth of an infant who is SGA. Mortality and morbidity are increased in SGA infants compared with those who are appropriate for gestational age. SGA infants at birth have many clinical problems that include impaired thermoregulation; difficulty in cardiopulmonary transition with perinatal asphyxia, pulmonary hypertension, hypoglycemia, polycythemia and hyperviscosity; impaired cellular immune function; and increased risk for perinatal mortality. SGA infants in their childhood and adolescence are at higher risk for impaired physical growth and neurodevelopment. Adolescents born SGA at term were reported to have learning difficulties with attention deficits. Cognitive performance is generally lower in SGA infants at the ages of 1 to 6 years compared with those whoe are appropriate for gestational age. Adults who were SGA infants could be at higher risk for ischemic heart diseases and essential hypertension (Figueras et al, 2007; Kaijser et al, 2008; Lapillonne et al, 1997; Norman and Bonamy, 2005; O’Keefe et al, 2003; Spence et al, 2007). FETAL MEMBRANES AND THEIR PATHOLOGY The fetal tissue–derived membrane structure surrounds the fetus and forms the amniotic cavity. This membrane, which lacks both vascular and nerve cells, is composed of an inner layer adjacent to the amniotic fluid and is called the amnion. The outer layer that is attached to the decidua is called the chorion. Amnion is composed of inner epithelial cells, and the mesenchymal cell layer is composed of fibroblast and an outer spongy layer. Intact, healthy fetal membranes are required for normal pregnancy outcome. Chorion is composed of an outer reticular cell layer composed of fibroblasts and macrophages and an inner cytotrophoblast layer. The elasticity and strength of these membranes are maintained by extracellular matrix proteins such as collagens, fibronectin, laminins, and the activity of MMP-2 and MMP-9 and their inhibitors until the initiation of parturition when the membranes are susceptible to rupture. During parturition, when contractions begin or membranes rupture, MMP activity in the amnion and chorion increases with a concurrent fall in tissue inhibitors of metalloproteinases. This change is followed by apoptosis in the amnion epithelial and chorion trophoblast layers of fetal membrane. Interestingly, some evidence suggests that fetal membranes have antimicrobial activity and are known to express TLR-2 and TLR-4, which are pattern recognition receptors and help in initiating a protective host response to infection. The histopathology of amnion and chorion includes infections, amniotic fluid contaminants, and fetal diseases. In addition to the membranes, whose infection can lead to chorioamnionitis, another vulnerable portal for infection to occur is the placental intervillous space and fetal villi that provide hematogenous access. Hematogenous sources of infection are typically associated with inflammation of villi (villitis) and intervillous space (intervillositis). Viral pathogens (cytomegalovirus, HIV, herpes simplex virus) 50 PART II Fetal Development commonly produce hematogenous infection of the placenta in addition to bacteria, spirochetes, fungi, and protozoa (Gersell, 1993; Goldenberg et al, 2000; Lahra and Jeffery, 2004). UMBILICAL CORD The connecting cord from the developing embryo or fetus to the placenta is the umbilical cord, or funiculus umbilicalis. During prenatal development in humans, the normal umbilical cord contains two umbilical arteries and one umbilical vein buried within Wharton’s jelly. The umbilical vein supplies the fetus with oxygenated blood from the placenta while the arteries return the deoxygenated, nutrient-depleted blood to the placenta. In the fetus, the umbilical vein branches into the ductus venosus and another branch that joins the hepatic portal vein. Shortly after parturition, physiologic processes cause the Wharton’s jelly to swell with the collapse of blood vessels, resulting in a natural halting of the flow of blood. Within the infant, the umbilical vein and ductus venosus close and degenerate into remnants known as the round ligament of the liver and the ligamentum venosum, while the umbilical arteries degenerate into what is known as medial umbilical ligaments. Abnormalities associated with the umbilical cord can affect both the mother and the child. Pathology of umbilical cord is generally grouped as congenital remnants, infections, meconium, and masses. Abnormalities that have clinical significance are nuchal cord, single umbilical artery, umbilical cord prolapse, umbilical cord knot, umbilical cord entanglement, vasa previa, and velamentous cord insertion. Common intrauterine infections can result in the umbilical cord being invaded by fetal cells and bacteria infiltrated from the decidua to amniotic fluid, or they can elicit fetal inflammatory response. Umbilical cord inflammation, known as funisitis or vasculitis, poses a higher risk for development of neurologic compromise in the fetus. Funisitis is predictive of a lower median Bayley psychomotor developmental index in infants. Meconium pigment at high concentrations can damage the umbilical cord by triggering apoptosis of smooth muscle cells. Vascular necrosis caused by meconium is associated with oligohydramnios, low Apgar scores, and significant neurodevelopmental delay. Interruption of normal blood flow in the cord can cause prolonged hypoxia in utero. Clamping of the umbilical cord within minutes of birth is hospitalbased obstetric practice. A Cochrane review studying the effects of the timing of umbilical cord clamping in hospitals showed that infants whose cord clamping occurred later than 60 seconds after birth had a significantly higher risk of neonatal jaundice requiring phototherapy. However, randomized, controlled studies have shown that delayed cord clamping in preterm infants reduces the incidence of intraventricular hemorrhage and late-onset sepsis. Furthermore, premature clamping can increase the risk of ischemia and hypovolemic shock, which can lead to fetal complications (McDonald and Middletone, 2008; Mercer et al, 2006). SUGGESTED READINGS Aluvihare VR, Kallikourdis M, Betz AG: Regulatory T cells mediate maternal tolerance to the fetus, Nat Immunol 5:266-271, 2004. Ashkar AA, Di Santo JP, Croy BA: Interferon γ contributes to initiation of uterine vascular modification, decidual integrity, and uterine natural killer cell maturation during normal murine pregnancy, J Exp Med 192:259-269, 2000. Baker DJP: In utero programming of chronic disease, Clin Sci 95:115-128, 1998. Christensen BC, Houseman EA, Marsit CJ, et al: Aging and environmental exposures alter tissue-specific DNA methylation dependent upon CpG island context, PLoS Genet 5:e1000602, 2009. Goldenberg RL, Culhane JF, Iams JD, et al: Epidemiology and causes of preterm birth, Lancet 371:75-84, 2008. Hanna J, Goldman-Wohl D, Hamani Y, et al: Decidual NK cells regulate key developmental processes at the human fetal-maternal interface, Nat Med 12:1065-1074, 2006. Kalkunte S, Mselle TF, Norris WE, et al: VEGF C facilitates immune tolerance and endovascular activity of human uterine NK cells at the maternal-fetal interface, J Immunol 182:4085-4092, 2009. Moffett A, Loke C: Immunology of placentation in eutherian mammals, Nat Rev Immunol 6:584-594, 2006. Murphy SP, Hanna NN, Fast LD, et al: Evidence for participation of uterine ­natural killer cells in the mechanisms responsible for spontaneous preterm labor and delivery, Am J Obstet Gynecol 200:308, 2009. Paria BC, Reese J, Das SK, et al: Deciphering the cross-talk of implantation: advances and challenges, Science 296:2185-2188, 2002. Roberts JM, Hubel CA: Is oxidative stress the link in the two–stage model of preeclampsia? Lancet 354:788-789, 1999. Slattery MM, Morrison JJ: Preterm delivery, Lancet 360:1489-1497, 2002. Thadhani R, Sachs BP, Epstein FH, et al: Circulating angiogenic factors and the risk of preeclampsia, N Engl J Med 350:672-683, 2004. Complete references and supplemental color images used in this text can be found online at www.expertconsult.com C H A P T E R 6 Abnormalities of Fetal Growth Rebecca Simmons Fetal growth and size at birth are critical in determining mortality and morbidity, both immediately after birth and in later life. Normal fetal growth is determined by a number of factors, including genetic potential, the ability of the mother to provide sufficient nutrients, the ability of the placenta to transfer nutrients, and intrauterine hormones and growth factors. The pattern of normal fetal growth involves rapid increases in fetal weight, length, and head circumference during the last half of gestation. During the last trimester, the human fetus accumulates significant amounts of lipid. The birthweight for gestational measurements among populations has been shown to increase over time; therefore, standards for normal fetal growth require periodic reevaluation for clinical relevance. These increases in birthweight for gestational age over time are attributed to improvements in living conditions and maternal nutrition and changes in obstetric management. Variations in fetal growth have been identified in diverse populations and are associated with geographic locations (sea level versus high altitude), populations (white, African American, Latino), maternal constitutional factors, parity, maternal nutrition, fetal gender, and multiple gestations. In this chapter, we discuss these factors in greater detail and critically review the long-term effects of abnormal fetal growth. DEFINITIONS Most approaches for defining fetal growth use gestational age-based norms. The duration of pregnancy has become an integral component of prenatal growth assessment, and all currently prevailing definitions of fetal growth are specific for gestational age. Assessing the gestational age accurately, however, can be challenging. Any error in dating will lead to misclassification of the infant, which can have significant clinical implications. In many instances, the method of gestational age determination has contributed to variations in the gestational age–specific reference growth curves. For example, some nomograms are based on approximating the gestational age to the nearest week, whereas others use the completed weeks. The birthweight charts are also affected by other variables that may limit their reliability. Many of these variables, such as fetal gender, race, parity, birth order, parental size, and altitude, contribute to the normal biologic variations in human fetal growth. There is continuing controversy regarding whether the reference growth charts should be customized by multiple variables or developed from the whole population. The customized approach predicts the optimal growth in an individual pregnancy and therefore specifically defines suboptimal growth for that pregnancy. However, it has been argued that such an approach can lead to a profusion of standards and might not contribute to improving the outcome of infants who are small for gestational age (SGA). In recognition of the utility of a national standard, a population-based reference chart for fetal growth has been developed from all the singleton births (more than 3 million) in the United States in 1991 (Alexander, 1996). More recently, a similar national population-based fetal growth chart, which is also sex specific, has been developed in Canada (Kramer et al, 2001). There is insufficient evidence about whether one approach is superior to the other in improving the perinatal outcome. There is no universal agreement on the classification of an SGA infant. Various definitions appear in the medical literature, making comparisons between studies difficult. In addition, investigators have shown that the prevalence of fetal growth restriction varies according to the fetal growth curve used (Alexander et al, 1996). The most common definition of SGA refers to a weight less than the 10th percentile for gestational age or birthweight less than 2 standard deviations (SDs) from the mean. Some investigators also use measurements less than the 3rd percentile to define SGA. However, these definitions do not make a distinction among infants who are constitutionally small, growth restricted and small, and not small but growth restricted relative to their potential. As an example, as many as 70% of fetuses with a weight less than the 10th percentile for gestational age at birth are small simply because of constitutional factors such as female sex or maternal ethnicity, parity, or body mass index; they are not at high risk of perinatal mortality or morbidity. In contrast, true fetal growth restriction is associated with numerous perinatal morbidities. This has clinical relevance to perinatologists and neonatologists, because many of the tiniest premature neonates in the neonatal intensive care units are probably growth restricted. McIntire et al (1999) reported a threshold of increased adverse outcomes in infants born with measurements less than the 3rd percentile and suggested that this level of restriction represents a clinically relevant measurement. Other researchers have found higher rates of neonatal complications when the 15th percentile of birthweight is used as a cutoff level (Seeds and Peng, 1998). There is an important distinction in identifying the fetus with intrauterine growth restriction (IUGR), and the fetus that is constitutionally small (i.e., SGA). IUGR is a condition in which a fetus is unable to achieve its genetically determined potential size and represents a deviation and a reduction in the expected fetal growth pattern. IUGR complicates approximately 5% to 8% of all pregnancies and 38% to 80% of all neonates with low birthweight (LBW). This discrepancy underscores the fact that no uniform definition of IUGR exists. Even when a normal intrauterine growth pattern is established for a population, somewhat arbitrary criteria are used to define growth restriction. 51 52 PART II Fetal Development PATTERNS OF ALTERED GROWTH Neonates with intrauterine growth retardation can be classified as demonstrating either symmetrical or asymmetrical growth. Infants with symmetric IUGR have reduced weight, length, and head circumference at birth. Weight (and then length) of infants with asymmetric growth retardation is affected, with a relatively normal or “headsparing” growth pattern. Factors intrinsic to the fetus in general cause symmetrical growth restriction, whereas asymmetric IUGR is often associated with maternal medical conditions such as preeclampsia, chronic hypertension, and uterine anomalies. Asymmetric patterns generally develop during the third trimester, a period of rapid fetal growth. However, now that fetal surveillance is more common, asymmetric growth restriction is often diagnosed in the second trimester. Furthermore, many extremely premature neonates (<750 g) probably have IUGR. Factors that are well recognized to limit the growth of both the fetal brain and body include chromosomal anomalies (e.g., trisomies), congenital infections (toxoplasmosis, rubella, cytomegalovirus, herpes simplex, malaria, HIV, and parvovirus), dwarf syndromes, and some inborn errors of metabolism. Cardiac and renal structural anomalies are common fetal conditions associated with SGA. These conditions retard fetal growth primarily by impaired cell proliferation. Recognized causes of IUGR are listed in Table 6-1. Etiologies of Fetal Growth Restriction The epidemiology of fetal growth restriction varies internationally (Keirse, 2000). In developed countries, the most frequently identified cause of growth restriction is smoking, whereas in developing countries, maternal nutritional TABLE 6-1 Causes of Intrauterine Growth Restriction Genetic Inheritance, chromosomal abnormalities, fetal gender Maternal ­constitutional effects Low maternal pre-pregnancy weight, low pregnancy weight gain, ethnicity, socioeconomic status, history of intrauterine growth restriction Nutrition Low pre-pregnancy weight (body mass index), low pregnancy weight gain, malnutrition (macronutrients, micronutrients), maternal anemia Infections TORCH infections (toxoplasmosis, rubella, cytomegalovirus, syphilis) Decreased O2-carrying capacity High altitude, maternal congenital heart disease, hemoglobinopathies, chronic anemia, maternal asthma Uterine and ­placental anatomy Abnormal uterine anatomy, uterine fibroid, vascular abnormalities (single umbilical artery, velamentous umbilical cord insertion, twin-twin transfusion), placenta previa, placental abruption Uterine and ­placental ­function Maternal vasculitis (system lupus erythematosus), decreased uteroplacental perfusion, maternal illness (preeclampsia, chronic hypertension, diabetes, renal disease) Toxins Tobacco, ethanol, lead, arsenic factors (prepregnancy weight, maternal stature) and infections (malaria) are the leading identified causes (Krampl, 2000; Robinson et al, 2000). In addition, in developing countries there is a direct correlation between the incidence of LBW (<2500 g) and IUGR. In developing countries, the high incidence of infants with LBW is almost exclusively caused by the incidence of IUGR. Data from developed countries show the opposite, with rates of LBW being explained almost exclusively by prematurity rates. In the United States, the cause of IUGR is identified in approximately 40% of cases, with the remaining cases labeled as idiopathic. Numerous factors have been identified as influencing size at birth. In a simplified manner, these factors can be grouped as fetal, placental, or maternal in origin. These factors will be discussed in detail in the following sections. Fetal Causes of Growth Restriction Fetal factors affecting growth include gender, familial genetic inheritance, chromosomal abnormalities, and dysmorphic syndromes. In one large, population-based study, the frequency of IUGR among infants with congenital malformations was 22%. The majority of the infants affected had chromosomal abnormalities. Other studies have similarly found that fetal growth restriction is more common among infants with malformations. Fetal gender also influences size, with male infants showing greater intrauterine growth than female infants (Glinianaia et al, 2000; Skjaerven et al, 2000; Thomas et al, 2000). Placental Causes of Growth Restriction In mammals, the major determinant of intrauterine growth is the placental supply of nutrients to the fetus (Fowden et al, 2006). In many species, fetal weight near term is positively correlated to placental weight, as a proxy measure of the surface area for maternal-fetal transport of nutrients. Fetal weight near term is positively correlated to placental weight, and the nutrient transfer capacity of the placenta depends on its size, morphology, blood flow, and transporter abundance (Fowden et al, 2006). In addition, placental synthesis and metabolism of key nutrients and hormones influences the rate of fetal growth (Fowden and Forhead, 2004). Changes in any of these placental factors can, therefore, affect intrauterine growth; however, the fetus is not just a passive recipient of nutrients from the placenta. The fetal genome exerts a significant acquisitive drive for maternal nutrients through adaptations in the placenta, particularly when the potential for fetal and placental growth is compromised. Placental maturation at the end of pregnancy is associated with an increase in substrate transfer, a slowing (but not cessation) of placental growth, and a plateau in fetal growth near term (Fox, 1997). Abnormalities of placental growth, senescence, and infarction have been shown to affect fetal growth. The placentas from pregnancies complicated by poor fetal growth have a higher incidence of vascular damage and abnormalities (Pardi et al, 1997). Fetal size and placental growth are directly related, and placentas from pregnancies yielding growth-restricted infants demonstrate a higher incidence of smallness and abnormality than do those from CHAPTER 6 Abnormalities of Fetal Growth pregnancies with appropriately grown infants. The difference in size is seen even in a comparison of placentas associated with growth-restricted infants and those associated with appropriate for gestational age (AGA) infants of the same birthweight (Heinonen et al, 2001). Placental growth in the second trimester correlates with placental weight and function and, thus, with weight at birth (Godfrey et al, 1996). Clinical conditions associated with reduced placental size (and subsequent reduced fetal weight) include maternal vascular disease, uterine anomalies (uterine fibroids, abnormal uterine anatomy), placental infarctions, unusual cord insertions, and abnormalities of placentation. Multiple gestations are associated with greater risk for fetal growth restriction. The higher risk stems from crowding and from abnormalities with placentation, vascular communications, and umbilical cord insertions. Divergence in fetal growth appears from approximately 30 to 32 weeks in twin gestation compared with singleton pregnancies (Alexander et al, 1996; Glinianaia et al, 2000; ­Skjaerven et al, 2000). Others have identified differences in fetal growth between twins and singletons as occurring earlier in gestation, at approximately 21 weeks’ gestation (Devoe and Ware, 1995). Larger effects on fetal growth are seen with increasing number of fetal multiples. Abnormalities in placentation are also more common with multiple gestations (Benirschke, 1995). Monochorionic twins can share placental vascular communication (twin-twin transfusion), leading to fetal growth restriction during gestation. Fetal competition for placental transfer of nutrients raises the incidence of growth restriction and discordance in growth between fetuses. The rate of birthweights less than the 5th percentile is higher in monochorionic twins. Placental growth is restricted in utero because of limitation in space, leading to a higher incidence of placenta previa in multiple-gestation pregnancies. In addition, abnormalities in cord insertions (marginal and velamentous cord insertions) and occurrence of a single umbilical artery are more frequently found in multiple gestations. The higher incidence of growth restriction in multiple-gestation pregnancies is strongly associated with monochorionic gestations, the presence of vascular anastomoses, and discordant fetal growth (Hollier et al, 1999; Sonntag et al, 1996; Victoria et al, 2001). Placentas of smaller fetuses with discordant growth are significantly smaller than those of their larger twin counterparts (Victoria et al, 2001). Investigators have shown an effect of altitude on fetal growth, with infants born at high altitudes having lower birthweights (Galan et al, 2001). Differences in fetal growth are detected from approximately 25 weeks’ gestation with pregnancies at 4000 m. In these high-altitude pregnancies, the abdominal circumference is most affected (Krampl et al, 2000). At tremendously high altitudes, the incidence of LGA infant births is markedly reduced. In the United States (and at less severe altitudes), infants born at higher altitudes are lighter at birth, but those differences are not pronounced. Interestingly, investigators have shown that adaptation to high altitude during pregnancy is also possible. Tibetan infants have higher birthweights than infants of more recent immigrants of ethnic Chinese living at the same high-altitude (2700 to 4700 m) region of Tibet (Moore et al, 2001). Tibetan infants also have less IUGR than do infants born to more recent immigrants to the area. 53 Maternal Causes of Growth Restriction Maternal health conditions associated with chronic decreases in uteroplacental blood flow (maternal vascular diseases, preeclampsia, hypertension, maternal smoking) are associated with poor fetal growth and nutrition. Preeclampsia has been shown to be associated with fetal growth restriction (Ødegård et al, 2000; Spinillo et al, 1994; Xiong et al, 1999). Investigators have shown that the extent of growth restriction correlates with the severity and the onset during pregnancy of the preeclampsia. Ødegård et al (2000) showed that fetuses exposed to preeclampsia from early in pregnancy had the most serious growth restriction, and more than half of these infants were born SGA. Chronic maternal diseases (cardiac, renal) may decrease the normal uteroplacental blood flow to the fetus and thus may also be associated with poor fetal growth (Spinillo et al, 1994). Maternal constitutional factors have a significant effect on fetal growth. Maternal weight (pre-pregnancy), maternal stature, and maternal weight gain during pregnancy are directly associated with maternal nutrition and correlate with fetal growth (Clausson et al, 1998; Doctor et al, 2001; Goldenberg et al, 1997; Mongelli and ­Gardosi, 2000). Numerous studies show that these findings are often confounded by highly associated cultural and socioeconomic factors. The woman with a previous SGA infant has a higher risk of a subsequent small infant (Robinson et al, 2000). Investigators have shown a higher incidence of SGA infants to be associated with lower levels of maternal education (Clausson et al, 1998). Parity of the mother also affects fetal size; nulliparous women having a higher incidence of SGA infants (Cnattingius et al, 1998). A large population-based study in Sweden found that women who were older than 30 years, were nulliparous, or had hypertensive disease were at increased risk of preterm and term growth-restricted infants. Studies have shown differential fetal growth for women of diverse ethnicities, with Latina and white women having higher rates of LGA infants, and African American women having a higher incidence of SGA infants ­(Alexander et al, 1999; Collins and David, 2009; Fuentes-Afflick et al, 1998). These gender and ethnic differences in birthweight become pronounced after 30 weeks’ gestation (Thomas et al, 2000). Investigators in California have shown that U.S.-born black women have higher rates of prematurity and LBW infants than do foreign-born black women. Other researchers have found that even among women with low risk of LBW infants (married, age 20 to 34 years, 13 or more years of education, adequate prenatal care, and absence of maternal health risk factors and tobacco or alcohol use), the risk of delivering an SGA infant is still higher for African American women than for white women (Alexander et al, 1999; Collins and David, 2009). It is unclear whether these differences in fetal growth are caused by inherent differences or by differential exposure to environmental factors. Maternal nutrition and supply of nutrients to the fetus affect fetal growth significantly, primarily in developing countries (Doctor et al, 2001; Godfrey et al, 1996; ­Neggers et al, 1997; Robinson et al, 2000; Zeitlin et al, 2001). Although numerous factors interact with and affect fetal 54 PART II Fetal Development development, maternal malnutrition is assumed to be a major cause of IUGR in developing countries. Furthermore, pre-pregnancy weight may be a potential marker for intergenerational effects on infant weight in developing countries. A woman’s birthweight has been shown to correlate with her infant’s weight as well as the placental weight during pregnancy. In the United States, Strauss and Dietz (1999) report that low maternal weight gain in the second and third trimesters is associated with a twofold risk of IUGR, whereas poor maternal weight gain in the first trimester has no such effect on fetal growth (Strauss and Dietz, 1999). These investigators also showed that older women (older than 35 years) and smokers were at increased risk of IUGR associated with lower weight gains in late pregnancy. Teen pregnancy represents a special condition in which fetal weight is highly influenced by maternal nutrition. Teen mothers (younger than 15 years) have been shown to have a higher risk for delivering a growth-restricted infant (Ghidini, 1996). Teen pregnancies are complicated by the additional nutritional needs of a pregnant mother, who is still actively growing, as well as by socioeconomic status of pregnant teens in developed countries (Scholl and ­Hediger, 1995). Maternal nutrition and maternal weight gain are adversely affected by inadequate or poorly balanced intake in conditions such as alcoholism, drug abuse, and poverty. The effects of micronutrients on pregnancy outcomes and fetal growth have been less well studied. Maternal intake of certain micronutrients has also been found to affect fetal growth. Zinc deficiency has been associated with fetal growth restriction and other abnormalities, such as infertility and spontaneous abortion (Jameson, 1993; Shah and Sachdev, 2001). In addition, dietary intake of vitamin C during early pregnancy has been shown to be associated with an increase in birthweight (Mathews et al, 1999). Others have shown strong associations between maternal intake of folate and iron and infant and placental weights (Godfrey et al, 1996). In developing countries, the effects of nutritional deficiencies during pregnancy are more prevalent and easier to detect. Rao et al (2001) have estimated that one third of infants in India are born weighing less than 2500 g, mainly because of maternal malnutrition. These investigators have shown significant associations between infant birthweight and maternal intake of milk, leafy greens, fruits, and folate during pregnancy. Although toxins such as cigarette smoke and alcohol have a direct effect on placental function, they may also affect fetal growth through an associated compromise in maternal nutrition. Other environmental toxins (lead, arsenic, mercury) are associated with IUGR and are believed to affect fetal growth by entering the food chain and depleting body stores of iron, vitamin C, and possibly other nutrients (Iyengar and Nair, 2000; Srivastava et al, 2001). Numerous studies have shown associations between birthweight and maternal intake of macronutrients and micronutrients, but the effects of nutritional supplements used during pregnancy on fetal growth are equivocal (de Onis et al, 1998; Jackson and Robinson, 2001; Rush, 2001; Say et al, 2003). This finding is underscored by the results of a recent, large, double-blind, randomized controlled trial including 1426 pregnancies in rural Burkina Faso (Roberfroid et al, 2008). Pregnant women were randomly assigned to receive either iron and folic acid or the ­UNICEF-WHO-UNU* international multiple micronutrient preparation (UNIMMAP) daily until 3 months after delivery, with the UNIMMAP thereafter in both groups. Birthweight was increased by 52 g, and length was increased by 3.6 mm. Unexpectedly, the risk of perinatal death was marginally significantly increased in the UNIMMAP group (odds ratio, 1.78; 95% confidence interval, 0.95 to 3.32; p = 0.07). Maternal socioeconomic status and ethnicity have also been identified as risk factors for IUGR and poor health outcomes in infants. Numerous investigators have shown a significant effect of socioeconomic status on birth outcomes, including fetal growth restriction, in both developing and developed countries (Wilcox et al, 1995). In the United States, low levels of maternal and paternal education, certain maternal and paternal occupations, and low family income are associated with lower birthweights in children of African American and white women (Parker et al, 1994). In a large population-based study in Sweden, investigators have shown a higher incidence of fetal growth restriction in association with low maternal education (Clausson et al, 1998). Researchers have also shown that rates of compromised birth outcome are higher among African American women than among Mexican American and non-Hispanic white women (Collins and Butler, 1997; Frisbie et al, 1997; Thomas et al, 2000). Some of these studies also show that the risk of IUGR is higher in women without medical insurance. In the United States, the incidence of IUGR is significantly higher among African American women than among white women; this higher incidence is seen even among African American women with higher socioeconomic status (Alexander et al, 1999). In a study in Arizona, the incidence of IUGR was found to be lower in Mexican American women than in white women (Balcazar, 1994). Other researchers have shown that Mexican-born immigrants in California have better perinatal outcomes (including birthweight) than African Americans and U.S.-born women of Mexican descent (Fuentes-Afflick et al, 1998). The reasons for this apparent paradox are unclear, but one postulate is the tendency of recent immigrants to maintain the favorable nutritional and behavioral characteristics of their country of origin (Guendelman and English, 1995). These studies support the speculation that the differences in fetal growth between groups do not reflect inherent differences in fetal growth, but rather stem from inequalities in nutrition, health care, and other environmental factors (Keirse, 2000; Kramer et al, 2000). Smoking Cigarette smoking is consistently found to adversely affect intrauterine growth in all studies in which this factor is considered. In developed countries, cigarette smoking is the single most important cause of poor fetal growth (Kramer et al, 2000). The incidence of IUGR in smokers *United Nations Children’s Fund, World Health Organization, United Nations University. CHAPTER 6 Abnormalities of Fetal Growth is estimated to be threefold to 4.5-fold higher than in nonsmokers (Nordentoft et al, 1996). Cigarette smoking has a significant effect on abdominal circumference and fetal weight, but not on head circumference (Bernstein et al, 2000). Lieberman et al (1994) reported that cigarette smoking also appears to have a dose-dependent effect on the incidence of IUGR, with this effect being seen especially with heavy smoking and smoking during the third trimester. These investigators have shown that if women stop smoking during the third trimester, their infants’ birthweights are indistinguishable from those of infants born to the normal population. Other researchers have shown that even a reduction in smoking is associated with improved fetal growth (Li et al, 1993; Walsh et al, 2001). Numerous potential causes of the effects of smoking on fetal growth have been suggested, including direct effects of nicotine on placental vasoconstriction, decreased uterine blood flow, higher levels of fetal carboxyhemoglobin, fetal hypoxia, adverse maternal nutritional intake, and altered maternal and placental metabolism (Andres and Day, 2000; Pastrakuljic et al, 1999). SHORT-TERM OUTCOMES IUGR alters many physiologic and metabolic functions in the fetus and neonate that result in a number of morbidities. A large cohort study of 37,377 pregnancies found a fivefold to sixfold greater risk of perinatal death for both preterm and term fetuses that had IUGR (Cnattingius et al, 1998; Lackman et al, 2001; Mongelli and Gardosi, 2000). Predictive factors for perinatal mortality in preterm fetuses with IUGR reveals that of all antenatal factors examined, only oligohydramnios and abnormal umbilical artery Dopplers with absent or reversed diastolic flow were predictive of perinatal mortality (Scifres et al, 2009). Although the growth-restricted fetus may show symmetric or asymmetric growth at birth, it is unclear whether the proportionality of the fetus with IUGR truly affects outcomes or is related to the timing or the severity of the insult. Lin et al (1991) found that symmetric IUGR resulted in higher levels of prematurity and higher rates of neonatal morbidity. In contrast, Villar et al (1990) have shown that infants with asymmetric IUGR have higher morbidity rates at birth. They found that infants with low ponderal index measurements (which they defined as Weight ÷ Length3) had a higher risk of low Apgar scores, long hospitalization, hypoglycemia, and asphyxia at birth than infants with symmetric IUGR. There is evidence to suggest that infants with asymmetric IUGR show better gains in weight and length in the postnatal period than symmetrically restricted infants (May et al, 2001). Other investigators propose that IUGR represents a continuum, with symmetric IUGR occurring as the severity of growth retardation increases. Data also suggest that the more severe the growth restriction, the worse the neonatal outcomes, including risk of stillbirth, fetal distress, neonatal hypoglycemia, hypocalcemia, polycythemia, low Apgar scores, and mortality (Kramer et al, 1990; Spinillo et al, 1995). Fetal growth restriction is associated with intrauterine demise. Almost 40% of term stillbirths and 63% of preterm stillbirths are SGA (Mongelli and Gardosi, 2000). Both short-term and long-term effects of abnormalities 55 in SGA fetuses have been described. Perinatal mortality for intrauterine SGA infants is higher overall than that for appropriately grown term and preterm infants (Clausson et al, 1998). The risk of perinatal death is estimated to be fivefold to sixfold greater for both preterm and term fetuses with IUGR (Lackman et al, 2001). Overall, intrauterine death, perinatal asphyxia, and congenital anomalies are the main contributing factors to the higher mortality rate in SGA infants. The effects of acute fetal hypoxia may be superimposed on chronic fetal hypoxia, and placental insufficiency may be an important etiologic factor in these outcomes. Investigators have described higher incidences of low Apgar scores, umbilical artery acidosis, need for intubation at delivery, seizures on the first day of life, and sepsis in SGA infants (McIntire et al, 1999). The incidence of adverse perinatal effects correlates with the severity of the growth restriction, the highest rates of respiratory distress syndrome, metabolic abnormalities, and sepsis being found in the most severely growth-restricted infants (Spinillo et al, 1995). As previously described, Villar et al (1990) reported that infants with asymmetric IUGR and low ponderal index measurements had a higher risk of low Apgar scores, hypoglycemia, asphyxia, and long hospitalization. Preterm infants with growth abnormalities have a much higher risk of adverse outcomes. Preterm SGA infants have a higher incidence of a number of complications, including sepsis, severe intraventricular hemorrhage, respiratory distress syndrome, necrotizing enterocolitis, and death, than do normally grown preterm infants (Gortner et al, 1999; McIntire et al, 1999; Simchen et al, 2000). In addition, SGA infants have a higher incidence of chronic lung disease at corrected gestational ages of 28 days and 36 weeks. Neonatal hypoglycemia and hypothermia occur more frequently in growth-restricted infants (Doctor et al, 2001). These metabolic abnormalities presumably occur from decreased glycogen stores, inadequate lipid stores, and impaired gluconeogenesis in the growth-restricted neonate. Growth-restricted neonates have inadequate fuel stores and are at increased risk for hypoglycemia during fasting, and these risks are increased in preterm SGA infants. Infants with IUGR also have a higher incidence of hypocalcemia, with the incidence correlating strongly with the severity of growth restriction (Spinillo et al, 1995). Developmental Outcomes: Early Childhood Neurologic outcomes, including intellectual and neurologic function, are affected by growth restriction. Overall, neurologic morbidity is higher for SGA infants than for AGA infants. Without identified perinatal events, SGA infants have a higher incidence of long-term neurologic or developmental handicaps. Investigators have found the incidence of cerebral palsy to be greater in IUGR infants than in a population with normal fetal growth (Blair and Stanley, 1990; Spinillo et al, 1995; Uvebrant and Hagberg, 1992). SGA infants born at term appear to have double or triple the risk for cerebral palsy, between 1 to 2 per 1000 live births and 2 to 6 per 1000 live births (Goldenberg et al, 1998). The rate of cerebral palsy is also higher in preterm growth-restricted infants than in preterm infants with appropriate fetal growth (Gray et al, 2001). At 7 years 56 PART II Fetal Development of age, children whose birth was associated with hypoxiarelated factors had a higher risk for adverse neurologic outcomes. Infants with symmetric IUGR, or perhaps more severe restriction, were at higher risk than infants with asymmetric IUGR. Other researchers have shown higher rates of learning deficits, lower intelligence quotient scores, and increased behavioral problems in children with a history of fetal growth restriction, even at 9 to 11 years of age (Low et al, 1992). LONG-TERM CONSEQUENCES: THE DEVELOPMENTAL ORIGINS OF ADULT DISEASE Programming The period from conception to birth is a time of rapid growth, cellular replication and differentiation, and functional maturation of organ systems. These processes are highly sensitive to alterations in the intrauterine milieu. The term programming describes the mechanisms whereby a stimulus or insult at a critical period of development has lasting or lifelong effects. The “thrifty phenotype” hypothesis proposes that the fetus adapts to an adverse intrauterine milieu by optimizing the use of a reduced nutrient supply to ensure survival; but because this adaptation favors the development of certain organs over that of others, it leads to persistent alterations in the growth and function of developing tissues (Hales and Barker, 1992). In addition, although the adaptations may aid in survival of the fetus, they become a liability in situations of nutritional abundance. Epidemiology It has been recognized for nearly 70 years that the early environment in which a child grows and develops can have long-term effects on subsequent health and survival ­(Kermack, 1934). The landmark cohort study of 300,000 men by Ravelli et al (1976) showed that men who were exposed in utero to the effects of the Dutch famine of 1944 and 1945 during the first half of gestation had significantly higher obesity rates at the age of 19 years. Subsequent studies demonstrated relationships among LBW, the later development of cardiovascular disease (Barker et al, 1989), and impaired glucose tolerance (Fall et al, 1995) in men in England. Men who were smallest at birth (2500 g) were nearly sevenfold more likely to have impaired glucose tolerance or type 2 diabetes than those who were largest at birth. Barker et al (1993) also found a similar relationship between lower birthweight and higher systolic blood pressure and triglyceride levels. Valdez et al (1994) observed a similar association between birthweight and subsequent glucose intolerance, ­hypertension, and hyperlipidemia in a study of young adult Mexican American and non-Hispanic white men and women participants in the San Antonio Heart Study. Normotensive individuals without diabetes whose birthweights were in the lowest tertile had significantly higher rates of insulin resistance, obesity, and hypertension than subjects whose birthweights were normal. In the Pima Indians, a population with extraordinarily high rates of type 2 diabetes, McCance et al (1994) found that the development of diabetes in the offspring was related to both extremes of birthweight. In their study, the prevalence of diabetes in subjects 20 to 39 years old was 30% for those weighing less than 2500 g at birth, 17% for those weighing 2500 to 4499 g, and 32% for those weighing 4500 g or more. The risk of developing type 2 diabetes was nearly fourfold higher for those whose birthweight was less than 2500 g. Other studies of populations in the United States (Curhan et al, 1996), Sweden (Lithell et al, 1996; ­McKeigue et al, 1998), France (Jaquet et al, 2000; Leger et al, 1997), Norway (Egeland et al, 2000), and Finland (Forsen et al, 2000) have all demonstrated a significant correlation between LBW and the later development of adult diseases. Studies controlling for the confounding factors of socioeconomic status and lifestyle have further strengthened the association between LBW and a higher risk of coronary heart disease, stroke, and type 2 diabetes. In 1976, the Nurses’ Health Study was initiated, and a large cohort of American women born from 1921 to 1946 established. The association between LBW and increased risks of coronary heart disease, stroke, and type 2 diabetes remained strong even after adjustment for lifestyle factors such as smoking, physical activity, occupation, income, dietary habits, and childhood socioeconomic status (Rich-Edwards et al, 1999). Role of Catch-up Growth Many studies have suggested that the associations between birth size with later disease can be modified by body mass index (BMI) in childhood. The highest risk for the development of type 2 diabetes is among adults who were born small and become overweight during childhood (Eriksson et al, 2000). Insulin resistance is most prominent in Indian children who were SGA at birth, but had a high fat mass at 8 years of age (Bavdekar et al, 1999). Similar findings were reported in 10-year-old children in the United Kingdom (Whincup et al, 1997). In a Finnish cohort, adult hypertension was associated with both lower birthweight and accelerated growth in the first 7 years of life. In contrast, in two preliminary studies from the United Kingdom, catch-up growth in the first 6 months of life was not clearly related to blood pressure in young adulthood, although birthweight was (McCarthy et al, 2001). Interpreting the findings of these studies is complicated by the vague definitions of catch-up growth. The term, which can encompass either the first 6 to 12 months only or as much as the first 2 years after birth, usually refers to realignment of genetic growth potential after IUGR. This definition allows for fetal growth retardation at any birthweight; large fetuses can be growth retarded in relation to their genetic potential. However, postnatal factors can obviously affect infant growth in the first few months of life. For example, breastfeeding appears to protect against obesity later in childhood, but breastfed infants usually exhibit higher body mass during the first year of life than formula-fed infants. Although it is likely that accelerated growth confers an additional risk to the growth-retarded fetus, these conflicting results demonstrate the need for additional, carefully designed studies to determine how childhood growth rates affect the later development of cardiovascular disease and type 2 diabetes. CHAPTER 6 Abnormalities of Fetal Growth Size at Birth, Insulin Secretion, and Insulin Action The mechanisms underlying the association between size at birth and impaired glucose tolerance or type 2 diabetes are unclear. A number of studies in children and adults have shown that nondiabetic or prediabetic (abnormal glucose tolerance) subjects with LBW are insulin resistant and thus are predisposed to development of type 2 diabetes (Bavdekar et al, 1999; Clausen et al, 1997; Flanagan et al, 2000; Hoffman et al, 1997; Leger et al, 1997; Li et al, 2001; Lithell et al, 1996; McKeigue et al, 1998; Phillips et al, 1994; Yajnik et al, 1995). IUGR is known to alter the fetal development of adipose tissue, which is closely linked to the development of insulin resistance (Lapillonne et al, 1997; Widdowson et al, 1979). In a well-designed case-control study of 25-yearold adults, Jaquet et al (2000) demonstrated that individuals who were born SGA at 37 weeks’ gestation or later had a significantly higher percentage of body fat (15%). Insulin sensitivity, after adjustment for BMI or total fat mass, was markedly impaired in these SGA subjects. There were no significant differences between the SGA and control groups regarding parental history of type 2 diabetes, cardiovascular disease, hypertension, or dyslipidemia. Of importance when generalizing the findings to other populations, the causes of IUGR in these subjects were gestational hypertension (50%), smoking (30%), maternal short stature (7%), congenital anomalies (7%), and unknown (6%). The adverse effect of IUGR on glucose homeostasis was originally thought to be mediated through programming of the fetal endocrine pancreas. Growth-retarded fetuses and newborns have been shown to have a reduced population of pancreatic β cells (Van Assche et al, 1977). LBW has been associated with reduced insulin response after glucose ingestion in young men without diabetes; however, a number of other studies have found no effect of LBW on insulin secretion in humans (Clausen et al, 1997; Flanagan et al, 2000; Lithell et al, 1996). However, none of these earlier studies adjusted for the corresponding insulin sensitivity, which has a profound effect on insulin secretion. Jensen et al (2002) measured insulin secretion and insulin sensitivity in a wellmatched population of 19-year-old, glucose-tolerant white men whose birthweights were either less than the 10th percentile (i.e., SGA) or between the 50th and 75th percentiles (controls). To eliminate the major confounding factors, such as “diabetes genes,” the researchers ensured that none of the participants had a family history of diabetes, hypertension, or ischemic heart disease. They found no differences between the groups in regard to current weight, BMI, body composition, and lipid profile. When data were controlled for insulin sensitivity, insulin secretion was found to be lower by 30%. However, insulin sensitivity was normal in the SGA subjects. These investigators hypothesized that defects in insulin secretion precede defects in insulin action, and that SGA individuals demonstrate insulin resistance once they accumulate body fat. Epidemiologic Challenges The data described in the preceding section suggest that LBW is associated with glucose intolerance, type 2 diabetes, and cardiovascular disease. However, the question 57 remains whether these associations reflect fetal nutrition or other factors that contribute to birthweight and the observed glucose intolerance. Because of the retrospective nature of the cohort identification, many confounding variables were not always recorded, such as lifestyle, socioeconomic status, education, maternal age, parental build, birth order, obstetric complications, smoking, and maternal health. Maternal nutritional status, either directly in the form of diet histories, or indirectly in the form of BMI, height, and pregnancy weight gain, were usually not recorded. Instead, birth anthropometric measures were used as proxies for presumed undernutrition in pregnancy. Size at Birth Cannot Be Used as a Proxy for Fetal Growth Birthweight is determined by the sum of multiple known and unknown factors, including gestational age, maternal age, birth order, genetics, maternal pre-pregnancy BMI, and pregnancy weight gain, plus multiple environmental factors, such as smoking, drug use, infection, and maternal hypertension. Some of these determinants may be related to susceptibility to adult disease, and others may not. Conversely, some prenatal determinants of adult outcomes may not be related to fetal growth. A good example of how size at birth may potentially be a proxy for an underlying causal pathway is the hypothesis that essential hypertension in the adult is caused by a congenital nephron deficit (Brenner and Chertow, 1993). This study shows that kidney volume is smaller in adults who were thinner at birth, after adjustment for current body size. In contrast, maternal cigarette smoking is a good example of a prenatal exposure that restricts fetal growth, but to date no association has been found between cigarette smoking and adverse long-term outcome in offspring. Genetics versus Environment Several epidemiologic and metabolic studies of twins and first-degree relatives of patients with type 2 diabetes have demonstrated an important genetic component of diabetes (Vaag et al, 1995). The association between LBW and risk of type 2 diabetes in some studies could theoretically be explained by a genetically determined reduced fetal growth rate. In other words, the genotype responsible for type 2 diabetes may itself restrict fetal growth. This possibility forms the basis for the fetal insulin hypothesis, which suggests that genetically determined insulin resistance could result in insulin-mediated low growth rate in utero as well as insulin resistance in childhood and adulthood (Hattersley et al, 1999). Insulin is one of the major growth factors in fetal life, and monogenic disorders that affect the fetus’s insulin secretion or insulin resistance also affect fetal growth (Elsas et al, 1985; Froguel et al, 1993; Hattersley et al, 1998; Stoffers et al, 1997). Mutations in the gene encoding glucokinase have been identified that result in LBW and maturity-onset diabetes of the young. Such mutations are rare, and no analogous common allelic variation has been discovered, but it is likely that some variations exist that, once identified, will help to explain a proportion of the cases of diabetes in LBW subjects. 58 PART II Fetal Development There is obviously a close relationship between genes and the environment. Maternal gene expression can alter the fetal environment, and the maternal intrauterine environment also affects fetal gene expression. An adverse intrauterine milieu is likely to have profound long-term effects on the developing organism that might not be reflected in birthweight. Cellular Mechanisms Fetal malnutrition has two main causes: poor maternal nutrition and placental insufficiency. In the extensive literature about the fetal origins hypothesis, these two concepts have not been discerned clearly. Such a distinction is necessary, because maternal nutrition has probably been adequate in the majority of populations in which the hypothesis has been tested. Only extreme maternal undernutrition, such as occurred in the Dutch famine, reduces the birthweight to an extent that could be expected to raise the risk of adult disease (Lumey et al, 1995). To a lesser extent but equally important is the LBW in populations with low resources, resulting in maternal undernutrition. Overall, in most populations it is reasonable that placental insufficiency has been a main cause of LBW. The oxygen and nutrients that support fetal growth and development rely on the entire nutrient supply line, beginning with maternal consumption and body size, but extending to uterine perfusion, placental function, and fetal metabolism. Interruptions of the supply line at any point could result in programming of the fetus for the future risk of adult diseases. The intrauterine environment influences development of the fetus by modifying gene expression in both pluripotential cells and terminally differentiated, poorly replicating cells. The long-range effects on the offspring (into adulthood) are determined by which cells are undergoing differentiation, proliferation, or functional maturation at the time of the disturbance in maternal fuel economy. The fetus also adapts to an inadequate supply of substrates (e.g., glucose, amino acids, fatty acids, and oxygen) through metabolic changes, redistribution of blood flow, and changes in the production of fetal and placental hormones that control fetal growth. The fetus’s immediate metabolic response to placental insufficiency is catabolism, consuming its own substrates to provide energy. A more prolonged reduction in availability of substrates leads to slowed growth, which enhances the fetus’s ability to survive by reducing the use of substrates and lowering the metabolic rate. Slowed growth in late gestation leads to disproportionate organ size, because the organs and tissues that are growing rapidly at the time are affected the most. For example, placental insufficiency in late gestation can lead to reduced growth of the kidney, which is developing rapidly at that time. Reduced replication of kidney cells can permanently reduce cell numbers, because there seems to be no capacity for renal cell division to catch up after birth. Substrate availability has profound effects on fetal hormones and on the hormonal and metabolic interactions among the fetus, placenta, and mother. These effects are most apparent in the fetus of the mother with diabetes. Higher maternal concentrations of glucose and amino acids stimulate the fetal pancreas to secrete exaggerated amounts of insulin and stimulate the fetal liver to produce higher levels of insulin-like growth factors. Fetal hyperinsulinism stimulates the growth of adipose tissue and other insulin-responsive tissues in the fetus, often leading to macrosomia. However, many offspring of mothers with diabetes with fetal hyperinsulinism are not overgrown by usual standards, and many with later obesity and glucose intolerance were not macrosomic at birth (Pettitt et al, 1987; Silverman et al, 1995). These observations suggest that birthweight is not a good indication of intrauterine nutrition. MACROSOMIA Excessive fetal growth (macrosomia, being large for gestational age) is found in 9% to 13% of all deliveries and can lead to significant complications in the perinatal period (Gregory et al, 1998; Wollschlaeger et al, 1999). Maternal factors associated with macrosomia during pregnancy include increasing parity, higher maternal age, and maternal height. In addition, the previous delivery of an infant with macrosomia, prolonged pregnancy, maternal glucose intolerance, high pre-pregnancy weight or obesity, and large pregnancy weight gain have all been found to raise the risk of delivering an infant with macrosomia (Mocanu et al, 2000). Maternal complications of macrosomia include morbidities related to labor and delivery. Prolonged labor, arrest of labor, and higher rates of cesarean section and instrumentation during labor have been reported. In addition, the risks of maternal lacerations and trauma, delayed placental detachment, and postpartum hemorrhage are higher for the woman delivering an infant with macrosomia (Lipscomb et al, 1995; Perlow et al, 1996). Complications of labor are more pronounced in primiparous women than in multiparous women (Mocanu et al, 2000). The neonatal complications of macrosomia include traumatic events such as shoulder dystocia, brachial nerve palsy, birth trauma, and associated perinatal asphyxia. Other complications for the neonate are elevated insulin levels and metabolic derangements, such as hypoglycemia and hypocalcemia (Wollschlaeger et al, 1999). In a large population-based study in the United States, macrosomia (defined as birthweight greater than 4000 g) was detected in 13% of births. Of these, shoulder dystocia was noted in 11% (Gregory et al, 1998). Macrosomia is often not detected during pregnancy and labor. The clinical estimation of fetal size is difficult and has significant false-positive and false-negative rates. Ultrasonography estimates of fetal weight are not always accurate, and there are a wide range of sensitivity estimates for the ultrasound detection of macrosomia. In addition, there is controversy regarding how to define macrosomia and which ultrasound measurement is most sensitive in predicting macrosomia. Smith et al (1997) demonstrated a linear relation between abdominal circumference and birthweight. They showed that the equations commonly used for estimated fetal weight have a median error rate of 7%, with greater errors seen with larger infants. Using receiver operating characteristics curves to measure the diagnostic accuracy of ultrasound, CHAPTER 6 Abnormalities of Fetal Growth O’Reilly-Green and Divon (1997) reported sensitivity and specificity rates of 85% and 72%, respectively, for estimation of birthweight exceeding 4000 g. In their study, the positive predictive value (i.e., a positive test result represents a truly macrosomic infant) was approximately 49%. Chauhan et al (2000) found lower sensitivity for the use of ultrasound measurement of abdominal and head circumference and femur length (72% sensitivity), similar to the sensitivity of using clinical measurements alone (73%). Other investigators have shown that clinical estimation of fetal weight (43% sensitivity) has higher sensitivity and specificity than ultrasound evaluation in predicting macrosomia (Gonen et al, 1996). In a retrospective study, Jazayeri et al (1999) showed that ultrasound measurement of abdominal circumference of greater than 35 cm predicts macrosomia in 93% of cases and is superior to measurements of biparietal diameter or the femur. Other researchers have reported that an abdominal circumference of more than 37 cm is a better predictor (Al-Inany et al, 2001; Gilby et al, 2000). Numerous investigators have also questioned whether antenatal diagnosis improves birth outcomes in macrosomic infants. Investigators indicate the low rates of specificity of antenatal tests resulting in high rates of false-positive results (Bryant et al, 1998, O’Reilly-Green and Divon, 1997). Antenatal identification of macrosomia or possible macrosomia can lead to a higher rate of cesarean section performed for infants with normal birthweights (Gonen et al, 2000; Mocanu et al, 2000; Parry et al, 2000). Macrosomia is a risk factor for shoulder dystocia, but the majority of cases of shoulder dystocia and birth trauma occur in infants with macrosomia (Gonen et al, 1996). A retrospective study of infants weighing more than 4200 g at birth showed a cesarean section rate of 52% in infants predicted antenatally to have macrosomia, compared with 30% in infants without such an antenatal prediction. The antenatal prediction of fetal macrosomia is also associated 59 with a higher incidence of failed induction of labor and no reduction in the rate of shoulder dystocia (Zamorski and Biggs, 2001). Using retrospective data from a 12-year period, Bryant et al (1998) estimated that a policy of routine cesarean section for all infants with estimated fetal weight greater than 4500 g would require between 155 and 588 cesarean sections to prevent a single case of permanent brachial nerve palsy. SUMMARY This chapter has described many identified biologic and genetic factors associated with fetal growth and with abnormalities of fetal growth. Physicians are limited in the ability to identify a causative agent in every case. Modification of fetal growth is possible and occurs from diverse influences such as socioeconomic status, maternal nutrition, and maternal constitutional factors. Abnormal fetal growth influences acute perinatal outcomes and health during infancy, childhood, and adulthood. In schools of public health, students are taught to search “up river” for solutions to health problems. Solutions for ill health in adulthood may reside in the identification of methods to improve the health of the fetus. SUGGESTED READINGS Alexander GR, Himes JH, Kaufman R, et al: A United States national reference for fetal growth, Obstet Gynecol 87:163-168, 1996. Dahri S, Snoeck A, Reusens-Billen B, et al: Islet function in off-spring of mothers on low-protein diet during gestation, Diabetes 40:115-120, 1991. Wilson MR, Hughes SJ: The effect of maternal protein deficiency during pregnancy and lactation on glucose tolerance and pancreatic islet function in adult rat offspring, J Endocrinology 154:177-185, 1997. Zeitlin J, Ancel P, Saurel-Cibizolles M, Papiernik E: The relationship between IUGR and preterm delivery: an empirical approach using data from a European case-control study, Brit J Obstet Gynecol 107:750-758, 2000. Complete references used in this text can be found online at www.expertconsult.com C H A P T E R 7 Multiple Gestations and Assisted Reproductive Technology Kerri Marquard and Kelle Moley EPIDEMIOLOGY OF MULTIPLES Since the birth of the first in vitro fertilization (IVF) baby in 1978, the numbers of IVF clinics, ovarian stimulation cycles, and live births from assisted reproductive technology (ART) have all steadily increased. Between 1998 and 2003, total births in the United States increased by 4%, while ART births grew by 67% (Dickey, 2007). The growing use of ART, in addition to delayed childbearing until age-related fertility issues become apparent, has contributed greatly to multiple birth rates. In 2006 ART infants accounted for 1% of all U.S. births, but represented 17% of twins and 38% of triplets or greater (Centers for Disease Control and Prevention et al, 2008; Sunderam et al, 2009). Three percent of all U.S. births are multiples, yet in 2003 ART multiple live birth rates from fresh nondonor, donor oocyte, and frozen embryo transfer cycles were 34%, 40%, and 25%, respectively (Centers for Disease Control and Prevention et al, 2008). The assisted conception and spontaneous rates associated with twins, triplets, and higher-order multiples (HOMs; i.e., four or more fetuses) are as follows. In 2003, 62.7%, 16.3%, and 21% of twins were conceived naturally, from ART, and from non-ART ovulation induction, respectively. For triplets in the same year, 17.7% were natural conceptions, 45.4% were from ART, and 36.9% were from non-ART ovulation induction. HOM rates from spontaneous conception, ART, and non-ART ovulation induction were 8%, 30%, and 62.4%, respectively (Dickey, 2007). Given the maternal, perinatal, and neonatal complications associated with multiples, the goal of infertility treatment is one healthy child. Multifetal pregnancies drastically affect individuals, families, and the public health system. Of particular importance in both maternal and fetal outcomes are fetal number and placentation. et al, 1995) At 10 to 14 weeks’ gestation, ultrasound criteria correctly diagnosed chorionicity in 99.3% of cases as confirmed by postpartum placental examination and gender (Carroll et al, 2002). Assessment of zygosity and chorionicity in 237 same-sex twins with physical likeness questionnaires, DNA analysis, and placental inspection accurately diagnosed 96% of twin pairs (Forget-Dubois et al, 2003). ZYGOSITY AND CHORIONICITY Zygosity and placentation affect fetal morbidity and mortality in multifetal pregnancies. Dizygotic twins (DZTs), which comprise 67% of spontaneous twins, arise from the fertilization of two separate eggs by different sperm (Gibbs et al, 2008), and with few exceptions they lead to a dichorionic diamniotic arrangement in which the placenta can be separate or fused. A rare case of dizygotic monochorionic (MC) diamniotic (DA) twins has been reported (Souter et al, 2003). The overall DZT rate varies from 4 to 50 per 1000 worldwide: 1.3% in Japan, 7.1% to 11% in the United States, 8.1% in India, 8.8% in England and Wales, and 49% in Nigeria (Table 7-1) (MacGillivray, 1986). Risk factors for DZT include advancing maternal age, increased parity, female relatives with DZT, taller height, and larger body mass index (Hoekstra et al, 2008; MacGillivray, 1986). The true incidence of monozygotic twins (MZTs) is difficult to ascertain because of its rarity, inaccuracies in diagnosis, and lack of confirmatory studies at birth, but spontaneous MZT rates are estimated to occur in 0.3% to 0.5% of all pregnancies and in 30% of all twins (Bulmer, 1970; MacGillivray, 1986). Unlike DZT, it is unclear whether MZT is related to genetics or environment (Bortolus et al, 1999; Hoekstra et al, 2008), although familial components may play a role (Hamamy et al, 2004). Chorionicity in monozygotic gestations is determined by the DIAGNOSING ZYGOSITY AND CHORIONICITY Determining zygosity and chorionicity is important medically, genetically, and psychosocially for the individual and family. More immediately, the chorion-amnion arrangement is crucial to antepartum management in cases of one fetal demise or selective reduction, and because of potential associated problems such as twin-twin transfusion syndrome (TTTS), growth discordance, intrauterine growth restriction (IUGR), congenital anomalies, and cord accidents. Diagnosing zygosity is possible using ultrasound markers including the number of placental sites, thickness of dividing membrane, the lambda sign, and fetal gender in addition to postpartum placental examination, physical similarity questionnaires, blood type, and DNA analysis (Hall, 2003; Ohm Kyvik and Derom, 2006; Scardo 60 TABLE 7-1 Twinning Rates per 1000 Births by Zygosity Country Nigeria Monozygotic Dizygotic Total 5.0 49 54 African American 4.7 11.1 15.8 Caucasian 4.2 7.1 11.3 England and Wales 3.5 8.8 12.3 India 3.3 8.1 11.4 Japan 3.0 1.3 4.3 United States From MacGillivray I: Epidemiology of twin pregnancy, Semin Perinatol, 10:4-8, 1986; and Cunningham FG, Hauth JC, Wenstrom KD, et al, editors: Williams obstetrics, ed 22, New York, 2005, McGraw-Hill. CHAPTER 7 Multiple Gestations and Assisted Reproductive Technology Zygosity Dizygotic or Monozygotic Monozygotic Day of division 0–3 days 0–3 days 4–8 days 8–13 days Fetal membranes 2 Amnions, 2 Chorions, 2 Placentas 2 Amnions, 2 Chorions, 1 Placenta 2 Amnions, 1 Chorion, 1 Placenta 1 Amnion, 1 Chorion, 1 Placenta A 61 B C D FIGURE 7-1 Placentation and membranes based on timing of embryonic division. A, Two amnions, two chorions, and separate placentas from the division of either a dizygotic or monozygotic embryo within 3 days of fertilization. B, Two amnions, two chorions, and one fused placenta from the division of either a dizygotic or monozygotic embryo within 3 days of fertilization. C, Two amnions, one chorion, and one placenta from monozygotic embryonic cleavage, days 4 to 8 after fertilization. D, One amnion, one chorion, and one placenta from a monozygotic embryo splitting, days 8 to 13 after fertilization. (Modified from Gibbs R, Karlan B, Haney A, et al: Danforth’s obstetrics & gynecology, ed 10, Philadelphia, 2008, Lippincott, Williams & Wilkins.) timing of the embryonic division (Figures 7-1 and 7-2) (Benirschke and Kim, 1973; Hall, 2003). In 18% to 36% of MZTs, the zygote divides within 72 hours of fertilization resulting in dichorionic (DC) DA gestation (the placenta can be separate or fused); 60% to 75% split between days 4 and 8, leading to an MC-DA unit, and 1% to 2% separate between days 8 and 13, leading to an MC monoamniotic (MA) pregnancy. Embryonic division after day 13 results in conjoined twins with an MC/MA placenta (Cunningham et al, 2005; Gibbs et al, 2008; Hall, 2003). Although the majority of ART MZTs are MC/DA, any of the three MZ placental arrangements can transpire after ART, implying that the timing and mechanism of embryonic splitting are variable (Aston et al, 2008; ­Knopman et al, 2009). Monozygotic DCDA twins can occur after inner cell mass (ICM) splitting and atypical hatching in a blastocyst embryo (Meintjes et al, 2001; Van Langendonckt et al, 2000). In addition, MA twinning may be increased after IVF (Alikani et al, 2003) and zona pellucida (ZP) manipulation (Slotnick and Ortega, 1996). Contributing factors to MZT in ART include ICM damage and other ZP abnormalities (Hall, 2003). INCREASE IN MONOZYGOTIC TWINS WITH ASSISTED REPRODUCTIVE TECHNOLOGY The first reported association between ART and MZT (Yovich et al, 1984) preceded numerous accounts of similar findings. The majority (>90%) of ART twins are dizygotic (Gibbs et al, 2008) secondary to transferring multiple embryos; however, the rate of MZTs per pregnancy after fertility treatment is higher (0.9% to 4.9%) (Alikani et al, 2003; Blickstein et al, 2003; Elizur et al, 2004; Knopman et al, 2009; Papaniklaou et al, 2009; Schachter et al, 2001; Sharara and Abdo, 2010; Sills et al, 2000; Vitthala et al, 2009; Wenstrom et al, 1993) versus the general population (0.3% to 0.5%) (Bulmer, 1970; MacGillivray, 1986). Several theories to explain the mechanism responsible for elevated MZTs with ART have been proposed. AGE Maternal age affects fertility and reproductive outcomes. Spontaneous dizygotic twinning increases with advancing maternal age (Bortolus et al, 1999; Bulmer, 1970; ­Mac­Gillivray, 1986), but the connection between age and MZTs is controversial. Some studies reported trends toward elevated MZT rates in older women (Abusheikha et al, 2000; Alikani et al, 2003; Bulmer, 1970), whereas others found no association with increasing maternal age and MZT (Bortolus et al, 1999; Skiadas et al, 2008), and one study found that the MZT risk doubles in women younger than 35 years (Knopman et al, 2009). Overall, the correlation between age and MZT in ART remains unclear. ZONA PELLUCIDA MANIPULATION The ZP, an acellular protein surrounding the ovum, provides a species-specific sperm barrier and decreases polyploidy by inhibiting penetration by multiple sperm (Speroff and Fritz, 2005). Components of both ART and non-ART procedures are capable of modifying this barrier. Stimulation protocol, elevated follicle-stimulating hormone or estradiol levels, and prolonged culture conditions can change ZP thickness (Loret et al, 1997). In addition, ZP manipulations performed during IVF all potentially affect MZT risk. Manipulation of the ZP in IVF occurs via both intracytoplasmic sperm injection (ICSI) and AH (assisted hatching). The injection of one sperm into a mature oocyte (i.e., ICSI) is most commonly performed for male factor infertility. AH is achieved with an artificial breach in the ZP by laser, chemical, or mechanical methods and is indicated in patients with a poor prognosis (Practice Committee of Society for Assisted Reproductive Technology and Practice Committee of American Society for Reproductive Medicine, 2008c). Some studies indicate no association with ICSI or AH and MZT (Behr et al, 2000; Elizur et al, 2004; Knopman et al, 2009; Meldrum et al, 1998; Milki et al, 2003; Sills et al, 2000), whereas others imply that 62 PART II Fetal Development Embryo Yolk sac A Inner cell cell mass Embryo Amniotic cavity B Yolk sac Two-cell embryo Blastocoel Amniotic cavity Chorionic cavity Embryo C Yolk sac FIGURE 7-2 Types of monozygotic placentation. A, Dichorionic diamniotic pregnancy. B, Monochorionic diamniotic pregnancy. C, Monochorionic monoamniotic pregnancy. (Adapted from Hall JG: Twinning, Lancet 362:735-743, 2003; and Benirschke K, Kim CK: Multiple pregnancy, N Engl J Med 288:1276-1284, 1973.) ZP manipulation increases the risk of MZT (Alikani et al, 1994; Saito et al, 2000; Schieve et al, 2000; Skiadas et al, 2008; Vitthala et al, 2009), particularly if multiple embryos undergo hatching before transfer (Alikani et al, 2003; Schieve et al, 2000). The results of a Cochrane review on the risk of multiples with AH and ICSI reported increased multiples with both ICSI and AH, and elevated MZT with AH versus no manipulation (0.8%) (Das et al, 2009). Alterations in ZP thickness or abnormal ICM splitting or embryo hatching in both animal (Cohen, 1991; Malter and Cohen, 1989) and human embryos may explain this increase in MZT after ZP manipulation (Alikani et al, 1994; Cohen, 1991; Malter and Cohen, 1989; Sheen et al, 2001). BLASTOCYST TRANSFER After oocyte retrieval and insemination, embryos undergo intrauterine transfer at either the cleavage stage (day 2 to 3 after retrieval) or the blastocyst stage (day 5 to 6 after retrieval). Blastocyst-stage embryo transfer produces higher pregnancy rates (29.4% vs 36%) (Blake et al, 2007) and may lower overall multiple rates (Frattarelli et al, 2003), but evidence supports the increased incidence of MZTs (Behr et al, 2000; Chang et al, 2009; Jain et al, 2004; Knopman et al, 2009; Peramo et al, 1999; Practice Committee of American Society for Reproductive Medicine and Practice Committee of Society for Assisted Reproductive Technology, 2008a; Sheiner et al, 2001; Skiadas et al, 2008; Wright et al, 2004) compared with cleavage-stage embryos. One institution initially noted increased MZT after blastocyst transfer (5.6% vs 2%; Milki et al, 2003), but a follow-up study 3 years later demonstrated similar MZT rates between blastocyst and day 3 embryos (2.3% vs 1.8%), indicating that changes in culture media and an experienced embryology team may affect the rate of MZTs (Moayeri et al, 2007). Culture media or prolonged culture may influence MZTs with blastocyst transfer in both human and animal models. Elevated glucose levels in extended culture and subsequent glucose-induced apoptotic remodeling of the ICM (Cassuto et al, 2003; Menezo and Sakkas, 2002), and ICM splitting in extended culture in murine and human embryos are both explanations for the phenomenon of MZTs in blastocysts (Chida, 1990; Hsu and Gonda, 1980; Payne et al, 2007). Whether MZT rates are elevated with blastocyst transfer (Behr et al, 2000; Chang et al, 2009; Jain et al, 2004; Knopman et al, 2009; Peramo et al, 1999; Practice Committee of American Society for Reproductive Medicine and Practice Committee of Society for Assisted Reproductive Technology, 2008a; Sheiner et al, 2001; Skiadas et al, 2008; Wright et al, 2004), or similar to cleavagestage transfer (Frattarelli et al, 2003; Papanikolaou et al, 2006, 2010; Sharara and Abdo, 2010), the explicit role of culture media remains to be determined. Another environmental factor that may affect MZTs is temperature variation in frozen-thawed embryo cycles. Although minute evidence links frozen embryo transfers and temperature fluctuations to MZTs (Belaisch-Allart et al, 1995; Faraj et al, 2008; Toledo, 2005), most studies show no difference between fresh and frozen-thawed embryos and multiple rates (Blickstein et al, 2003) or MZT rates (Alikani et al, 2003; Knopman et al, 2009; Sills et al, 2000a, 2000b). OVARIAN STIMULATION Human studies of ovarian stimulation with clomiphene citrate and gonadotropins reveal a higher rate (1.2%) of MZTs compared with the expected rates in the general 63 CHAPTER 7 Multiple Gestations and Assisted Reproductive Technology population (Derom et al, 1987). MZT incidences of 1.5% after ovulation induction with gonadotropins, 0.72% after IVF, and 0.87% with IVF ICSI-AH suggests that gonadotropins elevate MZTs regardless of ZP manipulation (Schachter et al, 2001). Ovulation induction medications may cause uneven hardening of the ZP and atypical blastocyst hatching thereby increasing the chance of MZTs (Derom et al, 1987). FETAL COMPLICATIONS ASSOCIATED WITH MULTIPLES Singleton pregnancies after assisted conception have increased complications, including preterm delivery (<37 weeks’ gestation), low birthweight (LBW) (Helmerhorst et al, 2004; Schieve et al, 2007), prolonged hospital stay (Schieve et al, 2007), cesarean deliveries, neonatal intensive care unit (NICU) admission, and mortality compared with spontaneous singletons (Helmerhorst et al, 2004). Some risk persists in an IVF singleton pregnancy, even after spontaneous reduction from two to three initial heartbeats to one heartbeat (Luke et al, 2009). Although singleton IVF births are associated with morbidity and mortality, assisted-conception multiple gestations comprise the majority of adverse maternal, perinatal, and neonatal outcomes. Multiple pregnancies account for a small percentage of overall live births, but are responsible for a disproportionate amount of morbidity and mortality, largely because of intrauterine growth restriction and prematurity (Garite et al, 2004), including 13% of all preterm deliveries, 21% of all LBW infants, and 25% of all very LBW infants (Robinson et al, 2001). Compared with a singleton pregnancy, fetal and maternal complications are elevated in twins, and pregnancies with three fetuses or more have even greater morbidity and mortality rates (Table 7-2) (ACOG Practice Bulletin #56, 2004; Albrecht and Tomich, 1996; Elliott and Radin, 1992; Ettner et al, 1997; Grether et al, 1993; Kiely et al, 1992; Luke, 1994; Luke and Keith, 1992; Luke et al, 1996; Martin et al, 2003; Mauldin and Newman, 1998; McCormick et al, 1992; Newman et al, 1989; Seoud et al, 1992). Higher fetal number correlates with increased risk of growth restriction, earlier delivery, LBW, neonatal intensive care unit admission, length of stay, risk of major handicap and cerebral palsy, and death in first year of life (ACOG Practice Bulletin #56, 2004; Gardner et al, 1995; Garite et al, 2004). The average gestational ages for twin, triplet, and quadruplet deliveries are 35.3, 32.2, and 29.9 weeks, respectively (ACOG Practice Bulletin #56, 2004), corresponding to NICU admission rates fivefold higher in twins and 17-fold higher in triplets and HOMs (Ross et al, 1999). Twins have an increased risk of intrauterine fetal demise (fourfold), intraventricular hemorrhage, sepsis, necrotizing enterocolitis, respiratory distress syndrome and neonatal death (sixfold) versus singletons, and surviving infants of preterm multifetal pregnancies have higher rates of developmental handicap (Gardner et al, 1995). A review of 100 triplet gestations (88 with assisted conception) revealed that 78% experienced preterm labor (PTL), 14% delivered before 28 weeks, 5% had congenital anomalies, and 9.7% died in the perinatal period (Devine et al, 2001). TABLE 7-2 Morbidity and Mortality by Fetal Number Characteristic Twins Triplets Average birthweight 2347 g 1687 g Quadruplets 1309 g Average gestational age at delivery 35.3 wk 32.2 wk 29.9 wk Percentage with growth restriction 14-25 50-60 50-60 Percentage ­requiring admission to neonatal intensive care unit 25 75 100 Average length of stay in neonatal intensive care unit 18 days 30 days 58 days Percentage with major handicap - 20 50 Risk of cerebral palsy 4 times more than singletons 17 times more than singletons - Risk of death by age 1 year 7 times higher than singletons 20 times higher than singletons - From ACOG Practice Bulletin #56: Multiple gestation: complicated twin, triplet, and high-order multifetal pregnancy, Obstet Gynecol 104:869-883, 2004. Similar to ART singleton versus spontaneous singleton outcomes, ART multiples may have higher morbidity compared with spontaneous multiples. Assisted-­conception twins are at increased risk for LBW (Luke et al, 2009), preterm delivery (Luke et al, 2009; Nassar et al, 2003), cesarean delivery (Helmerhorst et al, 2004; Nassar et al, 2003), NICU admission (Helmerhorst et al, 2004; ­Nassar et al, 2003; Pinborg et al, 2004), longer length of stay, respiratory distress syndrome (Nassar et al, 2003), and birthweight discordance (Pinborg et al, 2004) versus spontaneously conceived twins. Oftentimes with IVF, because of multiple embryos transferred or MZ splitting, there are multiple heartbeats on an initial ultrasound examination that ultimately spontaneously reduce; however, these pregnancies are still at risk for an adverse outcome. In twin IVF cycles with two initial heartbeats on early ultrasound versus three heartbeats that spontaneously reduced to two heartbeats, pregnancies with three heartbeats on an early examination had increased rates of preterm delivery (35%) and LBW (47%) (Luke et al, 2009). In contrast, other studies suggest comparable outcomes in assisted conception and spontaneous multiples. Rates of pregnancy-induced hypertension, gestational diabetes mellitus (GDM), preterm premature rupture of membranes (PPROM), placenta previa, placental abruption (Nassar et al, 2003), congenital malformations (Nassar et al, 2003; Pinborg et al, 2004), and mortality (Pinborg et al, 2004) were similar in IVF twins versus spontaneous twins. Similarly, morbidity and mortality in ART triplets versus spontaneous triplets were comparable regarding the rates of PPROM, PTL, pregnancy-induced hypertension, GDM, gestational age at delivery, birthweight, and NICU admissions (Fitzsimmons et al, 1998). Besides fetal number, another important factor in pregnancy outcome is placental arrangement. MC multiples 64 PART II Fetal Development attributed to increased prenatal diagnosis and fetal surveillance (Allen et al, 2001; Heyborne et al, 2005; Rodis et al, 1997). TABLE 7-3 Incidence of Twin Pregnancy Zygosity and Chorionicity With Corresponding Complications Complication (%) Type of Twinning Dizygotic Delivery Incidence <37 Weeks’ Perinatal (%) IUGR Gestation Mortality 67-70 25 40 10-12 Monozygotic 30-33 40 50 15-18 Diamniotic-­ dichorionic 18-36 30 40 18-20 Diamniotic-­ monochorionic 60-75 50 60 30-40 Monoamnioticmonochorionic 1-2 40 60-70 30-70 From Cunningham FG, Hauth JC, Wenstrom KD, et al, editors: Williams obstetrics, ed 22, New York, 2005, McGraw-Hill; originally from Manning FA: Fetal biophysical profile scoring. In Fetal medicine: principles and practices, Norwalk, Conn, 1995, Appleton & Lange. IUGR, Intrauterine growth restriction. experience higher rates of morbidity and mortality, largely because of placental factors (Table 7-3) (Cunningham et al, 2005; Gaziano et al, 2000; Hack et al, 2008; Manning, 1995). When Dube et al (2002) studied different chorionicity–zygosity groups (monozygotic monochorionic [MZMC], di zygotic dichorionic [DZDC], and monozygotic dichorionic [MZDC]) they found smaller birthweight, and IUGR, congenital anomalies, and perinatal death in the MZMC versus DZDC twins, whereas MZDC and DZDC risks were similar, implying that poor outcomes are related more to chorionicity than zygosity. Negative outcomes such as cerebral palsy, mental retardation, and death, measured at 1 year of life in MC twins, are elevated (10%) compared with DC twins (3.7%), the majority of which are caused by complications from TTTS (Minakami et al, 1999). Both placental asymmetry and abnormal vascular anastomosis affect fetal morbidity and mortality. IUGR and growth discordance afflict both DC and MC pregnancies, but MC twins are more likely to have cord abnormalities, unequal placental distribution, and TTTS (ClearyGoldman and D’Alton, 2008). TTTS, ranging in severity from oligohydramnios to hydrops and fetal death (ClearyGoldman and D’Alton, 2008; Gaziano et al, 2000) occurs in 15% to 32% of MC pregnancies (Hack et al, 2008; Minakami et al, 1999). Mortality rates vary from 22% to 100% (Bajoria et al, 1995; Cleary-Goldman and D’Alton, 2008; Hack et al, 2008), and surviving infants are at risk for long-standing adverse neurologic outcomes (ClearyGoldman and D’Alton, 2008). MC-MA twins occur in 1 in 10,000 pregnancies, but they suffer the highest risk of perinatal morbidity and mortality (Cordero et al, 2006). Similar to other MC twins, MC-MA twins are susceptible to TTTS, growth discordance, IUGR, preterm delivery, and congenital anomalies, but they also face the unique complication of cord entanglement. These factors historically account for perinatal mortality rates of 30% to 70%; however, lower morbidity and mortality rates (8% to 23%) reported in recent articles (Allen et al, 2001; Cordero et al, 2006; Heyborne et al, 2005; Rodis et al, 1997; Roque et al, 2003) may be MATERNAL COMPLICATIONS Approximately 80% of multiples experience antepartum complications versus 25% of singletons (Norwitz et al, 2005), and hospitalization for hypertensive disorders, PTL, PPROM, placental abruption, and postpartum hemorrhage are elevated sixfold (ACOG Practice Bulletin #56, 2004). Mothers with two or more fetuses are at increased risk for myocardial infarction, left ventricular heart failure, pulmonary edema, GDM, operative vaginal or cesarean delivery, hysterectomy, blood transfusion, longer hospital stay, and the three major causes of maternal mortality: post partum hemorrhage, venous thromboembolism, and hypertensive disorders (Walker et al, 2004). Stratified by fetal number, plurality correlates with maternal morbidity where quadruplets and other HOMs experience significantly increased maternal morbidity versus twins and triplets (Wen et al, 2004). Hypertensive disorders occur in 12% to 20% of twins, triplets, and quadruplets compared with 6.5% of singletons, HELLP syndrome increases with higher numbers of fetuses (Day et al, 2005), and twins with preeclampsia experience more complications than singletons with preeclampsia (Sibai et al, 2000). Results of a triplet cohort showed that 96% had maternal complications, 96% required antenatal hospitalization, one in four were diagnosed with preeclampsia, and 44% encountered postpartum complications (Devine et al, 2001). PSYCHOSOCIAL FACTORS As fetal number in an assisted conception pregnancy increases, parents report decreased quality of life, increased social stigma, increased difficulty meeting material family needs (Ellison et al, 2005; Roca de Bes et al, 2009), increased depression (Ellison et al, 2005; Olivennes et al, 2005; Sheard et al, 2007), decreased marital satisfaction (Roca de Bes et al, 2009), increased fatigue (Sheard et al, 2007) and stress (Golombok et al, 2007; Olivennes et al, 2005; Sheard et al, 2007). Part of this challenge lies in the fact that there are 168 hours in 1 week, but adequately caring for 6-month-old triplets and household activities requires 197.5 hours per week (Bryan, 2003). Multiple fetuses themselves face an increased risk of long-term disabilities that contribute to increased parental fatigue and depression, and overall siblings of multiples are more at risk for behavioral issues (Bryan, 2003). Psychosocial consequences between naturally conceived multiples versus IVF multiples might differ. On the one hand, parents of IVF multiples can experience significantly more stress, increased child difficulty (Cook et al, 1998; Glazebrook et al, 2004), and increased dysfunctional parent-child interactions (Glazebrook et al, 2004) compared with spontaneous multiples. On the other hand, both ART and non-ART twin conceptions are more stressful for parents, creating higher levels of anxiety and depression compared with singletons (Vilska CHAPTER 7 Multiple Gestations and Assisted Reproductive Technology et al, 2009). Regardless of conception mode, multiples potentially have negative psychosocial effects on parents and families. COST Multiple gestations economically influence both the family and society. Preterm delivery, LBW, and postdischarge hospitalization are increased in multiple fetuses, all of which have a role in short- and long-term cost (Cuevas et al, 2005). Annually, neonatal health care consumes $10.2 billion in the Unites States, 57% of which comes from preterm infants (<37 weeks’ gestation) who comprise less than 10% of live births (St. John et al, 2000). According to gestational age, the mean initial hospital charge for infants born between 26 and 28 weeks’ gestation is approximately $240,000 compared with approximately $4800 for a term infant. By birthweight, infants weighing less than 1250 g cost approximately $250,000 compared with infants weighing more than 2500 g, who cost $5800 (Cuevas et al, 2005). ART multiples and their associated comorbidities have a significant role in health care expenditures. In the United Kingdom, IVF-induced multiples account for 27% of pregnancies yearly, but represent 54% of expenditures (Ledger et al, 2006); this is due largely to IVF triplets and twins, which are remarkably more costly than IVF singletons from both neonatal and maternal standpoints. Estimated maternal cost ratios for IVF singleton:twin and singleton:triplet are approximately 1:1.94 and 1:3.96, respectively, and neonatal cost ratios for IVF singleton:twin and singleton:triplet are 1:16 and 1:109, respectively (Ledger et al, 2006). The cost for IVF triplets from diagnosis through 1 year of life is tenfold (Ledger et al, 2006), and IVF twin cost up to 6 weeks postpartum is fivefold versus IVF singletons (Lukassen et al, 2004). ART singletons are more costly than spontaneous singletons, possibly because of increased rates of LBW (Chambers et al, 2007) or the underlying infertility contributing to these outcomes (Koivurova et al, 2004). An Australian study comparing IVF singletons, twins, and HOMs to control counterparts revealed no significant cost differences between ART twins and non-ART twins or between ART HOMs and non-ART HOMs, but combined neonatal-maternal cost was 57% higher for ART births than for non-ART births (Chambers et al, 2007). Because of the extreme economic cost of multiples, one proposed mechanism to reduce multifetal pregnancies is single embryo transfer (SET). Elective SET in first-cycle IVF patients costs less than double embryo transfer (DET) (Fiddelers et al, 2006; Gerris et al, 2004), and although SET produced lower live birth rates than DET in an unselected population (20.8% versus 39.6%) (Fiddelers et al, 2006), when stratified to less than 38 years age, SET and DET generate similar live birth rates (Gerris et al, 2004). DECREASING THE RISK OF MULTIPLES ART procedures and the rate of multiple pregnancies are rising concordantly. Although ART triplets and the incidence of HOMs has declined since 1996, ART twin rates 65 are unchanged.2 Primary forms of preventing multiples include canceling ovulation induction cycles or converting to IVF, and in IVF cycles limiting the number of embryos transferred. Worldwide differences exist in medical practice and laws regarding restrictions on the number of embryos transferred. The American Society for Reproductive Medicine (ASRM) and the Society for Assisted Reproductive Technology established transfer guidelines to assist in determining the appropriate embryo number in an attempt to decrease multiples. Recommended limits are based on age, prognosis, and embryo stage and further differentiate between a favorable patient (first IVF cycle, good quality embryos, number of embryos for potential cryopreservation, successful past IVF cycles) and less favorable conditions. To date there are no embryo transfer guidelines for frozen embryo cycles. Transfer recommendations by the ASRM are not legally binding, and they are subject to interpretation or adjustment based on clinical experience and unique patient instances (Practice Committee of Society for Assisted Reproductive Technology and Practice Committee of American Society for Reproductive Medicine, 2008b). Reduction of the number of embryos transferred and the incidence of triplets or greater in women aged 37 years or less from 1996 to 2003 may be attributed to a change in the 1998-1999 ASRM guidelines (Stern et al, 2007). The most important factor involved in creating multiple fetuses is the number of embryos transferred. Given the emotional, physical, and financial burden of ART and the concern that only one embryo might lower pregnancy rates, multiples fetuses were previously accepted as a known risk in an effort to ensure a pregnancy. More recently, however, practitioners across the globe are stressing the impact of multiples and are encouraging SET (Gerris, 2005). A Swedish study in women aged less than 36 years undergoing their first or second IVF cycle with two or more good-quality embryos were randomized to DET or SET followed by frozen-thawed embryo transfer if unsuccessful. Live birth rates were lower in the SET-alone versus the DET group, but SET followed by frozen-thawed embryo transfer resulted in a 38.8% live birth rate and a 0.8% multiple rate, compared with a 42.9% live birth rate and a 33.1% multiple rate in the DET group, showing that with SET pregnancy rates were acceptable and multiple pregnancy rates were significantly lower (Thurin et al, 2004). Women aged 36 to 39 years may also be candidates for SET, because similar live birth rates between SET and DET and significantly higher cumulative multiple rates occur in the DET (16.6%) versus SET (1.7%) groups (Veleva et al, 2006). Superb cryopreservation technique with frozen embryo transfer after SET in the appropriate patient lowers multiple fetus rates (Gerris, 2005) and leads to comparable cumulative LBR compared with DET (Veleva et al, 2009). Consideration of multiples from non-ART ovulation induction by controlled ovarian hyperstimulation also merits discussion. Twenty-two percent of twins, 40% of triplets, and 71% of HOMs in 2004 were a result of nonART ovulation induction (Dickey, 2009). During gonadotropin stimulation, follicular growth is supervised via ultrasound examination, and estradiol levels are monitored in an attempt to minimize overstimulation. Ovulation 66 PART II Fetal Development induction risk factors for multiples have revealed that HOMs are positively correlated with gonadotropin dose and stimulation length, estradiol levels greater than 1000 pg/mL, and seven or more follicles measuring 10 mm or greater, whereas negative predictors were age less than 32 years, lower body mass index, and a higher number of prior treatment cycles (Dickey, 2009). Techniques to reduce the chance of multiple fetuses include minimizing gonadotropin dose, canceling the cycle by discontinuing medications for excess follicles or high estradiol levels, or converting to IVF (Dickey, 2009; Nakhuda and Sauer, 2005); however, the specific criteria warranting cycle cancelation are not uniform between centers (Practice Committee of the American Society for Reproductive Medicine, 2006). MULTIFETAL PREGNANCY REDUCTION As the rates of ART procedures, multiple fetuses, and prematurity-related sequelae have increased, so has the use of selective reduction. Primary prevention of multiple fetuses by limiting the number of embryos transferred or canceling an overstimulated ovulation induction cycle is optimal; however, in reality multifetal pregnancies continue to occur. Multifetal pregnancy reduction (MFPR) provides another option to enhance overall survival and decrease the risk of fetal or neonatal morbidity and mortality by decreasing pregnancy loss rates and prematurity (Evans and Britt, 2005). First developed in the 1980s, selective termination of one or more fetuses is performed to reduce the final fetal number. The majority of patients reduce to twins, followed by singletons; few reduce to triplets (Stone et al, 2008). A discussion of the ethical, medical, and psychosocial factors involved in MFPR are important counseling points for any patient undergoing ovulation induction (Committee on Ethics, 2007). Improvements in MFPR techniques have enhanced success rates such that quadruplet or triplets reduced to twins have equal outcomes compared with natural twins (Evans and Britt, 2005; Evans et al, 2001). Success rates correlate with both beginning and ending fetal number (Evans et al, 2001). The average loss rate in one series of 1000 MFPR cases was 4.7% (Stone et al, 2008). Loss rates are higher after reducing to a singleton versus reducing to twins, but twins overall have higher morbidity than do singletons (Evans and Britt, 2005, 2008; Evans et al, 2001; Stone et al, 2008). Reduction of twins to singletons may be considered given a lower loss rate after reduction versus continuing with twins (Evans et al, 2004). Benefits after MFPR are apparent in preterm delivery rates, because half of twins and almost 90% of singletons are delivered full term, and 95% were delivered after 24 weeks’ gestation in one series (Stone et al, 2008). Although beneficial in certain cases, MFPR is not without medical and psychological risk. It might not be an option for some women; therefore primary prevention should be the focus for reducing the risk of multiple fetuses. SUMMARY Over the last 30 years, advances in ART have helped countless infertile couples achieve a pregnancy. The percentage of ART live births will likely continue on an upward trend because of increased accessibility of ART and delayed childbearing. These factors in addition to ART techniques will continue to contribute to multiple gestation rates. Multiple gestations are associated with increased maternal, fetal, and neonatal complications that generate a medical, psychological, and economic burden to families and society. Efforts to decrease multifetal pregnancies and prematurity-related sequelae include prevention-based practice policies and further knowledge regarding the mechanisms involved with MZT and ART. SUGGESTED READINGS ACOG Practice Bulletin #56: Multiple gestation: complicated twin, triplet, and high-order multifetal pregnancy, Obstet Gynecol 104:869-883, 2004. Aston KI, Peterson CM, Carrell DT: Monozygotic twinning associated with assisted reproductive technologies: a review, Reproduction 136:377-386, 2008. Chambers GM, Chapman MG, Grayson N, et al: Babies born after ART treatment cost more than non-ART babies: a cost analysis of inpatient birth-admission costs of singleton and multiple gestation pregnancies, Hum Reprod 22:31083115, 2007. Dickey RP: The relative contribution of assisted reproductive technologies and ovulation induction to multiple births in the United States 5 years after the Society for Assisted Reproductive Technology/American Society for Reproductive Medicine recommendation to limit the number of embryos transferred, Fertil Steril 88:1554-1561, 2007. Ellison MA, Hotamisligil S, Lee H, et al: Psychosocial risks associated with multiple births resulting from assisted reproduction, Fertil Steril 83:1422-1428, 2005. Evans MI, Britt DW: Fetal reduction, Semin Perinatol 29:321-329, 2005. Hack KE, Derks JB, Elias SG, et al: Increased perinatal mortality and morbidity in monochorionic versus dichorionic twin pregnancies: clinical implications of a large Dutch cohort study, BJOG 115:58-67, 2008. Hall JG: Twinning, Lancet 362:735-743, 2003. Knopman J, Krey LC, Lee J, et al: Monozygotic twinning: an eight-year experience at a large IVF center, Fertil Steril 94:502-510, 2010. Practice Committee of the American Society for Reproductive Medicine: Multiple pregnancy associated with infertility therapy, Fertil Steril 86(Suppl 1): S106-S110, 2006. Norwitz ER, Edusa V, Park JS: Maternal physiology and complications of multiple pregnancy, Semin Perinatol 29:338-348, 2005. Thurin A, Hausken J, Hillensjo T, et al: Elective single-embryo transfer versus double-embryo transfer in vitro fertilization, N Engl J Med 351:2392-2402, 2004. Complete references and supplemental color images used in this text can be found online at www.expertconsult.com C H A P T E R 8 Nonimmune Hydrops Scott A. Lorch and Thomas J. Mollen An infant with hydrops has an abnormal accumulation of excess fluid. The condition varies from mild, generalized edema to massive edema, with effusions in multiple body cavities and with peripheral edema so severe that the extremities are fixed in extension. Fetuses with severe hydrops may die in utero; if delivered alive, they may die in the neonatal period from the severity of their underlying disease or from severe cardiorespiratory failure. The first description of hydrops in a newborn, in a twin gestation, may have appeared in 1609 (Liley, 2009). ­Ballantyne (1892) suggested that the finding of hydrops was an outcome for many different causes, in contrast to the belief at that time that hydrops was a single entity. Potter (1943) first distinguished between hydrops secondary to erythroblastosis fetalis and nonimmune hydrops, by describing a group of infants with generalized body edema who did not have hepatosplenomegaly or abnormal erythropoiesis. Potter’s description of more than 100 cases of hydrops included two sets of twins in which one had hydrops and the other did not, thus presenting the first description of twintwin transfusion syndrome. With the nearly universal use of anti-D globulin and refinement of the schedule and doses for its administration, the occurrence of immune-mediated hydrops has steadily declined, such that later studies found that immune-mediated causes accounted for only 6% to 10% of all cases of hydrops (Heinonen et al, 2000; Machin, 1989). The reported incidence of nonimmune hydrops in the general population has been highly variable, ranging from 6 per 1000 pregnancies in a high-risk referral clinic in the United Kingdom between 1993 and 1999 (Sohan et al, 2001) to 1 in 4000 pregnancies (Norton, 1994); other published rates are 6 per 1000 pregnancies (Santolaya et al, 1992), 1.3 per 1000 pregnancies (Wafelman et al, 1999), and 1 per 1700 pregnancies (Heinonen et al, 2000). However, all the published studies come from single institutions, with the at-risk populations ranging from that of a highrisk pregnancy clinic to infants in a neonatal intensive care unit. No study has monitored all pregnant women in one geographic area to calculate the true population incidence of nonimmune hydrops, especially monitoring infants who died in utero. Geography also affects the incidence; several causes of nonimmune hydrops, such as α-thalassemia, are more common in certain areas of the world. Finally, the incidence of nonimmune hydrops may be rising because of the more routine use of ultrasound investigation in the late first trimester of pregnancy (Iskaros et al, 1997). ETIOLOGY Nonimmune hydrops has been associated with a wide range of conditions (Table 8-1). In many of these conditions, edema formation results from one of the following possible processes: ll Elevated central venous pressure, in which the cardiac output is less than the rate of venous return ll ll ll Anemia, resulting in high-output cardiac failure Decreased lymphatic flow Capillary leak The actual pathophysiology of hydrops for many of the conditions in Table 8-1, however, is still not understood. The most common causes of nonimmune hydrops are chromosomal, cardiovascular, hematologic, thoracic, infectious, and related to twinning (Abrams et al, 2007; Bellini et al, 2009; Wilkins, 1999). As with reported incidence rates, the relative contribution of these causes varies by study. The studies that focus on early fetal presentation of hydrops (postconceptional age of less than 24 weeks’ gestation) have found that chromosomal abnormalities, such as Turner syndrome and trisomies 13, 18, and 21, are the causes of 32% to 78% of all cases of hydrops (Boyd et al, 1992; Heinonen et al, 2000; Iskaros et al, 1997; McCoy et al, 1995; Sohan et al, 2001). For infants whose hydrops becomes evident after 24 weeks’ gestation, cardiovascular and thoracic causes are most prevalent, with rates ranging between 30% and 50% (Machin, 1989; McCoy et al, 1995; Shan et al, 2001). Studies from Asia have noted a higher percentage of cases from hematologic causes, probably because of the higher rates of α-thalassemia in the population (Lin et al, 1991; Nakayama et al, 1999). The percentage of infants with “idiopathic” hydrops, or hydrops of unknown etiology, varies from 5.2% to 50%, depending on the ability of the clinicians to complete their diagnostic evaluation and the inclusion of fetal deaths in the analysis (Bellini et al, 2009; Heinonen et al, 2000; Iskaros et al, 1997; Machin, 1989; McCoy et al, 1995; Nakayama et al, 1999; Santolaya et al, 1992; Sohan et al, 2001; Wafelman et al, 1999; Wy et al, 1999). Yaegashi et al (1998) used enzyme-linked immunosorbent assay and polymerase chain reaction techniques to improve the detection of parvovirus infection. In both their own institution, and in eight other series of patients, these investigators found evidence of parvovirus infection in 15% to 19% of all infants previously diagnosed with idiopathic hydrops. It is likely that, as there is increased understanding of and testing for many of the conditions listed in Table 8-1, the number of infants diagnosed with idiopathic, nonimmune hydrops will continue to decline. PATHOPHYSIOLOGY NORMAL FLUID HOMEOSTASIS Abnormal body fluid homeostasis is the underlying cause of edema, whether local or generalized. To understand the pathogenesis of hydrops, the clinician must consider the forces underlying normal fluid homeostasis. The regulation of net fluid movement across a capillary membrane depends on the Starling forces, which were first described by E.H. Starling in 1896 (Starling, 1896). Flow between intravascular and interstitial fluid compartments is 67 68 PART II Fetal Development TABLE 8-1 Conditions Associated With Hydrops Fetalis Condition Type Specific Conditions Condition Type Specific Conditions Hemolytic anemias Alloimmune, Rh, Kell, α-Chain hemoglobinopathies (homozygous α-thalassemia) Red blood cell enzyme deficiencies (glucose phosphate isomerase deficiency, glucose-6-phosphate dehydrogenase) Nervous system lesions Absence of corpus callosum Encephalocele Cerebral arteriovenous malformation Intracranial hemorrhage (massive) Holoprosencephaly Fetal akinesia sequence Other anemias Fetomaternal hemorrhage Twin-twin transfusion Diamond-Blackfan Pulmonary conditions Cardiac conditions Premature closure of foramen ovale Ebstein anomaly Hypoplastic left or right heart Subaortic stenosis with fibroelastosis Cardiomyopathy, myocardial fibroelastosis Atrioventricular canal Myocarditis Right atrial hemangioma Intracardiac hamartoma or fibroma Tuberous sclerosis with cardiac rhabdomyoma Cystic adenomatoid malformation of the lung Mediastinal teratoma Diaphragmatic hernia Lung sequestration syndrome Lymphangiectasia Renal conditions Urinary ascites Congenital nephrosis Renal vein thrombosis Invasive processes and storage disorders Tuberous sclerosis Gaucher disease Mucopolysaccharidosis Mucolipidosis Supraventricular tachycardia Atrial flutter Congenital heart block Chromosome ­abnormalities Trisomy 13, trisomy 18, trisomy 21 Turner syndrome XX/XY Bone diseases Hemangioma of the liver Any large arteriovenous malformation Klippel-Trenaunay syndrome Idiopathic infantile arterial calcification Osteogenesis imperfecta Achondroplasia Asphyxiating thoracic dystrophy Gastrointestinal ­conditions Bowel obstruction with perforation and meconium peritonitis Small bowel volvulus Other intestinal obstructions Prune-belly syndrome Tumors Neuroblastoma Choriocarcinoma Sacrococcygeal teratoma Hemangioma or other hepatic tumors Congenital leukemia Cardiac tumors Renal tumors Maternal or placental conditions Maternal diabetes Maternal therapy with indomethacin Multiple gestation with parasitic fetus Chorioangioma of placenta, chorionic vessels, or umbilical vessels Toxemia Systemic lupus erythematosus Miscellaneous Neu-Laxova syndrome Myotonic dystrophy Cardiac arrhythmias Vascular malformations Vascular accidents Thrombosis of umbilical vein or inferior vena cava Recipient in twin-twin transfusion Infections Cytomegalovirus, congenital hepatitis, human parvovirus, enterovirus, other viruses Toxoplasmosis, Chagas disease Coxsackie virus Syphilis Leptospirosis Lymphatic abnormalities Lymphangiectasia Cystic hygroma Noonan syndrome Multiple pterygium syndrome Congenital chylothorax Idiopathic determined by the balance among (1) capillary hydrostatic pressure, (2) serum colloid oncotic pressure, (3) interstitial hydrostatic pressure or tissue turgor pressure, and (4) interstitial osmotic pressure, which depends on lymphatic flow. The Starling equation defines the relationship among these forces and their net effect on net fluid movement, or filtration, across a semipermeable membrane (such as the capillary membrane) as: Filtration = K[(Pc − Pt ) − R(Op − Ot )] where K = capillary filtration coefficient, representing the extent of permeability of a membrane to water and thus describing capillary integrity; Pc = capillary hydrostatic pressure; Pt = interstitial hydrostatic pressure or tissue turgor pressure; R = reflection coefficient for a solute, representing the extent of permeability of the capillary wall to that solute; Op = plasma oncotic pressure as determined by plasma proteins and other solutes; and Ot = interstitial osmotic pressure (Figure 8-1). Although an abnormality of any of the components of this equation may, in theory, result in the accumulation of edema fluid, the fetal-placental unit presents a unique physiologic condition that effectively eliminates two of the factors, assuming unimpeded fetal-placental flow and an appropriately functioning maternal-placental interface. Because approximately 40% of fetal cardiac output is allocated to the placenta, there is rapid transport of water CHAPTER 8 Nonimmune Hydrops Ot Pt Interstitium Pc Op Capillary FIGURE 8-1 Starling forces and net effect on fluid homeostasis. Arrows represent net effect of movement of fluid across the capillary membrane for each factor under normal conditions. Pc, Capillary hydrostatic pressure; Pt, interstitial hydrostatic pressure or tissue turgor pressure; Op, plasma oncotic pressure as determined by plasma proteins and other solutes; Ot, interstitial osmotic pressure. between the fetus and mother. Any condition resulting in elevated fetal capillary hydrostatic pressure or low plasma colloid oncotic pressure would likely cause the net flow of water from fetal villi in the placenta to the maternal blood stream, where it can be effectively eliminated. This elimination of fluid would counteract the accumulation of interstitial fluid by the fetus. Although the placenta of a fetus with hydrops is also edematous, these changes are believed to occur with, and not before, fetal fluid accumulation. DERANGEMENTS IN FLUID HOMEOSTASIS Diamond et al (1932) suggested three possible mechanisms that might be relevant in infants with hydrops: anemia, low colloid osmotic pressure with hypoproteinemia, and congestive heart failure with hypervolemia. Others have reviewed these potential mechanisms (Phibbs et al, 1974), which remain among the central hypotheses addressed by investigators in this area. The causes of hydrops appear to be multifactorial, with mechanisms that produce elevated central venous pressure (CVP), capillary leakage, and impaired lymphatic drainage all contributing to its development. Infants with alloimmune hydrops (and several of the nonimmune hydrops conditions as well) have significant anemia. It has been proposed that anemia leads to congestive heart failure with increased hydrostatic pressure in the capillaries, causing vascular damage that results in edema. However, the hematocrit values of infants with and without hydrops overlap significantly, suggesting that anemia alone is not the complete explanation. A rapidly lowered hemoglobin concentration results in greater cardiac output to maintain adequate oxygen delivery. This output results in higher oxygen demands by the myocardium, which may be difficult to meet because of the anemia. The hypoxic myocardium can become less contractile and less compliant, with ventricular stiffness causing increased afterload to the atria. High-output congestive heart failure may then exist, resulting in elevated CVP. Raised CVP leads to increased capillary filtration pressures and impairment of lymphatic return (Weiner, 1993). In addition, reduced compliance of a right ventricle may result in flow reversal in the inferior vena cava, which may in turn cause end-organ damage to the liver, with consequent hypoalbuminemia and portal hypertension enhancing formation of both edema and ascites. Hydrops has been produced in fetal lambs 69 (Blair et al, 1994) in which the hemoglobin content was lowered in 12 fetuses through exchange transfusion using cell-free plasma; six became hydropic. Anemia developed more rapidly with a higher CVP in fetuses with hydrops than in the fetuses without hydrops. In the most severely anemic fetuses, it is probable that decreased oxygen transport causes tissue hypoxia, which in turn increases capillary permeability to both water and protein. These changes in capillary permeability also likely contribute to the development of hydrops. Infants who have erythroblastosis and hydrops seem to demonstrate a correlation between serum albumin concentration and the severity of hydrops (Phibbs et al, 1974). Initial therapy after birth, however, tends to rapidly raise the serum albumin value toward normal, and with diuresis the albumin concentrations normalize. This finding suggests that hypoalbuminemia may be the result of dilution rather than the cause of hydrops. To elucidate the role of isolated hypoproteinemia in the genesis of hydrops, Moise et al (1991) have induced hypoproteinemia in sets of twin fetal lambs. One twin from each set underwent serum protein reduction through repeated removal of plasma and replacement with normal saline; the other twin served as the control. Over 3 days, plasma protein concentrations were reduced by an average of 41%, with a 44% reduction in colloid osmotic pressure, in experimental subjects. No fetuses became edematous, and total body water content values were similar in experimental and control animals. Thus hypoproteinemia alone was insufficient to cause hydrops fetalis over the course of the study. Transcapillary filtration probably increased with hypoproteinemia, but was compensated by lymphatic return. Human fetuses with hypoproteinemia as a result of nephrotic syndrome or analbuminemia rarely experience hydrops, further supporting the hypothesis that hypoproteinemia alone is not sufficient to cause hydrops. Hypoproteinemia may, however, lower the threshold for edema formation in the presence of impaired lymphatic return or increased intravascular hydrostatic pressures. The most commonly diagnosed causes of nonimmune hydrops that appears in fetuses older than 24 weeks’ gestation are cardiac disorders. Any state in which cardiac output is lower than the rate of venous return results in an elevated CVP. Increased CVP raises capillary filtration pressures and, if high enough, restricts lymphatic return. Both of these mechanisms may then contribute to interstitial accumulation of fluid. Structural cardiac causes of elevated CVP include right-sided obstructive lesions and valvular regurgitation. The most common and easily reversible cause of nonimmune hydrops is supraventricular tachycardia (SVT). In general, cardiac output rises with heart rate. At the increasingly high rates seen in SVT, however, cardiac output plateaus and then diminishes. The heart rates observed with SVT are often associated with decreased cardiac output. Impaired cardiac output results in elevated CVP, which can give rise to edema through mechanisms discussed previously (Gest et al, 1990). Myocardial hypoxia (most often caused by severe anemia) and myocarditis (usually infectious) reduce both the contractility and compliance of the myocardium and can also cause an increase in CVP. A fourth factor that contributes to hydrops is decreased lymph flow. If the rate of fluid filtration from plasma 70 PART II Fetal Development to tissues exceeds the rate of lymph return to the central venous system, then edema and effusions may form. A structural impediment or increased CVP that opposes lymphatic return to the heart can impair lymph flow. To determine the effects of alterations in CVP on lymphatic return, Gest et al (1992) applied an opposing hydrostatic pressure to the thoracic duct in fetal lambs; they inserted catheters into the thoracic ducts of 10 fetuses. Varying the height of the catheter altered the thoracic duct outflow pressure. Thoracic duct flow was nearly constant over the physiologic range of CVP, but sharply decreased at elevated pressures; therefore lymphatic flow may be reduced or essentially blocked in pathologic states associated with elevated CVP. TABLE 8-2 Antenatal Investigation of Fetal Hydrops Area Testing Maternal History, including: Age, parity, gestation Medical and family histories Recent illnesses or exposures Medications Complete blood count and indices Blood typing and indirect Coombs antibody screening Hemoglobin electrophoresis Kleihauer-Betke stain of peripheral blood Syphilis, TORCH, and parvovirus B19 titers Anti-Ro and anti-La, systemic lupus erythematosus preparation Oral glucose tolerance test Glucose-6-phosphate dehydrogenase, pyruvate kinase deficiency screening Fetal Serial ultrasound evaluations Middle cerebral artery peak systolic velocity Limb length, fetal movement Echocardiography Amniocentesis Karyotype Alpha-fetoprotein Viral cultures; polymerase chain reaction analysis for toxoplasmosis, parvovirus 19 Establishment of culture for appropriate metabolic or DNA testing Lecithin-to-sphingomyelin ratio to assess lung maturity Fetal blood sampling Genetic testing Complete blood count Hemoglobin analysis Immunoglobulin M test; specific cultures Albumin and total protein measurements Measurement of umbilical venous pressure Metabolic testing PRENATAL DIAGNOSIS The initial presentation of fetal hydrops varies by report. Watson and Campbell (1986) found that two thirds of prenatally diagnosed cases were discovered on routine ultrasonographic examinations, and one third was referred for evaluation because of suspected polyhydramnios. Graves and Baskett (1984) reported that hydrops was more commonly discovered after referral for polyhydramnios, fetus large for dates, fetal tachycardia, or pregnancy-induced hypertension. Despite the underlying cause of hydrops or the clinical presentation, the prenatal diagnosis is made via the ultrasonographic finding of excess fluid in the form of ascites, pleural or pericardial effusions, skin edema, placental edema, or polyhydramnios. Several definitions for ultrasonographic diagnosis based on quantity and distribution of excess fluid have been proposed. One widely accepted set of criteria consists of the presence of excess fluid in any two of the previously listed compartments. Because this definition is based on the presence of excess fluid alone, the degree of severity is generally subjective. Swain et al (1999) outlined a multidisciplinary approach to the evaluation and management of the mother and fetus with hydrops. Table 8-2 provides recommendations for the investigation of fetal hydrops. Patient history should focus on ethnic background, familial history of consanguinity, genetic or congenital anomalies, and complications of pregnancy, including recent maternal illness and environmental exposures. Maternal disorders such as diabetes, systemic lupus erythematosus, myotonic dystrophy, and any type of liver disease should also be noted. Initial laboratory investigation includes blood typing and a Coombs’ test to rule out immune-mediated hydrops. Other blood tests are a screening for hemoglobinopathies, a Kleihauer-Betke test to eliminate fetal-maternal hemorrhage, and testing for the TORCH diseases (i.e., toxoplasmosis, other infections, rubella, cytomegalovirus infection, and herpes simplex), including syphilis and parvovirus B19. Further evaluation is directed at identifying possible causes (Forouzan, 1997). Rapid evaluation is necessary to determine whether fetal intervention is possible and to estimate the prognosis for the fetus. Many conditions, such as arrhythmias, twin-twin transfusion, large vascular masses, and congenital diaphragmatic hernias and other chest-occupying lesions, are discovered during the initial ultrasonographic evaluation (Coleman et al, 2002). Middle cerebral artery peak systolic velocity measurement can Adapted from Swain S, Cameron AD, McNay MB, et al: Prenatal diagnosis and management of nonimmune hydrops fetalis, Aust N Z J Obstet Gynaecol 39:285-290, 1999. TORCH, Toxoplasmosis, other infections, rubella, cytomegalovirus, and herpes simplex. aid in detecting the presence of fetal anemia (HernandezAndrale et al, 2004). If the initial ultrasonic examination is not helpful in identifying a cause, it may be helpful to repeat it at a later date to reassess fetal anatomy, monitor progression of the hydrops, and evaluate well-being of the fetus. Fetal echocardiography should also be performed to evaluate for cardiac malformations and arrhythmia. Amniotic fluid can be obtained for fetal DNA analysis, cultures, and lecithin-to-sphingomyelin ratio to assess lung maturity. Fetal blood sampling allows for other tests, such as a complete blood cell count, routine chemical analyses, DNA analysis, bacterial and viral cultures, metabolic studies, and serum immunoglobulin measurements. PRENATAL MANAGEMENT The goals of antenatal evaluation of fetal hydrops depend on the underlying cause. In diagnoses in which therapy is futile, the goal is to avoid unnecessary invasive testing and cesarean section. The prognosis should be discussed frankly with the parents, who should be given the option of terminating the pregnancy. If the underlying cause is CHAPTER 8 Nonimmune Hydrops amenable to fetal therapy, the risks and benefits of such therapy, as well as the warning that diagnostic error is possible, should be discussed with the family. SVT is one of the most common known causes of nonimmune hydrops, and it is the most amenable to treatment (Huhta, 2004; Newburger and Keane, 1979). Usually the mother is given antiarrhythmic agents, and the fetus is monitored closely for resolution of the SVT. Digoxin is most commonly administered, although other antiarrhythmics have been used, such as sotolol or flecainide, because transplacental transfer of digoxin may be impaired in the setting of hydrops. In extreme circumstances, such as fetal tachyarrhythmia refractory to maternal treatment, direct fetal administration of antiarrhythmic agents via percutaneous umbilical blood sampling or intramuscular injection, although untested and highly risky, has met with some success. If anemia is the cause of hydrops, transfusions of packed red blood cells may be administered to the fetus. Often a single transfusion reverses the edema, although serial transfusions may be necessary. Parvovirus B19 (Anand et al, 1987) and fetal-maternal hemorrhage are examples of diagnoses that are amenable to this therapy. Other diagnoses involving anemias that are refractory to transfusions, such as α-thalassemia, may require neonatal stem cell transplantation. Transfusions should be given, with the use of ultrasonographic guidance, into the intraperitoneal space or umbilical vein. Blood instilled into the abdominal cavity is taken up by lymphatics, but elevated CVPs present in hydropic fetuses may impair this uptake. If uptake of intraperitoneal blood is incomplete, treatment for the hydrops is less successful; degeneration of the remaining hemoglobin may create a substantial bilirubin load, necessitating phototherapy or exchange transfusion after the infant is delivered. Surgery continues to evolve as a promising therapy for select cases of fetal hydrops (Azizkhan and Crombleholme, 2008; Kitano et al, 1999). Fetal lung lesions such as congenital cystic adenomatoid malformation (CCAM) and pulmonary sequestration can, in the most extreme cases, result in mediastinal shift, pulmonary hypoplasia, cardiovascular compromise, and hydrops. A recent review of 36 fetuses with CCAM found higher rates of hydrops in infants whose mass-to-thorax ratio was greater than 0.56, in whom the lesion had a cystic predominance, or in whom the hemidiaphram was everted (Vu et al, 2007). Adzick et al (1998) reported on the outcome of 175 cases of fetal lung masses, including 134 cases of CCAM and 41 of extralobar pulmonary sequestration. The 76 fetuses with CCAM lesions without associated hydrops were all managed expectantly (maternal transport to a high-risk center, planned delivery near term, and resection in the newborn period). CCAMs frequently involute and may disappear before delivery; therefore there were no deaths in the nonhydropic group of fetuses in this study. Twentyfive fetuses with hydrops were managed expectantly, and all 25 died before or after preterm labor at 25 to 26 weeks’ gestation. These results highlight the fact that fetuses with lung lesions leading to hydrops have high mortality rates. Thirteen fetuses with CCAM and associated hydrops underwent open fetal resection or lobectomy. Eight survived and were reported as healthy at 1 to 7 years of follow-up. Maternal morbidities related to fetal intervention ranged from uterine wound infection with dehiscence to 71 mild postoperative interstitial pulmonary edema, which was treated with diuretics. High morbidity and mortality rates in severe twintwin transfusion with associated hydrops led to multiple international trials of laser photocoagulation of interfetal vascular connections. Although the trials met with varying results, metanalysis involving the three major trials demonstrates improvement in perinatal and neonatal outcomes. However, the current level of evidence is limited in the reported effect on neurodevelopmental outcomes in survivors. A 2008 Cochrane review recommends considering treatment with laser coagulation at all stages of twin-twin transfusion (Roberts et al, 2008). Fetal intervention has met with some success in other diagnoses with associated hydrops. Thoracoamniotic shunts for large unicystic lesions and pleuroamniotic shunts for hydrothorax have reportedly enhanced survival in extreme cases. Similarly, in cases of massive urinary ascites, urinary diversion via peritoneal shunts has been reported, but with a poor long-term prognosis (Crombleholme et al, 1990). However, as with other invasive interventions, there are potential risks with fetal surgery. A review of the reproductive outcomes of future pregnancies after a pregnancy complicated by maternal-fetal surgery found a complication rate of 35%: 12% affected by uterine dehiscience, 6% with uterine rupture, 3% requiring hysterectomy; and 9% with hemorrhage requiring transfusion (Wilson et al, 2004). These longer-term complications suggest that the potential benefit of fetal surgical intervention must be balanced by the potential complications of the procedure experienced by the mother. In cases in which the cause can be corrected by appropriate care at the time of delivery, such as elimination of a chorioangioma, and in cases in which no cause can be ascertained, close observation for fetal demise is the focus of prenatal management. Many cases of nonimmune hydrops manifest in the third trimester as preterm labor. It is difficult to decide whether to attempt tocolysis and delay delivery so as to allow the potentially beneficial administration of steroids before birth or to deliver the fetus immediately. If tocolysis is possible, expectant management should include usual biophysical testing, although fetal decompensation may be difficult to measure. Abnormal fetal heart tracings, oligohydramnios, decreased fetal movement, and poor fetal tone are all ominous signs. There is no indication to prolong pregnancy beyond attainment of a mature lung profile unless available evidence indicates improvement or resolution of the hydrops. NEONATAL EVALUATION Table 8-3 summarizes the diagnostic evaluations recommended for newborn infants with nonimmune hydrops of unknown cause. INTENSIVE CARE OF THE INFANT WITH HYDROPS FETALIS After successful resuscitation, including intubation, administration of surfactant, and placement of umbilical catheters, the clinical management can address both the cause and the complications of hydrops. Morbidity and 72 PART II Fetal Development TABLE 8-3 Diagnostic Evaluation of Newborns With Nonimmune Hydrops System Type of Evaluation Cardiovascular Echocardiogram, electrocardiogram Pulmonary Chest radiograph, pleural fluid examination Hematologic Complete blood cell count, differential platelet count, blood type and Coombs’ test, blood smear for morphologic analysis Gastrointestinal Abdominal radiograph, abdominal ultrasonography, liver function tests, peritoneal fluid examination, total protein and albumin levels Renal Urinalysis, blood urea nitrogen and ­creatinine measurements Genetic Chromosomal analysis, skeletal radiographs, genetic consultation Congenital ­infections Viral cultures or serologic testing, including TORCH agents and parvovirus Pathologic Complete autopsy, placental examination Adapted from Carlton DP, McGillivray BC, Schreiber MD: Nonimmune hydrops fetalis: a multidisciplinary approach, Clin Perinatol 16:839-851, 1989. TORCH, Toxoplasmosis, other infections, rubella, cytomegalovirus, and herpes simplex. mortality may result from the hydropic state, the underlying conditions giving rise to hydrops, or both. A fetus with hydrops that is delivered prematurely is subject to the additional complications of prematurity. If there is massive ascites or pleural effusions, initial resuscitation may require thoracentesis or peritoneal tap. Because of pulmonary edema, infants with hydrops are susceptible to pulmonary hemorrhage and require high levels of positive end-expiratory pressure. RESPIRATORY MANAGEMENT Virtually all infants with hydrops require mechanical ventilation because of pleural and peritoneal effusions, pulmonary hypoplasia, surfactant deficiency, pulmonary edema, poor chest wall compliance caused by edema, or persistent pulmonary hypertension of the newborn. The presence of persistent pleural effusions may necessitate the placement of chest tubes. Ascites may also compress the diaphragm and impair lung expansion. Breath sounds, chest movement, blood gas levels, and radiographs must all be monitored frequently, so that ventilator support can be reduced in response to improvements in lung compliance and water clearance. Pneumothoraces and pulmonary interstitial emphysema remain potential complications as long as ventilator support is continued. Infants who need a prolonged course of ventilation, particularly those born prematurely, may develop bronchopulmonary dysplasia. Chronic lung disease results in a longer and more complicated hospital course and contributes to the late mortality of hydrops. FLUID AND ELECTROLYTE MANAGEMENT A primary goal of fluid management is resolution of the hydrops itself. Maintenance fluids should be restricted, with volume boluses given only in response to clear signs of inadequate intravascular volume. The hydropic newborn has an excess of free extracellular water and sodium. Fluids given during resuscitation further increase the amount of water and sodium that must be removed during the immediate neonatal period. Initial maintenance fluids should contain minimal sodium. Serum and urine sodium levels, urine volume, and daily weights should be monitored carefully to guide administration of fluids and electrolytes. Urinary sodium levels may help differentiate between hyponatremia caused by hemodilution and urinary losses. CARDIOVASCULAR MANAGEMENT Shock may be a prominent feature of patients with hydrops. Hydropic infants may have hypovolemia as a result of capillary leakage, poor vascular tone, and impaired myocardial contractility from hypoxia or infection, impaired venous return caused by shifting or compression of mediastinal structures, or pericardial effusion. Adequate intravascular volume must be maintained, and correctable causes of impaired venous return should be addressed. Peripheral perfusion, heart rate, blood pressure, and acid-base status should be monitored carefully. CLINICAL COURSE AND OUTCOME Despite improvements in diagnosis and management, mortality from nonimmune hydrops remains high. Reported survival rates for all fetuses diagnosed antenatally with hydrops range from 12% to 24% (Heinonen et al, 2000; McCoy et al, 1995; Negishi et al, 1997). Higher survival rates have been reported in infants born alive, but the highest rates are still only 40% to 50% (Wy et al, 1999). Improved ultrasonography techniques and earlier testing may actually lead to lower survival rates as hydrops is diagnosed in more first-trimester infants. These infants are more likely to have chromosomal abnormalities that are incompatible with survival, but were previously not included in populations of hydropic fetuses. The best predictor of survival is the cause of the hydrops and the gestational age of the child at delivery. Highest survival rates are seen in infants with parvovirus infection, chylothorax, or SVT. The lowest survival rates are for hydrops from chromosomal cause, although the figures may be biased because a significant number of the pregnancies in such cases are terminated (Heinonen et al, 2000; Sohan et al, 2001). A recent review of 598 patients with nonimmune hydrops found other risk factors for increased mortality including younger gestational age, lower 5-minute Apgar score, and the need for increased respiratory support (Abrams et al, 2007). A smaller study from Taiwan also found that lower albumin levels were associated with a higher mortality rate (Huang et al, 2007). Interventions to improve outcomes in hydrops are limited by the rarity of the disease. Carlton et al (1989) reported on a group of 36 infants with nonimmune hydrops and noted that 90% of the infants who died within 24 hours had pleural effusions, compared with only 50% of those who survived. More than one third of the infants in this study required thoracentesis in the delivery room to aid in lung expansion. All the infants who lived more than 24 hours were treated with mechanical ventilation and CHAPTER 8 Nonimmune Hydrops received supplemental oxygen; they needed ventilation for an average of 11 days (range, 2 to 48 days). Most hydropic infants lose a minimum of 15% of their birthweight, and some lose as much as 30%. Ordinarily, diuresis begins on the second or third day after birth and continues for a period of 2 to 4 days. Once the edema has resolved, the infants have normal levels of circulating protein and eventually recover from their apparent capillary leak syndrome. No specific management strategies during the neonatal period, such as the use of high-frequency oscillatory ventilation, have been shown to improve outcome, although the published studies are powered to detect small survival differences (Wy et al, 1999). For infants who survive the immediate neonatal period, long-term outcomes appear to be excellent. Nonimmune hydrops by itself does not seem to lead to residual developmental delay. A small study from Japan found that 13 of 19 surviving infants with nonimmune hydrops had normal development at 1 to 8 years (Nakayama et al, 1999). The six infants with mild or severe delays in this study had other morbidities, such as extreme prematurity, structural cardiac lesions, or chromosomal anomalies. Thus 73 long-term morbidities from nonimmune hydrops appear to result from the underlying cause of the hydrops, gestational age at delivery, and complications arising immediately after delivery. SUGGESTED READINGS Abrams ME, Meredith KS, Kinnard P, et al: Hydrops fetalis: a retrospective review of cases reported to a large national database and identification of risk factors associated with death, Pediatrics 120:84-89, 2007. Anand A, Gray ES, Brown T, et al: Human parvovirus infection in pregnancy and hydrops fetalis, N Engl J Med 316:183-186, 1987. Azizkhan RG, Crombleholme TM: Congenital cystic lung disease: contemporary antenatal and postnatal management, Pediatr Surg Int 24:643-657, 2008. Huhta JC: Guidelines for the evaluation of heart failure in the fetus with or without hydrops, Pediatr Cardiol 25:274-286, 2004. Bellini C, Hennekam RCM, Fulcheri E, et al: Etiology of nonimmune hydrops fetalis: a systematic review, Am J Med Genet 149A:844-851, 2009. Machin GA: Hydrops revisited: literature review of 1414 cases published in the 1980s, Am J Med Genet 34:366-390, 1989. Swain S, Cameron AD, McNay MB, et al: Prenatal diagnosis and management of nonimmune hydrops fetalis, Aust N Z J Obstet Gynaecol 39:285-290, 1999. Complete references used in this text can be found online at www.expertconsult.com P A R T III Maternal Health Affecting Neonatal Outcome C H A P T E R 9 Endocrine Disorders in Pregnancy Gladys A. Ramos and Thomas R. Moore DIABETES IN PREGNANCY Currently, 17 million people in the United States have a form of diagnosed diabetes. Alarmingly, the data for 2003 to 2006 indicate that approximately 10.2% (11.5 million) of women older than 20 years have diabetes. Data indicate that new cases of type 2 diabetes mellitus are occurring at an increasing rate among American Indian, African American, Hispanic, and Latino children and adolescents (http://diabetes.niddk.nih.gov/dm). Continued immigration among populations with high rates of type 2 diabetes mellitus and the effects of changes in diet (increases in number of calories and fat content) and lifestyle (sedentary) portend marked rises in the percentage of patients with preexisting diabetes who will become pregnant in the future. There is also an epidemic of childhood obesity currently under way in the United States, with approximately 23 million (30%) children and youth who are overweight. This trend will have a profound effect on obstetrics and pediatric practice in the future. Expanded efforts to reach the populations at risk are necessary if a significant increase in maternal and neonatal morbidity is to be avoided (Persson and Hanson, 1998). Depending on the population surveyed, abnormalities of glucose regulation occur in 3% to 8% of pregnant women. Although more than 80% of this glucose intolerance arises only during pregnancy (gestational diabetes) and involves relatively modest episodes of hyperglycemia, the attendant fetal and newborn morbidity is disproportionate. Compared with weight-matched controls, infants of diabetic mothers (IDMs) have double the risk of serious birth injury, triple the likelihood of cesarean section, and quadruple the incidence of admission to a newborn intensive care unit. Studies indicate that the magnitude of risk of these maloccurrences is proportional to the level of maternal hyperglycemia. Therefore, to some extent, the excessive fetal and neonatal morbidity of diabetes in pregnancy is preventable or at least reducible through meticulous prenatal and intrapartum care. MATERNAL-FETAL METABOLISM IN NORMAL PREGNANCY AND DIABETIC PREGNANCY Normal Maternal Glucose Regulation With each meal, a complex combination of maternal hormonal actions, including the secretion of pancreatic insulin, glucagon, somatomedins, and adrenal catecholamines, ensures an ample but not excessive supply of glucose to the mother and fetus during pregnancy. The key effects of pregnancy on maternal metabolic regulation are as follows: ll ll ll ll Because the fetus continues to draw glucose from the maternal bloodstream across the placenta, even during periods of fasting, the tendency toward maternal hypoglycemia between meals becomes increasingly marked as pregnancy progresses and fetal glucose demand grows. Placental steroid and peptide hormone production (estrogens, progesterone, and chorionic somatomammotropin) rises linearly throughout the second and third trimesters, resulting in a progressively increasing tissue resistance to maternal insulin action. Progressive maternal insulin resistance requires a significant augmentation in pancreatic insulin production (more than twofold nonpregnant levels) during feeding to maintain euglycemia. Twenty-four–hour mean insulin levels are 30% higher in the third trimester than in the nonpregnant state. If pancreatic insulin output is not adequately augmented, maternal hyperglycemia and then fetal hyperglycemia result. The severity of hyperglycemia and its timing depend on the relative inadequacy of insulin production. Fetal Effects of Maternal Hyperglycemia Congenital Anomalies A major threat to IDMs is the possibility of a life-threatening structural anomaly. In the normoglycemic pregnancy, the risk of a major birth defect is 1% to 2%. Among women with pregestational diabetes, the risk of a fetal structural anomaly is fourfold to eightfold higher. In a recent cohort study of 2359 pregnancies in women with pregestational diabetes, the rate of anomalies was more than doubled. Major congenital anomalies occurred in 4.6% overall with 4.8% for type 1 diabetes mellitus and 4.3% for type 2 diabetes mellitus. This is a significant increase over the expected rate of birth defects in the general population (approximately 1.5%). Neural tube defects in IDM were increased 4.2-fold (Figure 9-1), and congenital heart disease by 3.4-fold. Prenatal diagnosis of these anomalies was accomplished in 65% of neonates (Macintosh et al, 2006). The typical defects and their frequency of occurrence, 75 76 PART III Maternal Health Affecting Neonatal Outcome experience with 105 diabetic patients, finding an overall malformation rate of 13.3%. A recent study conducted in the United Kingdom from 1991 to 2000 in patients with type 1 diabetes mellitus found similar results (Temple et al, 2002). Adverse outcome was significantly higher in the poor control group (HbA1c ≥7.5) than in the fair control group (HbA1c <7.5), with a ninefold increase in the congenital malformation rate (relative risk, 9.2; 1.1 to 79.9) (Temple et al, 2002). For a woman with an HbA1c value of less than 7.1%, the risk of delivering a malformed infant was equivalent or slightly less than that for the normoglycemic population. However, the anomaly rate rose progressively with increasing HbA1c, 14% with an HbA1c value of 7.2% to 9.1%, 23% with an HbA1c value of 9.2% to 11.1%, and 25% with an HbA1c value of greater than 11.2%. Pathogenesis FIGURE 9-1 Newborn with caudal regression syndrome, macrosomia, and respiratory distress. The mother had type 1 diabetes and a glycosylated hemoglobin concentration of 13.5% when first seen for prenatal care at 12 weeks’ gestation. ( From Creasy RK, Resnik R, editors: Maternal-fetal medicine: principles and practice, ed 2, Philadelphia, 1989, WB Saunders.) TABLE 9-1 Congenital Malformations in Infants of Mothers With Insulin-Dependent Diabetes Appropriate Risk Ratio Risk (%) All cardiac defects 18× 8.5 All central nervous system anomalies 16× 5.3 Anomaly Anencephaly 13× – Spina bifida 20× – All congenital anomalies 8× 18.4 Becerra JE, Khoury MJ, Cordero JF, et al: Diabetes mellitus during pregnancy and the risks for specific birth defects: a population based case-control study, Pediatrics 85:1, 1990. noted in a prospective study of infants with major malformations, are listed in Table 9-1. The majority of lesions involve the central nervous and cardiovascular systems, although other series have reported an excess of genitourinary and limb defects (Cousins, 1991). There is no increase in birth defects among offspring of diabetic fathers, nondiabetic women, or women in whom gestational diabetes develops after the first trimester. These findings suggest that glycemic control during embryogenesis is a critical factor in the genesis of diabetes-associated birth defects. In a study by Miller et al (1981), the frequency of congenital anomalies was proportional to the maternal glycohemoglobin (HbA1c) value in the first trimester (rate of anomalies 3.4% with HbA1c <8.5%, and 22.4% with HbA1c >8.5%). Lucas et al (1989) reported a similar The specific mechanisms by which hyperglycemia disturbs embryonic development are incompletely elucidated, but reduced levels of arachidonic acid and myo-inositol and accumulation of sorbitol and trace metals in the embryo have been demonstrated in animal models (Pinter et al, 1986). Fetal hyperglycemia may promote excessive formation of oxygen radicals in the mitochondria of susceptible tissues, leading to the formation of hydroperoxides, which inhibit prostacyclin. The resulting overabundance of thromboxanes and other prostaglandins may then disrupt vascularization of developing tissues. In support of this theory, the addition of prostaglandin inhibitors to mouse embryos in culture medium prevents glucose-induced embryopathy. Furthermore, the addition of dietary antioxidants in the form of high doses of vitamins C and E decreased fetal dysmorphogenesis to nondiabetic levels in rat pregnancy and rat embryo culture (Cederberg and Eriksson, 2005; El-Bassiouni et al, 2005). Prevention Because the critical period for teratogenesis is the first 3 to 6 weeks after conception, normal glycemic control must be instituted before pregnancy to prevent these birth defects. Several clinical trials of meticulous preconception glycemic control in women with diabetes have resulted in malformation rates equivalent to those in the general population (Fuhrmann et al, 1983). A recent metaanalysis of these trials demonstrated that the pooled risk of malformations was lower in women with preconception care compared with those without preconception counseling (Ray et al, 2001). The threshold level of glycemic control, as evidenced by the HbA1c value, necessary to normalize a patient’s risk of congenital anomalies appears to be a nearnormal value. Thus any elevation of the HbA1c above normal increases the risk of teratogenesis proportionately. Macrosomia Fetal overgrowth is a major problem in pregnancies complicated by diabetes, leading to unnecessary cesarean sections and potentially avoidable birth injuries. A 1992 study of birthweights in the previous 20 years indicated that 21% of infants with birthweights of 4540 g or greater were born to mothers who were glucose intolerant, a rate clearly disproportionate to the only 2% to 5% of gravidas with some form of diabetes (Shelley-Jones et al, 1992). Thus the CHAPTER 9 Endocrine Disorders in Pregnancy problem of abnormal fetal growth in diabetic pregnancy remains an important clinical challenge. Macrosomia is defined variously as birthweight above the 90th percentile for gestational age or birthweight greater than 4000 g; it occurs in 15% to 45% of diabetic pregnancies. Excessive fetal size contributes to a greater frequency of intrapartum injury (shoulder dystocia, brachial plexus palsy, and asphyxia). Macrosomia is also a major factor in the higher rate of cesarean delivery among diabetic women. Because the risk of macrosomia is fairly constant for all classes of diabetes, it is likely that first-trimester metabolic control has less of an effect on fetal growth than does glycemic regulation in the second and third trimesters. Growth Dynamics IDMs with macrosomia follow a unique pattern of in utero growth compared with fetuses in euglycemic pregnancies. During the first and second trimesters, differences in size between fetuses born to diabetic and nondiabetic mothers are usually undetectable with ultrasound measurements. After 24 weeks, however, the growth velocity of the IDM fetus’ abdominal circumference typically begins to rise above normal (Ogata et al, 1980). Reece et al (1990) demonstrated that the IDM fetus has normal head growth, despite marked degrees of hyperglycemia. Landon et al (1989) have reported that although head growth and femur growth of IDM fetuses were similar to those of normal fetuses, abdominal circumference growth significantly exceeded that of controls beginning at 32 weeks’ gestation (abdominal circumference growth in IDM fetuses is 1.36 cm/week, versus 0.901 cm/week in normal subjects). Morphometric studies of the IDM newborn indicate that the greater growth of the abdominal circumference is caused by deposits of fat in the abdominal and interscapular areas. This central depositing of fat is a key characteristic of diabetic macrosomia and underlies the pathology associated with vaginal delivery in these pregnancies. Acker et al (1986) showed that although the incidence of shoulder dystocia is 3% among infants weighing more than 4000 g, the incidence in infants from diabetic pregnancies who weigh more than 4000 g is 16%. Finally, despite our emphasis on birthweight, this alone may not be a sensitive measure of fetal growth. Catalano et al (2003) conducted body composition studies on infants born to mothers with diabetes and found that even when appropriate for gestational age, these infants have increased fat mass and percent body fat compared with a normoglycemic control group. Childhood Effects Higher growth velocity, begun in fetal life during a pregnancy complicated by diabetes, may extend into childhood and adult life. Silverman et al (1995) reported follow-up of IDMs through age 8 years in which half the infants weighed more than the 90th percentile for gestational age at birth. By age 8 years, approximately half of the IDMs weighed more than the heaviest 10% of the nondiabetic children. The asymmetry index was 30% higher in diabetic offspring than in the controls by age 8 years. These investigators also showed that offspring with diabetes have permanent derangement of glucose-insulin kinetics, resulting in a higher incidence of impaired glucose tolerance. In addition, Dabelea et al (2000) also reported that the mean 77 adolescent body mass index was 2.6 kg/m2 greater in sibling offspring of diabetic pregnancies compared with the index siblings born when the mother previously had normal glucose tolerance. Pathophysiology The pathophysiology of excessive fetal growth is complex and reflects the delivery of an abnormal nutrient mixture to the fetoplacental unit, regulated by an abnormal confluence of growth factors. Pedersen (1952) hypothesized that maternal hyperglycemia stimulates fetal hyperinsulinemia, which in turn mediates acceleration of fuel utilization and growth. The features of the abnormal growth in diabetic pregnancy include excessive adipose deposition, visceral organ hypertrophy, and acceleration of body mass accretion (Ogata et al, 1980). Data from the Diabetes in Early Pregnancy project suggest that maternal metabolic control is a critical factor leading to fetal macrosomia (Jovanovic-Peterson et al, 1991). In this study, in which meticulous glycemic care was maintained in early pregnancy and beyond, fetal weight did not correlate significantly with fasting glucose levels. During the second and third trimesters, however, postprandial blood glucose levels were strongly predictive of both birthweight and the overall percentage of macrosomic infants. With postprandial glucose values averaging 120 mg/dL, approximately 20% of infants had macrosomia; a 30% rise in postprandial levels to 160 mg/dL resulted in a predicted percentage of macrosomia of 35%. In contrast, Persson et al (1996) showed that fasting glucose concentrations account for 12% of the variance in birthweight and correlated best with estimates of neonatal fat. Similarly, Uvena-Celebrezze et al (2002) found that the strongest correlation was between fasting glucose and neonatal adiposity rather than postprandial measures. The Pedersen hypothesis (Pedersen, 1977) presumes that abnormal fuel milieu in the maternal bloodstream is reflected contemporaneously in the fetal compartment: Maternal hyperglycemia = Fetal hyperglycemia. Studies by Hollingsworth and Cousins (1981) have confirmed much of Pedersen’s hypothesis and note the following features of normal pregnancy: ll ll ll Maternal fasting blood glucose levels decline from approximately 85 mg/dL to 75 mg/dL. Mean blood glucose also declines. At night, maternal glucose levels drop markedly as the fetus continues to draw glucose stores from the maternal circulation. Postprandial peaks in maternal blood glucose rarely exceed 120 mg/dL at 2 hours or 130 mg/dL at 1 hour. In addition, in diabetic pregnancies: ll ll If maternal glucose levels surge excessively after a meal, the consequent fetal hyperglycemia is accompanied by fetal pancreatic beta-cell hyperplasia and hyperinsulinemia. Fetal hyperinsulinemia, lasting only episodically for 1 to 2 hours, has detrimental consequences for fetal growth and well-being, in that it (1) promotes storage of excess nutrients, resulting in macrosomia, and (2) drives catabolism of the oversupply of fuel, using energy and depleting fetal oxygen stores. 78 PART III Maternal Health Affecting Neonatal Outcome 8 6 Adjusted odds ratio 7 1 hour Glucose Fasting category mg/dL 6 1 2 3 4 5 6 7 5 4 3 75 75–79 80–84 85–89 90–94 95–99 100 1 hour mg/dL 2 hour mg/dL 106 106–132 133–155 156–171 172–193 194–211 212 90 91–108 109–125 126–139 140–157 158–177 178 3 4 5 6 7 Glucose category FIGURE 9-2 Adjusted odds ratios for cord C-peptide levels at >90th percentile at different glucose categories based on a 2-hour oral glucose challenge test. (Adapted from HAPO Study Cooperative Research Group, Metzger BE, Lowe LB, et al: Hyperglycemia and adverse pregnancy outcomes, N Engl J Med 358:1991-2002, 2008.) ll 5 1 hour Glucose Fasting category mg/dL 1 2 3 4 5 6 7 4 3 75 75–79 80–84 85–89 90–94 95–99 100 2 hour 1 hour mg/dL 2 hour mg/dL 106 106–132 133–155 156–171 172–193 194–211 212 90 91–108 109–125 126–139 140–157 158–177 178 2 2 1 Fasting 2 hour Adjusted odds ratio Fasting 1 2 Episodic fetal hypoxia stimulated by episodic maternal hyperglycemia leads to an outpouring of adrenal catecholamines, which in turn causes hypertension, cardiac remodeling, and cardiac hypertrophy. The Hyperglycemia and Adverse Pregnancy Outcomes (HAPO) study provided additional evidence in support of the Pedersen hypothesis. In this well-designed, multicenter prospective trial, women underwent a 75-g oral glucose challenge test between 28 and 32 weeks’ gestation. Providers were blinded to test results if the fasting value was 105 mg/dL or less and the 2-hour plasma glucose level was 200 mg/dL or less. The women did not receive any therapeutic intervention. The study found a continuous association between maternal glucose values, cord C-­peptide levels (Figure 9-2) and birthweight above the 90th percentile (Figure 9-3) (HAPO Study Cooperative Research Group et al, 2008). A followup analysis of the same cohort evaluating neonatal anthropometric measurements, found a link between maternal hyperglycemia, cord C-peptide levels and neonatal adiposity (HAPO Study Cooperative Research Group et al, 2009). This study suggested that maternal hyperglycemia results in neonatal adiposity that is mediated by fetal insulin production (C-peptide level >90th percentile) (HAPO Study Cooperative Research Group et al, 2009). Prevention of Macrosomia Because macrosomic fetuses are at an increased risk for immediate complications related to birth injury and for potential long-term consequences such as late childhood obesity and insulin resistance, measures for prevention of macrosomia have been recommended. As described previously, fetal hyperinsulinemia, which acts as a fetal growth factor, occurs in response to fetal hyperglycemia, which in turn reflects the maternal hyperglycemic condition. Therefore, measures that promote consistent maternal euglycemia may prevent macrosomia. Several prospective trials have shown that strict maternal glycemic control using insulin and dietary therapy and fastidious blood glucose monitoring can reduce the incidence of macrosomia 1 1 2 3 4 5 6 7 Glucose category FIGURE 9-3 Adjusted odds ratios for macrosomia at >90th percentile at different glucose categories based on a 2-hour oral glucose challenge test. (Adapted from HAPO Study Cooperative Research Group, Metzger BE, Lowe LB, et al: Hyperglycemia and adverse pregnancy outcomes, N Engl J Med 358:1991-2002, 2008.) (Coustan and Lewis, 1978; Langer et al, 1994; Thompson et al, 1990). Langer et al (1994) compared the outcomes of diabetic pregnancies managed conventionally (four blood glucose measurements per day) or intensely (seven blood glucose measurements per day). Fasting blood glucose values were maintained between 60 and 90 mg/dL and 2-hour postprandial values at less than 120 mg/dL. Outcomes were compared to nondiabetic control pregnancies. The rate of infants born weighing more than 4000 g was 14% in the conventionally managed group, 7% in the intensely managed group, and 8% in nondiabetic controls. Similarly, the rate of shoulder dystocia was 1.4% in the conventionally managed group, 0.4% in the intensely managed group, and 0.5% in the control group. Thus, like the reduction of congenital anomalies in mothers with diabetes by means of first-trimester euglycemia, strict glycemic control in the second and third trimesters may reduce the fetal macrosomia rate to near baseline. Fetal Hypoxic Stress As noted previously, episodic maternal hyperglycemia promotes a fetal catabolic state in which oxygen depletion occurs. Several fetal metabolic adaptive responses to this episodic hypoxia occur. For example, the drop in fetal oxygen tension causes stimulation of erythropoietin, red cell hyperplasia, and elevation in fetal hematocrit. Polycythemia can lead to poor circulation and postnatal hyperbilirubinemia. Profound episodic hyperglycemia in the third trimester causing severe fetal hypoxic stress has been theorized as the cause of sudden intrauterine fetal demise in poorly controlled diabetes. CLASSIFYING AND DIAGNOSING DIABETES IN PREGNANCY The classification system for diabetes in pregnancy recommended by White has been replaced by a scheme based on the pathophysiology of hyperglycemia and developed by the National Diabetes Data Group (NDDG) in 1979 (Hare and White, 1980). The two types are summarized 79 CHAPTER 9 Endocrine Disorders in Pregnancy in Table 9-2. This nomenclature is useful because it categorizes patients according to the underlying pathogenesis of their diabetes—­insulin-deficient (type 1) and insulinresistant (type 2 and gestational). One must remember that the diagnosis of gestational diabetes applies to any woman who is found to have hyperglycemia during pregnancy. A certain percentage of such women actually have type 2 diabetes, but the diagnosis cannot be confirmed until postpartum testing. Pregestational Diabetes Patients with type 1 diabetes mellitus typically exhibit hyperglycemia, ketosis, and dehydration in childhood or adolescence. Often the diagnosis is made during a hospital admission for diabetic ketoacidosis and coma. Rarely is the diagnosis of type 1 diabetes mellitus made during pregnancy. Conversely, it is not unusual for women with a tentative diagnosis of gestational diabetes to be found to have overt, type 2 diabetes mellitus after delivery. The American Diabetic Association has outlined three criteria for diagnosing type 2 diabetes mellitus in nonpregnant subjects. They include the finding of a casual plasma glucose of 200 mg/dL or greater, a fasting plasma glucose of 126 mg/dL or greater, or a 2-hour glucose value of 200 mg/dL or greater on a 75-g, 2-hour glucose tolerance test (GTT). Diagnostic criteria are listed in Box 9-1. Gestational Diabetes Gestational diabetes mellitus (GDM) is defined as glucose intolerance that begins or is first recognized during pregnancy (American Diabetes Association, 2002). Almost TABLE 9-2 Classification of Diabetes Mellitus Type Old ­Nomenclature Clinical Features Type 1 Juvenile-onset diabetes Insulin-deficient, ketosisprone; virtually all patients with type 1 diabetes mellitus are insulin dependent Type 2 Adult-onset diabetes Insulin-resistant, not ­ketosis-prone; few patients with type 2 diabetes mellitus are truly insulin dependent Gestational — Occurs during and resolves after pregnancy; insulin-­ resistant; not ketosis-prone uniformly, GDM arises from significant maternal insulin resistance, a state similar to type 2 diabetes mellitus. In many cases, GDM is simply preclinical type 2 diabetes mellitus unmasked by the hormonal stress imposed by the pregnancy. Although GDM complicates no more than 5% to 6% of pregnancies in the United States, the prevalence of GDM in specific populations varies from 1% to 14% (American Diabetes Association, 2002). Clinical recognition of GDM is important because therapy—including medical nutrition therapy, insulin when necessary, and antepartum fetal surveillance—can reduce the welldescribed perinatal morbidity and mortality associated with GDM. Traditionally, universal screening for GDM has been recommended (Metzger, 1991). However, the Fourth International Workshop-Conference on Gestational Diabetes and the American College of Obstetricians and Gynecologists have now indicated that either a risk-factor approach or universal screening can be considered (American College of Obstetricians and Gynecologists, 2001; Metzger and Coustan, 1998). This recommendation is based on the findings of Sermer et al (1994), who reported the results of screening Canadian women at 26 weeks’ gestation with the 100-g, 3-hour GTT. They identified several risk factors as significantly increasing the likelihood of GDM, among which were maternal age of 35 years or more, body mass index higher than 22 kg/m2, and Asian or other ethnicity than white. Women with one or no risk factors had a 0.9% risk of GDM, whereas the risk for those with two to five factors was 4% to 7%. By limiting screening for GDM to patients with more than one risk factor, these investigators were able to reduce testing by 34% while retaining a sensitivity rate of approximately 80%, with a false-positive result rate of 13%. Therefore in patients meeting all criteria listed in Table 9-3 and Box 9-2, it may be cost effective to avoid screening. Currently a multicenter trial is underway by National Institute of Child Health and Human Development Maternal Fetal Medicine Unit Network in which women with GDM will be randomized to standard therapy versus no therapy (Landon et al, 2009). This randomized study will address whether identification and treatment of GDM decrease perinatal morbidity. Notwithstanding these findings, multiple studies from more heterogeneous U.S. populations have demonstrated TABLE 9-3 Oral Glucose Tolerance Test for Gestational Diabetes Venous Plasma Glucose Level* BOX 9-1 C  riteria for the Diagnosis of Type 2 Diabetes Mellitus ll ll ll Symptoms of diabetes (polyuria, polydipsia, and unexplained weight loss) and a casual plasma glucose level >200 mg/dL (11.1 mmol/L). Casual is defined as any time of day without regard to time of last meal, or Fasting glucose level >126 mg/dL (7.0 mmol/L). Fasting is defined as no caloric intake for at least 8 hours, or Two-hour plasma glucose >200 mg/dL (11.1 mmol/L) during a 75-g oral glucose challenge test. Adapted from the American Diabetes Assoication: Clinical practice recommendations: ­Standards of medical care for diabetes, 2007, Diabetes Care 30:s4-s41, 2007. Fasting value 100-g Glucose Load 75-g Glucose Load mg/dL mg/dL mmol/L mmol/L 95 5.3 95 5.3 1-hr value 180 10.0 180 10.0 2-hr value 155 8.6 155 8.6 3-hr value 140 7.8 — — *Test prerequisites: 1-hr, 50-g glucose challenge result >135 mg/dL; overnight fast of 8 to 14 hours; carbohydrate loading for 3 days, including >150 g carbohydrate; seated and not smoking during the test; two or more values must be met or exceeded; either a 2-hr (75-g glucose) or 3-hr (100-g glucose) test can be performed. 80 PART III Maternal Health Affecting Neonatal Outcome BOX 9-2 C  riteria for Low Risk of Gestational Diabetes* ll ll ll ll ll ll Age <25 years Normal prepregnancy body weight No first-degree relatives with diabetes Not a member of an ethnic group at high risk for GDM No history of GDM in prior pregnancy No history of adverse pregnancy outcome Data from Metzger BE, Coustan DR: Summary and Recommendations of the Fourth International Workshop-Conference on Gestational Diabetes Mellitus. The Organizing Committee, Diabetes Care 21(Suppl 2):B161-B167, 1998. GDM, Gestational diabetes mellitus. *Screening for GDM may be omitted only if all criteria are met. the inadequacy of risk factor based screening of patients for GDM. Lavin et al (1981) noted that if only those with risk factors were screened, the percentage of GDM cases detected was similar to the detection rate in those without risk factors (1.4%). A later study (Weeks et al, 1995), which assessed the effect of screening only patients with risk factors, reported that selective screening would have failed to detect 43% of cases of GDM. Moreover, 28% of the women with undiagnosed GDM would have required insulin and had a several-fold higher risk of cesarean section because of macrosomia. Universal screening should be performed in women in ethnic groups at a higher risk for glucose intolerance during pregnancy, namely those of Hispanic, African, Native American, South or East Asian, Pacific Islands, or Indigenous Australian ancestry. For simplicity of administration, universal screening, with the possible exception of the lowest risk category, is probably best. The universal screening method for GDM has been shown to result in earlier diagnosis and improved pregnancy outcomes, including lower rates of macrosomia and a decrease in neonatal admissions to neonatal intensive care units (Griffin et al, 2000). The timing of screening for GDM is important. Because maternal insulin resistance rises progressively during pregnancy, screening too early can miss some patients who will become glucose intolerant later. Screening too late in the third trimester can limit the time during which metabolic interventions can take place. Thus, risk factors for GDM should be assessed at the initial prenatal visit. Factors that should lead to a first-trimester glucose challenge test are listed in Box 9-3. In the remaining patients, screening should be performed with the use of 50 g of glucose at 26 to 28 weeks’ gestation. Various threshold levels for the 50-g glucose challenge are in use, including 140 mg/dL, 135 mg/dL, and 130 mg/dL. The sensitivity of the GDM testing regimen depends on the threshold value used. The most commonly used threshold, 140 mg/dL, detects only 80% of patients with GDM and results in requiring a 3-hour oral GTT in approximately 10% to 15% of patients. Using a challenge threshold of 135 mg/dL improves sensitivity to more than 90% but increases the number of 3-hour oral GTTs by 42% (Ray et al, 1996). Thus, the clinician encountering a newborn with multiple stigmata of an IDM, yet whose mother had a negative diabetes screening test result, should realize that this result during pregnancy does not rule out GDM. This is also why every patient who delivered an BOX 9-3 Indications for the First-Trimester 50-g Glucose Challenge ll ll ll ll ll ll ll Maternal age >25 years Previous infant >4 kg Previous unexplained fetal demise Previous pregnancy with gestational diabetes Strong immediate family history of type 2 or gestational diabetes mellitus Maternal obesity (>90 kg) Fasting glucose level >140 mg/dL (7.8 mmol/L) or random glucose reading >200 mg/dL (11.1 mmol/L) infant with macrosomia in a prior pregnancy should be screened early in all subsequent pregnancies. Definitive diagnosis of gestational diabetes is made with a GTT. Either 100 g of glucose and 3 hours of testing, or 75 g of glucose and 2 hours of testing can be used. The diagnostic criteria are shown in Box 9-1. Two or more values must be met or exceeded for the diagnosis of GDM to be made. A GTT should be performed after overnight fasting and with modest carbohydrate loading before the test. PERINATAL COMPLICATIONS OF DIABETES DURING PREGNANCY Fetal Morbidity and Mortality Perinatal Mortality Perinatal mortality in diabetic pregnancy has decreased 30-fold since the discovery of insulin in 1922 and the advent of intensive obstetric and infant care in the 1970s (Figure 9-4). Improved techniques for maintaining maternal euglycemia have led to later timing of delivery and have reduced the incidence of iatrogenic respiratory distress syndrome (RDS). Nevertheless, the currently reported perinatal mortality rates among women with diabetes remain approximately twice those observed in nondiabetic women (Table 9-4). Congenital malformations, RDS, and extreme prematurity account for most perinatal deaths in contemporary diabetic pregnancy. Figure 9-5 shows the different rates of RDS in diabetic and euglycemic pregnancies. In the past decade, fewer intrauterine deaths have been reported, probably reflecting more careful fetal monitoring. Nevertheless, intrapartum asphyxia and fetal demise remain persistent problems. Birth Injury Birth injury, including shoulder dystocia (Keller et al, 1991) and brachial plexus trauma, is more common among IDMs, and macrosomic fetuses are at the highest risk (Mimouni et al, 1992). Shoulder dystocia, defined as difficulty in delivering the fetal body after expulsion of the fetal head, is an obstetric emergency that places the fetus and mother at great risk. Shoulder dystocia occurs in 0.3% to 0.5% of vaginal deliveries among normal pregnant women; the incidence is twofold to fourfold higher in women with diabetes, probably because the hyperglycemia in a diabetic pregnancy causes the fetal shoulder and abdominal widths CHAPTER 9 Endocrine Disorders in Pregnancy 60 50 40 30 50 40 30 10 18901922 1926- 1946- 1956- 1966- 1970- 1973- 19761945 1955 1965 1969 1972 1975 1981 TIME PERIODS FIGURE 9-4 Perinatal mortality rate (percentage) among infants of diabetic mothers from 1890 to 1981. ( From Creasy RK, Resnik R, editors: Maternal-fetal medicine: principles and practice, ed 2, Philadelphia, 1989, WB Saunders. Data from Craigin EB, Ryder GH: Obstetrics: a practical textbook for students and practitioners, Philadelphia, 1916, Lea and Febiger; DeLee JB: The principles and practice of obstetrics, ed 3, Philadelphia, 1920, WB Saunders; Jorge CS, Artal R, Paul RH, et al: Antepartum fetal surveillance in diabetic pregnant patients, Am J Obstet Gynecol 141:641-645, 1981; Pedersen J: The pregnant diabetic and her newborn, ed 2, Baltimore, 1977, Williams and Wilkins; and Williams JW: Obstetrics: a textbook for the use of students and practitioners, New York, 1925, D Appleton.) TABLE 9-4 Perinatal Mortality Rates (No. of Deaths per 100 Births) in Diabetic and Normal Pregnancies Mortality Mothers With Gestational Diabetes Mothers With Preexisting Diabetes Healthy Mothers* Fetal 4.7 10.4 Neonatal 3.3 12.2 5.7 4.7 Perinatal 8.0 11.6 10.4 *California data for 1986, corrected for birthweight, sex, and race. to become massive (Nesbitt et al, 1998). This relationship was investigated by Athukorala et al, (2007), who found a strong association with fasting hyperglycemia such that with each 1-mmol increase in the fasting value in the oral glucose-tolerance test there was an increasing relative risk (RR) of 2.09 (95% confidence interval [CI], 1.03 to 4.25) for shoulder dystocia. Although half of shoulder dystocias occur in infants of normal birthweight (2500 to 4000 g), the incidence of shoulder dystocia is 10-fold higher (5% to 7%) among infants weighing 4000 g or more and rises to 31% for infants whose mothers have diabetes (Gilbert et al, 1999; see also Chapter 15 for a discussion of complicated deliveries and Chapter 64 for a discussion of the neurologic consequences of birth injury.) Neonatal Morbidity and Mortality For a complete discussion of neonatal morbidity and ­mortality, see Chapter 94. Diabetic 20 10 0 20 0 60 NEW PERINATAL AND NEONATAL TECHNOLOGY RATE OF RDS (%) PERINATAL MORTALITY (percent) PRE INSULIN 70 ERA 81 Non-Diabetic 30 31 32 33 34 35 36 37 38 39 40 GESTATIONAL AGE (WEEKS) FIGURE 9-5 Rate of respiratory distress syndrome (RDS) versus gestational age. Improved management of maternal glycemic control permits delaying delivery until after 38 weeks’ gestation, when the risk of RDS approaches that in nondiabetic pregnancy. ( From Moore TR: A comparison of amniotic fluid fetal pulmonary phospholipids in normal and diabetic pregnancy, Am J Obstet Gynecol 186:641-650, 2002.) Polycythemia and Hyperviscosity Polycythemia (defined as central venous hemoglobin concentration >20 g/dL or hematocrit >65%) is not uncommon in IDMs and is apparently related to glycemic control. Widness et al (1990) demonstrated that hyperglycemia is a powerful stimulus to fetal erythropoietin production, probably mediated by decreased fetal oxygen tension. Neonatal polycythemia may promote vascular sludging, ischemia, and infarction of vital tissues, including the kidneys and central nervous system. Neonatal Hypoglycemia Approximately 15% to 25% of neonates delivered from women with diabetes during gestation will develop hypoglycemia during the immediate newborn period (Alam et al, 2006). This complication is usually much milder and less common in the infant of a woman whose insulindependent diabetes is well controlled throughout the entire pregnancy and who exhibits euglycemia during labor and delivery. Unrecognized postnatal hypoglycemia can lead to neonatal seizures, coma, and brain damage; therefore, it is imperative that the nurseries receiving IDMs have a protocol for frequent monitoring of the infant’s blood glucose level until metabolic stability is ensured. Hyperbilirubinemia The risk of hyperbilirubinemia is higher in IDMs than in normal infants. There are multiple causes of hyperbilirubinemia in IDMs, but prematurity and polycythemia are the primary contributing factors. Increased destruction of red blood cells contributes to the risks of jaundice and kernicterus. This complication is usually managed using phototherapy, but exchange transfusions may be necessary for marked bilirubin elevations. Hypertrophic and Congestive Cardiomyopathy In some infants with macrosomia of mothers with poorly controlled diabetes, a thickened myocardium and significant septal hypertrophy have been described (Gutgesell et al, 1976). Although the prevalence of myocardial 82 PART III Maternal Health Affecting Neonatal Outcome hypertrophy in IDMs may exceed 30% at birth, almost all cases have resolved by 1 year of age (Mace et al, 1979). Hypertrophic cardiac dysfunction in a newborn IDM often leads to respiratory distress, which may be mistaken for hyaline membrane disease. IDMs with cardiomegaly may have either congestive or hypertrophic cardiomyopathy. Echocardiograms show a hypercontractile, thickened myocardium, often with septal hypertrophy disproportionate to the ventricular free walls. The ventricular chambers are often smaller than normal, and there may be anterior systolic motion of the mitral valve, producing left ventricular outflow tract obstruction. The pathogenesis of hypertrophic cardiomyopathy in IDMs is unclear, although it is recognized to be associated with poor maternal metabolic control. There is evidence that the fetal myocardium is particularly sensitive to insulin during gestation, and Susa et al (1979) reported a doubling of cardiac mass in hyperinsulinemic fetal rhesus monkeys. The myocardium is known to be richly endowed with insulin receptors. Recently maternal insulin-like growth factor-1 (IGF-1), which is elevated in suboptimally controlled diabetic pregnancies, has been shown to be significantly elevated among neonates with asymmetrical septal hypertrophy. Because IGF-1 does not cross the placenta, it can exert its action through binding to the IGF-1 receptor on the placenta (Hayati et al, 2007). Halse et al (2005) noted that B-type natriuretic peptide, a marker for congestive cardiac failure, is elevated in neonates whose mothers had poor glycemic control during the third trimester. IDMs can also have congestive cardiomyopathy without hypertrophy. Echocardiography shows the myocardium to be overstretched and poorly contractile (Jaeggi et al, 2001). This condition is often rapidly reversible with correction of neonatal hypoglycemia, hypocalcemia, and polycythemia; it responds to digoxin, diuretics, or both. In contrast, treatment of hypertrophic cardiomyopathy with an inotropic or diuretic agent tends to further decrease the size of the ventricular chambers and leads to obstruction of blood flow. Prenatal echocardiogram can identify septal hypertrophy and other structural abnormalities, but routine fetal echocardiogram in diabetics has not been proved to be costeffective or to improve outcomes (Bernard et al, 2009). Respiratory Distress Syndrome Since the 1970s, improved maternal management and better protocols for timing of delivery have resulted in a dramatic decline in the incidence of RDS from 31% to 3%. Nevertheless, respiratory dysfunction in the newborn IDM continues to be a common complication of diabetic pregnancy; it may be because of surfactant deficiency or another form of pulmonary distress. Surfactant production occurs late in diabetic pregnancies. Studies of fetal lung ion transport in the diabetic rat by Pinter et al (1991) demonstrated decreased fluid clearance and a lack of thinning of the lung’s connective tissue in diabetic rats compared with controls. In humans, Kjos et al (1990) noted respiratory distress in 18 of 526 infants delivered after diabetic gestations (3.4%). Surfactant-deficient airway disease accounted for fewer than one third of cases, with transient tachypnea, hypertrophic cardiomyopathy, and pneumonia responsible for the majority. As a result, the near-term infant of a mother with poorly controlled diabetes is more likely to have neonatal RDS than the infant of a mother without diabetes at the same gestational age. This circumstance further compounds the diabetic infant’s metabolic and cardiovascular difficulties after birth. The fetus without diabetes achieves pulmonary maturity at a mean gestational age of 34 to 35 weeks. By 37 weeks’ gestation, more than 99% of normal newborn infants have mature lung profiles as assessed by phospholipid assays. In a diabetic pregnancy, however, it is unwise to assume that the risk of respiratory distress has passed until after 38.5 weeks’ gestation (Moore, 2002). Any delivery contemplated before 38.5 weeks’ gestation for other than the most urgent fetal and maternal indications should be preceded by documentation of pulmonary maturity through amniocentesis. OBSTETRIC COMPLICATIONS Pregnancy complicated by diabetes is subject to a number of obstetric disorders, including ketoacidosis, preeclampsia, polyhydramnios, and abnormal labor, at higher rates than in nondiabetic pregnancy. Preeclampsia Preeclampsia is an unpredictable multisystem disorder in which maternal neurologic, renal, and cardiovascular status can decline precipitously, which can threaten fetal health through placental ischemia and abruptio placentae. Preeclampsia is more common among women with diabetes, occurring two to three times more frequently in women with pregestational diabetes than in nondiabetic women (Moore et al, 1985; Sibai et al, 2000). However, the risk of developing preeclampsia is proportional to the duration of diabetes before pregnancy and the existence of nephropathy and hypertension; preeclampsia develops in more than one third of women who have had diabetes for more than 20 years. Patients with type 2 diabetes mellitus without complications have a risk profile similar to that of patients without diabetes, but the risk of hypertensive complications is 50% higher in women with evidence of renal or retinal vasculopathy than in those with no hypertension. The rate of fetal death is higher in women with diabetes and preeclampsia. In the patient with diabetes and chronic hypertension, preeclampsia may be difficult to distinguish from near-term blood pressure elevations. The onset is typically insidious and not confidently recognized until it is severe. Renal function assessment (creatinine, blood urea nitrogen, uric acid, and 24-hour urine collection) should be performed each trimester in women with evidence of pregestational diabetes and vascular disease. When a patient with diabetes experiences preeclampsia, she should be evaluated for delivery. If signs of severe disease are present (e.g., blood pressure >160 mm Hg systolic and 110 mm Hg diastolic, neurologic symptoms, or significant renal dysfunction), delivery should be performed promptly. In mild cases, the patient may be observed if the fetal lungs are immature. Preeclampsia after 38 weeks’ gestation, however, is appropriate grounds for initiating delivery. CHAPTER 9 Endocrine Disorders in Pregnancy Polyhydramnios Polyhydramnios is defined as excess amniotic fluid. The precise clinical definition varies, encompassing the recording of more than 2000 mL of amniotic fluid at delivery and various measures of amniotic fluid pocket depths as observed on ultrasonography. In practice, polyhydramnios is usually diagnosed when any single vertical pocket of amniotic fluid is deeper than 8 cm (equivalent to the 97th percentile) or when the sum of four pockets, one from each quadrant of the uterus (amniotic fluid index), exceeds approximately 24 cm (95th percentile; Moore and Cayle, 1990). The principal cause of hydramnios in diabetic pregnancies is usually poor glycemic control, although fetal gastrointestinal anomalies (e.g., esophageal atresia) need to be excluded. A rapid increase in fundal height should prompt a thorough ultrasound examination by a skilled examiner. The main clinical problems associated with hydramnios are fetal malposition and preterm labor. Management of the patient with hydramnios is predominantly symptomatic, focused on improving glycemic control and preventing premature labor. Enhanced patient awareness of contractions and the signs and subtle sensations of preterm labor is essential. MANAGEMENT Preconception Management Preconception counseling and a detailed medical risk assessment are recommended for all women with overt diabetes as well as for those with a history of gestational diabetes in a previous pregnancy. The significant effects on the maternal and neonatal complications of diabetic pregnancy cannot be realized until meticulous preconception metabolic control is achieved in all women contemplating pregnancy. The important elements to be considered in preconception counseling of patients with diabetes are the patient’s level of glycemic control; current status of the patient’s retinal and renal health; and any medications being taken, especially antihypertensive or thyroid medications. A realistic assessment of the patient’s risk of complications during pregnancy, including worsening of renal or ophthalmologic function, should be provided. Preconception management should lead to a comprehensive program of glucose control. The major goals of the prepregnancy metabolic program are as follows: ll ll ll ll Establishing a regimen of frequent, regular monitoring of capillary blood glucose levels Adopting an insulin dosing regimen that results in a smooth interprandial glucose profile (fasting blood glucose value 90 to 99 mg/dL, 1-hour postprandial glucose level of less than 140 mg/dL or 2-hour postprandial glucose level less than 120 mg/dL, no reactions between meals or at night) Bringing HbA1c level into the normal range Developing family, financial, and personal resources to assist the patient if pregnancy complications require that she lose work time or assume total bed rest This preconception care has been shown to decrease congenital anomalies and to result in fewer hospitalizations, 83 fewer infants requiring intravenous glucose after delivery, and a substantial reduction in total costs (Herman et al, 1999). Prenatal Metabolic Management The goals of glycemic monitoring, dietary regulation, and insulin therapy in diabetic pregnancy are to prevent the postnatal sequelae of diabetes in the newborn—macrosomia, shoulder dystocia, and postnatal metabolic instability. These measures must be instituted early and aggressively if they are to be effective. Principles of Dietary Therapy Because women with diabetes have inadequate insulin action after feeding, the goal of dietary therapy is to avoid single, large meals and foods with a large proportion of simple carbohydrates. Three major meals and three snacks are prescribed. The use of nonglycemic foods that release calories into the gut slowly also improves metabolic control. Nutritional therapy should be supervised by a trained professional who performs formal dietary assessment and counseling at several points during the pregnancy. The dietary prescription should provide adequate quantity and distribution of calories and nutrients to meet the needs of the pregnancy and support achieving the plasma glucose targets that have been established. For obese women (body mass index >30 kg/m2), a 30% to 33% restriction in caloric intake (to 25 kcal per kilogram of actual weight per day or less) has been shown to reduce hyperglycemia and plasma triglycerides with no increase in ketonuria. Moderate restriction of dietary carbohydrate intake to 35% to 40% of calories has been shown to reduce maternal glucose levels and improve maternal and fetal outcomes (Major et al, 1998). In a nonrandomized study, subjects with low carbohydrate intake (<42% of calories) had lower requirements for insulin for glucose control and significantly lower rates of macrosomia and cesarean deliveries for cephalopelvic disproportion and macrosomia. Recently, a randomized trial was performed in which 958 women with gestational diabetes were randomized to usual prenatal care or treatment (Landon, 2009). Women randomized to treatment underwent nutritional counseling, diet therapy, and insulin if indicated. Among those who underwent treatment there was lower mean birthweight, neonatal fat mass, rates of large for gestational age and macrosomic (>4000 g) infants. There was also a trend toward lower cord C-peptide levels in the treatment group. Maternal outcomes were significant for lower rates of cesarean delivery, preeclampsia, and shoulder dystocia (Landon et al, 2009). Principles of Glucose Monitoring The availability of chemical test strips for capillary blood glucose measurements has revolutionized the management of diabetes, and their use should now be considered the standard of care for pregnancy monitoring. The discipline of measuring and recording blood glucose levels before and after meals may have the effect of improving glycemic control (Goldberg et al, 1986). Controversy exists as to whether the target glucose levels to be maintained during diabetic pregnancy should be 84 PART III Maternal Health Affecting Neonatal Outcome designed to limit macrosomia or to closely mimic nondiabetic pregnancy profiles. The Fifth International Workshop Conference on Gestational Diabetes (Metzger et al, 2007) recommended the following: fasting plasma glucose less than 90 to 99 mg/dL (5.0 to 5.5 mmol/L) and 1-hour postprandial plasma glucose less than 140 mg/dL (7.8 mmol/L), or 2-hour postprandial plasma glucose less than 120 to 127 mg/dL (6.7 to 7.1 mmol/L). The glycemic control profiles from Cousins et al (1980) were derived from highly controlled studies in which volunteer subjects were fed test meals with specific caloric content on a rigid schedule. Parretti et al (2001) profiled normal pregnant women twice monthly, preprandially, and postprandially during the third trimester. Testing was conducted with capillary glucose meters, and the women followed an ad libitum diet. The data demonstrate that fasting and premeal plasma glucose levels are usually less than 80 mg/dL and often less than 70 mg/dL. Peak postprandial plasma glucose values rarely exceed 110 mg/dL. Yogev et al (2004) used a sensor that monitored interstitial fluid glucose levels to obtain continuous glucose information from pregnant women without diabetes and found similar results to those of Parretti et al (2001). The range of normal glucose levels in nondiabetic pregnancy is summarized in Table 9-5. Postprandial values must be assessed because they have the strongest correlation with fetal growth (JovanovicPeterson et al, 1991). The Diabetes in Early Pregnancy Study found that postprandial glucose levels were strongly predictive of both birthweight and the overall percentage of macrosomic infants. With postprandial glucose values averaging 120 mg/dL, approximately 20% of infants were macrosomic; a 30% rise in postprandial levels to 160 mg/dL resulted in a predicted percentage of macrosomia of 35%. Similar results were reported by de Veciana et al (1995). Compared with the group who performed preprandial glucose monitoring, the group performing postprandial glucose monitoring demonstrated a greater mean change in the HbA1c value (3.0% vs. 0.6%; p <0.001), lower birthweights (3469 g vs. 3848 g; p = 0.01), and lower rates of both neonatal hypoglycemia (3% vs. 21%, p = 0.05) and macrosomia (12% vs. 42%; p = 0.01). A typical schedule involves performing blood glucose checks upon rising in the morning, 1 or 2 hours after breakfast, before and after lunch, before and after dinner, and before bedtime. The goal of physiologic glycemic control in pregnancy, however, is not met by simply avoiding hypoglycemia. The data summarized here regarding fetal macrosomia and postnatal morbidity emphasize the key role of excessive postprandial excursions in blood glucose values. Therefore, close attention must be paid to preprandial and postprandial glycemic profiles. Principles of Insulin Therapy No available insulin delivery method approaches the precise secretion of the hormone from the human pancreas. The therapeutic goal of exogenous insulin therapy during pregnancy is to achieve diurnal glucose excursions similar to those of nondiabetic pregnant women. Normal pregnant women maintain postprandial blood glucose excursions within a relatively narrow range (70 to 120 mg/dL). As pregnancy progresses, the fasting and between-meal blood glucose levels drop progressively lower as a result of the continual uptake of glucose from the maternal circulation by the growing fetus. Any insulin regimen for pregnant women must be designed to avoid excessive unopposed insulin action during the fasting state. Insulin type and dosage frequency should be individualized. Use of regular insulin before each major meal helps to limit postprandial hyperglycemia. To provide basal insulin levels between feedings, a longer-acting preparation is necessary, such as isoprotane insulin (NPH) or insulin zinc (Lente). Typical subcutaneous insulin dosing regimens are two thirds of total insulin in the morning, of which two thirds as intermediate-acting and one third as regular insulin. The remaining third of the total insulin dose is given in the evening, with 50% as short-acting insulin before dinner and 50% as intermediate-acting given at bedtime. The use of an insulin pump for type 1 diabetes mellitus during pregnancy has become more widespread (Gabbe et al, 2000). An advantage of this approach is the more physiologic insulin release pattern that can be achieved with the pump. Ultrasonography has also been used to direct insulin management (Rossi et al, 2000). Kjos et al (2001) showed that serial normal fetal abdominal circumference measurements can be used to avoid insulin therapy without increasing neonatal morbidity. Oral Hypoglycemic Therapy Historically, insulin has been the mainstay of therapy for gestational diabetes because of early reports that oral hypoglycemic drugs are a potential cause of fetal anomalies and neonatal hypoglycemia. Sulfonylurea compounds are contraindicated during pregnancy because of a high level of transplacental penetration and clinical reports of prolonged and severe neonatal hypoglycemia (Zucker and Simon, 1968). An increased rate of congenital malformations, particularly ear anomalies, has been reported from a small case-control study (Piacquadio et al, 1991). TABLE 9-5 Ambulatory Glucose Values in Pregnant Women With Normal Glucose Tolerance No. of Subjects Fasting (mg/dL) Postprandial Level at 60 min (mg/dL) Parretti et al, 2001 51 69 (57-81) 108 (96-120) — Yogev et al, 2004 57 75 (51-99) 105 (79-131) 110 (68-142)* Study Postprandial Peak (mg/dL) Adapted from Metzger BE, Buchanan TA, Coustan DR, et al: Summary and recommendations of the Fifth International Workshop-Conference on Gestational Diabetes Mellitus, Diabetes Care 30:S251-S260, 2007. *The time of the peak postprandial glucose concentration was 70 minutes (range, 44 to 96 minutes). CHAPTER 9 Endocrine Disorders in Pregnancy However, when Towner et al (1995) evaluated the frequency of birth defects in patients who took oral hypoglycemic agents during the periconception period, they noted that first-trimester HbA1c level and duration of diabetes were strongly associated with fetal congenital anomalies, but that use of oral hypoglycemic medications was not. Glyburide, a second-generation sulfonylurea, has been shown to cross the placenta minimally in laboratory studies (Elliott et al, 1994) and in a large clinical trial. The prospective, randomized trial conducted by Langer et al (2000) compared glyburide and insulin in 404 women with gestational diabetes and showed equivalently excellent maternal glycemic control and perinatal outcomes. Beyond this single, encouraging study, experience with glyburide during pregnancy is limited (Coetzee and Jackson, 1985; Lim et al, 1997). Chmait et al (2004), reporting experience with 69 patients with gestational diabetes who were given glyburide, found a failure rate of 19% (>10% glucose values above target). Glyburide failure rate was higher in women receiving a diagnosis earlier in pregnancy (20 vs. 27 weeks’ gestation; p <0.003) and whose average fasting glucose in the week before starting glyburide was higher (126 vs. 101 mg/dL). Following the publication of the randomized control trial, several retrospective series have been published comprising 504 glyburide-treated patients, summarized recently by Moore (2007). Jacobson et al (2005) performed a retrospective cohort comparison of glyburide and insulin treatment of gestational diabetes. The insulin group (n = 268) consisted of those diagnosed in 1999 through 2000 and the glyburide group (n = 236) was diagnosed in 2001 through 2002. Glyburide dosing was begun with 2.5 mg in the morning and increased by 2.5 to 5.0 mg weekly. If the dose exceeded 10 mg daily, twice daily dosing was considered. If glycemic goals were not met on a maximum daily dose of 20 mg, treatment was changed to insulin. There were no statistically significant differences in gestational age at delivery, mode of delivery, birthweight, large for gestational age (LGA), or percent macrosomia. The rate of preeclampsia doubled in the glyburide group (12% vs 6%; p = 0.02). Women in the glyburide group also had significantly lower posttreatment fasting and postprandial blood glucose levels. The glyburide group was also superior in achieving target glycemic levels (86% vs. 63%; p <0.001). The failure rate (transfer to insulin) was 12%. The study size, however, was insufficient to detect less than a doubling of the rate of macrosomia/LGA and a 44% increase in neonatal hypoglycemia. Conway et al (2004) reported a retrospective cohort of 75 glyburide-treated patients with GDM. Good glycemic control was achieved by 84% of the subjects with glyburide, and treatment for 16% was switched to insulin. The rate of fetal macrosomia was similar between women successfully treated with glyburide and those who received insulin (11.1% vs. 8.3%; p = 1.0), and mean birthweight was also similar. Of note, a nonsignificantly higher proportion of infants in the glyburide group required intravenous glucose infusions because of hypoglycemia (25.0% vs. 12.7%; p = 0.37). Currently there is a growing acceptance of glyburide use as a primary therapy for GDM (Coustan, 2007). 85 OTHER AGENTS Metformin is frequently used in patients with polycystic ovary syndrome and type 2 diabetes mellitus to improve insulin resistance and fertility (Legro et al, 2007). Metformin therapy has been demonstrated to improve the success of ovulation induction (Vandermolen et al, 2001) and may reduce first-trimester pregnancy loss in women with polycystic ovary syndrome (Jakubowicz et al, 2002). However, the effects of continuing metformin treatment during pregnancy are currently being studied. Older studies evaluating the efficacy and safety of the treatment of pregestational and gestational diabetics with metformin raised concerns regarding a higher perinatal mortality, a higher rate of preeclampsia, and failure of therapy (Coetzee and Jackson 1979; Hellmuth et al, 2000). However, the metformin-treated women were older, more obese, and treated later in pregnancy. A more recent cohort study of metformin in pregnancy by Hughes and Rowan (2006) included 93 women with metformin treatment (only 32 continued until delivery) and 121 controls. There was no difference in perinatal outcomes between the groups. Glueck et al (2002) compared women without diabetes but with polycystic ovary syndrome who conceived while taking metformin and continued the agent through delivery (n = 28) to matched women without metformin therapy (n = 39). Gestational diabetes developed in 31% of women who did not take metformin versus 3% of those who did (odds ratio, 0.115; 95% CI, 0.014 to 0.938). Recently a large randomized controlled trial was performed comparing metformin to insulin for the treatment of gestational diabetes (Rowan et al, 2008). This study was powered to detect a 33% increase in composite outcome (neonatal hypoglycemia, respiratory distress, need for phototherapy, birth trauma, 5-minute American Pediatric Gross Assessment Record score of less than 7, or prematurity) in neonates born to mothers treated with metformin. Seven hundred fifty-one women with gestational diabetes between 20 and 30 weeks’ gestation were randomized to metformin or insulin. Of these, 363 women were assigned to metformin and 370 were assigned to insulin. Forty-six percent of women receiving metformin required the addition of insulin to obtain adequate glycemic control. There were no differences in the rate of the primary composite outcome. There was a lower rate of severe neonatal hypoglycemia in the metformin-treated group and no differences in neonatal anthropometric measurements. There was, however, a higher rate of prematurity in the metformin-treated group (12.1%) versus the insulin group (7.6%). A follow-up study is currently under way to assess the offspring of these women at 2 years of age. Prenatal Obstetric Management The overall strategy for managing a diabetic pregnancy in the third trimester involves two goals: (1) preventing stillbirth and asphyxia and (2) monitoring growth of the fetus to select the proper time and route of delivery to minimize maternal and infant morbidity. The first goal is accomplished by testing fetal well-being at frequent intervals, and the second through ultrasonographic monitoring of fetal size. 86 PART III Maternal Health Affecting Neonatal Outcome Periodic Biophysical Testing of the Fetus A variety of biophysical tests of the fetus are available to the clinician, including fetal heart rate testing, fetal movement assessment, ultrasound biophysical scoring, and fetal umbilical Doppler studies. Most of these tests, if applied properly, can be used with confidence to provide assurance of fetal well-being while awaiting fetal maturity; they are summarized in Table 9-6. Testing should be initiated early enough to avoid significant risk of stillbirth, but not so early that the risk of a false-positive result is high. In patients with poor glycemic control or significant hypertension, testing should begin as early as 28 weeks’ gestation. In lower-risk patients, most centers begin formal fetal testing by 34 to 36 weeks’ gestation. Counting of fetal movements is performed in all pregnancies from 28 weeks’ gestation onward. Assessing Fetal Growth Monitoring of fetal growth continues to be a challenging and highly inexact process. Although the current tools, consisting of serial plotting of fetal growth parameters, are superior to earlier clinical estimations, accuracy is still ±15%. Single and multiple longitudinal assessments of fetal size have been attempted. Calculation of Estimated Fetal Weight Several polynomial formulas using combinations of head, abdominal, and limb measurements have been proposed to predict the weight of the macrosomic fetus from ultrasonography parameters (Ferrero et al, 1994; Tongsong et al, TABLE 9-6 Tests of Fetal Well-Being Reassuring Result Test Frequency Counting of fetal movements Every night from 28 weeks’ ­gestation Ten movements in <60 min Comment Performed in all patients Nonstress test Twice weekly Two heart rate accelerations in 20 min Begin at: 28-34 wk in patients with insulindependent diabetes, 36 wk in patients with diet-controlled gestational diabetes Contraction stress test Weekly No heart rate decelerations in response to ≤3 contractions in 10 min Same as for nonstress test Ultrasound biophysical profile Weekly Score of 8 in 30 min The following findings are given 2 points each: 3 movements 1 flexion 30 sec breathing 2 cm amniotic fluid 1994). Unfortunately, in such formulas, small errors in individual measurements of the head, abdomen, and femur are typically multiplied together. In the obese fetus, the inaccuracies are further magnified. Bernstein and Catalano (1992) observed that a significant correlation exists between the degree of error in the ultrasonographically estimated fetal weight and the percentage of body fat on the fetus (r = 0.28; p <0.05). Perhaps this problem explains why no single formula has proved to be adequate in identifying the macrosomic fetus (Tamura et al, 1985). Shamley and Landon (1994) reviewed the relative accuracy of the various available formulas. Approximately 75% of the fetal weight predictions were within 10% of actual birthweight, with sensitivity for detecting macrosomia varying greatly (11% to 76%). In another study, McLaren et al (1995) found that 65% of weight estimates based on a simple abdominal circumference and femur length formula were within 10% of actual weight. A similar accuracy was achieved with more complex models (53% to 66% of estimates within the range). The formula developed by Shepard et al (1982), which uses biparietal diameter and abdominal circumference, is readily available in textbooks and is used most commonly in current ultrasonographic equipment software. The formula has accuracy levels similar to the statistics quoted previously. Serial Estimated Fetal Weight Assessments Because prediction of fetal weight from a single set of measurements is inaccurate, serial estimates showing the trend of ultrasonographic parameters (typically made every 1.5 to 3 weeks) might theoretically offer a better estimate of actual weight percentile. A comparison of the efficacy of serial estimated fetal weight calculations to a single measurement, however, did not show better predictive accuracy. Larsen et al (1995) reported that predictions based on the average of repeated weight estimates, on linear extrapolation from two estimates, or on extrapolation by a secondorder equation fitted to four estimates were no better than the prediction from the last estimate before delivery. Similar findings (that a single estimate is as accurate as multiple assessments) were reported by Hedriana and Moore (1994). Choosing the Timing and Route of Delivery Timing of delivery should be selected to minimize maternal and neonatal morbidity and mortality. Delaying delivery to as near as possible to the due date helps to maximize cervical ripeness and improve the chances of spontaneous labor and vaginal delivery. Yet the risks of fetal macrosomia, birth injury, and fetal death rise as the due date approaches (Rasmussen et al, 1992). Although earlier delivery at 37 weeks’ gestation might reduce the risk of shoulder dystocia, the higher rates of failed labor inductions and poor neonatal pulmonary status at this time must be considered. Therefore an optimal time for delivery of most diabetic pregnancies is between 39.0 and 40 weeks’ gestation. Delivery of a diabetic patient before 39 weeks’ gestation without documentation of fetal lung maturity should be performed only for compelling maternal or fetal reasons. Fetal lung maturity should be verified in such cases from the presence of more than 3% phosphatidyl glycerol or the equivalent in amniotic fluid as ascertained from an amniocentesis CHAPTER 9 Endocrine Disorders in Pregnancy specimen. After 39 weeks’ gestation, the obstetrician can await spontaneous labor if the fetus is not macrosomic and results of biophysical testing are reassuring. In patients with gestational diabetes and superb glycemic control, continued fetal testing and expectant management can be considered until 41 weeks’ gestation (Lurie et al, 1992). Given the previous data, the decision to attempt vaginal delivery or perform a cesarean section is inevitably based on limited data. The patient’s obstetric history from previous pregnancies, the best estimation of fetal weight, the fetal adipose profile (abdomen larger than head), and results of clinical pelvimetry should all be considered. A policy of elective cesarean section for suspected fetal macrosomia (ultrasonographically estimated fetal weight of greater than 4500 g) would require 443 cesarean deliveries to avoid one permanent brachial plexus injury (Rouse et al, 1996). Most large series of diabetic pregnancies report a cesarean section rate of 30% to 50%. The best means by which this rate can be lowered is early and strict glycemic control in pregnancy. Conducting a long labor induction in the patient with a large fetus and marginal pelvis may increase rather than decrease morbidity and costs. Intrapartum Glycemic Management Maintenance of intrapartum metabolic homeostasis is essential to avoid fetal hypoxemia and promote a smooth postnatal transition. Strict maternal euglycemia during labor does not guarantee newborn euglycemia in infants with macrosomia and long-established islet cell hypertrophy. Nevertheless, the use of a combined insulin and glucose infusion during labor to maintain maternal blood glucose in a narrow range (80 to 110 mg/dL) during labor is a common and reasonable practice. Typical infusion rates are 5% dextrose in lactated Ringer’s solution at 100 mL/hr and regular insulin at 0.5 to 1.0 U/hr. Capillary blood glucose levels are monitored hourly in such patients. For patients with diet-controlled gestational diabetes in labor, avoiding dextrose in all intravenous fluids normally maintains excellent blood glucose control. After 1 to 2 hours, no further assessments of capillary blood glucose are typically necessary. Neonatal Management Neonatal Transitional Management One of the metabolic problems common to IDMs is hypoglycemia, which is related to the level of maternal glycemic control over the 6 to 12 weeks before birth. Neonatal hypoglycemia is most likely to occur between 1 and 5 hours after birth, as the rich supply of maternal glucose stops with ligation of the umbilical cord and the infant’s levels of circulating insulin remain elevated. These infants therefore require close monitoring of blood glucose concentration during the first hours after birth. IDMs also appear to have disorders of catecholamine and glucagon metabolism as well as diminished capability to mount normal compensatory responses to hypoglycemia. In the past, IDMs were treated with glucagon; however, this treatment frequently results in high blood glucose levels that trigger insulin secretion and repeated cycles of hypoglycemia and hyperglycemia. Current 87 recommendations, therefore, consist of early oral feeding, when possible, with infusion of intravenous glucose. Ordinarily, blood glucose levels can be controlled satisfactorily with an intravenous infusion of 10% glucose. If greater amounts of glucose are required, bolus administration of 2 mL/kg of 10% glucose is recommended. Close monitoring to correct hypoglycemia while avoiding hyperglycemia and consequent stimulation of insulin secretion is important. Breastfeeding Most authorities prefer to maintain strict monitoring of newborn IDM glucose levels for at least 4 to 6 hours, which frequently necessitates admission to a newborn special care unit. IDMs who are delivered atraumatically and are well oxygenated, however, can be kept with their mothers while undergoing close glycemic monitoring for the first 1 to 2 hours of life. This approach permits early breastfeeding, which may reduce the need for intravenous glucose therapy. Summary Intensive management of women with glucose intolerance during pregnancy has resulted in markedly improved pregnancy outcomes. Despite these advances, care of the IDM continues to require vigilance and meticulous monitoring with a full understanding of the quality of the glycemic milieu in which the infant developed. DISORDERS OF THE THYROID INCIDENCE Thyroid disorders, in general, are more common in women than in men and represent a common endocrine abnormality during pregnancy. Both hyperthyroidism and hypothyroidism in the mother put the infant at risk and require careful management by the perinatal-neonatal team. Table 9-7 presents an overview of the approach to infants who are thought to be at risk for abnormal thyroid function because of maternal thyroid abnormalities. The most frequently described problem is the syndrome of postpartum thyroiditis, which has been reported to complicate as many as 5% of all pregnancies. The diagnosis of thyroid disease in pregnancy is complicated by the natural changes that occur in immunologic status of the mother and fetus and that variously affect the assessment of any of the autoimmune thyroid disorders. MATERNAL-FETAL THYROID FUNCTION IN PREGNANCY Maternal Thyroid Function Several pregnancy-related physiologic conditions affect maternal thyroid function, and the appropriate interpretation of thyroid function test results during pregnancy must take these normal physiologic factors into account. One important modification that occurs during pregnancy is the estrogen-dependent increase in thyroid-binding globulin. This results in an increase in total thyroxine and total triiodothyronine levels throughout pregnancy. The 88 PART III Maternal Health Affecting Neonatal Outcome TABLE 9-7 Approaches for Infants Judged to be at Risk for Abnormal Thyroid Function Possible Thyroid Abnormality Cord Blood Analyses Assessment at Birth Assessment at 2-7 Days of Life Congenital hyperthyroidism because mother has any of the following: Graves’ disease with hyperthyroidism and may have been treated with PTU, methimazole, 131I, iodides, or surgery; history of Graves’ disease; history of Hashimoto’s disease T4, TSH, TSAb Physical examination for intrauterine growth restriction, goiter, exophthalmos, tachycardia, bradycardia, size of anterior fontanel, synostosis, congenital anomalies; determination of gestational age by dates, ultrasonography during pregnancy, or Dubovitz examination; neurologic examination; determination of bone age (knee); for selected cases: electrocardiogram, auditory and visual evoked potentials, motor conduction, velocity tests, skull radiographs T4, TSH, TSAb (if available) ­ eterminations: d If results normal, give no treatment; observe and repeat T4, T3, TSH determinations at 7-10 days*; if results indicate hypothyroidism, repeat T4, TSH; if results are abnormal, begin treatment at 7-10 days†; if results indicate hyperthyroidism, begin treatment with PTU (8 mg/kg) and propranolol Congenital or early childhood hypothyroidism because mother has any of the following: Graves disease with excessive PTU ­therapy; Hashimoto’s disease; acute ­(subacute) ­thyroiditis; familial genetic defect in thyroxine synthesis; treatment with iodides or lithium for nonthyroidal illness; exposure to 131I while pregnant T4, TSH, ThyAb Same as for hyperthyroidism T4, TSH, ThyAb determinations: If results indicate hypothyroidism, perform ultrasound scan to define presence, size, location of thyroid tissue; if hypothyroidism confirmed, begin treatment with levothyroxine sodium (Synthroid), 0.05 mg/day Adapted from Creasy RK, Resnik R, editors: Maternal-fetal medicine: principles and practice, ed 2, Philadelphia, 1989, WB Saunders. I, Radioactive iodine; PTU, propylthiouracil; T4, thyroxine; TSAb, thyroid-stimulating antibody; T3, triiodothyronine; ThyAb, thyroid antibody; TSH, thyroid-stimulating hormone. *Graves disease does not develop at 7 to 10 days in children whose mothers receive PTU. levels of unbound (free) thyroxine (FT4) and free triiodothyronine (FT3), as well as levels of thyroid-stimulating hormone (TSH), in general remain unchanged. However, human chorionic gonadotropin, a second factor in pregnancy that may modify thyroid function, has a stimulatory effect on the thyroid and may transiently effect the FT4, FT3, and TSH levels in the first trimester and early in the second trimester. This stimulatory affect of human chorionic gonadotropin rarely causes aberrations of the thyroid function parameters into the thyrotoxic range (American College of Obstetricians and Gynecologists, 2001). Fetal Thyroid Function The fetal thyroid actively concentrates iodide after 10 weeks’ gestation, releases thyroxine (T4) after 12 weeks’ gestation, and becomes responsive to pituitary TSH at 20 weeks’ gestation. Although maternal TSH does not cross the placenta, maternal thyroid hormones and thyrotropin-releasing hormone are transferred to the fetus throughout gestation. Early studies found that cord blood thyroid function testing of neonates with congenital thyroid agenesis revealed hormone levels that were 30% of normal (Vulsma et al, 1989), suggesting a maternal source. Recent studies show that by 4 weeks after conception, small amounts of T4 and triiodothyronine (T3) from the maternal origin are found in the fetal compartment, with T4 levels increasing throughout gestation. Free T4 levels reach concentrations of biologic significance in the adult by midgestation. Studies using rat models have demonstrated that this thyroid hormone is important for corticogenesis early in the pregnancy (Morreale de Escobar et al, 2004). Transplacental transfer of thyroidstimulating immunoglobulin (TSI) may occur, causing fetal thyrotoxicosis. Other substances that may be transferred from the maternal compartment to the fetal compartment and affect fetal thyroid function are iodine, a radioactive isotope of iodine, propylthiouracil (PTU), and methimazole. The fetal effects of these agents are reviewed later. Hyperthyroidism Hyperthyroidism occurs in approximately 0.2% of pregnancies and results in a significant increase in the prevalence of both low-birthweight delivery and a trend toward higher neonatal mortality. The most common cause of thyrotoxicosis (85% of cases) in women of child-bearing age is Graves’ disease; other causes are acute (or subacute) thyroiditis (transient), Hashimoto’s disease, hydatidiform mole, choriocarcinoma, toxic nodular goiter, and toxic adenoma. Graves’ disease has a peak incidence during the reproductive years, but patients with the disorder may actually have remissions during pregnancy, followed by postpartum exacerbations. The unique feature of these pregnancies is that the fetus may also be affected, regardless of the mother’s concurrent medical condition. Thyroid function is difficult to evaluate in the fetus, and the status of the fetus may not correlate with that of the mother. Diagnosis The differential diagnosis of thyrotoxicosis becomes more difficult during pregnancy because normal pregnant women may have a variety of hyperdynamic signs and symptoms—intolerance to heat, nervousness, irritability, emotional lability, increased perspiration, tachycardia, and anxiety. Laboratory data are also difficult to evaluate CHAPTER 9 Endocrine Disorders in Pregnancy because total serum thyroxin values are normally elevated during pregnancy as a result of estrogen-induced increases in thyroxine-binding globulin. Therefore if thyroxinebinding globulin is increased, then T3 resin uptake may be in the euthyroid to slightly increased range in a patient who has true hyperthyroidism. Hollingsworth (1989) has reviewed the assessment of thyroid function tests in nonpregnant and pregnant women, along with the differential diagnosis of hyperthyroidism during pregnancy. Pathogenesis of Graves’ Disease The pathogenesis of Graves’ disease is not completely understood, but it probably represents an overlapping spectrum of disorders that are characterized by the production of polyclonal antibodies. It has been appreciated since the 1960s (Sunshine et al, 1965) that abnormal TSIs, which appear to be immunoglobulin G, are present in pregnant women with Graves’ disease and cross the placenta easily to cause neonatal hyperthyroidism in some infants (McKenzie and Zakarija, 1978). The clinical spectrum of Graves’ disease in utero is broad and may result in stillbirth or preterm delivery. Some affected infants have widespread evidence of autoimmune disease, including thrombocytopenic purpura and generalized hypertrophy of the lymphatic tissues. Thyroid storm can occur shortly after birth, or the infant may have disease that is transient in nature, lasting from 1 to 5 months. Infants born to mothers who have been treated with thioamides may appear normal at birth, but demonstrate signs of thyrotoxicosis at 7 to 10 days of age, when the effect of thioamide suppression of thyroxine synthesis is no longer present. The measurement of thyroid-stimulating antibodies is useful in predicting whether the fetus will be affected. Management of the Mother Because radioactive iodine therapy is contraindicated during pregnancy, treatment of the pregnant woman with thyrotoxicosis involves a choice between antithyroid drugs A B 89 and surgery. The therapeutic goal is to achieve a euthyroid, or perhaps slightly hyperthyroid, state in the mother while preventing hypothyroidism and hyperthyroidism in the fetus. Either PTU or methimazole can be used to treat thyrotoxicosis during pregnancy. Because methimazole therapy can be associated with aplasia cutis in the offspring of treated women, and because PTU crosses the placenta more slowly than methimazole, PTU has become the drug of choice for use during pregnancy. Ordinarily, thyrotoxicosis can be controlled with doses of 300 mg per day. Once the disorder is under control, however, it is important to keep the dose as low as possible, preferably less than 100 mg daily, because this drug crosses the placenta and blocks fetal thyroid function, possibly producing hypothyroidism in the fetus. In women with cardiovascular effects, the use of betablockers may be appropriate to achieve rapid control of thyrotoxicosis. Because administration of propranolol to pregnant women has been associated with intrauterine growth restriction and impaired responses of the fetus to anoxic stress as well as postnatal bradycardia and hypoglycemia, the doses must be closely controlled. Iodides have also been used, particularly in combination with betablocking agents, to control thyrotoxicosis. Long-term iodide therapy presents a risk to the fetus. Because of the inhibition of the incorporation of iodide into thyroglobulin, a large, obstructive goiter can develop in the fetus. Surgery during pregnancy is best reserved for cases in which the mother is hypersensitive to antithyroid drugs, compliance with medication is poor, or drugs are ineffective in controlling the disease. Effects on the Newborn Approximately 1% of infants born to mothers with some level of thyrotoxicosis have thyrotoxicosis (Figure 9-6). Assessment of fetal risk in utero includes measurement of TSIs, with the expectation that if the titers are high, there is a higher risk of thyrotoxicosis. Additional assessment of the fetus should pay particular attention to elevated resting C FIGURE 9-6 A, Hypothyroid 21-year-old mother who experienced Graves’ disease at age 7 years and was treated by subtotal thyroidectomy. She was given maintenance therapy with daily levothyroxine sodium (Synthroid; 0.15 mg) throughout pregnancy. B, Her infant girl was born at term with severe Graves’ disease, goiter, and exophthalmos that persisted for 6 months. C, The child was healthy at 20 months old. ( From Creasy RK, Resnik R, editors: Maternal-fetal medicine: principles and practice, ed 2, Philadelphia, 1989, WB Saunders.) 90 PART III Maternal Health Affecting Neonatal Outcome heart rate and poor fetal growth. Daneman and Howard (1980) reported on the outcome of nine infants with neonatal thyrotoxicosis and noted normal growth, but a high incidence of craniosynostosis and intellectual impairment. It may be necessary to treat the asymptomatic mother with thioamides and propranolol (and thyroid replacement) during pregnancy to treat the infant and prevent serious neonatal morbidity and long-term problems. Mothers with thyrotoxicosis who are taking normal doses of thioamide can safely breastfeed their infants, although thioamide appears in breast milk in low amounts. Currently there does not appear to be any long-term adverse outcome for infants whose mothers have received PTU during pregnancy. HYPOTHYROIDISM Hypothyroidism complicates about 1 to 3 per 1000 pregnancies. The leading cause of hypothyroidism in pregnancy is Hashimoto’s thyroiditis, which is a chronic autoimmune thyroiditis characterized by painless inflammation and enlargement of the thyroid gland (Casey and Leveno, 2006). Other causes of primary hypothyroidism include iodine deficiency, thyroidectomy, or ablative radioiodine therapy for hyperthyroidism. Secondary causes of hypothyrodism include Sheehan’s syndrome caused by obstetric hemorrhage leading to pituitary ischemia, necrosis and abnormalities in all pituitary hormones, lymphocytic hypophysitis, and hypophysectomy (Casey and Leveno, 2006). Women with over hypothyroidism are at increased risk of pregnancy complications, such as a higher rate of miscarriage, preeclampsia, placental abruption, growth restriction, and stillbirth (Casey and Leveno, 2006). Diagnosis The finding of a goiter may be associated with cases of Hashimoto’s thyroiditis or iodine deficiency. The signs and symptoms of hypothyroidism are usually insidious and easily confused with those of normal pregnancy including fatigue, cold intolerance, cramping, constipation, weight gain, hair loss, insomnia, and mental slowness. The serum TSH level is an accurate and widely clinically available test to diagnose hypothyroidism. If the serum TSH is elevated, a free T4 level should be obtained. In the classic definition of hypothyroidism, the serum TSH is elevated and the free T4 is low. Other forms of hypothyroidism have also been described, including subclinical hypothyroidism, which is defined as an elevated TSH with a normal free T4, or hypothyroxinemia defined as a normal TSH but a low free T4; these have no clinical significance to the mother, but may be associated with neonatal effects discussed later in the neonatal neurologic development section. Management of the Mother Levothyroxine is the treatment of choice. Adults with hypothyroidism require approximately 1.7 μg/kg of body weight and should be initiated on full replacement (www. thyroidguidelines.org). The goal of therapy is normalization of the TSH level; therefore the TSH is checked at 4- to 6-week increments, and the dose of levothyroxine is adjusted by 25- to 50-μg increments. Neonatal Neurologic Development In humans, early epidemiologic data from iodine-­deficient areas of Switzerland suggested a link between mental retardation in the children of women with abnormal thyroid function (Gyamfi et al, 2009). Studies performed by Haddow et al (1999) found that in women with overt, untreated hypothyroidism, the intelligence quotient (IQ) points of children aged 7 to 9 years (using the Wechsler Intelligence Scale IQ test) were 7 points lower in cases than in controls (p = 0.005). The percentage of children with IQ scores less than 85 was higher in the cases than in controls (19% versus 5%; p = 0.007; Haddow et al, 1999). It has been demonstrated that early transplacental passage of thyroid hormone is important for normal neurodeveloment of the fetus. T3 is made from the conversion of maternal T4. If maternal T4 levels are low, fetal T3 in the brain will be low even if the maternal and fetal serum T3 are normal (Morreale de Escobar et al, 2004). Studies involving rats—which, like humans, are dependent on maternal thyroid hormone early in development—have demonstrated that thyroid hormone receptor is present in the brain before neural tube closure, suggesting a biologic role (Morreale de Escobar et al, 2004). Furthermore in humans, thyroid hormone concentrations in the cerebral cortex at 20 weeks’ gestation are comparable to those found in adults (Morreale de Escobar et al, 2004). Lavado-Autric et al (2003) have demonstrated that in iodine-deficient rat pups, there is aberrant neuronal migration, blurring of the cytoarchitecture, and abnormal morphology in the somatosensory cortex and hippocampus. Early in human development there is expression of nuclear thyroid receptors, which are already occupied by T3, suggesting that normal maternal T4 levels are necessary for normal cortical development (Morreale de Escobar et al, 2004). Given the increased risk of adverse perinatal and neurodevelopmental outcomes, all women with overt hypothyroidism should be treated in pregnancy. However, controversy still exists as to whether subclinical hypothyroidism (defined as an elevated TSH level, but ­normal free T4) or hypothyroxinemia (defined as a normal TSH level, but a low free T4) warrant treatment in pregnancy. In 1969, Man and Jones were the first to evaluate offspring of mothers with hypothyroxinemia and found that they had lower IQ scores than normal controls or of those born to mothers with adequately treated hypothyroidism (Gyamfi et al, 2009). Studies performed by Pop et al (1999) in the iodine-deficient areas of the Netherlands have shown that free T4 levels below the 10th percentile at 12 weeks’ gestation were associated with lower scores on the Dutch version of the Bayley Scale of Infant Development at 10 months old. The study included women with a low free T4 (hypothyroxinemia) and excluded women with elevated TSH. A follow-up study on these same infants, tested in both motor and mental scores at 1 and 2 years of age, found significantly lower scores in infants born to mothers with low free T4 levels (Pop et al, 2003). Casey et al (2005) performed a study on CHAPTER 9 Endocrine Disorders in Pregnancy Parkland Hospital patients with subclinical hypothyroidism defined as a TSH at 97.5th percentile or higher and a normal free T4 level. Approximately 2.3% of women screened (404 women) were identified as having subclinical hypothyroidism; compared with normal controls, they had a higher incidence of placental abruption (RR, 3.0; 95% CI, 1.1 to 8.2) and preterm birth before 34 weeks’ gestation (RR, 1.8; 95% CI, 1.1 to 2.9) (Casey et al, 2005). The authors concluded that the reduction in IQ in children born to women with subclinical hypothyroidism may be caused by prematurity. Based on the available animal and clinical data, the American Association of Clinical Endocrinologists, the American Thyroid Association, and the Endocrine Society recommend universal screening for all pregnant women. However, the American College of Obstetricians and Gynecologist (2001) recommend that screening should be performed only in women who have risk factors, such as pregestational diabetes, or who are symptomatic. Universal screening is not recommended by the American College of Obstetricians and Gynecologist, given that decision and cost effectiveness studies on the effects of such a strategy are currently lacking. Furthermore, data are lacking regarding therapy dosing, efficacy, or whether medication should be stopped after pregnancy in otherwise asymptomatic women with subclinical hypothyroidism and hypothyroxinemia. A multicenter randomized trial is currently underway to examine whether screening and treatment of hypothyroxinemia or subclinical hypothyroidism have a long-term effect on neurodeveloment of offspring (http://clinicaltrials.gov; study identifier number NCT00388297). 91 SUGGESTED READINGS Alam M, Raza SJ, Sherali AR, et al: Neonatal complications in infants born to diabetic mothers, J Coll Physicians Surg Pak 16:212-215, 2006. American College of Obstetricians and Gynecologists: ACOG Practice Bulletin. Clinical management guidelines for obstetrician-gynecologists: gestational diabetes, Obstet Gynecol 98:525-538, 2001. American Diabetes Association: Expert Committee on the Diagnosis and Classification of Diabetes Mellitus: Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus, Diabetes Care 25(Suppl 1): S5-S20, 2002. Casey BM, Dashe JS, Wells CE, et al: Subclinical hypothyroidism and pregnancy outcomes, Obstet Gynecol 105:239-245, 2005. Casey BM, Leveno KJ: Thyroid disease in pregnancy, Obstet Gynecol 108:12831292, 2006. Coustan DR: Pharmacological Management of Gestational Diabetes: an overview, Diabetes Care 30:S206-S208, 2007. Study Cooperative Research Group HAPO, Metzger BE, Lowe LP, et al: Hyperglycemia and adverse pregnancy outcomes, N Engl J Med 358:19912002, 2008. HAPO Study Cooperative Research Group: Hyperglycemia and adverse pregnancy outcomes: associations with neonatal anthropometrics, Diabetes 58:453-459, 2009. Landon MB, Thom E, Spong CY, et al: A multicenter, randomized trial of treatment for mild gestational diabetes, N Engl J Med 361:1339-1348, 2009. Langer O, Conway D, Berkus M, et al: A comparison of glyburide and insulin in women with gestational diabetes mellitus, N Engl J Med 343:1134-1138, 2000. Morreale de Escobar, Obregon MJ, Escobar del Rey F: Role of thyroid hormone during early brain development, Eur J Endocrinol 151(Suppl 3):U25-U37, 2004. Pop VJ, Browthers EP, Vader HL, et al: Maternal hypothyroxinaemia during early pregnancy and subsequent child development: a 3-year follow-up study, Clin Endocrinol (Oxf) 59:282-288, 2003. Ray JG, O’Brien TE, Chan WS: Preconception care and the risk of congenital anomalies in the offspring of women with diabetes mellitus:a meta-analysis, QJM 94:435-444, 2001. Rowan JA, Hague WM, Gao W, et al: Metformin versus insulin for the treatment of gestational diabetes, N Engl J Med 358:2003-2015, 2008. Temple R, Aldridge V, Greenwood R, et al: Association between outcome of pregnancy and glycaemic control in early pregnancy in type 1 diabetes: population based study, BMJ 325:1275-1276, 2002. Complete references and supplemental color images used in this text can be found online at www.expertconsult.com C H A P T E R 10 Maternal Medical Disorders of Fetal Significance: Seizure Disorders, Isoimmunization, Cancer, and Mental Health Disorders Thomas F. Kelly and Thomas R. Moore A significant spectrum of maternal medical disorders may complicate pregnancy. Some of these disorders, although readily manageable in nonpregnant patients, can be lethal to pregnant women. As a result, two questions arise for the specialist caring for a pregnant woman with a medical complication. First, is the condition affected by the patient’s normal adaptations to pregnancy? Second, how does the medical problem affect the woman and her fetus? Although many medical conditions during pregnancy can be managed much like they would be in a nonpregnant woman, there are usually nuances in care during gestation to which the obstetrician must be attuned and that will potentially affect the fetus and neonate. During pregnancy, any potential medical therapy should be considered carefully to minimize fetal risk. For example, thalidomide and diethylstilbestrol were prescribed in the past for morning sickness and recurrent miscarriages, respectively, on the basis of the reasonable hypotheses that maternal sedation would decrease nausea and increased estrogens would support the placenta and reduce the likelihood of first-trimester loss. Unfortunately, the use of both of these agents led to significant and tragic congenital anomalies in the offspring. In view of the profusion of new drugs available today, the importance of assessing the expected risk-to-benefit ratio for each medication prescribed to a pregnant woman is increasingly important. An example is phenytoin treatment for maternal seizures. If there is a less teratogenic alternative that is effective, it should be prescribed; however, for most women taking phenytoin, other agents are unable to control their seizures, and the risks to the fetus must be accepted. The altered pharmacokinetics of many drugs during pregnancy must be considered, because the dosing of familiar medications may have to be adjusted if toxicity is to be avoided. Classic examples are thyroid hormone replacement (a higher level of thyroid-binding globulin in pregnancy increases total thyroxin, but leaves the serum free thyroxin value unchanged) and aminoglycoside antibiotic therapy (increased glomerular filtration in pregnancy results in lower serum drug levels). This chapter discusses four maternal conditions that can influence fetal growth, development, and outcome: seizure disorders, red blood cell isoimmunization, cancer, and mental disorders. The basics of management, the potential effects of pregnancy on the condition, and the effects of the condition on the mother and fetus are considered. 92 MATERNAL SEIZURE DISORDERS Epilepsy is the most common major neurologic disorder in pregnancy. Approximately 18 million women are affected worldwide, and 40% of those are of childbearing age. The estimated prevalence in pregnancy is 0.2% to 0.7% (Chen et al, 2009). The pattern of maternal seizures ranges from complex partial to generalized tonic clonic (grand mal) and generalized absence (petit mal) seizures. Physiologically, seizures arise from paroxysmal episodes of abnormal brain electrical discharges; when associated with motor activity, they are termed convulsive. The effect of pregnancy on the frequency and severity of the seizure disorder has been difficult to ascertain because of limited prospective data. The International Registry of Antiepileptic Drugs and Pregnancy (EURAP) recently reported on more than 1800 patients whose seizure frequency and treatment were recorded. Fifty-eight percent of patients had no seizures during their pregnancy. When using first-trimester seizure activity as a reference, 64% had no change in frequency in the second and third trimester, 6% improved, and 12% deteriorated (EURAP Study Group, 2006). The only exception was that tonic-clonic seizures occurred more frequently in women using oxcarbazepine monotherapy; this has been confirmed by an Australian registry in which seizures occurred in 50% of pregnant women with epilepsy who were receiving therapy. However, in a subset that had no seizures for 12 months before pregnancy, there was a 50% to 70% reduced frequency during gestation (Vajda et al, 2008). In patients in whom higher numbers of seizures occur during gestation, decreased plasma concentrations of antiepileptic medications have been hypothesized as causative. The fall in plasma drug levels during pregnancy may be due in part to increased protein binding, reduced absorption, and increased drug clearance. The adequacy of prepregnancy seizure control can influence a patient’s course during gestation. Patients whose seizures were poorly controlled tended to have more frequent seizures during pregnancy, whereas patient who had no seizures for 2 years before pregnancy had only a 10% chance of experiencing seizures during gestation. These latter patients may be candidates for stopping therapy or considering monotherapy if they have previously required multiple antiepileptic drugs (Schmidt et al, 1983; Walker et al, 2009). CHAPTER 10 Maternal Medical Disorders of Fetal Significance: Seizure Disorders, Isoimmunization, Cancer, and Mental Health Disorders PERINATAL RISK For reasons that are not clear, women with seizures have more obstetric complications during pregnancy and a higher rate of poor perinatal outcomes. Rates of preeclampsia, preterm delivery, small-for-gestational-age infants, congenital malformations, cerebral palsy, and perinatal mortality have all been reported as higher in women with an antecedent seizure disorder (Lin et al, 2009; Nelson and Ellenberg, 1982). Pregnancy outcome is also greatly influenced by the mother’s socioeconomic status and age as well as by the prenatal care received. Earlier publications suggested an increased risk of congenital malformations in children of mothers with epilepsy even without prenatal use of antiepileptic drugs (Bjerkedal, 1982). More recent data appear to refute this, and the malformation risk apparently correlates with the number of medications used. A study comparing patients with epilepsy to matched controls revealed that women receiving no medication had no increased rate of congenital malformations; however, monotherapy was associated with and increased risk of embryopathy (odds ratio of 2.8). Furthermore the frequency was even higher with use of two or more drugs (odds ratio of 4.2) (Homes et al, 2001). Recent updates from five international registries have reported malformation rates ranging from 3.7% to 8.0% (with monotherapy) and 6% to 9.8% (with polytherapy) (Meador et al, 2008). Specific malformations include a fivefold rise in the rate of orofacial clefts (Friis et al, 1986), an increase in the rate of congenital heart disease, particularly with trimethadione (Friis and Hauge, 1985), and a 3.8% incidence of neural tube defects in fetuses exposed to valproic acid (Samrén et al, 1997). Facial abnormalities (e.g., midface hypoplasia) are not specific to any particular antiepileptic drug; they have been seen with phenytoin, carbamazepine, and trimethadione. Some antiepileptic medications can adversely affect postnatal cognitive development. Although conclusive data are lacking, there may be an increased adverse effect, particularly with valproate (Meador et al, 2008; Tomson and Battino, 2009). 93 fetal complications. Preconception counseling is preferable and should entail (1) adjusting medication doses into the therapeutic range, (2) attempting to limit the patient to one drug if possible, and (3) choosing an agent with the least risk of teratogenesis. Frank discussion of the various risks of each agent should be conducted, particularly the risks associated with valproic acid and trimethadione. Usually if the patient’s disease is adequately controlled with one agent, it rarely needs to be changed, because the risks of increasing seizure activity are believed to outweigh the potential for reducing congenital malformations. Patients taking antiepileptic medications should also take folic acid supplements (800 to 1000 μg) before conception, because inhibition of folate absorption has been proposed as a teratogenic mechanism, particularly with phenytoin. During gestation, the anticonvulsant levels should be checked monthly, and the dose should be adjusted accordingly, particularly with the use of lamotrigine, carbamazepine, and phenytoin (Harden et al, 2009). Although the evidence is less clear with other agents such as phenobarbital, valproate, primidone, and ethosuximide, serial level assessment should not be discouraged. Medications should not be changed unless they prove ineffective at the optimal serum level. If a patient reports greater seizure activity, the serum drug level should be checked immediately. A common reason for increased seizures is that the patient is not taking her medication, usually because she fears its teratogenicity. Mothers taking phenytoin, phenobarbital, or primidone may have a higher incidence of neonatal coagulopathy as a result of vitamin K–dependent clotting factor deficiency. Although maternal vitamin K supplementation in the third trimester may be reasonable, there is insufficient evidence to determine whether it will reduce neonatal hemorrhagic complications (Harden et al, 2009). RED BLOOD CELL ISOIMMUNIZATION The classic features of the fetal hydantoin syndrome are facial clefting, a broad nasal ridge, hypertelorism, epicanthal folds, distal phalangeal hypoplasia, and growth and mental deficiencies; however, these effects also result from the use of other antiseizure medications (Table 10-1). The postulated cause of this syndrome is the teratogenic action of a common epoxide intermediate of these medications. The hydantoin syndrome was found to develop in fetuses with inadequate epoxide hydrolase activity (Buehler et al, 1990). This enzymatic deficiency appears to be recessively inherited. It appears that preconception folic acid supplementation can reduce the risk of major congenital malformations in women taking antiepileptic medication (Harden et al, 2009). Hydrops fetalis, a condition associated with abnormal fluid collections in various body cavities of the fetus, was first described in 1892. The causes are many but can be divided into two categories: immune causes and nonimmune causes. In immune-mediated hydrops, circulating immunoglobulins lead to the destruction of fetal red blood cells and hemolytic anemia. Landsteiner and Weiner first elucidated the Rhesus factor (Rh) in 1940. Levine et al, showed pathogenesis of erythroblastosis to be due to maternal isoimmunization in 1941. Rh immune globulin was developed in the middle 1960s. Prophylaxis protocols using Rh immune globulin significantly reduced the incidence of D isoimmunization, increasing the relative frequency of alloimmunization against atypical red blood cell antigens such as E, Duffy, and Kell (Figure 10-1). The ­pathophysiology of immune-mediated fetal hemolytic disease is similar, regardless of the blood group antigen involved; therefore this discussion focuses on the Rh system. MANAGEMENT GENETICS Management of the pregnant patient with epilepsy is based on keeping her free of seizures. Theoretically, this goal reduces maternal physical risk and lowers the incidence of The Rh blood group actually represents a number of antigens, designated D, Cc, and Ee. The genes for these antigens, located on the short arm of chromosome 1, are FETAL HYDANTOIN SYNDROME 94 PART III Maternal Health Affecting Neonatal Outcome inherited in a set of three from each parent. The presence of D determines whether the individual is Rh positive, and the absence of D (there is no recessive allele, so “d” does not exist) yields the Rh-negative type. Rarely, a patient exhibits a Du variant and should be considered D positive. The combinations of these various antigens occur with different frequencies. For example, prevalence of Cde (41%) is higher than that of CDE (0.08%) (Lewis et al, 1971). Although the Rh phenotype is the result of D status, the various genotype combinations help to predict the zygosity of an individual. Approximately 45% of Rh-positive individuals are homozygous and therefore will always produce an Rh-positive offspring; 55% are heterozygous and may have an Rh-negative child if paired with an Rh-negative partner. TABLE 10-1 Clinical Features of the Fetal Hydantoin Syndrome Craniofacial abnormalities Broad nasal ridge Wide fontanel Low-set hairline Broad alveolar ridge Metopic ridging Short neck Ocular hypertelorism Microcephaly Cleft lip with or without palate Abnormal or low-set ears Epicanthal folds Ptosis of eyelids Coloboma Coarse scalp hair Limb abnormalities Smallness or absence of nails Hypoplasia of distal phalanges Altered palmar crease Digital thumb Dislocated hip Data from Briggs GC, Freeman RK, Yaffe SJ: Drugs in pregnancy and lactation, ed 7, Baltimore, 2005, Lippincott Williams & Wilkins. 50 45 40 1969 1996 35 There are no sex differences in the frequency of Rh negativity; however, racial variations are striking. Rh negativity is common in the Basque population (30% to 35%), but rare in Chinese, Japanese, and North American Indian populations (1% to 2%). The average incidence in white populations is approximately 15%. North American blacks have a higher incidence (8%) than African blacks (4%). PATHOGENESIS Before modern blood banking, Rh-negative patients became immunized from the transfusion of Rh-positive blood. In the case of atypical red blood cell antigens (e.g., Duffy, Kidd), blood transfusion is still a significant cause of isoimmunization. Currently, fetal transplacental hemorrhage is the primary cause of Rh isoimmunization. Rh immune globulin prophylaxis protocols have reduced but not eliminated this problem. Transplacental hemorrhage of fetal cells into the maternal circulation is surprisingly common, with 75% of women showing evidence of this event at some time during gestation (Bowman et al, 1986). Using sensitive Kleihauer-Betke testing, 0.01 mL or more fetal cells are found in 3%, 12% and 46% of women in the first, second and third trimesters, respectively. The amount of fetal blood is usually small, but approximately 1% of women have 5 mL. In 0.25% of women, 30 mL or more of fetal cells are noted after delivery. Obstetric events increase the chance of transplacental hemorrhage (Box 10-1). As little as 0.3 mL of Rh-positive blood produces immunization, and the risk is dose dependent. ABO blood group incompatibility between fetus and mother affords some protection, reducing the risk from 16% to 2%. The primary maternal immune response is slow and can take as long as 6 months to develop. The first appearance of immunoglobulin (Ig) M class anti-D antibodies is weak; they do not cross the placenta, but are soon followed by smaller IgG antibodies that are capable of traversing the placental barrier. Therefore the initial event causing sensitization rarely results in fetal hemolysis. A second transplacental hemorrhage leads to the more rapid and abundant amnestic IgG response that, in the presence of fetal D-positive cells, can cause significant hemolysis and fetal anemia. The severity of hemolytic disease is related to the maternal antibody titer, the affinity for the red blood Percent 30 25 20 15 10 5 0 RhD E C Kell Duffy MNS Kidd FIGURE 10-1 Differences in incidence of maternal red cell antibodies over time. (Adapted from Queenan JG, Smith BD, Haber JM, et al: Irregular antibodies in the obstetric patient, Obstet Gynecol 34:767-770, 1969; and Geifman-Holtzman O, Wojtowycz M, Kosmas E, et al: Female alloimmunization with antibodies known to cause hemolytic disease, Obstet Gynecol 89:272-275, 1997.) BOX 10-1 O  bstetric Events Associated With Increased Risk of Fetal-Maternal Hemorrhage ll ll ll ll ll ll ll ll ll Unexplained vaginal bleeding Abruptio placentae External breech version Amniocentesis Chorionic villus sampling, placental biopsy Umbilical cord sampling Manual removal of the placenta Abdominal trauma Ectopic pregnancy CHAPTER 10 Maternal Medical Disorders of Fetal Significance: Seizure Disorders, Isoimmunization, Cancer, and Mental Health Disorders cell membrane, and the ability of the fetus to compensate for the red blood cell destruction. Table 10-2 summarizes the severity levels of fetal and neonatal disease and their incidence. Most cases are mild and result in normal outcomes; the cord blood is strongly Coombs positive, but the infants do not exhibit significant anemia or hyperbilirubinemia. Moderate disease results from the red blood cell destruction and the greater production of indirect bilirubin. Although the mother is able to clear this product for the fetus in utero, the neonate is deficient in the liver glucuronyl transferase enzyme, leading to the buildup of this water-insoluble molecule. Albumin carries the indirect bilirubin, but if the binding capacity is exceeded, diffusion of the bilirubin into the fatty tissues occurs. Neural tissue is high in lipid content. Ultimate destruction of neurons can occur, resulting in kernicterus. Treatment depends on the recognition of the hyperbilirubinemia and usually entails phototherapy and possible exchange transfusion in the nursery (see Chapter 79). Severe disease occurs when the fetus is unable to produce sufficient red blood cells to compensate for the increased destruction of these cells. Extramedullary hematopoiesis, which is prominent in the liver, ultimately leads to enlargement, hepatocellular damage, and portal hypertension. This process is believed to be the etiology of placental edema and ascites. Albumin production diminishes because of hepatocellular damage and results in anasarca, giving rise to hydrops fetalis. The theory that hydrops is due to fetal heart failure no longer holds, as a result of observations that these infants are neither hypervolemic nor in failure. The relationship between fetal anemia and hydrops is variable, but most hydropic fetuses have hemoglobin levels less than 4 g/dL or have a hemoglobin concentration deficit greater than 7 g/dL (Nicolaides et al, 1988). MANAGEMENT Management of the sensitized patient, for both Rh and ­atypical red blood cell antigens, requires an understanding of the mechanism of disease and the skill and experience to TABLE 10-2 Classification of the Severity of Hemolytic Disease Severity Description Mild Indirect bilirubin result, 16-20 mg/dL No anemia No treatment needed 45-50 Fetal hydrops does not develop Moderate anemia Severe jaundice with risk of kernicterus unless treated after birth 25-30 Fetal hydrops develops in utero Before 34 weeks’ gestation After 34 weeks’ gestation 20-25 10-12 10-12 Moderate Severe predict its severity. Although management schemes follow some basic guidelines, successful management requires access to a blood bank with expertise in antibody typing and individuals skilled in prenatal diagnostic procedures (e.g., cordocentesis). Referral to experienced high-risk centers for the management of this problem is common in the United States. MONITORING At their first prenatal visit, all pregnant patients should undergo a blood type and antibody screen (indirect Coombs’ test), which identifies the Rh-negative woman and screens for the presence of anti-D antibody and other immunoglobulins that are associated with atypical red blood cell antigens and capable of causing fetal hemolytic disease (Table 10-3). Any positive result on antibody screening should be evaluated aggressively to identify the antibody and quantify its amount by titer. A consultation with a blood bank pathologist may be necessary for atypical antibodies. The amount of the anti-D antibody present according to indirect antibody titer is important. For a titer that is less than 16 and remains at that level throughout the pregnancy, most centers consider the fetus to be at negligible risk of hydrops or stillbirth. Each blood bank sets different standards for this critical titer, which depend on the assay used. At a titer of 16, the risk is 10%; at 128, it is 75%. For a woman with a previously affected fetus, no titer is predictive; therefore management based on its result may underestimate the severity of fetal disease. Once the patient has been identified as isoimmunized, obtaining her obstetric history is important. All of her prior pregnancies and their outcomes must be documented to attempt to elucidate the timing and cause of the sensitization and to assess the risk in the current pregnancy. In general, the condition is worse with each pregnancy. In a first sensitized pregnancy, the risk of hydrops is approximately 10%. More than 90% of patients who have delivered one hydropic baby will deliver another one subsequently. Paternal blood typing and Rh genotyping should be performed to calculate the fetal risk of Rh positivity. Given TABLE 10-3 Examples of Atypical Red Blood Cell Antigens Associated With Fetal Hemolytic Disease Incidence (%) Data from Bowman JM: Maternal blood group immunization. In Creasy R, Resnik R, editors: Maternal-fetal medicine: principles and practice, Philadelphia, 1984, WB Saunders. 95 Blood Group System Antigen Severity of Hemolytic Disease Kell K Mild to severe Duffy Fya Mild to severe Kidd JKa JKb Mild to severe Mild to severe MNSs M N S s U Mild to severe Mild Mild to severe Mild to severe Mild to severe P PP Mild to severe Public antigens Yta Moderate to severe PART III Maternal Health Affecting Neonatal Outcome the high percentage of heterozygosity in Rh-­positive individuals, assays utilizing free DNA in the maternal circulation are used to determine a potentially D+ fetus. Accuracy with this technique can approach 97% (Geifman-Holtzman et al, 2006). For atypical blood group immunization, a history of prior obstetric outcomes plus transfusions (a significant cause of such antibodies) and determination of the father’s antigen status are of similar importance. For example, a woman with an anti-Kell antibody titer of 128 would be at moderate risk of hydrops, unless the father of the baby were found to be Kell negative. No invasive procedures for isoimmunization should be performed until the father’s antibody status is established, unless the fetus’s paternity is in question or the partner is unavailable. Ultrasonographic screening to identify the prehydropic fetus is notoriously unreliable, in that significantly anemic fetuses may not be grossly hydropic. However, clues that have been proposed to suggest impending hydrops are polyhydramnios, skin thickening, early ascites (particularly around the fetal bladder), and placental thickening. Ultrasonographic measurements of the liver and spleen have been suggested as aids in predicting anemia in nonhydropic fetuses, but lack sensitivity and specificity (Bahado-Singh et al, 1998; Vintzileos et al, 1986). Antibody titers should be followed monthly to predict which fetus is at risk (in the absence of a history of a prior infant with hydrops). If the indirect antibody titer value is less than 16, the development of hydrops is unlikely. Most centers consider the critical value to be 16, because above this level, one cannot ensure the absence of hydrops. When a patient’s antibody titer value is equal to or greater than 16, more invasive testing is needed. Such testing is usually accomplished with amniocentesis and, less commonly, through direct fetal blood sampling. Amniocentesis is the standard screening technique for fetal anemia, because amniotic fluid contains hemolytic products excreted from the fetal kidneys and lungs, including bilirubin. The unconjugated form of bilirubin is secreted across the respiratory tree and therefore into the amniotic fluid. The supernatant is analyzed with a spectrophotometer, and the bilirubin peak corresponding to the level of absorbance (or optical density [OD]) of these hemoglobin products (450 nm) is quantified. The magnitude of the peak is calculated after subtracting the mean baseline value surrounding the peak. The difference from baseline to peak is the ∆OD450 value. The amount of the increase in the ∆OD450 value correlates reasonably well with severity of hemolysis. In 1961, Liley developed a graph that has been used to predict the severity of fetal hemolytic disease using the ∆OD450 value. The graph is divided into three zones. Zone I represents the lowest risk and indicates an unaffected fetus, whereas zone III strongly suggests a severely affected fetus, in which fetal hydrops and death can ensue if disease is not treated within the next 7 to 10 days. The Liley curve is authoritative in the third trimester, but its reliability before 26 weeks’ gestation has been questioned (Queenan et al, 1993). At less than 20 weeks’ gestation, the ∆OD450 value must be greater than 0.15 or less than 0.09 to be predictive of severe or mild disease, respectively, with a “gray zone” between the two values that is nonpredictive (Ananth and Queenan, 1989). Thus, the clinician must integrate clinical history and ultrasonography clues for possible impending hydrops or consider fetal blood sampling in a fetus at risk between 18 and 25 weeks’ gestation. Amniocytes can be obtained during the amniocentesis procedure. Through polymerase chain reaction analysis, fetal D antigen typing can be obtained as early as 14 weeks’ gestation (Dildy et al, 1996). This process allows the identification of the fetus that is not at risk for hemolytic disease caused by maternal antibody isoimmunization. Isoimmunization with atypical red blood cell antigens other than D can be detected similarly, reducing the amount of unnecessary invasive procedures on an otherwise unaffected fetus. A recent advance in ultrasound technology has dramatically altered the assessment of the potentially affected fetus. Doppler blood flow measurements in the umbilical vein (Iskarios et al, 1998) and in the middle cerebral artery have shown to be altered in anemia, with the latter revealing an elevated peak systolic velocity. Using a developed reference range with a cutoff of 1.5 multiples of the median, Mari (2000), was able to predict which nonhydropic fetuses had moderate or severe anemia (Figure 10-2). A prospective multicenter trial compared the use of middle cerebral artery Doppler against the “gold standard” amniotic fluid ∆OD450. One hundred sixty-five fetuses were studied, and almost half were found to have severe fetal anemia at cordocentesis. Middle cerebral artery Doppler was noted to have a sensitivity of 88%, a specificity of 85%, and an accuracy of 85%. Doppler outperformed ∆OD450 analysis using the Liley curve, with a 9% improvement in the accuracy. Furthermore, it has been suggested that 50% of invasive procedures could have been averted with the use of Doppler (Oepkes et al, 2006). Direct ultrasoundguided fetal umbilical blood sampling provides valuable 120 110 MCA peak systolic velocity 96 100 1.5 MoM 90 80 70 Median 60 50 40 30 20 10 15 17 19 21 23 25 27 29 31 33 35 37 39 Gestational age (wks) FIGURE 10-2 Peak velocity of systolic blood flow in the middle cerebral artery (MCA) with advancing gestation. The bottom curve indicates the median peak systolic velocity in the middle cerebral artery, and the top line indicates 1.5 multiples of the median (MoM; the threshold for significant fetal anemia). (Adapted from Mari G, Hanif F: Fetal Doppler: umbilical artery, middle cerebral artery, and venous system, Semin Perinatol 32:254, 2008.) CHAPTER 10 Maternal Medical Disorders of Fetal Significance: Seizure Disorders, Isoimmunization, Cancer, and Mental Health Disorders data for the fetus at risk, particularly after 18 weeks’ gestation. Cordocentesis is used most often when the risk determined from history, indirect antibody titer values, Liley curve comparisons, or ultrasonographic clues are significant. Cordocentesis also provides vascular access if fetal transfusion becomes necessary. Although most obstetricians are skilled in amniocentesis, cordocentesis is usually performed in tertiary centers by perinatologists. The latter procedure carries a higher risk of fetal loss and morbidity, and it is associated with a more significant chance of worsening maternal sensitization (Bowman and Pollock, 1994). THERAPY Fetal transfusion therapy is the mainstay of treatment for the severely affected but premature fetus. In most centers, for pregnancies beyond 33 weeks’ gestation with suspected fetal anemia, administration of steroids and delivery are preferable to an invasive procedure with significant morbidity and mortality. Intraperitoneal transfusions, the primary therapy in the past, are still a useful treatment. O-negative, tightly packed, irradiated blood is infused percutaneously into the fetal peritoneal cavity via an amniocentesis needle with real-time ultrasound guidance. Approximately 10 mL is transfused for each week of gestation beyond 20 weeks. Red blood cell absorption occurs promptly through the subdiaphragmatic lymphatics, although it may be erratic in the hydropic fetus. Overall survival rates with fetuses requiring transfusion approaches 89% (Van Kamp et al, 2005). However, hydrops was associated with a decreased survival rate of 78% (Van Kamp et al, 2001), with severely hydropic fetuses having a survival rate of 55%. Problems associated with intraperitoneal transfusions include injury to vascular or intraabdominal organs and the inability to obtain fetal blood type and blood count. This procedure should be avoided if possible in the hydropic fetus, because the resulting higher abdominal pressure can precipitate venous compression and lead to circulatory collapse. Direct intravascular transfusions are currently the first line of treatment. The advantages of this procedure include assessing the severity of fetal anemia and documenting the fetal blood type. Ultimate success does not appear to be affected by the presence of hydrops (Ney et al, 1991). The overall success rate is approximately 85%. Limitations are usually related to the procedure. The most accessible site usually requires an anterior placenta, in which the cord root can be visualized; posterior placentation makes the technique more difficult. In addition, risks include bleeding from the cord puncture site, fetal exsanguination, cord hematoma, rupture of membranes, and chorioamnionitis. Additional transfusions are required every 2 to 3 weeks to compensate for the falling hematocrit associated with fetal growth, the finite life of the transfused red blood cells, and ongoing hemolysis of the existing fetal erythrocytes. Ultimately the entire fetal blood supply is replaced with Rh-negative blood by successive transfusions. A major neonatal side effect is bone marrow depression, which may require postnatal transfusions for 1 to 3 months (De Boer et al, 2008). 97 As the pregnancy proceeds through the third trimester, frequent fetal testing is performed either with nonstress tests and amniotic fluid index or with biophysical profiles. Delivery is planned for 3 to 6 weeks before term, usually after demonstration of a mature fetal lung profile. Preterm delivery may be indicated in a severely affected fetus, regardless of lung profile, if the risk of transfusion is deemed to exceed the morbidity of a delivery of a nearterm yet premature baby. CANCER PRINCIPLES Cancer complicating pregnancy is rare, with an estimated frequency of 1 case per 1000 live births. The trend of delaying childbearing to later maternal age may have influenced this rate. The sites or types of cancer in pregnancy, in descending order of frequency, are cervical, breast, ovarian, lymphoma, melanoma, brain, and leukemia (Table 10-4) (Haas et al, 1984; Jacob and Stringer, 1990). Finding a malignancy during gestation poses a unique set of issues that must be addressed with care. Will the pregnancy accelerate the malignant process? Are the accepted therapies appropriate for the mother, and are they safe for the unborn fetus? Will delay of therapy adversely affect the mother? Should the pregnancy be terminated, or should the child be delivered prematurely to maximize treatment of the mother with no resultant risk to the child? Few conditions in pregnancy require as meticulous a multidisciplinary approach as cancer. Oncologists who are unaccustomed to interacting with a pregnant woman commonly wish to have the child delivered before giving definitive cancer therapy, for fear of the teratogenic risks to the fetus. Therefore input regarding the situation must be acquired not only from an oncologist but also from the obstetrician, perinatologist, pediatrician, neonatologist, and dysmorphologist. The patient and her family must be involved in decision making, being given information not only about the risks of the disease and its potential therapies, but also about the limitations of current knowledge about cancer in pregnancy and the uncertainties of outcomes. TABLE 10-4 Cancers that Can Complicate Pregnancy Site/Type Incidence (per number of gestations) Cervix 1:2000 to 1:10,000 Breast 1:3000 to 1:10,000 Melanoma 1:1000 to 1:10,000 Ovary 1:10,000 to 1:100,000 Colorectal 1:13,000 Leukemia 1:75,000 to 1:100,000 Lymphoma 1:1000 to 1:6000 Data from Pentheroudakis G, Pavlidis N, Castiglione M: Cancer, fertility and ­pregnancy: ESMO clinical recommendations for diagnosis, treatment and followup, Ann Oncol 20(Suppl 4):S178-S181, 2009. 98 PART III Maternal Health Affecting Neonatal Outcome SURGERY Patients can undergo indicated surgery during pregnancy safely. The risk of fetal loss does not seem to rise with uncomplicated anesthesia and surgery. However, if there are any complications from either the anesthesia or the operation, the risk of fetal and maternal mortality increases. For example, fetal loss is rarely related temporally to maternal appendectomy, but is common when the mother’s appendix has ruptured. The most comprehensive series of pregnant patients undergoing surgery was collected in Sweden (Mazze and Kallen, 1989). Although these researchers found a slight increase in the rates of low birthweight and neonatal death by 7 days of life, the rates of stillbirth and congenital malformation were similar to the outcomes expected without surgery. Anesthetic agents are not believed to be teratogenic. A few management guidelines help to optimize outcome for the pregnant patient undergoing surgery. If general anesthesia is used, the maternal airway must be protected to avoid aspiration. Gastrointestinal motility is reduced during pregnancy, and the stomach may contain significant residual contents after many hours without eating. A left lateral decubitus position on the operating table is preferred to maximize uteroplacental blood flow. If the fetus is viable, monitoring of the fetal heart rate should be performed to assist the anesthesia team in optimizing fetal status. Preoperative counseling with the patient is important to allow the surgical team to make appropriate choices in regard to any interventions for fetal distress. CHEMOTHERAPY Administering agents that impair cell division in pregnancy is a concern for both the mother and the care team. Often the patient is more concerned about this issue than about the underlying cancer. Cytotoxic chemotherapy should be avoided in the first trimester because of the high incidence of spontaneous abortion and the potential teratogenic effects on the fetus (Table 10-5). For a few agents with confirmed teratogenic effects (e.g., methotrexate, amethopterin, chlorambucil), chemotherapy must be avoided during organogenesis. The use of folic acid antagonists in first-trimester cancer treatment raises the specific problem of possible induction of neural tube defects, because these lesions are known to be folate sensitive. The literature regarding most other chemotherapeutic agents is limited, consisting of a few collected series; therefore these agents should be used cautiously, with their potential harm to the fetus balanced against their benefit to the maternal condition. Little is also known regarding the long-term outcomes of fetuses exposed to chemotherapeutic agents in utero. The National Cancer Institute in Bethesda, Maryland, maintains a registry in the hopes of determining the delayed effects. A small series of fetuses exposed to chemotherapeutic agents for acute leukemia revealed normal mental development with follow-up between 4 and 22 years (Aviles and Niz, 1988). The risk of teratogenicity does not appear to be higher with combination chemotherapy than with single-agent therapy (Doll et al, 1989). Studies performed thus far involve small numbers of patients, with power insufficient to show a statistic difference, but there seems not to be a trend. Low birthweight is found in approximately 40% of babies whose mothers received cytotoxic drugs during pregnancy (Nicholson, 1968). Theoretical consequences, such as bone marrow suppression, immune suppression, and anemia, could occur in the fetus. As a result, the timing of chemotherapy should account for the anticipated date of delivery. Data regarding safety for breastfeeding the neonate of a mother receiving cancer chemotherapy are limited. For this reason, the majority of agents are contraindicated in nursing mothers. TABLE 10-5 Common Chemotherapy Agents and Uses Class or Drug Risk Category Common Uses Alkylating Agents Busulfan Dm Leukemias Chlorambucil Dm Lymphomas, ­leukemias Cyclophosphamide Dm Breast, ovary, ­lymphomas, ­leukemias Melphalan Dm Ovary, leukemia, myeloma Procarbazine Dm Lymphomas 5-Fluorouracil D* Breast, ­gastrointestinal 6-Mercaptopurine Dm Leukemias Methotrexate Xm Trophoblastic disease, lymphomas, leukemias, breast 6-Thioguanine Dm Leukemias Bleomycin Dm Cervix, lymphomas Daunorubicin Dm Leukemias Doxorubicin Dm Leukemias, ­lymphomas, breast All-trans-retinoic acid X Leukemias l-Asparaginase Cm Leukemias Cisplatin Dm Ovary, cervix, sarcoma Hydroxyurea D Leukemias Prednisone C* Lymphomas, ­leukemias, breast Tamoxifen Dm Breast, uterus Paclitaxel Dm Breast, ovarian Vinblastine Dm Breast, lymphomas, choriocarcinoma Vincristine Dm Leukemias, ­lymphomas Antimetabolites Antibiotics Other Agents Data from Neoplastic diseases. In Cunningham FG, MacDonald PC, Gant NF, et al, editors: Williams obstetrics, ed 20, Stamford, Conn., 1997, Appleton and Lange; and Briggs GC, Freeman RK, Yaffe SJ: Drugs in pregnancy and lactation, ed 7, Philadelphia, 2005, Lippincott Williams & Wilkins. *Risk factor D if used in the first trimester. CHAPTER 10 Maternal Medical Disorders of Fetal Significance: Seizure Disorders, Isoimmunization, Cancer, and Mental Health Disorders RADIATION THERAPY The deleterious effects of irradiation on the fetus have been both theorized and actual. Fortunately, the amount of concern about the former far exceeds the incidence of the latter. Irradiation promotes genetic damage and thus the potential for congenital malformations. The risk depends on both dose and time (Table 10-6). A dose of less than 5 rad is believed to be of little consequence (Brent et al, 1989). If radiation exposure occurs before implantation, the adverse outcomes are usually a small increase in miscarriage. The major concern is high-dose radiation (>10 rad) received during the period of organogenesis (embryonic weeks 1 to 10). The central nervous system is the most radiation-sensitive organ, and the complications most often observed are microcephaly and mental retardation. Cataracts and retinal degeneration are also seen. After organogenesis is complete, there is still a risk of central nervous system abnormalities. However, the sequelae most often seen are skin changes and anemia. Because of the highly variable yet potentially grave consequences of irradiation greater than 10 rad, patients should be counseled accordingly, and termination of pregnancy should be offered as an alternative if exposure has occurred in the previable period (Orr and Shingleton, 1983). There are several other considerations for a pregnant woman undergoing radiation therapy. First, the dose used in estimating risk should be the amount that the fetus actually receives. For example, axillary or neck irradiation for lymphoma involves a lower direct fetal exposure than direct pelvic irradiation for cervical cancer. The latter treatment, if given in the second trimester, will likely cause fetal demise. Second, the magnitude of radiation scatter to the pelvis must be considered. External shielding does not prevent internal reflection of the ion beam. Third, the advancing size of the uterus actually increases the amount of radiation exposure of the fetus, because of the closer proximity of the nonpelvic irradiation. Therefore an 8-week-old fetus may actually receive a smaller radiation dose from supraclavicular irradiation than a 30-week-old fetus. Fourth, will the fetus concentrate the radiation, and therefore increase its dose? This is exemplified by the use of radioactive iodine (131I) for maternal thyroid conditions. TABLE 10-6 Radiation Dose Thresholds for Deleterious Fetal Effects Weeks Since LMP Fetal Dose (mGy) Potential Fetal Effects 2-4 (preimplantation) >50-100 Spontaneous ­abortion, but generally not ­malformation The actual rad dose is markedly higher in the fetus, because the fetal thyroid concentrates the iodine. Diagnostic tests such as radiography may also be associated with radiation exposure for the fetus (Table 10-7). The doses involved are usually much smaller than those used for cancer therapy. Nonetheless, the practitioner should limit the amount of radiographic testing if at all possible. Inadvertent imaging of a patient who is not known to be pregnant continues to occur despite sensitive pregnancy tests, thus creating significant concerns. Indicated radiographs should never be withheld because of pregnancy, but lead shielding of the patient’s abdomen and careful selection of the type of study should be performed to minimize the pelvic dose. Usually the amount of fetal exposure is much less than 5 rad, with no significantly greater risk of malformations. There does appear to be a slightly higher incidence of childhood cancer if the fetus is exposed to doses on the order of 10 mGy (Doll and Wakeford, 1997). CERVICAL CANCER Cervical cancer is the most common malignancy found in pregnancy. The incidence is approximately 1 in 2500 gestations. A Papanicolaou test smear should be performed for all patients at their first prenatal visit. However, approximately 30% of cervical cancers can be associated with negative cytologic smear results. Although the evaluation for TABLE 10-7 Estimated Radiation Dose to the Fetus With ­Common Diagnostic Imaging Procedures Test Fetal Dose (rad) Radiograph Upper extremity <0.001 Lower extremity <0.001 Upper gastrointestinal series (barium) 0.048-0.360 Cholecystography 0.005-0.060 Lumbar spine 0.346-0.620 Pelvis 0.040-0.238 Hip and femur series 0.051-0.370 Chest (two views) <0.010 Retropyelography 0.800 Abdomen (kidneys, ureter, bladder 0.200-0.245 Urography (intravenous pyelography) 0.358-1.398 Barium enema 0.700-3.986 Computed Tomography Scan Head <0.050 Chest 0.100-0.450 Abdomen (10 slices) 0.240-2.600 Abdomen and pelvis 0.640 2-8 (organogenesis) >200 Malformations Pelvis 0.730 8-15 100-1000 Severe mental ­retardation Lumbar spine 3.500 >15 >1000 Mental retardation Adapted from Doyle S, Messiou C, Rutherford JM, et al: Cancer presenting during pregnancy: radiological perspectives, Clin Radiol 64:857-871, 2009. LMP, Last menstrual period. 99 Other Ventilation perfusion scan 0.06-1.00 From Bentur Y: Ionizing and nonionizing radiation in pregnancy. In: Koren G: ­Medication safety in pregnancy and breastfeeding, Philadelphia, 2007, McGraw Hill. 100 PART III Maternal Health Affecting Neonatal Outcome an abnormal Papanicolaou test result should not be altered because of pregnancy, many physicians are reluctant for fear of cervical hemorrhage. Cervical biopsy remains the mainstay of diagnosis. The greater vascularity of the cervix during pregnancy predisposes bleeding. An experienced colposcopist may be able to defer actual biopsy in cases of possible visual findings of a noninvasive process. However, if cancerous invasion is suspected or if the physician is uncertain of the visual findings, biopsy is necessary. If microinvasive disease is confirmed by biopsy, cone biopsy is required to rule out frankly invasive disease. This procedure is undertaken with caution during pregnancy, because of the associated high rate of bleeding complications and miscarriage. Cervical conization may raise the risk of incompetence or preterm labor. The assistance of a gynecologic oncologist is preferred, given these unique sets of potential consequences. A shallow cone biopsy will reduce the risk of subsequent cervical weakness. The therapy for invasive cervical cancer is based on the stage of disease and the gestational age of the fetus. Therapy can involve external beam radiation, internal radiotherapy (brachytherapy), or surgery (Table 10-8). In most cases, delay of definitive therapy by 4 to 14 weeks may be acceptable. Pregnancy does not seem to accelerate the growth of the tumor. However, patient counseling is important. In the extremely previable gestation, the likelihood of achieving a safe gestational age for the fetus without worsening the stage or spread of the cancer in the mother must be balanced against parental desires based on ethical or religious beliefs. Conversely, it might be reasonable to delay definitive therapy until a time when delivery would not likely result in a long-term disability because of extreme prematurity. BREAST CANCER Breast cancer is the most common malignancy of women, with approximately 1 in 8 women affected in their lifetimes (Goldman and O’Hair, 2009). The incidence of breast cancer in pregnancy is estimated to be 10 to 30 per 100,000 pregnancies (Isaacs, 1995). Pregnancy does not seem to influence the actual course of the disease; however, there appears to be a higher risk of delay in diagnosis and a trend toward more advanced stages at diagnosis in pregnant women than in nonpregnant controls. The diagnostic procedures for breast cancer should not be altered during pregnancy. Any suspicious mass should undergo biopsy. Mammography, although discouraged for routine screening in pregnancy, can be used safely if indicated. The amount of radiation is negligible—­ approximately 0.01 cGy (Liberman et al, 1994). Mammography imaging may be hampered by physiologic changes because of pregnancy, and ultrasound examination may be a useful alternative. Metastatic evaluation may be hampered somewhat because of a reluctance to use bone and liver scans during pregnancy. Magnetic resonance imaging can be used safely in the second and third trimesters. Surgical therapy for breast cancer should not be delayed because of pregnancy. The risks of mastectomy and axillary node dissection appear to be low (Isaacs, 1995). Radiation therapy is usually not recommended during pregnancy because of the risk of beam scatter to the pregnant uterus. If the pregnancy is to ­continue and the patient has evidence of tumor invasion in the lymph nodes, adjuvant chemotherapy is often given. The timing of delivery should account for the following factors: ll When would the fetus have a reasonable chance for survival with a low risk of severe permanent morbidity? ll Can the number of cycles of chemotherapy be minimized with an earlier delivery? In addition, avoiding delivery just before or just after administration of ­chemotherapy is important to reduce the risk of immunosuppression and infection. ll How long can radiotherapy be delayed without increasing the risk of metastatic spread of the tumor? Approximately 10% of women treated for breast cancer become pregnant, the majority within 5 years of diagnosis. Data from small series suggest that pregnancy does not influence the rate of recurrences or of distal metastasis (Dow et al, 1994). However, women should be encouraged to delay childbearing for at least 2 to 3 years, which is the time of the highest rate of recurrence. Breastfeeding may be possible in women who have undergone conservative breast cancer surgery. TABLE 10-8 Treatment Options for Cervical Cancer in Pregnancy Gestational Age (weeks) Stage I to IIa Stage IIb to IIIb <20 4500 cGy Wide pelvic irradiation If no spontaneous abortion: modified radical hysterectomy If spontaneous abortion: brachytherapy or radical hysterectomy with lymphadenectomy 5000 cGy Wide pelvic irradiation If no spontaneous abortion: type II radical hysterectomy If spontaneous abortion: brachytherapy or cesarean section at fetal viability Subsequent wide pelvic radiation (±5000 cGy) and brachytherapy (±5000 cGy) >20 Cesarean section at fetal viability Subsequent wide pelvic radiation (5000 cGy) and brachytherapy (5000 cGy) or cesarean radical hysterectomy with lymphadenectomy Cesarean section at fetal viability Subsequent wide pelvic radiation (±5000 cGy) and brachytherapy (±5000 cGy) Data from Berman ML, Di Saia PJ, Brewster WR: Pelvic malignancies, gestational trophoblastic neoplasia and nonpelvic malignancies. In Creasy RK, Resnik R, editors: Maternalfetal medicine, ed 4, Philadelphia, 1999, WB Saunders. CHAPTER 10 Maternal Medical Disorders of Fetal Significance: Seizure Disorders, Isoimmunization, Cancer, and Mental Health Disorders OVARIAN CANCER Most ovarian cancer occurs in women older than 35 years. Delayed childbearing has been more widely accepted, exemplified by British birth rates doubling in women older than 30 years and tripling in women older than 40 years since 1975 (Palmer et al, 2009), in addition to a twofold increased birth rate among U.S. women older than 40 years since 1981 (Martin et al, 2009). It would not be surprising for the rate of ovarian and other cancers during pregnancy to increase. However, the current estimate of actual ovarian malignancies in pregnancy is low and estimated to range from 1 in 10,000 to 1 in 50,000 deliveries (Jacob and Stringer, 1990; Palmer et al, 2009). Whereas most ovarian cancers are epithelial in origin, borderline epithelial and germ cell tumors (dysgerminomas and malignant teratomas) are more common in pregnancy. The widespread use of ultrasonography, particularly in the first two trimesters, has been helpful in identifying adnexal masses. Fortunately, most are benign functional cysts. Actual malignancy is rare and is estimated at 5% of the ovarian masses found. The risk is higher in nonpregnant females, approaching 15% to 20%. Surgery for a suspected ovarian mass occurs in approximately 1 per 1000 pregnancies. Most procedures are performed not for suspected malignancy, but because of concern about torsion and rupture. The incidence of adnexal torsion ranges from 1% to 50%, and there appears to be a trend with increasing rates in masses greater than 6 cm (Yen et al, 2009). The maximal times of risk of these events are at the end of the first trimester, when the uterus elevates beyond the true pelvis, and at the time of delivery. The characterization of an ovarian process can be aided by ultrasonography or magnetic resonance imaging, but these modalities are not definitive. Ultrasound scoring systems that use size and character poorly predict malignancy, but have a better negative predictive value (Lerner et al, 1994). Although an ovarian cyst, particularly if it is simple in nature, is likely not malignant, the patient must be cautioned that histologic diagnosis is more definitive. Indications for surgical exploration include a complex mass, a persistent simple cyst 8 cm or larger, or one that is symptomatic (Leiserowitz, 2006). The optimal time for laparotomy is in the second trimester. At that time, there is minimal interference from the gravid uterus and less of a risk of fetal loss, and the theoretical concerns of teratogenic exposure to anesthetic agents are avoided. Some patients opt for more conservative management; they should be counseled that they have a 40% chance of needing urgent intervention with either surgery or percutaneous drainage (Platek, 1995). If a malignancy is confirmed at the time of laparotomy, treatment and staging are no different than for a nonpregnant woman. Frozen-section diagnosis, peritoneal washings, omentectomy, and subdiaphragmatic biopsy are performed. Depending on the cell type and the stage, treatment can range from removal of the affected adnexa to complete hysterectomy and bilateral oophorectomy. Chemotherapy may be given during pregnancy if necessary. Fortunately, most epithelial ovarian cancers found in pregnant women are usually of 101 a lower stage, with 59% of reported cases being stage I (Palmer et al, 2009). SURVIVORS OF CHILDHOOD CANCER Given the improvements of therapy for childhood cancer, a large number of these individuals have survived into adulthood. Some are unable to conceive because of high-dose radiation or cytotoxic chemotherapy. The risk of decreased fertility for patients exposed to pelvic radiation therapy may as high as 32% (Geogeseu et al, 2008). Those who remain fertile may have concerns regarding whether their treatment increases the risk of adverse pregnancy outcomes. Although data are limited, female cancer survivors treated with radiation therapy appear to have increased risks of premature delivery, low birthweight, and miscarriage. There is no evidence that female partners of male cancer survivors treated with radiation have these excess risks (Reulen et al, 2009). MENTAL DISORDERS Pregnancy can be a stressful process. At times, it may induce a psychotic event. Women experiencing a mental disorder during pregnancy who have no history of a mood disorder usually exhibit a milder constellation of symptoms. Serious disorders such as mania and schizophrenia that are antecedent to pregnancy may not be so benign. In women with all types of mental illness, and in previously nonaffected women, the postpartum state is a time of greater maternal risk. Ten percent to 15% of new mothers experience a depressive disorder (Weissman and Olfson, 1995). Furthermore, there appears to be an increased perinatal risk with mental disorders and pregnancy (Table 10-9). Women with preexisting mental illness have a higher recurrence risk in the puerperium. Patients with suspected mental illness should be assessed for substance abuse and thyroid dysfunction. A multidisciplinary approach is advantageous. If the patient’s mental competency is an issue, the caregiver should obtain legal assistance to be able to make medical decisions for the patient. DEPRESSION Depression ranks as the fourth leading cause of ­disability worldwide, and recognized prevalence appears to be increasing (Dossett, 2008). A study by Dietz et al (2007) suggests that the prevalence of depression during pregnancy or postpartum, as defined by onset within 3 to 6 months after delivery, to be approximately 10%. The predisposing risk factors for depression include early childhood loss, physical or sexual abuse, socioeconomic deprivation, genetic predisposition, and lifestyle stress caused by multiple roles (McGrath et al, 1990). These factors can exaggerate or prolong symptoms and, if not addressed, can lengthen the duration of depression. The obstetrician must be aware that life events such as miscarriage, infertility, and complicated pregnancy in patients with risk factors are likely to precipitate depression; therefore there is a low threshold for diagnosis and treatment of mood alterations in such patients. 102 PART III Maternal Health Affecting Neonatal Outcome TABLE 10-9 Impact of Psychiatric Illness on Pregnancy Outcome Impact on Outcome Illness Teratogenic Effects Obstetric Neonatal Treatment Options Anxiety disorders N/A Increased incidence of forceps deliveries, prolonged labor, ­precipitate labor, fetal distress, preterm delivery, and ­spontaneous abortion Decreased developmental scores and inadaptability; slowed mental development at 2 years of age Benzodiazepines Antidepressants Psychotherapy Major depression N/A Increased incidence of low ­birthweight, decreased fetal growth, and postnatal ­complication Increased newborn ­cortisol and catecholamine levels, infant crying, rates of ­admission to neonatal ­intensive care units Antidepressants Psychotherapy ECT Bipolar disorder N/A Increased incidence of low ­birthweight, decreased fetal growth, and postnatal ­complication Increased newborn cortisol and catecholamine levels, infant crying, rates of ­admission to neonatal intensive care units Lithium Anticonvulsants Antipsychotics ECT Schizophrenia Congenital malformations, especially ­cardiovascular Increased incidence of preterm delivery, low birthweight, small for gestational age, placental abnormalities and placental hemorrhage Increased rates of postnatal death Antipsychotics From ACOG Practice Bulletin: Clinical management guidelines for obstetrician-gynecologists number 92, April 2008 (replaces practice bulletin number 87, November 2007): use of psychiatric medications during pregnancy and lactation, Obstet Gynecol 111:1001-1020, 2008. ECT, Electroconvulsive therapy. Alternatively, perinatal loss experienced by a woman without predisposing risk factors will probably lead to a grief reaction or adjustment disorder, which may be misdiagnosed as depression. Chronic medical conditions that are associated with a high prevalence of depression and may occur in women of childbearing age include renal failure, cancer, AIDS, and chronic fatigue or pain. Antihypertensives, hormones, anticonvulsants, steroids, chemotherapeutics, and antibiotics can cause depression. Alcoholism and substance abuse may manifest as depression. Underlying personality disorders complicate the diagnosis of depression by confusing the clinical situation, in addition to contributing to the secondary effect of many physicians’ avoidance of patients suffering from such disorders. Therapeutic interventions for depression include psychotherapy and medication. Electroconvulsive therapy has been shown to be an effective and relatively safe treatment in refractory cases (Rabheru, 2001). However, there remains controversy and some degree of concern because of published series and case reports suggesting a risk of fetal cardiac arrhythmias, vaginal bleeding, and premature uterine contractions (Bhatia et al, 1999); therefore it should be reserved for refractory cases and performed in a setting with immediate access to obstetric care (Pinette et al, 2007; Richards, 2007). Treatment of depression is effective in approximately 70% of cases. Supportive treatment alone is rarely effective in major depressive episodes. Most antidepressant medications currently prescribed during pregnancy are selective serotonin reuptake inhibitors (SSRIs). SSRIs have an advantage over the tricyclic antidepressants by not causing orthostatic hypotension. Unfortunately, and although limited series suggest that the SSRIs are relatively safe, little is known about long-term consequences for children exposed to SSRIs in utero (Altshuler et al, 1996; Chambers et al, 1996; Karasu et al, 2000). Fluoxetine is the best-studied SSRI in terms of safety. Alternatively, paroxetine has been associated with an increased risk of congenital heart defects (Kallen et al, 2007). Although data remain inconsistent, they suggest avoiding paroxetine as a first-line agent. However, if a certain agent is controlling the patient’s symptoms, it would seem reasonable not to change medications for the sake of these concerns. A recent issue has been raised regarding the use of SSRIs and persistent pulmonary hypertension of the newborn; however, the incidence remains low at approximately 10 in 1000 fetuses exposed after 20 weeks’ gestation (Chambers et al, 2006). Currently, there is no consensus regarding the use of SSRIs during breastfeeding. Fluoxetine has an active metabolite with a long half-life and is found in higher concentrations in infants (Eberhard-Gran et al, 2006). Short-term neonatal effects have been reported, including increased crying, decreased sleep, and irritability, particularly with fluoxetine and citalopram. Reasonable guidelines regarding the use of SSRIs and other psychotropic medications are listed in Table 10-10. The long-term side effects are currently listed as “unknown” (Briggs et al, 2005; Dodd et al, 2000). The theoretical concerns are that such drugs may affect the developing central nervous system of the newborn and that abnormalities may not be readily apparent in the short term. Therefore SSRIs should be prescribed for a nursing mother only if the benefit clearly exceeds the risk, and after the patient has been counseled regarding the potential yet currently ill-defined risks. Given current medical knowledge, bottle feeding should be offered as an acceptable alternative if antidepressants must be used. CHAPTER 10 Maternal Medical Disorders of Fetal Significance: Seizure Disorders, Isoimmunization, Cancer, and Mental Health Disorders 103 TABLE 10-10 Summary of Current Knowledge of Drug Excretion into Breast Milk, Drug Concentrations in Infant Serum, Adverse Effects in the Child, and Breastfeeding Recommendations for Different Psychotropic Drugs Class or Drug Drug Transfer into Breast Milk Infant Plasma ­Concentrations Adverse Effects in the Child Breastfeeding ­ ecommendations R SSRIs Low transfer Low plasma concentrations Case reports of adverse effects in infants exposed to fluoxetine and citalopram Compatible with ­breastfeeding; however, fluoxetine and citalopram may not be drugs of first choice TCAs Low transfer Low plasma concentrations (except doxepin) No suspected immediate adverse effects observed (except doxepin) Compatible with breastfeeding; however, doxepin should be avoided Other antidepressants Limited data Limited data Limited data None able to be made Benzodiazepines Low transfer High plasma concentrations with longer acting drugs with active metabolites Case reports of CNS depression reported for diazepam Sporadic use of short-acting benzodiazepines unlikely to cause adverse effects Lithium Low transfer Dose received by the infant is high Limited data; some reports of toxicity in the infant Limited data; however, breastfeeding should be avoided Carbamazepine, sodium valproate Low transfer Low plasma concentrations Some case reports of various adverse effects in the infant Generally more compatible with breastfeeding than lithium Lamotrigine High transfer High plasma concentrations Limited data None able to be made Novel antipsychotics Moderate transfer Variable plasma ­concentrations Limited data None able to be made Modified from Eberhard-Gran M, Esklid A, Opjordsmoen S: Use of psychotropic medications in treating mood disorders during lactation: practical recommendations, CNS Drugs 20:187-198, 2006. CNS, Central nervous system; SSRI, selective serotonin reuptake inhibitor; TCA, tricyclic antidepressants. POSTPARTUM PSYCHOSIS A severe disorder, postpartum psychosis is fortunately rare, occurring in 1 to 4 per 1000 births (Weissman and Olfson, 1995). This condition is more worrisome than postpartum depression, because of the patient’s inability to discern reality from the periods of delirium. Patients at risk for postpartum psychosis may have underlying depression, mania, or schizophrenia. Other risks are younger age and family history. The recurrence rate is approximately 25%. The peak onset of symptoms is between 10 and 14 days after delivery. Recognition of this disorder is extremely important to the protection of the patient and her family. SCHIZOPHRENIA The prevalence of schizophrenia is approximately 1% in the general population (Myers et al, 1984); it is associated with delusions, hallucinations, and incoherence. Morbidity due to this mental illness is higher than that due to any other. There appears to be a genetic component to the etiology; schizophrenia develops in approximately 10% of offspring of an affected person. Concordance of schizophrenia in identical twins reaches 65%. There is some speculation and controversy as to whether low birthweight (Smith et al, 2001) and obstetric complications (Kendell et al, 2000) are associated with a higher rate of schizophrenia. Because the peak age of incidence is approximately 20 years and women are affected more often than men, it is unrealistic to assume that obstetricians will never encounter patients with schizophrenia. There appear to be higher rates of cesarean section and surgical vaginal delivery in affected patients (Bennedsen et al, 2001b). Children of women with schizophrenia may have a higher rate of sudden infant death syndrome and congenital malformations (Bennedsen et al, 2001a). However, it is difficult to ascertain whether these risks are independent of other factors such as smoking, poor socioeconomic status, and use of certain medications. Treatment is achieved primarily through the use of psychotropic medication. The potential for teratogenesis appears low with the older-generation medications in the phenothiazine class, but most data for this issue were derived from the use of lower doses given to patients with hyperemesis gravidarum. Antipsychotic medication does cross the placenta. Current recommendations include avoiding use in the first trimester if possible, the use of lower doses or higher-potency alternatives, and cessation of medication 5 to 10 days before delivery (Herz et al, 2000). The use of most antipsychotics in breastfeeding is associated with an unknown risk (Briggs et al, 2005). Lithium, used primarily in mania, is associated with a higher rate of Ebstein anomaly. Although the incidence of this consequence is low, either discontinuing the medication in the first trimester or continuing its use with careful counseling is a viable alternative. Fetal echocardiography should be performed in women who have used lithium in early pregnancy. 104 PART III Maternal Health Affecting Neonatal Outcome SUGGESTED READINGS ACOG Committee on Practice Bulletin–Obstetrics: ACOG Practice Bulletin: Clinical management guidelines for obstetrician-gynecologists, number 92, April 2008. Use of psychiatric medications during pregnancy and lactation, Obstet Gynecol 111:1001-1020, 2008. Bent RL: Saving lives and changing family histories: appropriate counseling of pregnant women and men and women of reproductive age, concerning the risk of diagnostic radiation exposures during and before pregnancy, Am J Obstet Gynecol 200:4-24, 2009. Cohn D, Ramaswamy B, Blum K, et al: Malignancy in pregnancy. In Creasy RK, Resnik R, Iams JD, editors: Maternal-fetal medicine, ed 6, Philadelphia, 2009, WB Saunders, pp 885-904. Moise KJ: Management of rhesus alloimmunization in pregnancy, Obstet Gynecol 112:164-176, 2008. Walker S, Permezal M, Berkovic S: The management of epilepsy in pregnancy, BJOG 116:758-767, 2009. Complete references and supplemental color images used in this text can be found online at www.expertconsult.com C H A P T E R 11 Hypertensive Complications of Pregnancy Andrew D. Hull and Thomas R. Moore Hypertension is the most common medical problem in pregnancy, affecting 10% to 15% of all pregnant women. As the third most common cause of maternal mortality after thromboembolic disease and hemorrhage, hypertension accounts for almost 16% of maternal deaths in the United States (Berg et al, 2003). Complications arising from hypertensive disorders have profound effects on the fetus and neonate and thus are a major source of perinatal mortality and morbidity. Preeclampsia is also the primary cause of iatrogenic prematurity. CLASSIFICATION OF HYPERTENSIVE DISORDERS OF PREGNANCY Any discussion of hypertension and pregnancy must begin with a set of definitions. Although many classifications are in use worldwide, perhaps one of the more useful comes from the Report of the National High Blood Pressure Education Program Working Group on High Blood Pressure in Pregnancy (2000) (Table 11-1). Although this classification scheme appears to be somewhat pedantic, it is of paramount importance because pregnancy outcome varies according to the type of hypertension involved. For practical purposes, hypertension in pregnancy can be divided into the following categories: chronic hypertension, gestational hypertension, and preeclampsia. Hypertension is defined as a systolic blood pressure of 140 mm Hg or higher or a diastolic pressure of 90 mm Hg or higher, measured on two separate occasions. Korotkoff phase V (disappearance of sound) is used rather than Korotkoff phase IV (muffling of sound) to define diastolic pressure, because Korotkoff IV is poorly reproducible in pregnancy. The term severe hypertension identifies a population at significantly increased risk for stroke and cardiac decompensation, and it is defined as a systolic blood pressure of 160 mm Hg or a diastolic pressure of 110 mm Hg or higher. CHRONIC HYPERTENSION Up to 5% of pregnant women have chronic hypertension, which is diagnosed when hypertension is present before pregnancy or recorded before 20 weeks’ gestation. However, when hypertension is first noted in a patient after 20 weeks’ gestation, it may be difficult to distinguish chronic hypertension from pregnancy-induced hypertension or preeclampsia. In such cases, the precise diagnosis might not be made until after delivery. Hypertension that is first diagnosed during the second half of pregnancy and persists more than 12 weeks postpartum is diagnosed as chronic hypertension. Chronic hypertension has an adverse effect on pregnancy outcome. Women with the disorder are at higher risk for preterm delivery and placental abruption, and their fetuses are at risk for intrauterine growth restriction (IUGR) and demise (Ferrer et al, 2000). Superimposed preeclampsia complicates up to 50% of pregnancies in women with preexisting severe chronic hypertension (Sibai and Anderson, 1986), and it occurs before 34 weeks’ gestation in 50% of cases (Chappell et al, 2008). The adverse effects on fetal and maternal perinatal outcomes are directly related to the severity of the preexisting hypertension. When chronic hypertension is secondary to maternal renal disease, the risks of poor outcome are further increased, with as much as a tenfold rise in fetal loss rate (Jungers et al, 1997). Women with untreated severe chronic hypertension are at increased risk for cardiovascular complications during pregnancy, including stroke (Brown and Whitworth, 1999). The majority of cases of chronic hypertension seen in pregnancy are idiopathic (essential hypertension), but causes of secondary hypertension should always be sought because pregnancy outcome is worse in women with secondary hypertension. Renal disease (e.g., chronic renal failure, glomerulonephritis, renal artery stenosis), cardiovascular causes (coarctation of the aorta, Takayasu arteritis), and, rarely, Cushing disease, Conn syndrome, and pheochromocytoma should be excluded through physical examination, history, and more detailed testing if needed. All patients with chronic hypertension should be evaluated periodically with serum urea, creatinine, and electrolyte measurements, urinalysis, and 24-hour urine collection for protein and creatinine clearance determinations. Typically this assessment should be performed in each trimester and more frequently if the patient’s condition deteriorates. ANTIHYPERTENSIVE TREATMENT OF CHRONIC HYPERTENSION IN PREGNANCY Except in cases of severe hypertension, randomized trials have shown that antihypertensive treatment of chronic hypertension in pregnancy does not improve fetal outcome (Sibai and Anderson, 1986). Rates of preterm delivery, abruption, IUGR, and perinatal death are similar in treated and untreated women. Therefore treatment is usually reserved for patients whose hypertension places them at a significant risk of stroke (systolic blood pressure of 180 mm Hg or higher or diastolic pressure of 110 mm Hg or higher). Patients with less severe hypertension who were taking medications before conception might be able to discontinue therapy with close surveillance. The risk of superimposed preeclampsia is not changed by antihypertensive therapy, so its development should be tracked carefully. 105 106 PART III Maternal Health Affecting Neonatal Outcome TABLE 11-1 Classification of Hypertensive Disorders TABLE 11-2 Drugs Commonly Used to Treat Chronic of Pregnancy* Category Definition Chronic hypertension Hypertension present before pregnancy or diagnosed before 20 weeks’ gestation, or diagnosed for the first time during pregnancy that persists postpartum Gestational hypertension Transient if blood pressure returns to normal by 12 weeks after delivery, and preeclampsia was not diagnosed before delivery Chronic if blood pressure does not resolve by 12 weeks after delivery Preeclampsiaeclampsia Usually occurs after 20 weeks’ gestation Hypertension accompanied by proteinuria in a woman normotensive before 20 weeks’ gestation Strongly suspected if nonproteinuric hypertension is accompanied by systemic symptoms such as headache, visual disturbance, abdominal pain, or laboratory abnormalities such as low platelet count and elevated liver enzyme values (HELLP syndrome) Preeclampsia ­superimposed on chronic hypertension Preeclampsia occurring in a chronically hypertensive woman Adapted from the Report of the National High Blood Pressure Education Program Working Group on High Blood Pressure in Pregnancy, Am J Obstet Gynecol 183:S1-S22, 2000. HELLP, Hemolysis, elevated liver enzymes, and low platelets. *Hypertension is defined as a systolic blood pressure ≥140 mm Hg systolic or diastolic blood pressure ≥90 mm Hg. The choice of antihypertensive agent for use in pregnancy is governed by a desire to adjust blood pressure without having ill effects on the fetus. Because excessive lowering of maternal blood pressure below 140 mm Hg systolic or 90 mm Hg diastolic (140/90 mm Hg) can compromise uterine perfusion, with consequent slowing of fetal growth, fetal hypoxia, or both, the therapeutic goal is to maintain maternal pressures at 140 to 155 systolic and 90 to 105 diastolic. The drugs most commonly used in pregnancy are listed in Table 11-2. Methyldopa, a centrally acting antihypertensive agent, formerly was the most widely used drug in this setting. Many obstetricians remain faithful to the use of this agent because of extensive clinical and research experience demonstrating its safety for both mother and fetus during pregnancy (Report of the National High Blood Pressure Education Program Working Group on High Blood Pressure in Pregnancy, 2000). This agent does not impair uteroplacental perfusion and has a wide therapeutic margin before side effects are seen. However, methyldopa has the disadvantage of a rather slow onset of action with prolonged time to therapeutic effect (days), and compliance with methyldopa therapy may be impeded by side effects such as sedation in some patients. Labetalol is a mixed alpha1-adrenergic and beta1- and beta2-adrenergic blocking agent, and it is the most frequently used alternative to methyldopa. Some pure betablockers have been associated with a significant increase in the risk of IUGR (e.g., atenolol), and the mixed adrenergic blockade produced by labetalol is thought to mitigate this Hypertension in Pregnancy and Their Modes  of Action Drug Mode of Action Methyldopa Centrally acting antihypertensive Labetalol Mixed alpha- and beta-adrenergic blocker Nifedipine Calcium channel blocker Hydralazine Peripheral vasodilator Prazosin Alpha-blocker unwanted effect (Pickles et al, 1989). Labetalol is also used intravenously to manage severe hypertension accompanying preeclampsia. Calcium channel blockers (e.g., nifedipine) are used mainly as second-line drugs, usually in long-acting, extended-release forms. Calcium channel blockers appear to be as effective as methyldopa and labetalol with minimal fetal side effects (Levin et al, 1994). Hydralazine, a potent peripheral vasodilator, is frequently used intravenously to treat acute hypertensive emergencies in pregnancy (blood pressure of >160/110 mm Hg). Its role as an oral agent in the management of chronic hypertension is limited to a second- or third-line choice. Long-term use of hydralazine may be associated with a lupuslike syndrome in some patients. Prazosin, an alpha-adrenergic blocker, has been used as a second- or third-line drug in pregnant women whose hypertension is difficult to control or is severe with an early onset. This agent appears to be similar in efficacy to nifedipine in such a setting (Hall et al, 2000). Although diuretics are used extensively in adults with hypertension, there appears to be little role for them in the treatment of chronic hypertension in pregnancy. Diuretics have been alleged to reduce or prevent the normal plasma volume expansion seen in pregnancy (Sibai et al, 1984), an effect that theoretically might impede fetal growth, although the evidence for this is mixed. Most authorities restrict the use of diuretics in pregnant patients to those with cardiac dysfunction or pulmonary edema. Angiotensin-converting enzyme inhibitors and angiotensin receptor blockers should not be used during pregnancy. In the second and third trimesters, these agents are associated with malformation of the fetal calvarium, fetal renal failure, oligohydramnios, pulmonary hypoplasia, and fetal and neonatal death (Buttar, 1997). Angiotensinconverting enzyme inhibitors appear to be safe when taken in the first trimester (Steffensen et al, 1998), but a patient who conceives while taking an angiotensin receptor blocker or angiotensin-converting enzyme inhibitor should be switched to a safer alternative as soon as possible. Similar precautions apply to the use of angiotensin receptor blockers in pregnancy. ANTENATAL FETAL SURVEILLANCE IN CHRONIC HYPERTENSION As the third trimester progresses, patients with chronic hypertension are at an increasing risk of slowing of fetal growth and superimposed preeclampsia. Antenatal surveillance CHAPTER 11 Hypertensive Complications of Pregnancy in women with chronic hypertension should include careful screening for signs and symptoms of superimposed preeclampsia, which constitutes the greatest perinatal risk. Fetal growth should be followed with serial ultrasonography evaluations (every 3 to 6 weeks). All patients should perform fetal movement counts from 28 weeks’ gestation onward, and cases with suspected fetal growth impairment should be followed with twice-weekly non-stress tests with amniotic fluid index or weekly ultrasound biophysical profile. Although the optimum interval for these tests is controversial (every 3 to 7 days), and their role is unproven in the absence of fetal IUGR or other evidence of fetal compromise, most centers begin regular fetal biophysical testing at 32 to 34 weeks’ gestation and continue until delivery. If fetal growth tapers below expectations (typically sonographic estimated fetal weight [EFW] falls below the tenth percentile or abdominal circumference much smaller percentile than head), more intensive fetal surveillance is indicated. In cases with IUGR, serial sonography should be performed at 10- to 21-day intervals with attention paid to amniotic fluid volume, careful profiling of each biometric parameter, and cerebral and umbilical Doppler waveforms. Typical indications for delivery in the setting of IUGR and hypertension include no growth of the head and abdomen over a 10-day interval, severe oligohydramnios, biophysical score of less than 6, or reversal of end-diastolic velocity on the umbilical Doppler waveform. However, individualization of management in these cases is important. Women with renal impairment and chronic hypertension have a markedly higher risk of poor perinatal outcome than normotensive women and women with hypertension without renal impairment. In addition, moderate or severe renal disease (serum creatinine level ≥1.4 mg/ dL) may accelerate the loss of renal function during pregnancy (Cunningham, 1990; Hou, 1999). The incidence of impaired fetal growth is directly related to the degree of renal impairment, and women undergoing dialysis are at particular risk for fetal growth failure, preterm delivery, and fetal death, even with optimal management. Those who start dialysis during pregnancy are at the greatest risk, with only a 50% chance of a surviving infant (Hou, 1999). GESTATIONAL HYPERTENSION The diagnosis of gestational hypertension can be made with confidence only after delivery; it is defined as hypertension occurring in the second half of pregnancy in the absence of any other signs or symptoms of preeclampsia. Because a woman with apparent gestational hypertension at 36 weeks’ gestation can rapidly evolve into preeclampsia at 39 weeks’ gestation, the diagnosis of gestational hypertension should always evoke caution and vigilance. Only if the patient’s blood pressure returns to normal postpartum without development of signs of preeclampsia during the pregnancy should the final diagnosis of gestational hypertension be applied. During pregnancy, gestational hypertension is indistinguishable from preeclampsia in evolution. Therefore all patients with gestational hypertension should be regarded as being at risk for progression to preeclampsia. The earlier gestational hypertension is evident, the greater the risk of preeclampsia. When the diagnosis is made 107 before 30 weeks’ gestation, more than one third will develop preeclampsia, whereas the risk is less than 10% when the diagnosis is made after 38 weeks’ gestation. Decisions to treat patients with gestational hypertension with antihypertensive agents must be carefully considered, given the risk of concurrent preeclampsia and the lack of evidence supporting improved fetal outcome. Gestational hypertension tends to recur in subsequent pregnancies and predisposes women to hypertension in the future (Marin et al, 2000). PREECLAMPSIA-ECLAMPSIA Preeclampsia is one of the most enigmatic diseases affecting humans. Apparently unique to humans, preeclampsia has proved difficult to simulate in animal experiments. Despite years of intensive research, the underlying causes of the disease are only recently becoming clearer. It is evident that the clinical manifestations of preeclampsia arise from vascular endothelial dysfunction that ultimately may involve the central nervous, renal, hepatic, and cardiovascular systems. In its full-blown form, preeclampsia can produce a profound coagulopathy and liver, respiratory, or cardiac failure. The classic symptom triad of hypertension, proteinuria, and edema defines preeclampsia. Most classifications of preeclampsia no longer include edema, because this common finding affects approximately 80% of pregnant women near term. Preeclampsia is divided into mild and severe forms (Box 11-1). This distinction is important, because in the presence of severe disease at any gestational age, the only appropriate treatment option is delivery, whereas expectant management may be acceptable in a woman who has mild disease and is remote from term. Although the precise etiology of preeclampsia remains uncertain, numerous factors are associated with elevated the risk (Table 11-3). Up to 10% of primigravid patients have mild preeclampsia, and approximately 1% have severe disease. PREECLAMPSIA Etiology The most widely accepted theory for the pathophysiology of preeclampsia is based on a model of impaired placental implantation that results in placental hypoperfusion and hypoxia. The placenta then releases substances into the maternal circulation that adversely affect endothelial function, leading to the clinical syndrome of widespread vascular dysfunction, which is recognized as the syndrome of preeclampsia (Myers and Baker, 2002). Individual responses to the process of progressive vascular dysfunction vary in severity and timing in a manner that seems to have genetic, familial, and immunologic components. For example, preeclampsia occurring in a first-degree relative confers a fourfold increase in risk of the disease in siblings and children (Chesley and Cooper, 1986). Women born to mothers with preeclampsia have a higher risk. There is some evidence that the presence of certain genotypes, such as factor V Leiden and thrombophilia (de Vries et al, 1997), or metabolic defects such as hyperhomocystinemia secondary to methylenetetrahydrofolate reductase deficiency (Kupferminc et al, 1999), predispose women to 108 PART III Maternal Health Affecting Neonatal Outcome BOX 11-1 F  eatures of Mild and Severe ­Preeclampsia TABLE 11-3 Risk Factors for Development of Preeclampsia Factor Relative Risk MILD ll Systolic blood pressure ≥140 mm Hg or diastolic pressure of 90 mm Hg ll Proteinuria ≥300 mg/24 hr Primigravida 3 Age >40 years 3 African American race 1.5 SEVERE* ll Systolic blood pressure ≥160 mm Hg or diastolic pressure of 100 mm Hg ll Proteinuria ≥5 g/24 hr ll Elevated serum creatinine value ll Eclampsia ll Pulmonary edema ll Oliguria <500 mL/hr ll HELLP syndrome ll Intrauterine growth restriction ll Symptoms suggestive of end-organ involvement: headache, visual disturbance, epigastric or right upper quadrant pain Family history 5 Chronic hypertension 10 Chronic renal disease 20 Antiphospholipid syndrome 10 Insulin-dependent diabetes mellitus 2 Multiple gestation 4 Modified from ACOG practice bulletin: Diagnosis and management of preeclampsia and eclampsia. Number 33, January 2002. American College of Obstetricians and Gynecologists, Int J Gynaecol Obstet 77:67-75, 2002. HELLP, Hemolysis, elevated liver enzymes, and low platelets. *Any single feature in the severe definition satisfies criteria for the diagnosis of severe preeclampsia. preeclampsia, although a true candidate gene is yet to be established and probably never will be. Population studies have suggested that women exposed to the antigenic effects of sperm before conception have a lower rate of preeclampsia than do women who conceive with lesser degrees of exposure, although the evidence is inconclusive (Koelman et al, 2000). The endothelial dysfunction that characterizes preeclampsia (Roberts, 1999) manifests as greater vascular reactivity to circulating vasoconstrictors such as angiotensin, reduced production of endogenous vasodilators such as prostacyclin and nitric oxide (Ashworth et al, 1997), increased vascular permeability, and an increased tendency toward platelet consumption and coagulopathy. The end result is hypertension, proteinuria secondary to glomerular injury, edema, and a tendency toward extravascular fluid overload with intravascular hemoconcentration. Predictors Perhaps one of the most important contributions that prenatal care makes to maternal and fetal outcomes is the detection of preeclampsia and the prevention of eclampsia (Backe and Nakling, 1993; Karbhari et al, 1972). A wide variety of biochemical and physical tests has been proposed as screening tools for the early detection of preeclampsia (Dekker and Sibai, 1991). Most physical tests have been discredited, and even the most widely used biochemical tests have poor predictive values. Uric acid levels are elevated in many cases of preeclampsia, but the sensitivity of the measurement is low (Lim et al, 1998). Early detection of proteinuria is possible with the use of more sensitive tests, such as gel electrophoresis (Winkler et al, 1988), rather than conventional urinalysis, but such tests do not lend themselves to routine use. Clinicians should be aware of the limitations of routine urine testing for detection of proteinuria, with standard dipstick testing being notoriously inaccurate (Bell et al, 1999). Doppler ultrasonographic assessment of the vascular dynamics in the uterine arteries during the second trimester has been proposed as a valuable screening tool in populations in which obstetric ultrasonography is routine (Kurdi, 1998). Up to 40% of women who develop preeclampsia have abnormal waveforms, and this finding was reported to be associated with a sixfold rise in the risk of preeclampsia (Papageorghiou et al, 2002). Other researchers have obtained less impressive results (Goffinet et al, 2001). Recently the role of angiogenic factors in the pathophysiology of preeclampsia has been explored. Vascular endothelial growth factor (VEGF) and placental growth factor bind to Flt-1 and sFlt-1 receptors and have a critical role in angiogenesis and placental development. The interactions among FLt-1, VEGF, and placental growth factor promote angiogenesis and placental vasculogenesis, whereas those among sFLT-1, VEGF, and placental growth factor lead to the inactivation of those proteins and disordered angiogenesis and endothelial dysfunction. sFlt-1 levels have been found to be elevated in women with preeclampsia, and such elevated levels of sFlt-1 precede the features of clinical preeclampsia. Recent reviews of the state of the art in this area (Wang et al, 2009; Widmer et al, 2007) concluded that sFlt-1 and placental growth factor levels were significantly different after 25 weeks’ gestation between women destined to develop preeclampsia and those with a normal pregnancy course. Measurements of these factors earlier in pregnancy do not appear to have the same predictive value. There is no doubt that angiogenic factors are intimately involved in the pathophysiology of preeclampsia, but alterations in their levels do not seem to be the cause of the disease. Several large studies are ongoing to further explore the role of these factors and their potential use in the prediction of disease. On balance, no effective screening test to predict preeclampsia currently exists, and clinicians are faced with the necessity of diagnosing the disease early and managing it as adroitly as possible. Prevention If an accurate predictor of preeclampsia could be identified, the next logical step would be the application of a preventive or ameliorative treatment. Unfortunately attempts to identify an effective treatment have proved equally CHAPTER 11 Hypertensive Complications of Pregnancy difficult. Given the recognized association between vascular endothelial dysfunction and preeclampsia (in particular, vasoconstriction and excessive clotting in the maternal placental arteries), prostaglandin inhibitors have been viewed as a likely candidate for prophylaxis or treatment. Numerous trials (Duley et al, 2001) have been conducted with lowdose aspirin, based on the idea that the ability of aspirin to irreversibly inhibit production of the vasoconstrictive prostaglandin thromboxane would promote greater activity of prostacyclin, a vasodilatory prostaglandin. This ability of aspirin would help to maintain patency in the maternal placental vascular bed and limit or prevent the evolution of preeclampsia. Unfortunately, although a modest reduction in the frequency of preeclampsia (approximately 15%) was documented, no improvement in key measures of perinatal outcome was demonstrable in a metaanalysis of the results of available studies (Duley et al, 2001). Calcium supplementation was briefly in vogue as a preventive treatment in the 1990s, on the basis of the known vasodilatory effect of calcium and impressive results in earlier, small studies (Atallah et al, 2000); however, its worth was not supported in a metaanalysis (Atallah et al, 2000). Similarly, it has been suggested that antioxidants may have a role in preeclampsia prevention, but the only available trial to date showed mixed results, with improvements in biochemical indices in women receiving vitamins C and E, although perinatal outcomes were not different in treated and untreated groups (Chappell et al, 1999). Of concern was the finding that women in whom preeclampsia developed despite vitamin therapy had markedly worsened preeclampsia than controls in whom the disease developed. Therefore, at present, an ideal preventive measure for preeclampsia does not exist. Antepartum Management Given the current inability to predict or prevent preeclampsia, clinicians are left to address established disease and to try to prevent maternal and fetal morbidity. The division of established preeclampsia into mild and more severe forms is of great worth in determining management and minimizing morbidity (see Box 11-1). Mild disease is generally managed conservatively (bed rest and frequent fetal and maternal biophysical assessments) until term is reached or there is evidence of maternal or fetal compromise. The appearance of severe preeclampsia mandates delivery in all but highly selected cases regardless of gestational age. Patients with a diagnosis of mild preeclampsia should be evaluated for signs of maternal or fetal compromise, which would make their disease severe. Evaluation should include a 24-hour urine collection to evaluate for proteinuria; full blood count and platelet measurements; determination of serum uric acid, blood urea nitrogen, and creatinine levels; and evaluation of liver transaminases. Fetal size should be estimated with ultrasonography; the presence of IUGR (less than the tenth percentile) is a sign of severe preeclampsia. Patients with mild disease at 37 weeks’ gestation or more should be delivered, because prolonging pregnancy has no material benefit and increases the risks of maternal and fetal morbidity. Patients at earlier gestational stages should be closely monitored with sequential clinical 109 TABLE 11-4 Drugs for Acute Treatment of Hypertension in Severe Preeclampsia Drug Dosage Hydralazine 1-2 mg test dose 5-10 mg IV followed by 5-10 mg every 20 min as required, to a total of 30 mg Labetalol 10-20 mg IV followed by 20-80 mg every 10 min to a total of 300 mg Nifedipine 10 mg PO every 10-30 min up to three doses IV, Intravenous; PO, by mouth. and laboratory evaluations. Such monitoring often begins in the hospital and may be continued in an outpatient or home setting with appropriate supervision. If the clinical picture deteriorates or term is reached, the baby should be delivered. There is no evidence that antihypertensive therapy influences progression of preeclampsia, and its use may actually be dangerous by masking worsening hypertension. Fetal well-being should be evaluated until delivery by means of kick counts and regular non-stress tests or modified biophysical profiles. Patients with severe disease should be delivered. The only exception to this approach is the diagnosis of severe preeclampsia, in a patient remote from term (<28 weeks’ gestation), on the basis of proteinuria and transiently (unsustained) severe hypertension alone. Such patients may be managed conservatively under close supervision while antenatal corticosteroids are administered without adversely affecting maternal or fetal outcome (Sibai et al, 1990). There is no reason for conservative management in any other circumstance. The patient with severe preeclampsia at less than 24 weeks’ gestation should be offered termination of the pregnancy; all others should be delivered by the most expedient means. Cesarean section should be reserved for obstetric indications. Severe hypertension requires treatment with fast-acting antihypertensive agents if stroke and placental abruption are to be avoided. Intravenous hydralazine is well established as a first-line drug for this purpose, although there is a growing experience with other agents, including intravenous labetalol and oral nifedipine (Duley and Henderson-Smart, 2000a) (Table 11-4). The aim of treatment is to lower blood pressure into the mild preeclampsia range (140/90 mm Hg) to reduce the risk of stroke and other maternal cardiovascular complications. There is evidence to support the use of parenteral magnesium sulfate to prevent eclampsia in all cases of severe disease (Duley et al, 2003). Severe preeclampsia can manifest as classic disease with severe proteinuric hypertension, or it can cause atypical findings such as pulmonary edema or severe central nervous system symptoms, including blindness. More commonly, patients show evidence of microangiopathy leading to the hemolysis, elevated liver enzymes, and low platelets (HELLP) syndrome. The full-blown clinical syndrome of HELLP carries a significant maternal risk. Earlier reports suggested that the disease carries a grave prognosis (Weinstein, 1982). This suggestion remains true for florid clinical cases, but most patients now have “laboratory” HELLP and never experience major clinical features 110 PART III Maternal Health Affecting Neonatal Outcome of the syndrome because delivery is initiated before their condition deteriorates to that point. Preeclampsia and Fetal Risk Because the only recourse in severe preeclampsia is delivery, the disease has a corresponding effect on prematurity and its attendant complications. IUGR is not uncommon in severe preeclampsia, and there may be evidence of progressive deterioration in fetal well-being with worsening disease. Infants delivered at less than 34 weeks’ gestation will benefit from antenatal steroid therapy—even as little as 8 hours of therapy before delivery may have benefit. Many patients are able to deliver vaginally, but fetal compromise may preclude aggressive induction and mandate delivery by cesarean section. The incidence of respiratory distress syndrome is lower in infants of mothers with preeclampsia who are delivered preterm than in those of age-matched controls without antenatal steroid exposure (Yoon et al, 1980). Nonetheless, the morbidity of such infants is greater because of hypoxemic insults received in utero. Infants born to mothers with preeclampsia may also have thrombocytopenia or neutropenia, which further complicates their newborn course (Fraser and Tudehope, 1996). Intrapartum Management All women in labor with a diagnosis of preeclampsia should receive magnesium sulfate as seizure prophylaxis (Box 11-2). Although the absolute risk of seizure is low (1 in 2000 to 3000), the occurrence of seizures is unpredictable, and the efficacy of magnesium sulfate and margin of safety has been validated in multiple randomized trials (Duley et al, 2003). The mechanism of action of MgSO4 in the prevention of seizures is still unresolved, with various theories being advanced, including peripheral neuromuscular blockade, membrane stabilization, N-methyl D-aspartate (NMDA) receptor blocking activity, cerebral vasodilation, and calcium channel blocking action (Belfort et al, 2006). Blood pressure should be maintained in the mild preeclampsia range using intravenous antihypertensive agents (labetalol, hydralazine). Epidural anesthesia is indicated for pain control and to aid in blood pressure management. Vaginal delivery should be possible in most cases. Delivery by cesarean section should be reserved for obstetric indications. Careful attention to fluid balance should be maintained. After delivery, the preeclamptic process should begin to resolve rapidly. ECLAMPSIA Eclampsia is the occurrence of generalized tonic-clonic seizures in association with preeclampsia. It affects approximately 1 in 2500 deliveries in the United States and may be much more common in developing countries, affecting as many as 1% of parturients. Up to 10% of maternal deaths are due to eclampsia (Duley, 1992). BOX 11-2 M  agnesium Sulfate Therapy for Prevention of Eclampsia ll ll ll ll Bolus 4-6 g IV over 20 min Continuous infusion 1-2 g/hr Follow up levels every 6-8 hours to target 4-6 mEq/L Continue infusion 24 hours after delivery or 24 hours after seizure if seizure occurs despite magnesium therapy Most cases of eclampsia occur within 24 hours of delivery. Almost 50% of seizures occur before the patient’s admission to the labor and delivery department, approximately 30% are intrapartum, and the remainder are postpartum. There is a considerable drop in the risk of eclampsia by 48 hours postpartum, with seizures occurring in less than 3% of women beyond that time. Most patients have antecedent features that are suggestive of preeclampsia, although in some cases eclampsia may occur without warning. If eclampsia is left untreated, repetitive seizures become more frequent and of longer duration, and ultimately status eclampticus develops. Maternal and fetal mortality may be as high as 50% in severe cases, especially if the seizures occur while the patient is far from medical care. Randomized controlled trials have demonstrated the clear superiority of magnesium sulfate for the treatment of eclampsia over all other anticonvulsants (Duley and Gulmezoglu, 2002; Duley and Henderson-Smart, 2002b, 2002c). Intravenous magnesium sulfate is given as a 4-g bolus over 5 minutes followed by a maintenance infusion of 1 to 2 g/hr for 24 hours after delivery. Subsequent seizures can be treated with further bolus injections. In refractory cases, second-line treatment with other anticonvulsants may be required, or the patient may have to be paralyzed and their lungs ventilated. Delivery after an eclamptic seizure should take place in a controlled, careful manner. There is little to be added by performing an emergency cesarean section (Coppage and Polzin, 2002). The patient’s condition should be stabilized first. Vaginal delivery is possible in most cases, although cesarean delivery may be indicated if the status of the cervix is unfavorable or if fetal compromise is ongoing despite control of seizures and maternal stabilization. Infants born to mothers after eclampsia require careful observation after birth. SUGGESTED READINGS ACOG Practice Bulletin: Chronic hypertension in pregnancy, ACOG Committee on Practice Bulletins, Obstet Gynecol 98(Suppl l):177-185, 2001. Chappell LC, Enye S, Seed P, et al: Adverse perinatal outcomes and risk factors for preeclampsia in women with chronic hypertension: a prospective study, Hypertension 51:1002-1009, 2008. Myers JE, Baker PN: Hypertensive diseases and eclampsia, Curr Opin Obstet Gynecol 14:119-125, 2002. Wang A, Rana S, Karumanchi SA: Preeclampsia: the role of angiogenic factors in its pathogenesis, Physiology 24:147-158, 2009. Complete references used in this text can be found online at www.expertconsult.com C H A P T E R 12 Perinatal Substance Abuse Linda D. Wallen and Christine A. Gleason Substance abuse during pregnancy has been recognized as a problem for more than a century. Psychotropic substances, both legal (alcohol, cigarettes, and prescription drugs such as opioids and benzodiazepines) and illegal (opioids, amphetamines, cocaine, and marijuana), can cause obstetric, fetal, and neonatal complications. These complications include poor intrauterine growth, prematurity, abruptio placenta, fetal distress, spontaneous abortion, stillbirth, fetal (and maternal) cerebral infarctions and other vascular accidents, malformations, and neonatal neurobehavioral dysfunction. Although substance abuse occurs in all socioeconomic classes, illegal drug abuse is more frequently associated with unhealthy lifestyles, poor access to prenatal care, untreated health problems, poverty, stress, and psychological disorders. Because of these socioeconomic confounders as well as the confounders of polysubstance exposure and the influence of various postnatal environmental factors, it is often difficult to determine the effects of maternal use of one specific drug on the fetus and newborn. This chapter addresses the epidemiology of perinatal substance use and abuse; the effects of specific drugs on the fetus and newborn; maternal issues and their effects on the newborn; identification of pregnancies and babies at risk; neonatal management; and long-term effects and follow-up. The discussion will focus on abused substances that are known or suggested to be associated with significant perinatal and neonatal morbidity: alcohol, tobacco, nicotine, opioids, cocaine, marijuana, and methamphetamine. EPIDEMIOLOGY OF PERINATAL SUBSTANCE EXPOSURE PREVALENCE Prevalence rates for perinatal substance exposure have been determined by using a number of different definitions, survey methods, and drug use detection procedures (Lester et al, 2004). One of the most comprehensive geographically based prevalence studies on substance use and abuse by pregnant women was undertaken in California in the early 1990s by the Perinatal Substance Exposure Study Group (Vega et al, 1993). In that study, urine was collected at the time of delivery from more than 30,000 pregnant women. The prevalence rates for illicit drug use were 3.5% (overall), 8.8% (tobacco), and 7.2% (alcohol). The authors concluded that if these results could be extrapolated to the United States at large, an estimated 450,000 infants per year (11% of 4 million live births) would be exposed to alcohol, illicit drugs, or both in the days before delivery. Rates of perinatal substance exposure have not changed substantially over the past 20 years, although there is wide geographic variation. The U.S. Department of Health and Human Services Pregnancy Risk Assessment Monitoring System is designed to monitor maternal behaviors and experiences among women who deliver live-born infants. Data collected during 2000 to 2003 from 19 states revealed that during the last 3 months of pregnancy, tobacco use ranged from 4.9% to 27.5%, and alcohol use ranged from 2% to 8.7%. Data from a 2005-2006 National Survey on Drug Use and Health revealed that of pregnant women aged 15 to 44, 4% reported using illicit drugs, 16.5% reported using tobacco, and 12% reported using alcohol sometime within the past month (Substance Abuse and Mental Health Services Administration, 2007). These rates are significantly lower than the rates reported by women aged 15 to 44 years who were not pregnant (10% reported using illicit drugs, 29.5% reported using tobacco, and 53% reported using alcohol within the past month), confirming that the prevalence of substance use among pregnant women remains less than among nonpregnant women. The Maternal Lifestyle Study was developed in the early 1990s by the National Institute of Child Health and Human Development (NICHD) and National Institute on Drug Abuse (NIDA) to follow pregnant women known to be using opioids or cocaine and to follow their offspring. The study followed 11,800 pregnant women who received prenatal care during 1993 to 1995 at four teaching hospital sites. Women were followed from initial presentation throughout their offspring’s childhood, with a planned evaluation at 8 to 11 years. Meconium analyses were used to confirm maternal substance use; there was 66% agreement between meconium analyses and positive maternal reports. The prevalence of cocaine and opioid exposure was 10.7%; 98% of cocaine users also used other drugs; and only 2% of women used cocaine alone (Shankaran et al, 2007). EPIDEMIOLOGY OF SPECIFIC SUBSTANCES Nicotine—in the form of cigarettes, smokeless tobacco, and nicotine replacement patches—remains the substance used most often during pregnancy. Although cigarette smoking in the United States has decreased significantly over the last 20 years, 26% of reproductive-aged women still smoke, and 15% to 20% of women smoke during their pregnancies (Andres and Dar, 2000). Women who smoke during pregnancy are more likely to use opioids, alcohol, cocaine, amphetamines, and marijuana during pregnancy than women who do not smoke (Vega et al, 1993). Cigarette smoking has been associated with numerous perinatal complications, often in a dose-dependent fashion. Smoking has been shown to raise the risk of spontaneous abortion, stillbirth, fetal growth retardation, prematurity, perinatal mortality, and sudden infant death syndrome (Andres and Day, 2000; Kallen, 2001; Lambers and Clark, 1996; Tuthill et al, 1999). Cigarette smoking represents 111 112 PART III Maternal Health Affecting Neonatal Outcome the most influential and most common factor adversely affecting perinatal outcomes. In a 2008, Rogers “estimated that if all the women in the United States stopped smoking, there would be an 11% reduction in stillbirths and 5% reduction in neonatal deaths.” Before 1970, the detrimental effects of alcohol abuse during pregnancy were believed to be related only to drunkenness, such as an increased risk for accidents. There was a widely held belief that the placenta formed a protective barrier between alcohol and the fetus. This belief repudiated by studies in the United States (Jones and Smith, 1973) and France (Lemoine et al, 1967) describing fetal alcohol syndrome. These studies led to the 1989 U.S. federal law requiring that warning labels be placed on all alcoholic beverage containers regarding alcohol-related birth defects. Despite this extensive public health campaign designed to inform and warn women about the dangers of alcohol consumption during pregnancy, approximately 10% of pregnant women still report using alcohol (Centers for Disease Control and Prevention, 2009). A smaller percentage of these women are alcoholics, but their infants are at significantly higher risk for fetal alcohol syndrome or alcohol-related neurobehavioral disorders compared with the infants of nonalcoholic pregnant women who use alcohol. Opium derivatives have been used as analgesics for centuries and remain the most effective analgesics available. Opioids of clinical interest are morphine, heroin, methadone, meperidine, oxycodone, and codeine. Perinatal problems associated with opium were first reported in the late 1800s. Since the 1950s, heroin use, particularly among women, has been endemic in most major American cities. Compared with cocaine, marijuana, alcohol, and tobacco abuse, opioid addiction during pregnancy is rare (Shankaran et al, 2007). The prevalence of opioid use among pregnant women is reported to range from 1% to 2% (Vega et al, 1993; Yawn et al, 1994) to as much as 21% in a highly selected group of women (Behnke and Eyler, 1993; Nair et al, 1994; Ostrea et al, 1992b). One multicenter study found that the prevalence of opioid use varied by center and ranged from 1.6% to 4.5% at the different sites (Lester et al, 2001). In addition, these centers reported higher rates of opioid use by mothers of low-birthweight and very low-birthweight infants. Rates for heroin use are higher in metropolitan areas and cities and are more concentrated in northeastern and west coast cities. Opioid abuse is more common in groups of lower socioeconomic status, and women using opioids during pregnancy are more likely to use other drugs (Bauer, 1999; Brown et al, 1998; van Baar and de Graaff, 1994). Investigators have also reported that 93% of women identified as using opioids and cocaine during pregnancy had also used a combination of alcohol, nicotine, or marijuana (Bauer, 1999). Of the opioid drugs known to be abused during pregnancy, heroin and methadone have been studied the most extensively. Heroin can be ingested through smoking or by the intranasal or intravenous route. Reports from European countries suggest a trend away from intravenous injection of opioids (Hartnoll, 1994). The use of noninjectable heroin may reduce the risk of transmission of human immunodeficiency virus (HIV); however, its wider use ensures the emergence of new groups of heroin users for whom the risk of intravenous use is a major deterrent. The euphoria-producing effect of cocaine was exploited extensively in the United States in the late nineteenth and early twentieth centuries, when the agent was an active ingredient in a number of widely used over-the-counter elixirs and tonics. Cocaine use markedly decreased after the Harrison Narcotic Act of 1914 and the supervening Comprehensive Drug Abuse Prevention and Control Act of 1970, which classified cocaine as a schedule II drug (i.e., one of “high abuse potential with restricted medical use,” similar to opioids, barbiturates, and amphetamines). Cocaine’s reputation as a glamour drug, the widely held misconception that cocaine is not addictive, and the development and marketing of crack, a cheap version of cocaine, were major factors in the resurgence of drug use. Growing concern regarding the effects of maternal cocaine use on pregnancy outcomes was one of the reasons that the U.S. Congress passed the 1986 Narcotics Penalties and Enforcement Act, which imposed severe penalties on any person convicted of either possessing or distributing effects; however, this law did not appreciably alter the fact that cocaine and other stimulants had become the drugs of choice for women in the United States. In the 1990s, studies based on urine toxicology screening reported a prevalence of cocaine use among pregnant women of 5% in New York City, 1.1% in a geographic sample in California, and less than 0.5% in private hospitals in Denver, Colorado (Burke and Roth, 1993). Prevalence increases to 18% when both self-reporting and urine testing are used, and the highest prevalence rates are reported from studies using meconium testing. Amphetamines have surpassed cocaine as the primary illicit drugs used by pregnant women in many areas of California and other states. Methamphetamine (or “crystal”) has been the primary form abused, because it can be produced locally and fairly cheaply. Greater restrictions on the importing of cocaine have also contributed to resurgence in amphetamine use. Amphetamines have always been popular among adolescents, especially females, and accordingly women of child-bearing age are at high risk for perinatal abuse. A California study of drug-exposed infants in the social welfare system documented a higher prevalence of amphetamine use among white pregnant women than in women of other ethnicities (SagatunEdwards et al, 1995). The Infant Development, Environment, and Lifestyle (IDEAL) study chose four major U.S. cities with known methamphetamine abusing populations and reported methamphetamine use in 5.2% of pregnant women. As in previous reports, these women frequently used other substances: 25% used tobacco, 22.8% used alcohol, 6% used marijuana, and 1.3% used barbiturates (Arria et al, 2006). According to the 2003 National Survey on Drug Use and Health, marijuana is the most widely used illegal drug used in the United States, with approximately 14.6 million people reporting that they use the drug. In a 1986 study, the most frequent age of marijuana use for women included the childbearing ages, with 50% of 18- to 35-year-old women reporting that they used marijuana at least once and 8% reporting that they used marijuana a minimum of 10 of the past 30 days (Clayton et al, 113 CHAPTER 12 Perinatal Substance Abuse 1986). The 2002-2006 National Household Survey on Drug Abuse included information on drug use in the last 30 days from over 94,000 women aged 18 to 44 years, of whom 5017 were pregnant. Marijuana use was reported in 7.3% of nonpregnant women and 2.8% of pregnant women with moderate to heavy use (more than five times per month) in 1.8% of pregnant women compared with 3.7% in nonpregnant women. Marijuana use declined over the course of pregnancy, from 4.5% in the first trimester to 1.5% in the third trimester (Muhuri and Gfroerer, 2009). HEALTH POLICY In the 1980s, with rising cocaine use and the emergence of crack cocaine, national attention turned to drug use during pregnancy, with a general public outcry being the result. Children born to crack addicts were widely believed to be irrevocably damaged and public opinion was that mothers should be punished. As the foster care system became overwhelmed and as evidence to the contrary has emerged, public opinion regarding prenatal drug use has shifted more toward maternal treatment and prevention rather than punishment. As Lester et al (2004) stated, the initial overreaction of the public “in which drug-exposed children were characterized as irrevocably and irreversibly damaged” has shifted “to a perhaps equally premature excessive ‘sigh of relief’ that drugs such as cocaine do not have lasting effects, especially if children are raised in appropriate environments.” This statement has led to a change in the public discourse regarding health policy interventions for substance use and abuse during pregnancy. The focus of current policy is to provide appropriate medical care for substance-using pregnant women, including medical management of their chemical dependency and programs to decrease substance use during pregnancy. However, there remains the important step of recognizing that “the idea that illegal drugs are more harmful to the unborn fetus than legal drugs is incorrect” (­Thompson et al, 2009). Future research and intervention need to include programs to educate women of childbearing age of the significant effects of both legal and illegal substance use on the early-gestation fetus, starting even before pregnancy may be recognized. PERINATAL EFFECTS OF SPECIFIC DRUGS (Table 12-1) ETHANOL (ALCOHOL) Pharmacology and Biologic Actions Alcohol is a mood-altering substance that enhances the effects of the inhibitory neurotransmitter gamma-­ aminobutyric acid and lessens the effect of the excitatory neurotransmitter glutamate, thus acting as a central nervous system (CNS) depressant or sedative. Alcoholic beverages come in many forms, and for centuries they have been consumed for diverse reasons: celebrations, relaxation, religious ceremonies, and medicinal purposes (alcohol is an excellent sedative and tocolytic agent). The alcohol contained in alcoholic beverages is ethanol. Ethanol is absorbed in the digestive tract and into the body fat and bloodstream. Ethanol is metabolized to acetaldehyde by alcohol dehydrogenase (ADH), primarily in the liver. ADH is then metabolized to acetate by aldehyde dehydrogenase (ALDH) and eventually eliminated as water and CO2. Although acetaldehyde is short-lived, it can cause significant tissue damage, which is particularly evident in the liver, where most alcohol metabolism takes place. Pregnant women have slower rates of alcohol clearance, likely related to hormonal alterations in the activity of the alcohol-metabolizing enzymes; this leads to slower clearance of alcohol compared with nonpregnant women consuming the same amount of alcohol (Shankaran et al, 2007). Complications of Pregnancy Heavy drinking carries a higher risk of cardiovascular and hepatic complications in women compared with men, and the alcohol-associated mortality rate is also considerably higher (Smith and Weisner, 2000). These factors alone can clearly complicate a woman’s pregnancy. In addition, nutritional deficiencies and poor diet can affect general health and dentition, which can negatively affect a pregnancy. Alcoholism is a chronic disease that is often progressive and can be fatal. Pregnant alcoholics often have related medical disorders such as cirrhosis, pancreatitis, and alcohol-related neurologic problems. These disorders can affect the health and well-being of their fetus. TABLE 12-1 Enhanced Risk for Various Events or Processes after Substance Use during Pregnancy* Event or Process Ethanol Cigarettes Marijuana Opiates Cocaine Malformations + – – – – Amphetamines – Abortion/stillbirth + + ? + + + Intrauterine growth restriction + + – + + + Prematurity – + ? + + + Withdrawal ? + +/– + – – Central nervous system sequelae + ? + ? + ? Sudden infant death syndrome + + ? + ? ? Foster care + – – + + + +, Causes event or process; –, does not cause event or process; ?, not known whether agent causes event or process. *Although risk is increased, the risk ratio ranges for many from 1 to 2 for these associations. 114 PART III Maternal Health Affecting Neonatal Outcome Alcohol affects prostaglandin levels, increasing levels of its precursors in human placental tissue and thus affecting fetal development and parturition. In fact, researchers have used this knowledge to test the effect of aspirin, which inhibits alcohol-induced increases in prostaglandin levels, on reducing alcohol-induced fetal malformations in a mouse model (Randall, 2001). Specific obstetric complications of heavy drinking may relate to alterations in prostaglandin levels, including an increased risk for spontaneous abortion, abruptio placenta, and alcohol-related birth defects such as fetal alcohol syndrome. Discriminating Features short palpebral fissures Associated Features epicanthal folds low nasal bridge flat midface Fetal Alcohol Syndrome Fetal alcohol syndrome (FAS) was first described by Lemoine (1967), a Belgian pediatrician who observed a common pattern of birth anomalies in children born to alcoholic mothers in France. This description was followed by a landmark article by Jones and Smith (1973) reporting similar features in several children born to alcoholic mothers in the United States. It is unclear how much alcohol exposure is necessary to cause fetal teratogenicity, and even high consumption levels do not always result in the birth of a child with FAS (Abel and Hannigan, 1995). However, a woman with a previous affected child is at increased risk for having a child with FAS if she consumes alcohol during a subsequent pregnancy. The adverse effects of alcohol on the fetus are related to gestational age at exposure, the amount of alcohol consumed, and the pattern of consumption (e.g., binge drinking), maternal peak blood alcohol concentrations, maternal alcohol metabolism, and the individual susceptibility of the fetus. Studies show that maternal peak blood alcohol levels are affected by maternal nutrition, age, body size, and genetic disposition (Eckardt et al, 1998; Maier and West, 2001). In addition, various risk factors increase susceptibility to FAS, including advanced maternal age and confounding factors such as nonwhite race, poverty, and socioeconomic status (Abel, 1995; Bagheri et al, 1998; May and Gossage, 2001). In the United States, the incidence of FAS is tenfold higher for African Americans living in poverty than for white middle-class women (Abel, 1995). Despite the differences in incidence of FAS worldwide, reports consistently indicate poverty or socioeconomic status as major determinants of FAS (Abel, 1995; May et al, 2000). Features of FAS include characteristic facial dysmorphology (short palpebral fissures, midface hypoplasia, broad flat nasal bridge, flat philtrum, and thin upper lip; Figure 12-1), prenatal and postnatal growth deficiency, and variable CNS abnormalities. Skeletal anomalies, abnormal hand creases, and ophthalmologic, renal, and cardiac anomalies have been described in children with FAS, but less frequently than the facial dysmorphology and CNS abnormalities that include structural brain defects (e.g., dysgenesis of the corpus callosum and cerebellar hypoplasia), cognitive abnormalities, delayed brain development, and signs of neurologic impairment, including lifelong behavioral and psychosocial dysfunction. In 1996, the Institute of Medicine further defined the criteria for the diagnosis of FAS and proposed a new term— alcohol-related neurodevelopmental disorder (ARND). short nose minor ear anomalies indistinct philtrum micrognathia thin upper lip In the Young Child FIGURE 12-1 Facies in fetal alcohol syndrome. (From Streissguth AP, Little RE: Alcohol, pregnancy, and the fetal alcohol syndrome, ed 2, unit 5 of Alcohol use and its medical consequences: a comprehensive slide teaching program for biomedical education. Developed by Project Cash of the Dartmouth Medical School. Reproduced with permission from MilnerFenwick, Inc., Timonium, Michigan, 1994.) This term includes structural CNS and cognitive abnormalities in children with confirmed fetal exposure to alcohol. Unlike FAS, a diagnosis of ARND does not require the presence of facial or other physical abnormalities. In 2000, the American Academy of Pediatrics Committee on Substance Abuse published these new definitions with an explanatory drawing (Figure 12-2). The incidence of FAS in the United States has been estimated to vary from 1.95 to 5 cases per 1000 live births (Abel, 1995; American Academy of Pediatrics Committee on Substance Abuse and Committee on Children with Disabilities, 2000; Bertrand et al, 2005; Sampson et al, 1997). FAS is recognized more frequently in the United States than in other countries and is most common (4.3%) among women who report heavy drinking (Abel, 1995). Accurate incidence and prevalence rates of FAS are difficult to obtain because of wide variations in methodologies used for estimation of rates, and because the clinical diagnosis is often missed in the neonatal period. In fact, most cases (up to 89%) are not diagnosed until after a child is age 6 (Centers for Disease Control and Prevention [CDC], 1997). FAS is diagnosed from the history and physical findings. No laboratory tests are available for clinical use to quantify the extent of alcohol exposure during fetal life. There are also no clinical methods for validating maternal self-reporting of alcohol use, quantifying the level of fetal exposure, or predicting future disability after fetal exposure (Jones and Chambers, 1999). Koren et al (2002) have proposed meconium fatty acid ethyl ester levels as a potential biologic marker for fetal alcohol exposure. Whether this finding is shown to correlate with childhood outcomes remains to be studied. Investigators have shown that pediatricians fail to recognize FAS in the newborn and do not always inquire about alcohol exposure during pregnancy 115 CHAPTER 12 Perinatal Substance Abuse FAS with confirmed maternal exposure FAS without confirmed maternal exposure Partial FAS with confirmed exposure OR OR OR Alcohol-related birth defects (ARBD)† Alcohol-related neurodevelopmental disorder (ARND)† A B C D E F Confirmed Exposure to Alcohol Facial Anomalies Growth Retardation CNS Abnormalities Cognitive Abnormalities Birth Defects *Adapted from Fetal Alcohol Syndrome; Diagnosis, Epidemiology, Prevention, and Treatment. 1996;4–5. Letter designations in the figure indicate the following: A. Confirmed maternal alcohol exposure indicates a pattern of excessive intake characterized by substantial, regular intake or heavy episodic drinking. Evidence of this pattern may include frequent episodes of intoxication, development of tolerance or withdrawal, social problems related to drinking, legal problems related to drinking, engaging in physically hazardous beavior while drinking, or alcohol-related medical problems such as hepatic disease. B. Evidence of a characteristic pattern of facial anomalies that includes features such as short palpebral fissures and abnormalities in the premaxillary zone (e.g., flat upper lip, flattened philtrum, and flat midface). C. Evidence of growth retardation, including at least one of the following: • low birthweight for gestational age • decelerating weight over time not caused by nutrition • disproportional low weight to height D. Evidence of CNS neurodevelopmental abnormalities, including at least one of the following: • decreased cranial size at birth • structural brain abnormalities (e.g., microcephaly, partial or complete agenesis of the corpus callosum, cerebellar hypoplasia) • neurological hard or soft signs (as age appropriate), such as impaired fine motor skills, neurosensory hearing loss, poor tandem gait, poor eye-hand coordination E. Evidence of a complex pattern of behavior or cognitive abnormalities that are inconsistent with developmental level and cannot be explained by familial background or environment alone, such as learning difficulties; deficits in school performance; poor impulse control; problems in social perception; deficits in higher level receptive and expressive language; poor capacity for abstraction or metacognition; specific deficits in mathematical skills; or problems in memory, attention, or judgment F. Birth defects associated with alcohol exposure include: Cardiac Atrial septal defects Ventricular septal defects Aberrant great vessels Tetralogy of Fallot Skeletal Hypoplastic nails Shortened fifth digits Radioulnar synostosis Flexion contractures Camptodactyly Clinodactyly Pectus excavatum and carinatum Klippel-Feil syndrome Hemivertebrae Scoliosis Renal Aplastic, dysplastic, hypoplastic kidneys Horseshoe kidneys Ureteral duplications Ocular Strabismus Retinal vascular anomalies Refractive problems Secondary to small globes Auditory Conductive hearing loss Neurosensory hearing loss Other Virtually every malformation has been described in some patient with FAS. The etiologic specificity of most of these anomalies to alcohol teratogenesis remains uncertain. Hydronephrosis †Alcohol-related effects indicate clinical conditions in which there is a history of maternal alcohol exposure, and where clinical or animal research has linked maternal alcohol ingestion to an observed outcome. There are two categories, alcohol-related neurodevelopmental disorder and alcohol-related birth defects, which may co-occur. If both diagnoses are present, then both diagnoses should be rendered. FIGURE 12-2 Diagnostic classification of fetal alcohol syndrome and alcohol-related effects. (From the American Academy of Pediatrics Committee on Substance Abuse and Committee on Children with Disabilities: Fetal alcohol syndrome and alcohol-related neurodevelopmental disorders, Pediatrics 106:359. Reproduced with permission from the American Academy of Pediatrics, 2000.) 116 PART III Maternal Health Affecting Neonatal Outcome (Stoler and Holmes, 1999). One promising screening tool is the use of averaged cranial ultrasound images to examine the size and shape of the corpus callosum, which is typically dysgenic in FAS (Bookstein et al, 2005). Guidelines to aid in the earlier recognition and referral of infants and children with FAS and fetal alcohol spectrum disorder have been published recently (Bertrand et al, 2005; Hoyme et al, 2005). FAS is not a problem just for neonatologists. Adolescents who were exposed prenatally to alcohol have a different approach to alcohol than their nonexposed peers, with an increased risk for earlier use and subsequent alcohol abuse (Baer et al, 2003). One study in rodents suggests that fetal ethanol exposure increases ethanol intake later in life by making it smell and taste better (Youngentob and Glendinning, 2009). Children with FAS continue to have serious disabilities into adulthood (Streissguth et al, 1991; Streissguth, 1993). Although the facial features and growth restriction are no longer as distinctive as during childhood, mental retardation continues to have a significant effect. Adults with FAS have behavior, socialization, and communication dysfunction, and on average they function at the second- or third-grade level. A significant number of FAS patients do not achieve fully independent living. Earlier recognition and intervention for children with FAS and its variants may help to minimize eventual adulthood disabilities and help to prepare adolescents and young adults with the disorder for independent living (Bertrand et al, 2005). Fetal Growth Intrauterine growth restriction (IUGR) is one of the most consistent findings of prenatal exposure to alcohol (Hannigan and Armant, 2000). Growth deficit begins in utero and continues throughout childhood (American Academy of Pediatrics Committee on Substance Abuse and Committee on Children with Disabilities, 2000). The facial features and the growth restriction become less noticeable during adolescence and puberty (Streissguth et al, 1991; Streissguth, 1993). CIGARETTE SMOKING AND NICOTINE Pharmacology and Biologic Actions Cigarette smoke contains a complex mixture of approximately 4000 compounds, including nicotine and carbon monoxide, which can adversely affect the fetus (Lester et al, 2004). In rodents, nicotine releases chemicals in the reward center of the brain, which likely triggers the euphoria that smokers experience. Nicotine activates nicotinic acetylcholine receptors, and these receptors remain depressed for a longer time after their activation stops, which likely accounts for compulsive smoking (Cohen, 2007). Nicotine crosses the placenta and concentrates in fetal blood and amniotic fluid, where its levels significantly exceed maternal blood concentrations (Haustein, 1999). The serum concentration of cotinine, the primary metabolite of nicotine, is used to quantitate the level of smoking and fetal exposure. Cotinine has a half-life of 15 to 20 hours, and because its serum levels are tenfold higher than those of nicotine, this substance may represent a ­better marker for intrauterine exposure (Lambers and Clark, 1996). Complications of Pregnancy Although the exact mechanism of the adverse effects of smoking on pregnancy is unknown, cigarettes contain numerous potentially toxic compounds that affect fetal health in a number of ways. Nicotine and its metabolites can act as vasoconstrictors, and a study in pregnant rhesus monkeys demonstrated a nicotine-associated decrease in uterine blood flow (Suzuki et al, 1980), which might provide a partial explanation for the association between maternal cigarette smoking and low birthweight. Theories regarding mechanisms for the adverse effects of smoking on fetal health include direct vasoconstrictive effects of nicotine on uteroplacental blood flow, the induction of fetal hypoxia from carbon monoxide production, direct toxic effects and indirect effects of altered maternal nutritional intake, and altered maternal and placental metabolism (Andres and Day, 2000; Pastrakuljic et al, 1999). When pregnant women smoke cigarettes, the resulting increased levels of carbon monoxide cross the placenta and form carboxyhemoglobin in the fetus, with resulting hypoxemia (Lambers and Clark, 1996). Supporting this theory, serum erythropoietin levels are higher in tobacco smoke–exposed infants at delivery, a finding that is presumed to reflect fetal hypoxia (Beratis et al, 1999; Jazayeri et al, 1998). In addition to the fetal hypoxia theory, there have recently been studies demonstrating that nicotine may act as a developmental neurotoxin targeting nicotinic acetylcholine receptors (Lester et al, 2004; Levin and Slotkin, 1998) and may disturb protein metabolism during gestation, leading to decreased serum amino acids in umbilical cord blood (Jauniaux et al, 2001). Maternal smoking has also been show to affect the length of gestation in a dose-dependent manner, with a higher risk of preterm delivery (Jaakkola et al, 2001; Savitz et al, 2001) and a twofold increase in the incidence of placental abruption (Ananth et al, 1996). Perinatal mortality is increased in pregnant smokers, likely reflecting the increases in rates of prematurity, placental abruption, and placenta previa in women who smoke. Mothers who smoke during pregnancy commonly continue to smoke during their infants’ childhood. Asthma and recurrent otitis media are more common in infants who are exposed to passive smoking (Ey et al, 1995; Martinez et al, 1995). Fetal Growth The effect of smoking on fetal growth is significant and dose dependent (Kyrklund-Blomberg et al, 1998; Nordentoft et al, 1996). Studies have shown lower birthweights associated with levels of nicotine exposure, with a 1-g reduction in birthweight observed for every microgram per milliliter increase in maternal serum cotinine level (­Eskenazi et al, 1995; Perkins et al, 1997). Investigators have shown a dose-dependent relationship between the amount of smoking and the extent of fetal growth restriction and birthweight reduction (Horta et al, 1997; ­Jaakkola et al, 2001; Savitz et al, 2001; Sprauve et al, 1999). CHAPTER 12 Perinatal Substance Abuse Both reducing and ceasing cigarette smoking during pregnancy have been shown to be beneficial and to lead to improved fetal growth. Lieberman et al (1994) have shown that if pregnant women stop smoking during the third trimester, their infants’ weights are indistinguishable from those of a nonsmoking population. Other investigators report that even a modest reduction in smoking is associated with improved fetal growth (Li et al, 1993; Walsh et al, 2001). 117 Marijuana, the dried material from the hemp plant ­Cannabis sativa, is most often smoked but can be ingested with food. Its most active ingredient is delta-9-tetrahydrocannabinol (THC), which binds to cannabinoid receptors in the brain and modifies the release of several neurotransmitters. Its primary biologic effects include euphoria; relaxation; increased heart rate, blood pressure, and appetite; and impaired coordination, decision-making, short-term memory, concentration, and learning. Although controversial, there is increasing evidence that regular marijuana use can cause respiratory difficulties, cognitive impairments, withdrawal, and dependence (Khalsa et al, 2002). THC crosses the placenta and collects in the amniotic fluid. occurring endorphins and enkephalins (Vaccarino and Kastin, 2000). As modulators of the sympathoadrenal system, endogenous opioids are important during periods of diverse forms of stress. Activation of these receptors by the endogenous opioids has physiologic effects, including analgesia, drowsiness, respiratory depression, decreased gastrointestinal motility, nausea, vomiting, and alterations in the endocrine and autonomic nervous systems. Activation of these same endogenous opioid receptors by exogenous opioid drugs has similar clinical effects, producing euphoria, sleepiness, and decreased sensitivity to pain, as well as adverse effects such as constipation and nephrotic syndrome (ACOG, 2005). The use of opioid drugs can result in the development of tolerance, physiologic dependence, and addiction. Tolerance leads to a shortened duration of the action of opioids and a decrease in the intensity of the drug action, followed by the need for a higher dose to obtain the same clinical effect. Tolerance is believed to result from continued occupancy of the opioid receptor. Continuous administration of opioids, therefore, leads to the more rapid onset of tolerance (Anand and Arnold, 1994; Suresh and Anand, 2001). With physiologic dependence, there is a need for further drug administration to prevent withdrawal symptoms (agitation, dysphoria, temperature instability). Addiction is a more severe form of dependence that involves a complex pattern of drug-seeking behavior (Christensen, 2008). Complications of Pregnancy Complications of Pregnancy Marijuana use during pregnancy is inconsistently reported to have effects on birth outcomes. The inconsistency likely stems from the fact that marijuana is often used in combination with other substances, thus potentiating the risks for prematurity and low birthweight. Day et al (1991) reported a higher incidence of meconium-stained amniotic fluid and other pregnancy and delivery complications after maternal marijuana use, but subsequent studies have not replicated this finding. The IDEAL study did not find any relation between marijuana use and altered fetal growth parameters (Smith et al, 2006). Frequent use of marijuana was associated with a small decrement in birthweight in a study that did not control for other illicit drug use (Schempf, 2007). Obstetric complications associated with maternal use of opioids include a higher incidence of spontaneous abortion, premature delivery, preterm labor, abruptio placentae, chorioamnionitis, impaired fetal growth, and fetal distress. In women who use opioids during pregnancy, the incidence of preterm labor and premature delivery ranges from 25% to 41% (Chiriboga, 1993; Lam et al, 1992; Little et al, 1990). Maternal opioid use is also associated with higher rates of meconium-stained amniotic fluid, lower American Pediatric Gross Assessment Record (Apgar) scores, longer duration of membrane rupture (Gillogley et al, 1990), and increased incidence of syphilis and HIV infection at birth (Bauer, 1999). The rates of maternal complications of pregnancy are further increased when drug use is added to infection with HIV. Women infected with HIV who also use opioids, particularly methadone, have higher rates of miscarriage, preterm deliveries, and small-for-gestational-age infants, and more vaginal and urinary tract infections than women with HIV who do not use drugs (Mauri et al, 1995). The etiology of these opioid-associated pregnancy complications is multifactorial. Maternal lifestyle, malnutrition, infections, and polydrug effects are likely to result in poor perinatal outcomes, including poor intrauterine growth and prematurity. Because the drug supply is often episodic, the pregnant addict is subject to episodes of withdrawal and overdose, thereby subjecting the fetus to intermittent episodes of hypoxia in utero, hindering growth and raising the risk of spontaneous abortion, stillbirth, fetal distress, and prematurity. Infants born to mothers who are addicted to opioids are more MARIJUANA Pharmacology and Biologic Actions OPIOIDS (INCLUDING PRESCRIPTION PAIN KILLERS) Pharmacology and Biologic Actions Opioids are either drugs derived from the opium poppy or synthesized compounds that have similar biologic actions. The prototype opioid is morphine, and all opioids relate to it. For example, heroin (diacetylmorphine) exerts its effects by being metabolized to morphine, as does codeine, which is methylated morphine. Other opioids, such as methadone and oxycodone, are structurally unlike morphine but share its pharmacologic properties, because they stimulate similar opioid receptors. Specific opioid receptors (μ, δ, κ) have been identified in the nervous system and bowel that are activated by endogenous opioids, such as the naturally 118 PART III Maternal Health Affecting Neonatal Outcome likely to be of low birthweight, to be premature, and to suffer from infection and perinatal asphyxia (Christensen, 2008). Fetal Growth Initial reports from studies addressing the effects of maternal opioid use on the fetus suggested that infants exposed to opioids in utero have a higher incidence of IUGR (Lam et al, 1992) and smaller head circumference (Bauer, 1999; Boer et al, 1994). Hulse et al (1997) reported an association with heroin abuse and increased prematurity, low birthweight, and reduced fetal growth. However, more recent work controlling for confounding factors has not demonstrated a significant relationship between opioid use and prematurity, low birthweight, or IUGR (Minozzi et al, 2008; Sharpe and Kuschel, 2004). COCAINE Pharmacology and Biologic Actions Cocaine is a highly psychoactive stimulant with a long history of abuse. A naturally occurring anesthetic of the tropane family of alkaloids, cocaine is obtained from the Erythroxylon coca plant, which is indigenous to the mountain slopes of Central and South America. The coca leaf has been chewed or made into a stimulant tea by the natives of these areas to decrease fatigue and hunger. The pharmacologic actions of cocaine include inhibition of postsynaptic reuptake of norepinephrine, dopamine, and serotonin neurotransmitters by sympathetic nerve terminals, thus allowing higher concentrations of these neurotransmitters. In adults, cocaine binds strongly to neuronal dopamine reuptake transporters, thereby increasing postsynaptic dopamine at the mesolimbic and mesocortical levels and producing the addictive cycle of euphoria and dysphoria (Malanga and Kosofsky, 1999). Tryptophan uptake is similarly inhibited, altering serotonin pathways with resultant effects on sleep. Cocaine use leads to a sense of well-being, increased energy, increased sexual achievement, and an intense euphoria or “high.” The sympathomimetic action can have potentially devastating physiologic effects on the cardiovascular system. In adults, cocaine has been associated with cerebral hemorrhage, cardiac arrest, cardiac arrhythmias, myocardial infarction, intestinal ischemia, and seizures. Chronic use is associated with anorexia, nutritional problems, and paranoid psychosis and can ultimately result in neurotransmitter depletion and a “crash,” characterized by lethargy, depression, anxiety, severe insomnia, hyperphagia, and cocaine craving. Two forms of cocaine are commonly used—cocaine hydrochloride and cocaine base, which are either extracted by organic solvents or precipitated as “crack” through the use of ammonia and baking soda). Cocaine hydrochloride is a water-soluble white powder that is used orally, intranasally (“snorting”), or intravenously (“running”). Intravenous users are more likely to have a history of heroin abuse and often use the drug in combination with heroin (known as speedballing). Cocaine hydrochloride decomposes on heating and is, therefore, cocaine converted to the free base for inhalation. “Freebasing” involves extracting cocaine from aqueous solution into an organic solvent such as ether. Crack, the most widely available form of freebase, is almost pure cocaine; when smoked, it readily enters the bloodstream to produce levels similar to those occurring with intravenous use. Crack cocaine is popular in urban minority communities, where it may be smoked in combination with PCP (known as spacebasing). Crack smoking appears to be particularly reinforcing and is associated with compulsive use, binges, and acceleration of the addictive process. Cocaine and some of its metabolites readily cross the placenta and achieve pharmacologic levels in the fetus (Schenker et al, 1993). Amniotic fluid may serve as a reservoir for cocaine, and its metabolites and prolong exposure to vasoactive compounds. The extent to which cocaine or its metabolites are responsible for aberrant fetal growth, neurodevelopmental sequelae in exposed infants, and the range of congenital malformations reported in the literature may be less than suggested by uncontrolled case reports early in the cocaine-epidemic era. The confounding effects of increased use of multiple drugs, tobacco, alcohol, nutritional deficits, and decreased use of prenatal care among cocaine users make interpretation of the causal relationships between gestational cocaine exposure and intrauterine growth and subsequent neurobehavioral development difficult (Chiriboga, 1993). These identified confounders might serve to explain the reported effects attributed to cocaine in clinical series (Dempsey et al, 1996), although significant effects on the fetus and newborn have been reported in more recent studies that controlled for many confounders (Bada et al, 2002; Shankaran et al, 2007). Complications of Pregnancy Adverse perinatal outcomes associated with cocaine use are believed to be largely because of the vasoconstrictive effects of cocaine on uterine blood supply (Woods et al, 1987). An increase in maternal mean arterial blood pressure, a decrease in uterine blood flow, and a transient rise in fetal systemic blood pressure after an intravenous cocaine infusion have been described in fetal sheep along with significant fetal hypoxemia associated with changes in uterine blood flow (Moore et al, 1986; Woods et al, 1987). Maternal hypertension and intermittent fetal hypoxia contribute to the higher risks for abruptio placentae and IUGR seen in cocaine-exposed infants. To date, no well-defined cocaine-associated syndrome has been identified, and the teratogenic potential of cocaine remains controversial. Earlier reports had suggested that cocaine-exposed infants had a higher rate of limb reduction anomalies, heart defects, ocular anomalies, intestinal atresia or infarction, and other vascular disruption sequences. However, the preponderance of more recent data from multiple studies has failed to demonstrate higher rates of other congenital anomalies among cocaine-exposed infants (Behnke et al, 2001). Any association between fetal cocaine exposure and malformations is likely to be confounded by higher rates of maternal tobacco, marijuana, or alcohol use among the cocaineexposed groups. CHAPTER 12 Perinatal Substance Abuse Women who use cocaine during pregnancy are at higher risk for stillbirths, spontaneous abortions, abruptio placentae, IUGR, anemia and malnutrition, and maternal death from intracerebral hemorrhage. Cocaine directly stimulates uterine contractions because of its alpha-­adrenergic, prostaglandin, or dopaminergic effects, with resulting greater risk for fetal distress and premature deliveries. Abruptio placentae appears to be related to cocaine only when the drug is used shortly before delivery (Ostrea et al, 1992b). Pregnant women who use cocaine are also at high risk for premature labor, low-birthweight infants, premature rupture of the membranes, and perinatal infections. Cocaine use significantly increases the odds ratio for prematurity, low birthweight, premature rupture of membranes, and IUGR (Bada et al, 2002, 2005; Shankaran et al, 2007) as well as perinatal infections. Overall, because of the higher risks of premature delivery, the frequency of respiratory distress syndrome is greater in cocaine-exposed infants. Cocaine-exposed infants less frequently require surfactant administration and intubation for respiratory distress syndrome; however, the risks of bronchopulmonary dysplasia are similar in infants who have and those who have not been exposed to cocaine during gestation (Hand et al, 2001). Fetal Growth Infants exposed to cocaine in utero have lower birthweight, smaller birth length, and smaller head circumference (Bada et al, 2002, 2005; Behnke et al, 2001). Cocaine is hypothesized to reduce fetal growth via vasoconstriction of uteroplacental vessels with consequent decreased fetal substrate and oxygen delivery (Schempf, 2007). Several studies have shown a dose-response effect of cocaine exposure on fetal growth. In the Maternal Lifestyle Study, cocaine-exposed infants were 1 week younger in gestational age, and after controlling for confounders, cocaine exposure was associated with decrements in birthweight (151 g), length (0.71 cm), and head circumference (0.43 cm) at 40 weeks’ gestation (Shankaran et al, 2007). After adjusting for the effects of birthweight, gestational age, sex, maternal height, maternal weight gain, and other drug use, newborns with a high exposure to cocaine, as measured by radioimmunoassay of cocaine metabolites in maternal hair, had a disproportionately smaller head circumference even for their birthweight, resulting in “head wasting” (Bateman and Chiriboga, 2000). AMPHETAMINES Pharmacology and Biologic Actions Amphetamine (methylphenethylamine) was synthesized in 1887 and introduced in the United States in 1931. The N-methylated form, methamphetamine (or “crystal”), is increasingly abused because it readily dissolves in water for injection and it sublimates (converts directly from a solid to gas) when smoked (known as ice). The amphetamine isomers have similar clinical effects and can be distinguished only in the laboratory. Amphetamines were initially marketed for the treatment of obesity and narcolepsy 119 and continue to be used for the treatment of attention deficit disorders in children. Amphetamines are classified as schedule II drugs, like cocaine and narcotics. Amphetamines are taken orally, inhaled, or injected. The clinical effects and toxicity of these agents are often indistinguishable from those of cocaine. The primary difference is in the duration of action. The psychotropic effects of cocaine are of a short duration—5 to 45 minutes. The effects of amphetamines may last from 2 to 12 hours. Methamphetamine exposure has direct and indirect effects on the fetus, with increases in maternal blood pressure and restrictions in delivering nutrients and oxygen to the fetus (Smith et al, 2003). The clinical effects of amphetamines resemble those of cocaine. Like cocaine, amphetamines are sympathomimetics, and they potentiate the actions of norepinephrine, dopamine, and serotonin. In contrast to cocaine, amphetamines appear to exert their CNS effects primarily by enhancing the release of neurotransmitters from presynaptic neurons. Amphetamines can block reuptake of released neurotransmitters; they can also exert a weaker direct stimulatory action on postsynaptic catecholamine receptors. Complications of Pregnancy The medical and obstetric complications of amphetamine use are similar to those described for cocaine use. Amphetamine toxicity has been described as more intense and prolonged than cocaine toxicity. Visual, auditory, and tactile hallucinations are common, and microvascular damage has been seen in the brains of chronic users. Amphetamine withdrawal is characterized by prolonged periods of hypersomnia, depression, and intense, often violent paranoid psychosis. Obstetric complications include a higher incidence of stillbirth. Methamphetamine use is also associated with an increased incidence of premature delivery and placental abruption. Methamphetamine users who stop using earlier in gestation have rebound weight gain, suggesting that the anorexic effects are limited to continuous use (Smith et al, 2003). Like the pregnancies of cocaine users, the pregnancies of amphetamine users are characterized by poor prenatal care, sexually transmitted diseases, and cardiovascular problems including abruptio placentae and postpartum hemorrhage. The risk of cerebrovascular accidents is lower in pregnant amphetamine users than in pregnant cocaine users, but the mechanism for this difference is not understood. Perinatal problems associated with maternal amphetamine use include prematurity and IUGR (Smith et al, 2006). Fetal growth restriction, leading to smaller head circumference and lower birthweight, can result from the vasoconstrictive effects of norepinephrine or other vasoactive amines or from diminished maternal nutrient delivery as a consequence of the anorectic effect of amphetamine. Systemic effects from altered norepinephrine metabolism explain the transient bradycardia and tachycardia reported in exposed infants. Studies have failed to show consistent patterns of malformations in amphetamine-exposed infants, although several studies report cleft lip and cleft palate in association with amphetamine and methamphetamine exposure during early gestation (Plessinger, 1998). 120 PART III Maternal Health Affecting Neonatal Outcome Fetal Growth The IDEAL study followed 84 methamphetamineexposed and 1534 unexposed infants (confirmed by screening for meconium). Both groups included alcohol, tobacco, and marijuana use, but excluded opioids, PCP, and LSD (Arria et al, 2006; Smith et al, 2003, 2006). Cocaine and methamphetamine use occurred together in 13% of the methamphetamine users, and tobacco, alcohol, and marijuana use were also more frequent in the methamphetamine users. Methamphetamine-exposed infants were 3.5-fold more likely to be small for gestational age (less than the tenth percentile; 18% incidence). Methamphetamine contributed significantly to low birthweight, even after correcting for confounders such as low socioeconomic status, gestational age, and tobacco exposure. IDENTIFYING PREGNANCIES AND BABIES AT RISK PREGNANCIES Identification of perinatal substance abuse to intervene and protect the health and well-being of both mother and child has been a goal of practitioners for decades, ever since the scope of the problem became known—particularly for alcohol. Urine toxicology testing was initially believed to be the best approach, and the universal approach was used in many busy labor and delivery units, particularly in urban centers. In 1994, an American College of Obstetricians and Gynecologists (ACOG) Technical Bulletin concluded that urine toxicology testing had limited ability to detect substance abuse and therefore recommended against universal toxicology screening and for alternative screening methods. In 2003, the U.S. Congress passed the Keeping Children and Family Safe Act. This law requires each state (as a condition of receiving federal funds under the Child Abuse Prevention and Treatment Act) to develop policies and procedures designed “to address the needs of infants born and identified as being affected by illegal substance abuse or withdrawal symptoms resulting from prenatal drug exposure.” This law included a requirement that health care providers notify child protective services regarding prenatal substance exposure, but differed from the providers’ legal responsibility to report suspected child abuse or neglect because the former “shall not be construed to be child abuse” and “shall not require prosecution of the mother” (Washington State Department of Health, 2009). Each state was expected to develop their own guidelines for identification of at-risk pregnancies. A 2004 ACOG Committee Opinion stated that best practices included universal screening questions followed by brief interventions or referrals. A number of screening tools have been developed including tolerance, annoyed, cut down, eyeopener (T-ACE), tolerance, worried, eye-opener, amnesia, K/cut down (TWEAK), and parents, partner, past, pregnancy (assess for substance abuse, domestic violence) (4 P’s Plus) (Chang, 2001; Chasnoff et al, 2005; Sokol et al, 1989). Identification of at-risk pregnancies concentrated on the maxim “if you don’t ask, they won’t tell.” In 2008, ACOG issued a committee opinion entitled AtRisk Drinking and Illicit Drug Use: Ethical Issues in Obstetric and Gynecologic Practice, which reaffirmed the 2004 ACOG TABLE 12-2 Risk Indicators for Gestational Substance Exposure Maternal No prenatal care Precipitous labor Placental abruption Repeated spontaneous abortions Hypertensive episodes Severe mood swings Previous unexplained fetal demise Myocardial infarction or stroke Newborn Jittery with normal blood glucose level Marked irritability Unexplained seizures or apneic spells Unexplained IUGR NEC in an otherwise healthy term infant Neurobehavioral abnormalities Signs of neonatal abstinence syndrome IUGR, Intrauterine growth restriction; NEC, necrotizing enterocolitis. statement. The opinion stated that “as a result of intensive research in addiction over the past decade, evidence-based recommendations have been consolidated into a protocol for universal screening questions, brief intervention and referral to treatment.” This approach—in particular, the use of standard screening questionnaires—was strongly recommended for obstetricians. Therefore a history of drug and alcohol use should be routinely included in the initial contact with every pregnant patient. To be effective, the history taking must be nonjudgmental and must occur in the context of other lifestyle questions. When a positive history of use is obtained, intervention should begin immediately. The person taking the history should be prepared to offer preliminary counseling on risk reduction and concrete referrals for treatment programs, although access to drug programs is often restricted, inadequate, or delayed. Although the screening questionnaire approach works well for women who seek prenatal care, it is not useful for the significantly higher percentage of substance-using pregnant women who do not seek or actively avoid prenatal care. For these women, it is more helpful to identify maternal risk indicators for perinatal substance abuse at the time of delivery. Use of one of the screening questionnaires in addition to drug testing increases the likelihood of identifying at-risk pregnancies and neonates, allowing for earlier referral for treatment or specialized interventions. Several of these maternal risk indicators for perinatal substance abuse were identified in the 2004 ACOG statement (Table 12-2) (American College of Obstetricians and Gynecologists, 2004). BABIES It is important for practitioners to know the difference between a substance-exposed newborn and a substanceaffected newborn. The Washington State Department of Health defines them as follows: ll Substance exposed: ll Tests positive for substances at birth or ll Mother tests positive for substances at time of delivery or ll Is identified by medical practitioner as having been prenatally exposed to substances CHAPTER 12 Perinatal Substance Abuse Substance affected: ll Has withdrawal symptoms resulting from prenatal substance exposure or ll Demonstrates physical and behavioral signs that can be attributed to prenatal exposure to substances and is identified by a medical practitioner as affected For the identification of either of these groups of infants, assessment of both maternal and newborn risk indicators (see Table 12-2) is essential. If risk indicators suggest perinatal substance abuse, then consideration should be given to newborn drug testing. For urine testing, there is a poor correlation between maternal and newborn tests. The earliest newborn urine will contain the highest concentration of substances, but the first urination may be missed and urine output in the first day is often scant. However, some drug metabolites such as cocaine are present for 4 to 5 days, and marijuana metabolites may persist for weeks. The disadvantages of newborn urine drug testing are that it primarily reflects substance exposure during the preceding 1 to 3 days, and alcohol is nearly impossible to detect. Meconium drug testing (at term) reflects substance exposure during the second half of gestation, has a high sensitivity for opioids and cocaine, and can assess for more drugs than urine testing. The cost is similar to newborn urine testing, but it generally takes longer to obtain results. Other newborn drug tests include hair—which is costly and has a high sensitivity for cocaine, amphetamines, and opioids but not for marijuana—and umbilical cord segments, which is an evolving technology that is not widely available (Kuschel, 2007). ll PREGNANCY MANAGEMENT Ideally, perinatal substance use or abuse should be identified by universal screening procedures during prenatal visits; this gives the practitioner an opportunity to intervene, with the goal being prevention of significant obstetric and neonatal complications related to substance abuse. Screening procedures and counseling should include the following topics: Antepartum ll Initial screening for hepatitis, HIV, and tuberculosis (if not part of routine prenatal care) and ongoing screening for sexually transmitted infections ll Referring to methadone treatment program, if appropriate ll Discussing possible drug effects on the fetus and newborn ll Discussing contraception and prevention of sexually transmitted disease ll Discussing breastfeeding issues related to alcohol and drug use Intrapartum ll Effects of recent drug use on labor and fetal well-being ll Pain management (women in methadone or alternative opioid treatment programs will be less responsive to opioid pain medications) ll Intrapartum prophylaxis for HIV, herpes simplex virus infections ll Need for social services involvement Postpartum ll Breastfeeding issues (related to both drug use and infection) ll Contraception and pregnancy prevention including tubal ligation ll ll 121 Support for continuation in, or initiation of, a drug treatment program Child Protective Services notification MATERNAL METHADONE MAINTENANCE The potential benefits of maternal methadone maintenance are numerous (Kandall and Doberczak, 1999; Ward et al, 1999) and include the prevention of opioid withdrawal symptoms in the mother, better medical and prenatal care, improved health and growth of the fetus, and maintenance of opioid levels in the mother to decrease both the use of illicit drugs and the potential for perinatal infections. Methadone maintenance programs associated with comprehensive medical and psychosocial services for the pregnant woman are of additional benefit. Methadone maintenance has been shown to be associated with higher birthweight in some but not all studies (Brown et al, 1998; Kandall and Doberczak, 1999). Detoxification of a pregnant heroin user is infrequently attempted, because maternal drug withdrawal is believed to be associated with subsequent fetal withdrawal, fetal asphyxia, and spontaneous abortions (Barr and Jones, 1994). Dashe et al (1998) reported on a small study of opioid-using pregnant women undergoing safe maternal detoxification. Close to 60% of the women completed detoxification, but almost 30% resumed opioid use. McCarthy et al (1999) showed that women who reduced their methadone dose during pregnancy had infants with higher birthweights than a control group who continued on the same or increased methadone dose throughout pregnancy. In the United States, most pregnant, narcotic-addicted women are treated with daily methadone rather than a program of detoxification. Some authorities, however, have urged the reappraisal and reevaluation of the benefits of methadone maintenance in pregnancy (Brown et al, 1998; Hulse and O’Neil, 2001), and there are recent studies comparing methadone with buprenorphine in pregnancy (Jones et al, 2008; Minozzi et al, 2008). In the Netherlands, women enrolled in a methadone program had higher rates of prenatal care, which were associated with higher birthweights and reduced prematurity in the offspring (Soepatmi, 1994). When women in a methadone maintenance program were enrolled in an enhanced prenatal care program, their infants’ birthweights were significantly larger than those in the control group of women receiving regular methadone maintenance during pregnancy (Chang et al, 1992). Others have shown that higher methadone doses are associated with improved head circumference and increased gestational age at delivery (Hagopian et al, 1996). Using a metaanalysis design, Hulse et al (1997) found that low infant birthweight was associated with heroin use and that birthweights were improved with methadone treatment during pregnancy. These favorable outcomes are believed to be to the result of a stable intrauterine environment uncomplicated by periods of intoxication and withdrawal, as well as less stress and better nutrition in the mother. Several investigators have found that neonatal withdrawal symptoms, birthweight, length of pregnancy, and the number of days infants require treatment for abstinence do not correlate with maternal methadone dosage (Brown 122 PART III Maternal Health Affecting Neonatal Outcome et al, 1998; Finnegan, 1991; Madden et al, 1977; Rosen and Pippenger, 1976). In contrast, others have reported a correlation between the severity of neonatal withdrawal and maternal methadone dose (Dryden et al, 2009; Harper et al, 1977; Maas et al, 1990; Malpas et al, 1995). Studying maternal and neonatal serum levels of methadone does not help clarify this dilemma. Investigators have found no correlation between neonatal serum levels of methadone and the maternal methadone dose at delivery, the maternal serum levels, or the severity of withdrawal symptoms in the neonates (Harper et al, 1977; Mack et al, 1991). Other researchers have reported that neonatal signs of withdrawal correlate with the rate of decline of the neonatal plasma level during the first few days of life (Doberczak et al, 1993). There are no definitive guidelines for methadone doses during pregnancy, and there is continuing controversy over the most appropriate dose of methadone maintenance during pregnancy. The divergent findings noted previously have been used to argue either for weaning a pregnant woman to a low methadone maintenance dose or for attempting complete maternal detoxification during pregnancy. Some authorities believe in maintaining high methadone doses to keep the mother from “chipping” with additional street drugs, which would put her at risk for greater complications of pregnancy and for a higher risk of infections transmitted by intravenous use of drugs (HIV, hepatitis) or sexually transmitted disease. High-dose methadone maintenance ranges between 60 and 150 mg/day. Despite reports of the safe detoxification of pregnant women, there are still concerns that the fetus is placed at risk during maternal detoxification. The medical management of pregnant women who are addicted to opioids remains controversial (Christensen, 2008). There is a growing body of literature promoting breastfeeding for infants of mothers in methadone maintenance programs as both beneficial and safe. Studies have shown that breastfed infants tend to have less need for pharmacotherapy for neonatal abstinence syndrome, despite low and unpredictable levels of methadone in breast milk and in infant serum. Maternal serum levels, breast milk levels, and infant serum levels do not correlate with maternal methadone dose (Jansson et al, 2008). Buprenorphine (Suboxone, Subutex) Buprenorphine is an alternative opioid substitute that was first introduced in France in 1996. It is increasingly being used with or instead of methadone for the treatment of opioid addiction, because it has fewer autonomic side effects than methadone as well as improved compliance and treatment efficacy. Recent trials have compared methadone with buprenorphine treatment during pregnancy and found no difference in the neonatal outcome or incidence of withdrawal symptoms (Jones et al, 2008; Minozzi et al, 2008). Human Immunodeficiency Virus and Other Viral Infections Nationwide, intravenous drug abusers are the second largest risk group for HIV infection. Drug abusers also may be the primary source of infection for non–drug-using heterosexuals and children (Chamberland and Dondero, 1987). Seventy-five percent of cases of acquired immunodeficiency syndrome (AIDS) in children are perinatally acquired. The seropositivity rate varies across the country; the rate among female intravenous drug users in New York City and northern New Jersey is estimated at 50% to 70%, compared with 5% to 20% in California. Heroin and cocaine addicts often resort to prostitution to support their habits, and amphetamine and methamphetamine users often inject drugs several times daily. Alcohol decreases sexual inhibition, impairs judgment, and increases the incidence of unsafe sexual activity. Every infant born to a substance abuser should be evaluated for HIV infection, and universal precautions should be observed. The American Academy of Pediatrics Committee on Infectious Diseases (2009) recommends rapid HIV testing of any mother whose HIV status is not known, with appropriate consent as required by local law. Intravenous drug use places the woman at risk for multiple infectious complications, including cellulitis, thrombophlebitis, hepatitis, endocarditis, syphilis, gonorrhea, and AIDS. In a prospective study undertaken in Canada, a 5-year incidence of HIV seroconversion was 13.4%; the rate of conversion associated with injection of heroin or cocaine was 40% higher in women than in men (Spittal et al, 2002). Opioid abusers are also less likely to receive prenatal care or to obtain late prenatal care (Bauer, 1999). Heroin-addicted mothers are often poorly nourished, and iron-deficiency anemia is more common in pregnant opioid users than in nonusers. Bauer (1999) found that maternal hepatitis infections were fivefold higher in opioid-using women than in a control group of nonusers. Hepatitis C virus (HCV) is another chronic infectious condition that is spread by parenteral exposure to infected blood and can be perinatally acquired by the newborn. The most common risk factors for acquiring infection are injection drug use, having multiple sexual partners, or having received blood products before 1992. The 2009 American Academy of Pediatrics Red Book (American Academy of Pediatrics Committee on Infectious Diseases, 2009) states that “seroprevalence [of hepatitis C] among pregnant women in the United States has been estimated at 1% to 2%,” and the risk of perinatal transmission averages 5% to 6% from women who are HCV-RNA positive at the time of delivery. Maternal coinfection with HIV has been associated with increased risk of perinatal transmission of HCV. Antibodies to HCV and HCV RNA have been detected in colostrum, but the risk of HCV transmission is similar in breastfed and bottle-fed infants, so breastfeeding is currently allowed. NEONATAL MANAGEMENT AFTER GESTATIONAL SUBSTANCE ABUSE GENERAL Although at higher risk for medical complications, the majority of infants of drug-using women do not require intensive neonatal care; however, symptomatic infants often need more nursing care. Physical examination on admission should document a gestational age assessment, birthweight, head circumference, and length. Infants CHAPTER 12 Perinatal Substance Abuse should be examined carefully for evidence of malformations, dysmorphic facial features, or both. Studies such as electroencephalography and brain imaging may add diagnostic or prognostic information when physical or neurologic abnormalities are not clearly consistent with drug exposure, but these procedures are not indicated for most drug-exposed infants. If indicated by neonatal or maternal risk indicators, toxicology testing should be performed on neonatal urine, meconium, or both as soon as possible after birth. Infants whose mothers were not screened for HIV should undergo screening for perinatal HIV exposure and other infections such as hepatitis and syphilis, as clinically indicated. In some states, rapid testing of the newborn for HIV is “required by law if the mother refuses to be tested,” so that appropriate treatment of the infant can be started before 12 hours of age. Rapid screening detects only HIV1, the most common serotype of HIV in the United States, but it can miss HIV-2, so measurement of HIV antibodies should also be performed (American Academy of Pediatrics Committee on Infectious Diseases, 2009). Cocaine-exposed infants weighing more than 1500 g have longer hospital stays and increased need for therapies, procedures, intravenous fluid, and formula feeding. These infants also undergo more investigations for sepsis, more neonatal intensive care unit (NICU) admissions, and more social and family problems delaying discharge (Bada et al, 2002). In the Maternal Lifestyle Study, cocaine-exposed infants had a higher frequency of infection (odds ratio [OR], 3.1; 99% confidence interval [CI], 1.8 to 5.4) and neurologic signs and symptoms (adjusted OR, 1.7; 99% CI, 1.0 to 2.1). Neurologic signs were highest in the infants exposed to opioids and cocaine, but remained significantly increased in infants exposed to cocaine alone. Smoking also increased the risk for neurologic signs and symptoms (Shankaran et al, 2007). The association between cocaine exposure and fetal hypoxic ischemic episodes creates special concerns. Maternal cocaine use exposes infants to a higher than expected risk of problems with postasphyxial syndrome, and organ dysfunction from hypoxic-ischemic injury should be investigated and treated. Feedings in premature infants with cocaine exposure should be started cautiously, because premature infants exposed to cocaine may be at increased risk for necrotizing enterocolitis. In addition, after controlling for gender, gestational age, birthweight, maternal parity, ethnicity, and polydrug use, heavy cocaine use during pregnancy was associated with a slightly higher risk of subependymal hemorrhage (­Shankaran et al, 2007). Using the NICU Network Neurobehavioral Scale (NNNS), subtle differences in behavior were detected in drug-exposed infants (Lester et al, 2002; Smith et al, 2008). Cocaine-exposed infants showed lower arousal, and with heavy cocaine use they showed lower regulation and higher excitability than did unexposed infants (Lester et al, 2002). There were no stress or abstinence signs associated with cocaine exposure. Marijuana use was associated with more stress and abstinence signs and higher excitability scores (Lester et al, 2002). Low birthweight was also significantly correlated with poorer regulation and higher excitability. Methamphetamine exposure was also associated with increased stress signs, especially in first-trimester use. Heavy 123 methamphetamine use was related to lethargy, lower arousal, and increased physiologic stress (Smith et al, 2008). Neonatal neurologic abnormalities similar to a mild withdrawal syndrome, consisting of hypertonicity, irritability, and jitteriness, have been reported after in utero marijuana exposure, but without documented evidence of long-term sequelae (Cornelius et al, 1995). Finally, gestational nicotine exposure definitely altered the newborn neurobehavioral scores and has also been reported to elevate neonatal abstinence scores (Godding et al, 2004; Law et al, 2003). In the Maternal Lifestyle Study, only 100 women were identified as isolated users of opioids, and a similar number used cocaine and opioids. Transient but dramatic neurobehavioral signs are present in the first week of life as symptoms of opioid withdrawal (increased irritability, jitteriness, poor feeding, sweating, sneezing; Shankaran et al, 2007). See the discussion under Neonatal Abstinence Syndrome. BREASTFEEDING AND DRUG EXPOSURE Breastfeeding has the benefit of improved bonding, but the risks of HIV and continued drug exposure may outweigh this benefit. Women who wish to breastfeed despite these potential risks should undergo drug monitoring and sequential HIV antibody testing. Close observation of mother-infant interactions should be documented in the infant’s chart. Parenting and childcare skills should be stressed as part of the discharge education for the mother. All physician interactions with the family should be documented in detail. Most illicit drugs of abuse that are of low molecular weight and lipophilic, are readily excreted in breast milk, but have varying degrees of bioavailability (Howard and Lawrence, 1998). Cocaine has been detected in breast milk and Chasnoff et al described a two-week-old breastfed infant who had clinical signs of cocaine intoxification. Both the baby’s urine and the mother’s milk contained cocaine (Chasnoff et al, 1987). A widely-referenced report of cocaine seizures in a breastfed infant actually resulted from topical coacine applied to sore nipples, not from cocaine-laced breast milk (Chaney et al, 1988). Because of the potential risk of toxicity, breastfeeding is generally discouraged in women who are known abusers of these drugs and who are not willing to engage in substance abuse treatment and monitoring. Breastfeeding may be supported for women who are engaged in substance abuse treatment and who have received good prenatal care with a confirmed period of sobriety prior to delivery. Breastfeeding by women using methadone is recommended. Concentrations of methadone in human milk are low, and there are other significant advantages for the mother and infant (Jansson et al, 2008). Alcohol use while breastfeeding is not listed as a contraindication by the American Academy of Pediatrics, but excessive maternal alcohol intake during breastfeeding can be deleterious for the infant and should be avoided. Smoking in the postnatal period and during breastfeeding also has deleterious effects on the newborn. Smoking is associated with measurable levels of nicotine and cotinine in maternal breast milk. 124 PART III Maternal Health Affecting Neonatal Outcome NEONATAL ABSTINENCE SYNDROME Clinical Findings Classic neonatal withdrawal or abstinence syndrome consists of a wide variety of CNS signs of irritability, gastrointestinal and feeding problems (diarrhea, hyperphagia or poor feeding), autonomic signs of dysfunction (fever, sweating, sneezing), and respiratory symptoms (Tables 12-3 and 12-4). These symptoms are most often related to gestational opioid exposure, but are relatively nonspecific, with the differential diagnosis including infection, meningitis, hypocalcemia, hyponatremia, intracranial hemorrhage, seizures, and stroke. The signs of neonatal serotonin syndrome (or selective serotonin reuptake inhibitor withdrawal) may also mimic the signs of neonatal opioid abstinence syndrome (Boucher et al, 2008; Moses-Kolko et al, 2005). The timing of withdrawal signs from specific drug exposures can often be anticipated; for example, heroin withdrawal usually occurs within 24 hours of birth, whereas methadone withdrawal symptoms typically TABLE 12-3 Clinical Signs of Neonatal Withdrawal Syndrome (Narcotic Abstinence Syndrome) Central nervous system ­dysfunction Excoriation (from frantic ­movement) Hyperactive reflexes Increased muscle tone Irritability, excessive crying, high-pitched cry Jitteriness, tremulousness Myoclonic jerks Seizures Sleep disturbance Autonomic dysfunction Excessive sweating Frequent yawning Hyperthermia Respiratory symptoms Nasal stuffiness, sneezing Tachypnea Gastrointestinal and feeding disturbances Inadequate oral intake Diarrhea (loose, watery, ­frequent stools) Excessive sucking Hyperphagia Regurgitation These signs were abstracted from Finnegan L, Connaughton J, Kron R: Neonatal abstinence syndrome: assessment and management, Addict Dis 2:141-158, 1975. begin later, at approximately 48 to 72 hours after birth. The incidence of neonatal abstinence syndrome (NAS) in infants of women using heroin or methadone is high, with wide ranges reported between 16% and 90% (Agarwal et al, 1999; Boer et al, 1994; Maas et al, 1990; van Baar et al, 1994), and between 30% and 91% of infants with signs of NAS receive pharmacologic treatment for NAS with inpatient stays averaging 3 weeks (Dryden et al, 2009; Kuschel, 2007). Premature infants generally have milder signs of withdrawal and often show alternating periods of hyperactivity and lethargy, with tremors seen less commonly. The mortality rate for these infants is less than 1% (Boer et al, 1994). Death is rarely associated with withdrawal alone, but usually occurs as a consequence of prematurity, infection, and severe perinatal asphyxia. A number of evaluation tools are used to assess the severity of opioid withdrawal after birth. The neonatal abstinence score is a scale based on nursing observations of the severity of signs of withdrawal (Finnegan et al, 1975) and is the most widely used scale. The Lipsitz score was developed at the same time and is simpler to use, with a score greater than 4 indicating withdrawal. Green and Suffet (1981) introduced the Neonatal Narcotic Withdrawal Index as a rapid physician-based evaluation for neonatal signs of withdrawal. The use of these scoring systems allows more objective quantification of the severity of the infant’s withdrawal and the response to treatment. These scoring systems have shown good interobserver reliability and can improve clinicians’ ability to treat the withdrawing infant appropriately (Anand and Arnold, 1994; Franck and Vilardi, 1995). The goal of medical management of opioid withdrawal is to avoid serious symptoms of NAS, such as seizures, and to maintain the infant’s comfort while enabling the infant to feed, sleep, and gain weight in an appropriate manner. There are a number of reported threshold scores for initiating pharmacologic treatment (Finnegan NAS scores between 7 and 12) (Kuschel, 2007), but none of these choices have been examined in a scientific manner. The decision to begin treatment or to wean treatment should be influenced by the absolute score and other factors, such as the infant’s age, comorbidities, other conditions leading to abnormal behavior, and a daily evaluation of the abnormal clinical elements observed in the scoring system (Kuschel, 2007). Standard medical practice is to combine both developmental and behavioral methods with ­pharmacologic interventions as necessary to control symptoms and signs of narcotic abstinence. TABLE 12-4 Neonatal Neurobehavioral Symptoms after Fetal Drug Exposure Drug Alcohol Onset (days) Peak (days) 0-1 1-2 Duration Relative Severity 1-2 days Mild Likely NICU Admission No Symptoms ? Amphetamine 0-3 — 2-8 wk Mild No Neuro Cocaine 0-3 1-4 ? mo Mild No Neuro Heroin 0-3 3-7 2-4 wk Mild to severe Yes Neuro, Resp, GI Methadone 3-7 10-21 2-6 wk Mild to severe Yes Neuro, Resp, GI SSRI 0-3 1-3 2-10 days Mild to moderate No Neuro, Resp, GI Tobacco 0-1 1-2 2-3 days Mild No Neuro GI, Gastrointestinal symptoms including poor weight gain; Neuro, neurobehavioral symptoms; NICU, neonatal intensive care unit; Resp, respiratory symptoms; SSRI, selective serotonin reuptake ­inhibitor. CHAPTER 12 Perinatal Substance Abuse Treatment Treatment for opioid withdrawal should always begin with supportive, nonpharmacologic measures such as soothing including swaddling, rocking, decreased environmental stimulation, avoiding unnecessary handling and irritation, and progressing to pharmacologic management only when medically necessary. Between 30% and 91% of infants exhibiting signs of NAS will receive pharmacologic treatment (Kuschel, 2007). The mainstay of treatment for opioid withdrawal is the use of opioids, either alone or in combination with other medications (American Academy of Pediatrics Committee on Drugs, 1998; Osborn et al, 2002, 2005). Medication is titrated for each infant according to the severity of the signs of withdrawal and abstinence scoring. A recent survey in the United States showed that opioids are the most common medication used for narcotic withdrawal, and phenobarbital is frequently used or added for polydrug exposure (Sarkar and Donn, 2006). A metaanalysis of seven studies found that opioid treatment reduced the time to regain birthweight and the duration of supportive care compared with supportive care alone; however, the length of hospital stay was increased (Osborn et al, 2005). Phenobarbital was also shown to be superior to diazepam for treating NAS, but small, randomized, controlled trials comparing morphine to phenobarbital for treatment of NAS suggested that opioids are superior at decreasing treatment duration and lowering NAS scores (Ebner et al, 2007; Jackson et al, 2004; Osborn et al, 2002). Morphine is the opioid most often used in treating NAS. The only study comparing an oral preparation of morphine to dilute tincture of opium (DTO), a concentrated morphine solution that also contains alcohol, recommends oral morphine to avoid problems with dilution of highly concentrated DTO and the unwanted effects of the alcoholic extracts with various alkaloids (Langenfeld et al, 2005). A standard starting dose of morphine is 0.04 to 0.05 mg/kg given orally every 3 to 4 hours. This dose can be increased in increments of 0.05 to 0.1 mg until the symptoms are controlled. The usual dose for infants experiencing withdrawal at birth ranges from 0.08 to 0.2 mg every 3 to 4 hours (American Academy of Pediatrics Committee on Drugs, 1998; Anand and Arnold, 1994; Burgos and Burke, 2009; Levy and Sino, 1993). Higher doses may be needed to control significant physiologic signs of withdrawal such as diarrhea, pyrexia, hypertension, and significant hypertonicity. Although preparations of oral morphine are most commonly used in recent studies, methadone has also been reported as an option for treatment of NAS (Burgos and Burke, 2009; Kuschel, 2007). There is one retrospective report of methadone use for neonatal abstinence syndrome (Lainwala et al, 2005) and more extensive experience with methadone in the treatment of opioid withdrawal in older infants and children (Tobias, 2000; Tobias et al, 1990). Methadone has a long duration of action and can be administered by either the oral or the parenteral routes. The initial recommended methadone dose is 0.05 to 0.1 mg/kg followed by 0.025 to 0.05 mg/kg every 4 to 12 hours until abstinence scores are controlled. The total daily dose given to control symptoms is then divided into two doses 125 and given every 12 hours. Tobias et al (1990) showed that methadone could be given every 12 to 24 hours because of its longer half-life. There were no advantages to methadone over morphine in neonatal abstinence (Lainwala et al, 2005). In addition, sublingual buprenorphine has been reported as a treatment for NAS (Kraft et al, 2008). Phenobarbital has been used for signs of acute opioid withdrawal. Phenobarbital does not, however, reduce significant physiologic signs of withdrawal, such as diarrhea and seizures. At higher doses, phenobarbital has also been shown to impair infant sucking and cause excessive sedation. The doses of phenobarbital used are 5 to 20 mg/kg in the first 24 hours, followed by 2 to 4 mg/kg every 12 hours; the therapeutic blood level of phenobarbital for control of opioid withdrawal signs is not known. Combining oral morphine solution (i.e., DTO) and phenobarbital treatment was found to shorten duration of hospitalization and lessen the severity of withdrawal symptoms, compared with morphine treatment alone (Coyle et al, 2002, 2005). Compared with those treated with morphine alone, infants treated with morphine and phenobarbital were more interactive, had smoother movements, were easier to handle, and were less stressed. Dual treatment resulted in improved neurobehavioral organization during the first 3 weeks of life, which may indicate a more rapid recovery from opioid withdrawal (Coyle et al, 2005). In these studies, however, the infants were discharged home on phenobarbital therapy, from which they were slowly weaned throughout infancy. Some of these infants received phenobarbital for prolonged times. Clonidine, an α2-adrenergic receptor antagonist, is used treating opioid withdrawal symptoms in older children and adults. A recent randomized, controlled trial showed that adding clonidine to opioid treatment (i.e., DTO) significantly reduced the median length of therapy with opioids and the number of infants requiring high dose opioids, and it eliminated infants who met their definition of treatment failure (Agthe et al, 2009). During hospitalization, there were no significant adverse events (i.e., hypertension, hypotension, bradycardia, desaturations) related to clonidine use. A German group has also retrospectively reported their experience using clonidine and chloral hydrate for NAS compared to a larger group treated with morphine and phenobarbital; they also reported decreased duration of treatment, decreased length of stay, and reduction of symptoms in the clonidine group (Esmaeili et al, 2010). Before clonidine is used widely, a larger trial is indicated to investigate appropriate dosing regimens and to determine long-term safety. Once medications have been titrated to a level that controls the severity of opioid withdrawal and lowers the NAS scores, then tapering of the dosage should be started. A common method is to decrease the opioid dose by 10% to 20% of the highest dose, with continued surveillance of NAS scores to assure the infant tolerates the decrease. It is not unusual to note increased signs of opioid withdrawal during the medication tapering. The goal of weaning is to allow the infant to acclimate to a new and lower dose of medication while ensuring that he or she is comfortable and consolable and is able to sleep, eat, and gain weight appropriately. Objective measurements using established withdrawal scoring systems should be used to determine 126 PART III Maternal Health Affecting Neonatal Outcome the rate and efficacy of medication tapering. Most often opioid medications are weaned every 24 to 48 hours as long as NAS scores remain low and the infant’s clinical condition is unchanged (Burgos and Burke, 2009; Coyle et al, 2002; Jackson et al, 2004). Whether narcotics can safely be weaned more rapidly in infants has not been studied. One small study reported good results using oral morphine on an as-needed basis for elevated individual withdrawal scores, rather than providing regular doses of morphine every 3 to 4 hours (Ebner et al, 2007). No specific protocols for weaning phenobarbital or clonidine have been reported. The average length of hospital stay for infants with NAS who are treated with medications varies from 8 to 78 days, with a stay of 21 to 30 days being common (Coyle et al, 2002; Ebner et al, 2007; Jackson et al, 2004; Kuschel, 2007). Investigation of outpatient management of detoxification may result in a shorter hospital stay, but with a more prolonged duration of neonatal treatment ­(Kuschel, 2007). LONG-TERM EFFECTS OF PERINATAL SUBSTANCE ABUSE non–cocaine-exposed infants at birth, but there were no differences in weight between ages 1 and 6 years. For height, cocaine-exposed infants were shorter at birth through 2 years, but the difference disappeared by age 3, and cocaine-exposed infants had a smaller head circumference from birth to 1 year old, but this difference subsequently disappeared (Shankaran et al, 2007). A previous systematic review concluded that there is no consistent effect of cocaine exposure on physical growth in children younger than 6 years (Frank et al, 2001). Opioid-exposed infants also show no difficulties with postnatal growth through 3 years (Hunt et al, 2008). Infants with IUGR at birth remained significantly lighter (by 2.1 kg) and shorter (by 1.8 cm) at 6 years, and they continued to have a smaller head circumference (by 0.9 cm). There was no interaction between cocaine exposure and IUGR status (Shankaran et al, 2007). There were also no differences in medical diagnoses, blood pressure, hospitalizations, or overall health status associated with opioid exposure (Shankaran et al, 2007). The effect of cocaine on subsequent blood pressure is unclear. IUGR status at birth was significantly associated with hypertension after correcting for in utero drug exposures, maternal race, and child’s body mass index. SUDDEN INFANT DEATH SYNDROME IN DRUG-EXPOSED INFANTS NEUROBEHAVIORAL ABNORMALITIES The incidence of sudden infant death syndrome (SIDS) is greater in drug-exposed infants, although the increase appears less significant for cocaine-exposed infants (Chasnoff et al, 1989b; Kandall et al, 1993) than for methadone-exposed or heroin-exposed infants (Bauchner and Zuckerman, 1990). A metaanalysis of 10 studies demonstrated a 4.1 OR for SIDS among cocaine-exposed infants (Fares et al, 1997) compared with infants not exposed to perinatal drugs. However, after data were controlled for concurrent use of other drugs, the increased risk for SIDS could not be attributed to intrauterine cocaine alone, but was believed to be caused by exposure to other illicit drugs and smoking. There is a significant relationship between exposure to tobacco smoke and SIDS. A higher risk for SIDS is reported in infants exposed to maternal smoking with either antenatal or postnatal exposure (Blair et al, 1996; Mitchell et al, 1993; Schoendorf and Kiely, 1992). Additional evidence supporting a causal role includes a dose-response relationship (Rogers, 2008). With recent substantial decreases in SIDS risk (after “back to sleep”), maternal smoking is calculated to account for an increasing proportion of SIDS risk. Mitchell and Millerad (2006) suggest that one third of all SIDS deaths might be prevented by eliminating in utero exposure to maternal smoking. Despite the increased risks of SIDS among drug- and tobacco-exposed infants, home apnea monitoring is not indicated in the absence of other risk factors. POSTNATAL GROWTH AND MEDICAL PROBLEMS There are no documented long-term effects of substance exposure on postnatal growth. In the Maternal Lifestyle Study, cocaine-exposed infants were lighter than Detailed studies have identified both neurologic soft signs and a higher risk of cognitive, behavioral, and psychiatric problems linked to gestational substance exposure. However, significant confounding factors make it difficult to attribute specific deficits directly to substance exposures. No major neurologic deficits in motor development are found after in utero exposure to cocaine (Shankaran et al, 2007) or opioids (Hunt et al, 2008). Infants exposed to cocaine were reported to have lower motor skills at 1-month testing, but they displayed significant improvements over time. Both higher and lower levels of tobacco use were related to poor motor performance (Shankaran et al, 2007). Subtle neurobehavioral abnormalities have been reported in older studies and more recently using the NNNS to evaluate infants in the month after birth. Cocaine-exposed infants manifest a range of neurobehavioral abnormalities that were initially described as drug withdrawal, but are more likely caused by acute intoxication (Dempsey et al, 1996). Signs are present at birth and wane as cocaine and the metabolite, benzoylecgonine, are cleared from plasma. The infants are hypertonic, irritable, and tremulous (Chiriboga, 1993), and they may have abnormal crying, sleep, and feeding patterns. Tachycardia, tachypnea, and apnea have been noted in two blinded, controlled studies, with significant elevations in cardiac output, stroke volume, mean arterial blood pressure, and cerebral artery flow velocity resolving by day 2, which is consistent with an intoxicant effect of cocaine (van de Bor et al, 1990a, 1990b) Cocaine-exposed infants may have abnormal electroencephalograms or clinical seizures, perhaps the result of toxicity from the metabolite, benzoylecgonine (Konkol et al, 1994); however, neonatal seizures attributable directly to maternal use of cocaine are rare (Legido et al, 1992). CHAPTER 12 Perinatal Substance Abuse A number of studies suggest that exposure to exogenous opioids during fetal development may produce lifelong alterations in the developing brain. Infants who have been exposed to opioids in utero have a higher risk for fetal growth retardation and smaller head circumference than those who have not (Bauer, 1999). In addition, significant developmental and learning deficits have been described in both methadone-exposed and heroin-exposed children (Soepatmi, 1994; van Baar and de Graaff, 1994; van Baar et al, 1994), and Bunikowski et al (1998) have reported a higher incidence of abnormalities in intellectual performance, developmental retardation, and neurologic abnormalities in a group of opioid-exposed infants compared with a control group of infants. However, the treated infants in the study had an unusually high incidence of seizures during withdrawal treatment with phenobarbital alone, as well as a higher incidence of prematurity than the control infants. These important considerations temper the findings of that study. Early neurobehavioral changes have been associated with maternal methamphetamine use; heavy use was associated with lower arousal, increased lethargy, and increased CNS stress (Smith et al, 2008). However, a detailed analysis in a subgroup of the same population found that these same abnormalities on the NNNS were associated with maternal depression and that prenatal methamphetamine exposure was not associated with additional neurodevelopmental differences (Paz et al, 2009). Persisting behavioral, neurologic, and rearing problems are reported in children exposed to both cocaine and opioids (Chasnoff et al, 1989; Hunt et al, 2008). No significant differences in mean developmental scores were noted in a group of children exposed to cocaine plus polydrugs compared to a group without drug exposure (Chasnoff et al, 1992). Other investigators reported no differences between infants who have and those who have not been exposed to cocaine in mean cognitive, psychomotor, or language quotients at age 36 months (Kilbride et al, 2000). A recent report following a cohort of opioid exposed infants documented lower Bayley mental development index (MDI) scores at 18 months old, lower Stanford Binet Intelligence Scale scores at 3 years, and decreased scores on the Vineland Social Maturity Scale and Reynell Language Scale at 3 years (Hunt et al, 2008). However, a similar evaluation at 3 years showed no mental, motor, or behavioral deficits after controlling for birthweight and environmental risk factors (Messinger et al, 2004). The neurodevelopmental problems among children exposed to cocaine may occur either from a direct teratogenic drug effect during gestation or from the effects of the social environment in which the developing infant is reared. Singer et al (2002) reported that cocaine-exposed infants are twice as likely to have significant cognitive but not motor delays at 2 years, and they demonstrated a downward trend in mean developmental scores by 2 years, which is consistent with a deleterious effect of the environment, parental stimulation, socioeconomic status, or possibly other, indirect effects of drugs on the developing CNS. Other studies have not consistently demonstrated this association (Frank et al, 2001). In contrast, the Maternal Lifestyle Study, after controlling for covariates, found no differences in Bayley scores between cocaine-exposed, 127 opioid-exposed, and control infants for the first 3 years of life (Shankaran et al, 2007), although subtle effects on cognitive subscales were reported at 3 years (Messinger et al, 2004). Other investigators have shown normal development for opioid-exposed infants during the first 2 years of life after data have been controlled for socioeconomic status and birthweight (Bauer, 1999). Using regression analysis, researchers have shown that the amount of prenatal care obtained by the mother and the postnatal home environment were more predictive of the infant’s future intellectual performance. Conversely, the amount of maternal opioid use during pregnancy was not found to be predictive. Ornoy et al (2001) have shown that children exposed to heroin but adopted at an early age performed better on intelligence testing than did opioid-exposed infants who were raised in the homes of their biologic parents. The same investigators found a higher rate of attention deficit disorders in children exposed to opioids regardless of home environment, with the highest incidence in children who were raised in the homes of their biologic parents. Numerous studies point also to the importance of the home environment in optimizing child development. Variable outcomes of these studies may depend on the amount and type of drug use, or on other covariates such as nutritional status, poverty, psychosocial problems, and parental educational level. With the lack of major neurologic deficits, recent studies have focused on neurologic soft signs, nonfocal signs with no localized findings that are more often associated with cognitive deficits, and increased prevalence of attention deficit hyperactivity disorder and behavior problems. Neurologic soft signs including speech, balance, coordination, and tone were markedly stable over a 1-year period in 6- to 9-year-olds. Soft signs are a marker for increased risk of cognitive and psychiatric problems. In the Maternal Lifestyle Study, 23.5% of children exposed to cocaine had more than two soft neurologic signs, similar to rates in the comparison group. Soft signs were more common in infants born weighing less than 1500 g. Both cocaine and alcohol use significantly increased the incidence of soft signs in exposed infants weighing less than 1500 g, and this effect persisted after controlling for other substances used, birthweight, sex, and race (Shankaran et al, 2007). Childhood behavior problems were more common in infants with heavy cocaine exposure than in those with no or some cocaine exposure. Prenatal tobacco and alcohol exposure were also significantly associated with behavioral problems until 7 years, with a significant dose-response relationship (Shankaran et al, 2007). Evaluation at 7 years demonstrated behavior problems in cocaine- and substance-exposed infants that suggested direct effects resulting in neurobehavioral dysregulation, which was tracked through serial assessments beginning at an early age. This work on a model of early abnormalities predictive of later behavioral problems may allow for early identification and possible prevention of later behavioral problems (Lester et al, 2009). Follow-up studies have been reported at 9 to 15 years. Cognitive outcomes at 10 years were not abnormal in opioid-exposed children (Shankaran et al, 2007). Cocaine-exposed children were more likely to 128 PART III Maternal Health Affecting Neonatal Outcome be referred for special education services in school than unexposed children, with an estimated additional cost to the United States of $25,248,384 per year (Shankaran et al, 2007). Prenatal marijuana exposure is reported to be associated with deficiencies in executive function in 9 to 12 year olds (Fried and Smith, 2001; Fried et al, 1998), and recent reports have found that prenatal marijuana exposure has a significant effect on school-age intellectual performance (Goldschmidt et al, 2004, 2008; ­Richardson et al, 2002). In methamphetamine-exposed infants, neurodevelopmental abnormalities have been described to persist as late as 14 years (Cernerud et al, 1996). Intellectual capacity does not appear to be diminished among exposed infants. These children are described as exhibiting disturbed behavior, including hyperactivity, aggressiveness, and sleep disturbances. Eriksson et al (2000) reported that neurobehavioral abnormalities appear to be associated with the extent and duration of fetal exposure and with the severity of head growth restriction. In this study, children with the most severe problems were those born to mothers who abused amphetamines throughout pregnancy and were reared in homes with an addicted parent. Alterations in growth have been reported after prenatal exposure, with striking gender differences (Cernerud et al, 1996). Drug-exposed boys in Sweden were taller and heavier, and girls were smaller and lighter, than national standards. This finding suggests that fetal amphetamine exposure affects the onset of puberty and amphetamines may interfere with neurodevelopment of the adenohypophysis. Children of amphetamine abusers appear to be at high risk for social problems, including abandonment, abuse, and neglect. In two Swedish studies, only 22% of 10-year-old children who had been exposed to amphetamine in utero remained in the care of their biologic mothers, whereas 70% were in foster care (Cernerud et al, 1996; Eriksson and Zetterstrom, 1994). Considering all the data, it is still difficult to create a coherent picture and provide a prognosis for a newborn after gestational substance exposure. A recent report of volumetric magnetic resonance imaging in thirty five 12-year-old children exposed to cocaine in utero found smaller total parenchymal volumes, lower cortical gray matter volumes, and smaller head circumferences with prenatal substance exposure. The decreases were statistically significant only for prenatal cigarette exposure and for infants exposed to all substances studied (cocaine, tobacco, marijuana and alcohol) (Rivkin et al, 2008). As in other studies, exposure to multiple substances clearly has detrimental effects on the developing brain. SUMMARY The magnitude of observed perinatal outcomes after illicit maternal substance use pales in comparison to the established health and developmental risks associated with tobacco and alcohol exposure (Schempf, 2007). The greatest impact of illicit substance use may be the increased postnatal risks of neglect, maltreatment, and disruptions in the home environment. Health policy must be directed at reducing all these complex factors associated with perinatal substance abuse. SUGGESTED READINGS American Academy of Pediatrics Committee on Substance Abuse and Committee on Children with Disabilities: Fetal alcohol syndrome and alcohol-related neurodevelopmental disorders, Pediatrics 106:358-361, 2000. Burgos AE, Burke BL: Neonatal abstinence syndrome, NeoReviews 10:e222, 2009. Chasnoff IJ, McGourty RF, Bailey GW, et al: The 4P’s Plus screen for substance use in pregnancy: clinical application and outcomes, J Perinat 25:368-374, 2005. Jones HE, Martin PR, Heil SH, et al: Treatment of opioid-dependent pregnant women: clinical and research issues, J Subst Abuse Treat 35:245-259, 2008. Messinger DS, Bauer CR, Das A, et al: The Maternal Lifestyle Study: cognitive, motor, and behavioral outcomes of cocaine-exposed and opiate-exposed infants through three years of age, Pediatrics 113:1677-1685, 2004. Mitchell EA, Millerad J: Smoking and the sudden infant death syndrome, Rev Environ Health 21:81-103, 2006. Moses-Kolko EL, Bogen D, Perel J, et al: Neonatal signs after late in utero exposure to serotonin reuptake inhibitors, literature review and clinical applications, JAMA 293:2372-2383, 2005. Rogers JM: Tobacco and pregnancy: overview of exposure and effects, Birth Defects Res C Embryol Today 84:1-15, 2008. Schempf AH: Illicit drug use and neonatal outcomes: a critical review, Obstet Gynecol Surv 62:749-757, 2007. Shankaran S, Lester BM, Das A, et al: Impact of maternal substance use during pregnancy on childhood outcome, Semin Fetal Neonatal Med 12:143-150, 2007. Smith LM, LaGasse LL, Derauf C, et al: The infant development, environment, and lifestyle study: effects of prenatal methamphetamine exposure, polydrug exposure, and poverty on intrauterine growth, Pediatr 118:1149-1156, 2006. Thompson BL, Levitt P, Stanwood GD: Prenatal exposure to drugs: effects on brain development and implications for policy and education, Nat Rev Neurosci 10:303-312, 2009. Complete references and supplemental color images used in this text can be found online at www.expertconsult.com P A R T I V Labor and Delivery C H A P T E R 13 Antepartum Fetal Assessment Christian M. Pettker and Katherine H. Campbell A primary objective of obstetric care is the assessment and prevention of adverse fetal and neonatal outcomes. Maternal care is an integral step toward this goal. Optimization of the maternal state, through careful monitoring and treatment of chronic conditions such as diabetes or hypertension or acute states like preeclampsia or preterm labor, is one important facet of care to achieve desirable perinatal outcomes. Monitoring and management of the fetus, although a more obvious step toward this goal, are somewhat less straightforward. Fetal assessment demands a view into the intrauterine environment, which is somewhat inaccessible. Our ability to gain access to this space to gauge the needs and health of the fetus has improved dramatically with the developments in technology, as well as the increased understanding of fetal physiology over the past 50 years. As a result, perinatal morbidity and mortality have decreased substantially (Figure 13-1). In general, antepartum fetal assessment encompasses the screening and diagnosis of fetal disorders and fetuses that are at risk. Selecting appropriate patients at risk for adverse perinatal events can enhance the prediction of these events, although some tests may be appropriate even for a low-risk population. The assessment may allow for certain therapeutic options—often, timely delivery—to prevent fetal harm. The overall goal of these efforts is to reduce perinatal mortality, although the reduction of morbidities such as cerebral palsy or preventable birth injury is intertwined with this objective. In antenatal assessment in the third trimester, the prediction and detection of fetal acidemia and hypoxemia form a central principle underlying these efforts. It is important to make the distinction between antepartum and intrapartum fetal assessment. The latter is specifically related to monitoring the fetus during labor. The nature of labor affords certain advantages (e.g., dilation allows blood samples from the fetus) and restrictions (the lack of fluid after rupture of membranes creates difficulties for ultrasound examination) that do not occur in the antenatal period. As a result, this chapter focuses only on events and assessment preceding labor. GENERAL PRINCIPLES PRINCIPLES OF TESTING Many of the tests used for antepartum fetal assessment are screening tests that will lead to further testing allowing for diagnosis and decision-making; therefore it is important to note the principles guiding such tests. The outcome, principally perinatal morbidity and mortality, is a significant burden to both the individual and the overall health care system. The primary tools for assessment, ultrasound examination, and fetal heart rate monitoring, are generally easy, safe, and acceptable to patients. Screening has the potential to allow important and timely interventions, such as antenatal steroid administration or delivery. The predominant difficulty with fetal testing comes in the unproven utility of testing to improve outcomes. Furthermore, some tests, such as the nonstress test (NST), have high false-positive rates; therefore, when used as a diagnostic test (e.g., to decide on delivery), they can lead to the overuse of interventions. The specificity and sensitivity of the tests vary and the critical step to enhancing test performance is patient selection. The utility of fetal surveillance involves the judicious application of the tests in patients with specific risk profiles. FETAL PHYSIOLOGY AND BEHAVIOR The first trimester (≤14 weeks’ gestation) is mainly a time of system development and organogenesis. The ­hyperplastic enlargement during the first 11 weeks produces standard rates of growth, with deviation being rare. At the completion of the first trimester, the major organ ­systems have developed, allowing the opportunity during the second trimester to assess for anomalies in development. The second and third trimesters involve maturation of these systems. Because antenatal fetal assessment is primarily concerned with the prediction or detection of fetal hypoxemia and acidemia, the integration of the neurologic and cardiovascular systems, particularly as reflections of fetal acid-base status, is the cornerstone of this assessment. We are thus able to monitor the manifestations of hypoxemia and ­acidemia as shown by neurologic and cardiovascular changes. TECHNOLOGY The technology underpinning fetal assessment is ultrasound. Fetal heart rate monitoring during the antepartum period depends on a Doppler cardiogram; movements of the fetal heart, in particular the sounds of the valves, are detected by this monitor. The time between the beats is translated into a heart rate, which is then graphically represented on a chart over time. This process produces the 129 130 PART IV Labor and Delivery 35 White FMR Black FMR Fetal deaths per 1000 births 30 25 20 15 10 5 0 1950 1960 1970 1980 1985 1988 1989 1990 1991 1992 1993 1994 1995 1996 Year Cumulative risk of fetal death (percent) A 100 80 60 40 20 0 28–31 32–36 37–39 40–42 42+ Weeks’ gestation B FIGURE 13-1 Fetal mortality rate (FMR) in the United States, 1959 to 1996. (Adapted from McIlwaine GM, Dunn FH, Howat RC, et al: A routine system for monitoring perinatal deaths in Scotland, Br J Obstet Gynaecol 92:9-13, 1985.) fetal heart rate monitoring strip that becomes the NST or contraction stress test (CST). Contemporary ultrasound technology involves a wide array of features, including B-mode (basic imaging), M-mode (mapping the movement of structures over time), pulsed Doppler (demonstrating flow velocity in a particular area, such as a vessel), color Doppler (showing intensity and direction of flow through shades of red and blue), and power Doppler (a more sensitive form of colorized Doppler). INDICATIONS AND TIMING Most fetal testing protocols involve a stepwise approach, and the first step is the selection of the appropriate patient. Suggested assessments for low-risk pregnancies include one ultrasound examination for dating and one for the basic anatomic survey. Prenatal risk assessments for chromosomal disorders, such as first-trimester risk assessment with maternal serum analysis and fetal nuchal translucency assessment or the second-trimester maternal quadruple serum screen, are additional options. Whereas up to 30% of perinatal morbidity may occur in low-risk patients, routine fetal testing beyond that described previously in a lowrisk pregnancy is an ineffective use of resources. High-risk pregnancies are those at greater peril for perinatal morbidity and mortality. These pregnancies often have more justification for targeted or detailed anatomic ultrasound examinations and for regular assessment of fetal growth or heart rate assessment. Common conditions requiring increased fetal surveillance are shown in Box 13-1. Pregnancy dating should be confirmed at the earliest possible moment, and fetal anatomic screening is best accomplished in the second trimester, specifically at 18 to 20 weeks’ gestation, when visualization of the anatomic features is adequate. However, standards for the timing of antepartum fetal assessment to survey for fetal compromise do not exist. Certainly assessment with NST or biophysical profiles would have little utility before viability (approximately 24 weeks’ gestation). Guidelines for initiating fetal testing for specific indications are largely based on the risk of fetal loss at a particular gestational age. CHAPTER 13 Antepartum Fetal Assessment BOX 13-1 T  ypical Indications for Antepartum Fetal Assessment: High-Risk ­Pregnancies FETAL Abnormal fetal testing, fetal distress ll Anatomic anomaly ll Decreased fetal movement ll Heart block ll Intraamniotic infection ll Intrauterine growth restriction ll Multiple gestation ll Oligohydramnios ll MATERNAL-FETAL Abruptio placenta ll Anemia, fetal (e.g., parvovirus, Rh alloimmunization, NAIT) ll Abnormal serum screening (low PAPP-A, high MSAFP) ll Placenta previa, vasa previa ll Postterm or postdates premature rupture of fetal membranes ll Threatened preterm delivery ll MATERNAL Advanced maternal age ll Cardiac disease (severe) ll Cholestasis of pregnancy ll Diabetes, gestational ll Diabetes, pregestational ll Hemoglobinopathies ll HIV (receiving medication) ll Hypertension, chronic ll Hypertension, gestational ll History of IUFD ll Obesity ll Preeclampsia ll Pulmonary disease (severe) ll Renal disease ll Seizure disorder ll Substance abuse ll Systemic lupus erythematosis ll Thyroid disease ll Thrombophilia or thromboembolic disease ll IUFD, Intrauterine fetal demise; MSAFP, maternal serum alpha-fetoprotein; NAIT, neonatal ­alloimmune thrombocytopenia; PAPP-A, pregnancy associated plasma protein FETAL ASSESSMENT IN LOW RISK PREGNANCIES ULTRASOUND: PREGNANCY DATING The estimated date of delivery is defined at the beginning of pregnancy based on the best available information, including menstrual history, ultrasound data, and assisted reproduction technology. The median duration of a singleton pregnancy is 280 days (40 weeks) from the first day of the last menstrual period or 266 days (38 weeks) from the time of ovulation. Term is defined as 37 to 42 weeks’ (259 to 294 days) gestation. Given that the preterm and postterm periods are associated with increased risks to the fetus and newborn, pregnancy dating provides an approximate expectation for the completion of the pregnancy and serves as a basis for the efficient and appropriate use of fetal surveillance, testing, and treatment. Accurate pregnancy dating by ultrasound has been associated with 131 reduced diagnoses of growth restriction (Waldenstrom­ et al, 1992), reduced use of tocolysis for preterm labor (LeFevre et al, 1993), and a reduced need to intervene in postterm pregnancies (Neilson, 2000). In a spontaneous pregnancy in a woman with regular cycles and normal menstrual periods, the last menstrual period is often an accurate way of dating a pregnancy. Menstrual dating is less accurate in women who are taking oral contraceptives, were recently pregnant, or have irregular periods or intermenstrual bleeding. In these cases, and others in whom there is uncertainty, ultrasound dating in the first trimester is accurate and effective. A fetal pole may be seen beginning at 5 weeks’ gestation, and the fetal heartbeat should be visualized at 6 to 8 weeks’ gestation. In the first trimester, measurement of the crown-rump length is accurate to within 3 to 5 days; therefore this measurement should take priority in dating a pregnancy when the timing of the last menstrual period suggests a gestational age outside this range of variation (Drumm et al, 1976; Robinson and Fleming, 1975). A first-trimester ultrasound examination is indicated to confirm an intrauterine pregnancy (i.e., exclude ectopic pregnancy), confirm fetal viability, document fetal number, estimate gestational age, and evaluate the maternal pelvis and ovaries. In the second trimester, ultrasound dating is less accurate, but can nonetheless be helpful. Measurement of the biparietal diameter (BPD) of the fetal head, the most accurate parameter, can be accurate to within 7 to 10 days (Campbell et al, 1985; Waldenstrom et al, 1990). The BPD is also a parameter of choice because it is less affected by chromosomal anomalies, in particular Down syndrome (Cuckle and Wald, 1987). Usually in the second or third trimesters, several biometric measurements—such as cerebellar distance, femur and humerus length, and abdominal circumference—are recorded and a computerized algorithm can generate an estimated gestational age. ULTRASOUND: SECOND AND THIRD TRIMESTERS Perinatal ultrasound examination in the second and third trimesters can be classified broadly into three types: the basic or standard examination, the specialized (detailed) examination, and the limited examination. The standard examination (level I) includes the determination of fetal number, fetal viability, fetal position, gestational age, placental location, amniotic fluid volume, the presence or absence of a maternal pelvic mass, and the presence of gross fetal malformations (American College of Obstetricians and Gynecologists, 2009). Most pregnancies can be evaluated adequately by this basic examination. If the patient’s history, physical examination, or basic ultrasound examination suggest the presence of a fetal malformation, a specialized examination (level II) should be performed by a sonographer who is skilled in fetal evaluation. During a detailed ultrasound, which is best performed at 18 to 20 weeks’ gestation, fetal structures are examined in detail to identify and characterize any fetal malformation. In addition to identifying structural abnormalities, a specialized ultrasound examination can identify sonographic markers of fetal aneuploidy. In some situations, a limited 132 PART IV Labor and Delivery examination may be appropriate to answer a specific clinical question (such as fetal viability, amniotic fluid volume, fetal presentation, placental location, or cervical length) or to provide sonographic guidance for an invasive procedure (such as amniocentesis). Current debate centers on who should undergo sonographic examination and what type of evaluation these patients should have. Advocates of routine sonography cite several advantages of universal ultrasound evaluation, including more accurate dating of pregnancy and earlier and more accurate diagnosis of multiple gestation, structural malformations, and fetal aneuploidy (discussed later in ultrasound section). Opponents of routine sonographic examination argue that it is an expensive screening test ($100 to $250 for a standard examination) and that the cost is not justified by published research, which suggests that routine ultrasound examinations do not significantly change perinatal outcome (Crane et al, 1994; Ewigman et al, 1993; LeFevre et al, 1993) Second-trimester ultrasound examination is indicated in patients with uncertain dating, uterine size larger or smaller than expected for the estimated gestational age, a medical disorder that can affect fetal growth and development (e.g., diabetes, hypertension, collagen vascular disorder), a family history of an inherited genetic abnormality, or a suspected fetal malformation or growth disturbance (American College of Obstetricians and Gynecologists, 2009). In the United States, most patients undergo a standard examination at 18 to 20 weeks’ gestation to screen for structural defects. An understanding of normal fetal physiology is critical to the diagnosis of fetal structural anomalies. For example, extraabdominal herniation of the midgut into the umbilical cord occurs normally in the fetus at 8 to 12 weeks’ gestation and can be misdiagnosed as an abdominal wall defect. Placental location should be documented with the bladder empty, because overdistention of the maternal bladder or a lower uterine contraction can give a false impression of placenta previa. If placenta previa is identified at 18 to 22 weeks’ gestation, serial ultrasound examinations should be performed to follow placental location. Only 5% of placenta previa identified in the second trimester will persist to term (Zelop et al, 1994). The umbilical cord should also be imaged, and the number of vessels, placental insertion, and insertion into the fetus should be noted. The indications for third-trimester ultrasound examination are similar to that for second-trimester ultrasound. Fetal anatomy survey examinations and estimates of fetal weight become less accurate as gestational age increases, especially in obese women or in pregnancies complicated by oligohydramnios. However, fetal biometry and an anatomic survey should still be performed, because certain fetal anomalies, such as achondroplasia, will become evident for the first time later in gestation. FETAL MOVEMENT COUNTING Fetal movement (quickening) is typically perceived by the mother at 16 to 22 weeks’ gestation. Fetal hypoxemia is typically associated with a reduction in fetal activity; the fetus is essentially conserving energy and oxygen for vital activities. A typical procedure for fetal movement counting consists of having the patient record the interval taken to feel 10 fetal movements, usually after a meal when the fetus is more active. If 10 movements are not detected in 1 hour, further testing is often recommended. The data supporting fetal movement counting are mixed. A large international cluster randomized trial involving more than 68,000 patients demonstrated no benefit (Grant et al, 1989), and a Cochrane analysis found insufficient evidence to support this technique to prevent stillbirth (Mangesi and Hofmeyr, 2007). Fetal movement counting represents a low-technology screening test that can be applied easily to all pregnancies. Although its effectiveness in improving perinatal outcomes is debatable, it can be used as a costeffective first line strategy FETAL ASSESSMENT IN HIGH-RISK PREGNANCIES CARDIOTOCOGRAPHY Cardiotocography is the visual representation of fetal heart rate and uterine contractions. Fetal heart rate has been recognized as an important indicator of fetal status since the nineteenth century, with Lejumeau Kergaradec of Switzerland being credited with the first accounts of direct fetal auscultation and the uterine soufflé in 1821. Fetal heart rate monitoring is based on the principle that the fetal neurologic system, through its afferent and efferent networks, serves as a key mediator to demonstrate fetal well-being. Oxygenation, acidemia, and other vital functions are monitored by peripheral chemoreceptors and baroreceptors, which provide input on fetal status through afferent neurologic networks to the central nervous system (CNS). This information is processed by the CNS, and signals are conducted through efferent networks to produce peripheral changes, particularly to the heart via direct parasympathetic vagal neurons, direct sympathetic signals, or indirect sympathetic stimulation of catecholamine release. In this way, fetal cardiac activity can be seen as a surrogate for fetal oxygenation and acid-base status. For many years, assessment of the fetal heart rate was limited to the fetoscope, a direct stethoscope attributed to Adolphe Pinard in 1876. In 1957, Orvan Hess and Ed Hon at Yale University introduced electronic fetal heart rate monitoring as a window into the status of the fetus (Hon and Hess, 1957). This technology relied on direct monitoring through a scalp electrode; only years later would Doppler technology allow cardiac signals to be detected noninvasively. Fetal heart rate monitoring became a tool for fetal assessment as it was recognized that certain fetal heart rate patterns were associated with fetal compromise and poor fetal outcomes. The basic elements of a fetal heart rate strip are baseline, variability, accelerations, and decelerations. A baseline of 110 to 160 beats per minute is normal. Variability is determined by the irregular fluctuations in amplitude and frequency in the baseline, and variability of fewer than 6 beats per minute is often abnormal. Accelerations are classified as visually apparent abrupt increases that peak at 15 beats per minute or more above the baseline that last 15 seconds or longer. Fetal movements often coincide with fetal heart rate (FHR) accelerations. Finally decelerations, CHAPTER 13 Antepartum Fetal Assessment 133 TABLE 13-1 Interpretation of Antepartum Cardiotocography Term Characteristic Description Baseline Definition Mean fetal heart rate, rounded to increments of 5 beats/min (e.g. 140, 145); need baseline duration of ≥2 min during a 10-min segment, between periodic or episodic changes, to determine baseline Bradycardia <110 beats per minute for >10 min Tachycardia >160 beats per minute for >10 min Definition Fluctuations of the baseline heart rate; measured from peak to trough Absent Undetectable Minimal Undetectable to ≤5 beats/min Moderate 6-25 beats/min Marked >25 beats/min Definition Abrupt increase ≥15 beats/min lasting ≥15 s Variability Acceleration Deceleration Prolonged ≥2 min and <10 min (≥10 min is a baseline change) Definition Decreases in the fetal heart rate Variable Abrupt decrease onset to nadir <30 s; decrease ≥15 beats/min lasting ≥5 s to <2 min Early Late Prolonged Contractions Gradual decrease onset to nadir ≥30 s with contraction Gradual decrease onset to nadir ≥30 s; nadir of deceleration occurring after peak of contraction Decrease ≥15 beats/min lasting ≥2 min, but <10 min (≥10 min is a baseline change) Recurrent Occur with ≥50% of uterine contractions in any 20-min window Intermittent Occur with <50% of uterine contractions in any 20-min window Considerations Frequency, duration, intensity, and relaxation Normal ≤5 contractions per 10 minutes averaged over a 30-min window Tachysystole >5 contractions per 10 minutes averaged over a 30-min window; should always be qualified as to the presence or absence of associated FHR decelerations often classified as early, variable, or late, are decreases in the fetal heart rate that have specific pathologic and physiologic associations. Although primarily focused on intrapartum monitoring, the 2008 National Institute of Child Health and Human Development workshop report on fetal monitoring provides an excellent summary of the nomenclature and interpretation involved (Table 13-1) (Macones et al, 2008). NONSTRESS TEST A normal result of an NST is defined as a 20-minute fetal heart rate tracing that contains two heart rate accelerations lasting 15 seconds or longer that peak 15 beats or more above the baseline. Often this is called a reactive NST (Figure 13-2). Modifications are made in reference to gestational age. NSTs for fetuses at less than 32 weeks’ gestation are often considered reactive if the acceleration is 10 beats per minute or more above the baseline and lasts for at least 10 seconds. Furthermore, to account for the periodicity of 20- to 30-minute sleep cycles in the fetus, an NST that is not reactive over the first 20 minutes may be continued an additional 20 to 40 minutes. A nonreactive NST or an NST with specific abnormalities (e.g., high or low baseline, decelerations) should be followed by a biophysical profile (BPP). It is important to note that some abnormal states, such as a fetal CNS abnormality or maternal drug ingestion, may contribute to a nonreactive NST. In these cases, ultrasound examination may provide appropriate information to determine the diagnosis or required management. Falsely reassuring NSTs occur at a rate of 3 to 5 per 1000 tests, although this does not account for a baseline rate of unpreventable fetal deaths (Freeman et al, 1982b). The difficulty with the NST really lies in its lack of specificity for fetal death or compromise; the false-positive rate may be as high as 50% (Freeman et al, 1982b). The rather modest false-negative rate is likely because of the NST being a measurement of short-term hypoxemia. Indeed, longer-term fetal status can be measured through amniotic fluid assessment, because the amniotic fluid is correlated with fetal urinary output, which is a surrogate for renal perfusion. When combined with an assessment of amniotic fluid level, the false-negative rate of the NST is reduced to 0.8 per 1000, although a 60% false-positive rate remains (Miller et al, 1996). Indeed, when combined with the NST and amniotic fluid assessment—sometimes known as the modified biophysical ­profile—reduces the risk of fetal death to negligible levels in high-risk populations (Clark et al, 1989). For these reasons, the NST combined with amniotic fluid assessment is a modality of choice for monitoring the high-risk pregnancy. CONTRACTION STRESS TEST The CST assesses the fetal heart rate response in the presence of contractions. This test improves on the specificity and sensitivity of the NST by assessing the fetal response to stress. In fact, the CST preceded the NST, although the NST became more favorable because of fewer contraindications, ease of administration, and reduced time and supervision necessary. Compared with the NST, there is a much lower incidence of falsely reassuring tests (0.4 per 1000), representing an eightfold reduction in the risk of fetal loss in one study (Freeman et al, 1982a). 134 PART IV Labor and Delivery FIGURE 13-2 A reassuring nonstress test. Note two fetal heart rate accelerations exceeding 15 beats/min and lasting at least 15 seconds during the monitoring period. FIGURE 13-3 A contraction stress test. The fetal heart rate is plotted above the uterine contraction signal. Note the late deceleration after a contraction; this is a positive, or abnormal, test result. Contractions are stimulated by the administration of intravenous oxytocin or through maternal nipple stimulation. Of course, the CST is contraindicated in patients in whom contractions should not be provoked, such as threatened preterm delivery or preterm premature rupture of membranes, prior classical cesarean delivery, or placenta previa. A minimum of three contractions over a 10-minute period of continuous fetal heart rate assessment are necessary for a satisfactory test interpretation. An unsatisfactory test should be followed by continued testing with a modification of the mode of contraction stimulation. A negative (i.e., normal) test result demonstrates no late decelerations, whereas a positive test result shows late decelerations after 50% or more of contractions (Figure 13-3). A positive test result requires immediate further testing or evaluation, if not delivery. An equivocal test demonstrates late decelerations with less than 50% of contractions and requires further testing or monitoring. A test that encompasses a hyperstimulatory contraction pattern (e.g., five contractions within 10 minutes or contractions lasting longer than 90 seconds) is also considered equivocal and requires further testing. ULTRASOUND Although routine sonography for low-risk pregnant women is controversial, few would disagree that the benefits far outweigh the costs for high-risk patients. Given the higher risk for fetal complications such as anatomic anomalies or growth disturbances, a specialized examination is performed between 18 and 20 weeks’ gestation in most high-risk pregnancies. Additional ultrasound modalities are also available, including fetal echocardiography, three-dimensional (3D) sonography, and Doppler. Cardiac anomalies are the most common major congenital defects encountered in the antepartum period. A four-chamber view of the heart at the time of fetal anatomy survey at 18 to 20 weeks’ gestation will detect only 30% of congenital cardiac anomalies, although the detection rate can be increased to approximately 60% to 70% if the outflow tracts are adequately visualized (Kirk et al, 1994), but this still leaves 30% to 40% of all congenital cardiac anomalies undiagnosed. For this reason, fetal echocardiography should be performed by a skilled and experienced sonologist at 20 to 22 weeks’ gestation in all pregnancies at high-risk of a fetal cardiac CHAPTER 13 Antepartum Fetal Assessment anomaly; this includes pregnancies complicated by pregestational diabetes mellitus, a personal or family history of congenital cardiac disease regardless of the nature of the lesion or whether it has been repaired), maternal drug exposure (e.g., lithium and paroxitene) (Bérard et al, 2007), and pregnancies conceived by in vitro fertilization, but not if the pregnancy was conceived using clomiphene citrate or ovarian stimulation or intrauterine insemination alone (Olson et al, 2005). Compared with standard two-dimensional (2D) ultrasound, 3D ultrasound (or four-dimensional if fetal movements are included) allows for visualization of fetal structures in all three dimensions concurrently for the improved characterization of complex fetal structural anomalies and for storage of scanned images with 3D reconstruction at a later date or remote location (telemedicine). Unlike 2D ultrasound, 3D images are greatly influenced by fetal movements and are subject to more interference from structures such as fetal limbs, umbilical cord, and placental tissue. Because of movement interference, visualization of the fetal heart with 3D ultrasound is suboptimal. In addition to rapid acquisition of images that can be later reconstructed and manipulated, 3D ultrasound has other potential advantages: ll Surface rendering mode can provide clearer images of many soft tissue structures. Such images can improve the diagnosis of certain fetal malformations, especially craniofacial anomalies (cleft lip and palate, micrognathia, ear anomaly, facial dysmorphism, club foot, finger and toe anomalies), intracranial lesions, spinal anomalies, ventral wall defects, and fetal tumors. ll 3D ultrasound may be useful in early pregnancy by providing more accurate measurements of the gestational sac, yolk sac, and crown-rump length. It may also allow for a more accurate midsagittal view of the fetus for measuring nuchal translucency. ll 3D ultrasound can also be used to measure tissue volume. Preliminary data suggest that the assessment of cervical volume may predict the risk of cervical insufficiency (Rovas et al, 2005), and measurement of placental volume in the first trimester may predict fetuses at risk of intrauterine growth restriction (IUGR) (Schuchter et al, 2001). Despite these advantages and the fact that 3D ultrasound has been available since the early 1990s, it has yet to live up to its promises. Although 3D ultrasound is unlikely to replace standard 2D imaging in the near future, it is a valuable complementary modality in obstetric imaging. As the technology improves, it is likely that perinatal ultrasound will evolve to resemble computed tomography and magnetic resonance imaging. GROWTH ASSESSMENT Normal fetal growth is a critical component of a healthy pregnancy and the subsequent long-term health of the child. A systematic method of examination of the gravid abdomen was first described by Leopold and Sporlin (1894). Although abdominal examination has several limitations, particularly in the setting of maternal obesity, multiple pregnancy, uterine fibroids, or polyhydramnios, 135 it is safe, free, and well tolerated and may add valuable information to assist in antepartum management. Palpation is divided into four separate Leopold maneuvers. Each maneuver is designed to identify specific fetal landmarks or to reveal a specific relationship between the fetus and mother. For example, the first maneuver involves measuring fundal height. The uterus can be palpated above the pelvic brim at approximately 12 weeks’ gestation. Thereafter, fundal height should increase by approximately 1 cm per week, reaching the level of the umbilicus at 20 to 22 weeks’ gestation. Between 20 and 32 weeks’ gestation, the fundal height (in centimeters, from the superior edge of the pubic symphysis) is approximately equal to the gestational age (in weeks) in healthy women of average weight with an appropriately grown fetus. However, there is a wide range of normal fundal height measurements. One study has shown a 6-cm difference between the 10th and 90th percentiles at each week of gestation after 20 weeks (Belizan et al, 1978). Moreover, fundal height is maximal at approximately 36 weeks’ gestation, at which time the fetus drops into the pelvis in preparation for labor and the fundal height decreases. For these reasons, reliance on fundal height measurements alone will fail to identify more than 50% of fetuses with IUGR (Gardosi and Francis, 1999). Serial fundal height measurements by an experienced obstetric care provider are more accurate than a single measurement, and they will lead to an improved diagnosis of fetal growth restriction with reported sensitivities as high as 86% (Belizan et al, 1978). If the clinical examination is not consistent with the stated gestational age, an ultrasound examination is indicated to confirm gestational age and to establish a more objective measure of fetal growth. Ultrasound examination may also identify an alternative explanation for the discrepancy, such as multiple pregnancy, polyhydramnios, oligohydramnios, fetal demise, or uterine fibroids. For many years, obstetric sonography has used fetal biometry to define fetal size by weight estimation, although this approach has a number of key limitations. For example, regression equations used to create weight estimation formulas are derived primarily from cross-sectional data that rely on infants delivering within an arbitrary period of time after the ultrasound examination, and they assume that body proportions (i.e., fat, muscle, bone) are the same for all fetuses. Moreover, growth curves for healthy infants from 24 to 37 weeks’ gestation rely on data collected from pregnancies delivered preterm, which should not be regarded as normal pregnancies and are likely to be complicated by some element of uteroplacental insufficiency regardless of whether the delivery was spontaneous or iatrogenic. Despite these limitations, if the gestational age is well validated, the prevailing data suggest that prenatal ultrasound can be used to verify an alteration in fetal growth in 80% of cases and exclude abnormal growth in 90% of cases (Sabbagha, 1987). Sonographic estimates of fetal weight are commonly derived from mathematical formulas that use a combination of fetal measurements, especially the BPD, abdominal circumference (AC), and femur length (Hadlock et al, 1984). Whereas the BPD may be the most accurate indicator of gestational age in the second or third trimesters, fetuses gain weight in their abdomen, making the AC the 136 PART IV Labor and Delivery single most important measurement for fetal size. The AC is thus given more weight in these formulas. Unfortunately the AC is also the most difficult measurement to acquire, and a small difference in the AC measurement will result in a large difference in the estimated fetal weight (EFW). The accuracy of the EFW depends on a number of variables, including gestational age (in absolute terms, EFW is more accurate in preterm or IUGR fetuses than in term or macrosomic fetuses), operator experience, maternal body habitus, and amniotic fluid volume (measurements are more difficult to acquire if the amniotic fluid volume is low). Although objective, sonographic EFW estimations are not particularly accurate and have an error of 15% 20%, even in experienced hands (Anderson et al, 2007). Indeed, a sonographic EFW at term is no more accurate than a clinical estimate of fetal weight by an experienced obstetric care provider or the mother’s estimate of fetal weight if she has delivered before (Chauhan et al, 1992). Sonographic estimates of fetal weight must therefore be evaluated within the context of the clinical situation and balanced against the clinical estimate of fetal weight. Serial sonographic evaluations of fetal weight are more useful than a single measurement in diagnosing abnormal fetal growth. The ideal interval to evaluate fetal growth is every 3 to 4 weeks, with a minimum 10- to 14-day interval necessary to see significant differences. Because of the inherent error in fetal biometric measurements, more frequent ultrasound determinations of EFW may be misleading. Similarly, the use of population-specific growth curves, if available, will improve the ability of the obstetric care provider to identify abnormal fetal growth. For example, growth curves derived from a population that lives at high altitude, where the fetus is exposed to lower oxygen tension, will be different from those derived from a population at sea level. Abnormal fetal growth can be classified as insufficient (i.e., IUGR) or excessive (fetal macrosomia). The definition of IUGR has been a long-standing ­challenge for modern obstetrics. Distinguishing the healthy, constitutionally small-for-gestational-age fetus, defined as an EFW below the 10th percentile for a given week of gestation, from the nutritionally deprived, truly growth-restricted fetus has been particularly difficult. Fetuses with an EFW less than the 10th percentile are not necessarily pathologically growth restricted. Conversely, an EFW greater than the 10th percentile does not mean that an individual fetus has achieved its growth potential, and such fetuses may still be at risk of perinatal mortality and morbidity. As such, IUGR is best defined as either an EFW of less than the 5th percentile for gestational age in a well-dated pregnancy or an EFW of less than the 10th percentile for gestational age in a well-dated pregnancy with evidence of fetal compromise, such as oligohydramnios or abnormal umbilical artery Doppler velocimetry. Fetal growth restriction has traditionally been classified into asymmetric or symmetric IUGR. Asymmetric IUGR is characterized by normal head growth, but suboptimal body growth, and is seen most commonly in the third trimester. It is thought to result from a late pathologic event, such as chronic placental abruption leading to uteroplacental insufficiency, in an otherwise uncomplicated pregnancy and healthy fetus. In cases of symmetric IUGR, both the fetal head size and body weight are reduced, indicating a global insult that likely occurred early in gestation. Symmetric IUGR may reflect an inherent fetal abnormality (e.g., fetal chromosomal anomaly, inherited metabolic disorder, early congenital infection) or long-standing severe placental insufficiency caused by an underlying maternal disease (e.g., hypertension, pregestational diabetes mellitus, collagen vascular disorder). In practice, the distinction between asymmetric and symmetric IUGR is not particularly useful. Early and accurate diagnosis of IUGR coupled with appropriate intervention will lead to an improvement in perinatal outcome. If IUGR is suggested clinically and by ultrasound examination, thorough evaluations of the mother and fetus are indicated. Referral to a maternalfetal medicine specialist should be considered. Every effort should be made to identify the cause of IUGR and to modify or eliminate contributing factors. Up to 20% of cases of severe IUGR are associated with fetal chromosome abnormalities or congenital malformations, 25% to 30% are related to maternal conditions characterized by vascular disease, and a smaller proportion are the result of abnormal placentation. However, in a substantial number of cases (50% or more in some studies), the cause of the IUGR will remain uncertain even after a thorough investigation (Resnik, 2002). Fetal macrosomia is defined as an EFW (not birthweight) of 4500 g or greater, measured either clinically or by ultrasound, and is independent of gestational age, diabetic status, or actual birthweight (American College of Obstetricians and Gynecologists, 2000). Fetal macrosomia refers to a single cutoff EFW; this should be distinguished from the large-for-gestational age fetus, which is one in whom the EFW is greater than the 90th percentile for gestational age. By definition, 10% of all fetuses are large for gestational age at any given gestational age. Fetal macrosomia is associated with an increased risk of cesarean delivery, operative vaginal delivery, and birth injury to both the mother (including vaginal, perineal, and rectal trauma) and the fetus (orthopedic and neurologic injury) (American College of Obstetricians and Gynecologists, 2000, 2001; Kjos and Buchanan, 1999; Magee et al, 1993; O’Sullivan et al, 1973; Widness et al, 1985). Shoulder dystocia with resultant brachial plexus injury (Erb’s palsy) are a serious consequence of fetal macrosomia and are further increased in the setting of diabetes, because of the increased diameters in the upper thorax and neck of fetuses of mothers with diabetes. Fetal macrosomia can be determined clinically, by abdominal palpation using the Leopold’s maneuvers, or by ultrasound examination; these two techniques appear to be equally accurate (Watson et al, 1988). However, EFW measurements are less accurate in large (macrosomic) fetuses than in normally grown fetuses, and factors such as low amniotic fluid volume, advancing gestational age, maternal obesity, and the position of the fetus can compound these inaccuracies. Clinical examination has been shown to underestimate the birthweight by 0.5 kg or more in almost 80% of fetuses with macrosomia (Niswander et al, 1970). For these reasons, the prediction of fetal macrosomia is not particularly accurate, with a false-positive rate of 35% and a false-negative rate of 10% (Niswander et al, 1970; Watson et al, 1988). A number of alternative sonographic measurements have therefore been proposed CHAPTER 13 Antepartum Fetal Assessment 350 300 250 99th 200 95th AFI in an attempt to better identify the macrosomic fetus, including fetal AC alone, umbilical cord circumference, cheek-to-cheek diameter, and upper arm circumference; however, these measurements remain investigational and should not be used clinically. Despite the inaccuracy in the prediction of fetal macrosomia, an EFW should be documented either by clinical estimation or ultrasound examination in all women at high risk at approximately 38 weeks’ gestation. Suspected fetal macrosomia is not an indication for induction of labor, because induction does not improve maternal or fetal outcomes (American College of Obstetricians and Gynecologists, 2000). However, if the EFW is excessive, an elective cesarean delivery should be considered to prevent fetal and maternal birth trauma. Although controversy remains as to the precise EFW at which an elective cesarean delivery should be recommended, a suspected birthweight in excess of 4500 g in women with diabetes or 5000 g in women without diabetes is a reasonable threshold (American College of Obstetricians and Gynecologists, 2000, 2001, 2002). 137 Mean 5th 150 1st 100 50 0 16 18 20 22 24 26 28 30 32 34 36 38 40 42 Weeks gestation FIGURE 13-4 Amniotic fluid index (AFI; in mm) plotted against gestational age. The lines represent percentiles. AMNIOTIC FLUID ASSESSMENT Amniotic fluid plays a key role in the health and development of a growing fetus. Once considered an afterthought during the ultrasound examination of the fetus, evaluation of the amniotic fluid is now considered an integral part of ultrasound evaluation for fetal well-being. Amniotic fluid serves a number of important functions for the developing embryo and fetus. It provides cushioning against physical trauma; creates an environment free of restriction and or distortion, allowing for normal growth and development of the fetus; provides a thermally stable environment; allows the respiratory, gastrointestinal, and musculoskeletal tracts to develop normally; and helps to prevent infection (Hill et al, 1984). The chorioamnion acts as a porous membrane early in pregnancy, allowing the passage of water and solutes across the membrane; there is little contribution from the small embryo. As the pregnancy progresses into the late first trimester, the diffusion of fluid across the fetal skin occurs, increasing the volume of amniotic fluid. In the second half of the pregnancy, the main sources of amniotic fluid come from fetal kidneys and lungs. The primary sources for removal of fluid are from fetal swallowing and absorption into fetal blood perfusing the surface of the placenta. As more fluid is produced than is resorbed by the fetal-placental unit, the volume of amniotic fluid increases throughout the first 32 weeks of pregnancy (Figure 13-4). The volume peaks at approximately 32 to 33 weeks’ gestation, and at this gestational age equal amounts of fluid are produced and resorbed. After term, the amniotic fluid declines at a rate of 8% per week (Brace and Wolf, 1989). Because amniotic fluid plays a critical role in the normal development of a fetus, the assessment of amniotic volume is an essential component of the ultrasound evaluation for fetal well-being. Subjective estimates of the amniotic fluid volume have been validated, but two ultrasound measurements—amniotic fluid index (AFI) and maximum vertical pocket—have been developed to quickly and accurately assess the quantity of amniotic fluid surrounding the fetus. The AFI is a semiquantitative method for assessing the amniotic fluid volume with ultrasound. The gravid uterus is divided into four quadrants using the umbilicus, linea nigra, and external landmarks (Rutherford et al, 1987). The deepest amniotic fluid pocket is measured in each quadrant with the ultrasound transducer perpendicular to the floor. The four measurements are added together and the sum is regarded as the AFI. Pockets filled with umbilical cord or fetal extremities should not be used for generating the AFI (Hill, 1997). Researchers and clinicians have used a variety of measurements to define abnormalities in amniotic fluid volume. However, the normal range of the AFI most commonly used in clinical practice is 5 to 20 cm of fluid. Pregnancies with AFIs less than 5 cm can be described as having oligohydramnios, and pregnancies with measurements greater than 20 cm can be described as having polyhydramnios. The maximal vertical pocket is another semiquantitative method for assessing the fluid volume. The technique involves scanning the gravid uterus for the single deepest pocket of amniotic fluid that is free of umbilical cord and fetal parts and with the transducer perpendicular to the floor, measuring the pocket of fluid (Manning et al, 1981b). Currently this method is used mostly in multiple gestation pregnancies in which the AFI is not technically feasible. Oligohydramnios can be defined as a single measurement less than 2 cm. Polyhydramnios can be defined as a single measurement greater than 10 cm. BIOPHYSICAL PROFILE An NST alone might not be sufficient to confirm fetal well-being; in such cases, a biophysical profile (BPP) may be performed. The BPP refers to a sonographic scoring system performed over a 30- to 40-minute period designed to assess fetal well-being. The BPP was initially described for testing postterm fetuses, but has since been validated for use in both term and preterm fetuses (Manning et al, 138 PART IV Labor and Delivery 1981a, 1985, 1987; Vintzileos et al, 1983, 1987a, 1987b). Notably, BPP is not validated for use in active labor. The five variables described in the original BPP were: gross fetal body movements, fetal tone (i.e., flexion and extension of limbs), amniotic fluid volume, fetal breathing movements, and NST (summarized in Table 13-2) (Manning, 1989). More recently, however, BPP is interpreted without the NST. The individual variables of the BPP become apparent in healthy fetuses in a predictable sequence: fetal tone appears at 7.5 to 8.5 weeks, fetal movement at 9 weeks, fetal breathing at 20 to 22 weeks, and FHR reactivity at 24 to 28 weeks’ gestation. Similarly, in the setting of antepartum hypoxia, these characteristics typically disappear in the reverse order in which they appeared (i.e., FHR reactivity is lost first followed by fetal breathing, fetal movements, and finally fetal tone) (Vintzileos et al, 1987a). The amniotic fluid volume, which is composed almost entirely of fetal urine in the second and third trimesters, is not influenced by acute fetal hypoxia or acute fetal central nervous system dysfunction. Rather, oligohydramnios (decreased amniotic fluid volume) in the latter half of pregnancy and in the absence of ruptured membranes is a reflection of TABLE 13-2 Fetal Biophysical Profile Element Criterion (2 points for each element satisfied) Breathing ≥1 episode of breathing movements lasting 30 seconds Movement ≥3 discrete body or limb movements Tone ≥1 episode of active extension and flexion of limbs or trunk Amniotic fluid ≥1 pocket of amniotic fluid measuring ≥2 cm in two perpendicular planes Nonstress test ≥2 fetal heart rate accelerations lasting ≥15 seconds over 20 minutes chronic uteroplacental insufficiency, increased renal artery resistance leading to diminished urine output, or both (Oz et al, 2002); it predisposes to umbilical cord compression, thus leading to intermittent fetal hypoxemia, meconium passage, or meconium aspiration. Adverse pregnancy outcome (including nonreassuring FHR tracing, low Apgar score, and neonatal intensive care unit admission) is more common when oligohydramnios is present (Bochner et al, 1987; Morris et al, 2003; Oz et al, 2002; Tongsong and Srisomboon, 1993). Serial (weekly) screening of high-risk pregnancies for oligohydramnios is important, because amniotic fluid can become drastically reduced within 24 to 48 hours (Clement et al, 1987). Although each of the five features of the BPP are scored equally (2 points if the variable is present or normal, and 0 points if absent or abnormal, for a total of 10 points), they are not equally predictive of adverse pregnancy outcome. For example, amniotic fluid volume is the variable that correlates most strongly with adverse pregnancy events. The recommended management based on the BPP is summarized in Table 13-3 (Manning, 1989). A score of 8 to 10 out of 10 is regarded as reassuring; a score of 4 to 6 is suspicious and requires reevaluation, and a score of 0 to 2 suggests nonreassuring fetal testing—previously referred to as fetal distress (Manning et al, 1981a, 1981b). Evidence of nonreassuring fetal testing or oligohydramnios in the setting of otherwise reassuring fetal testing should prompt evaluation for immediate delivery (Vintzileos et al, 1987a, 1987b). DOPPLER Doppler velocimetry shows the direction and characteristics of blood flow, and it can be used to examine the maternal, uteroplacental, or fetal circulations. Because of placental capacitance, the umbilical artery is one of the few arteries that normally has forward diastolic flow, and it is one of the most frequently targeted vessels during TABLE 13-3 Interpretation and Management of Biophysical Profile Perinatal Morbidity or Mortality Within 1 week (no intervention) Management Normal <1/1000 No intervention 8/8 Normal — No intervention 8/10 (abnormal NST) Normal — 8/10 (abnormal amniotic fluid) Suspect chronic fetal ­compromise, renal anomaly, or rupture of membranes 89/1000 Rule out renal abnormality or rupture of membranes; consider delivery or prolonged observation if dictated by gestational age 6/8 (other) Equivocal, possible asphyxia Variable If fetus is mature, deliver; if immature, repeat test within 4-6 hours 6/8 (abnormal amniotic fluid) Suspect asphyxia 89/1000 Repeat 4-6 hours; consider delivery 4/8 Suspect asphyxia 91/1000 If ≥36 weeks’ gestation or documented pulmonary maturity, deliver immediately; if not, repeat within 4-6 hours 2/8 High suspicion of asphyxia 125/1000 Immediate delivery 0/8 High suspicion of asphyxia 600/1000 Immediate delivery Score Comment 10/10 No intervention Adapted from Manning FA, Morrison I, Harman CR, et al: Fetal assessment based on fetal biophysical profile scoring: experience in 19,221 referred high-risk pregnancies, Am J Obstet Gynecol 157:880, 1987. NST, Nonstress test. CHAPTER 13 Antepartum Fetal Assessment pregnancy. Umbilical artery Doppler velocimetry measurements reflect resistance to blood flow from the fetus to the placenta. Factors that affect placental resistance include gestational age, placental location, pregnancy complications (placental abruption, preeclampsia), and underlying maternal disease (chronic hypertension). Doppler velocimetry of umbilical artery blood flow provides an indirect measure of placental function and fetal status (Giles et al, 1985). Decreased diastolic flow with a resultant increase in systolic-to-diastolic ratio suggests increased placental vascular resistance and fetal compromise. Severely abnormal umbilical artery Doppler velocimetry (defined as absent or reversed diastolic flow) is an especially ominous observation and is associated with poor perinatal outcome, particularly in the setting of IUGR (Ducey et al, 1987; McCallum et al, 1978; Rochelson et al, 1987; Trudinger et al, 1991; Wenstrom et al, 1991; Zelop et al, 1996). The overall mortality rate for fetuses with absent or reversed flow may be near 30% (Karsdorp et al, 1994). It should be noted that abnormal Doppler studies are often seen in cases of anatomic anomalies or chromosomal abnormalities, which should be noted when managing a case. The role of ductus venosus and middle cerebral artery (MCA) Doppler in the management of IUGR pregnancies is not well defined. As such, urgent delivery should be considered in IUGR pregnancies when the results of umbilical artery Doppler studies are severely abnormal regardless of gestational age. However, it is unclear how to interpret these findings in the setting of a normally grown fetus. For these reasons, umbilical artery Doppler velocimetry should not be performed routinely on low-risk women. Appropriate indications include IUGR, cord malformations, unexplained oligohydramnios, suspected or established preeclampsia, and possibly fetal cardiac anomalies. Umbilical artery Doppler velocimetry has not been shown to be useful in the evaluation of a variety of high-risk pregnancies, including diabetic and postterm pregnancies, primarily because of its high false-positive rate (Baschat, 2004; Farmakides et al, 1988; Landon et al, 1989; Stokes et al, 1991). As such, in the absence of IUGR, obstetric management decisions are not usually made on the basis of Doppler velocimetry studies alone. Nonetheless, new applications for Doppler technology are currently under investigation. A recent application that has proved extremely useful is the noninvasive evaluation of fetal anemia resulting from isoimmunization. When a fetus develops severe anemia, cardiac output increases and there is a decline in blood viscosity, resulting in an increase in MCA blood flow, which can be demonstrated by measuring the peak velocity using MCA Doppler velocimetry (Mari et al, 1995). This demonstration can help the perinatologist to better 139 counsel such patients about the need for cordocentesis and fetal blood transfusion. Doppler studies of other vessels— including the uterine artery, fetal aorta, ductus venosus, and fetal carotid arteries—have contributed considerably to our knowledge of maternal-fetal physiology, but as yet have resulted in few clinical applications. SUMMARY There are a variety of testing modalities available to the obstetrician for assessing fetal well-being in the antepartum period, each with specific applications, advantages, and disadvantages. As such, it is difficult to apply generalized protocols to the assessment of the fetus. A stepwise approach entails applying the appropriate tests for low-risk patients and identifying those patients, from the results of those tests or from historical factors, for whom further testing is needed. Although many tests, including NST, fetal weight assessment, and uterine artery Doppler, may be somewhat nonspecific and may have misleading falsepositive rates, combining those tests with others increases the specificity. Test results that raise concerns require further investigation or active management. SUGGESTED READINGS American College of Obstetricians and Gynecologists (ACOG): Fetal macrosomia, Washington, DC, 2000, ACOG Practice Bulletin No. 22, ACOG. American College of Obstetricians and Gynecologists (ACOG): Gestational diabetes, Washington, DC, 2001, ACOG Practice Bulletin No. 30, ACOG. American College of Obstetricians and Gynecologists (ACOG): Shoulder dystocia, Washington, DC, 2002, ACOG Practice Bulletin No. 40, ACOG. American College of Obstetricians and Gynecologists (ACOG): Ultrasonography in pregnancy, Washington, DC, 2009, ACOG Practice Bulletin No. 101, ACOG. Anderson NG, Jolley IJ, Wells JE: Sonographic estimation of fetal weight: comparison of bias, precision and consistency using 12 different formulae, Ultrasound Obstet Gynecol 30:173-179, 2007. Baschat AA: Doppler application in the delivery timing of the preterm growthrestricted fetus: another step in the right direction, Ultrasound Obstet Gynecol 23:111-118, 2004. Brace RA, Wolf EJ: Normal amniotic fluid volume changes throughout pregnancy, Am J Obstet Gynecol 161:382-388, 1989. Freeman RK, Anderson G, Dorchester W: A prospective multi-institutional study of antepartum fetal heart rate monitoring: I. risk of perinatal mortality and morbidity according to antepartum fetal heart rate test results, Am J Obstet Gynecol143 :771-777, 1982. Freeman RK, Anderson G, Dorchester W: A prospective multi-institutional study of antepartum fetal heart rate monitoring: II. contraction stress test versus nonstress test for primary surveillance, Am J Obstet Gynecol 143:778-781, 1982. Macones GA, Hankins GD, Spong CY, et al: The 2008 National Institute of Child Health and Human Development workshop report on electronic fetal monitoring: update on definitions, interpretation, and research guidelines, Obstet Gynecol 112:661-666, 2008. Manning FA, Morrison I, Harman CR, et al: Fetal assessment based on fetal biophysical profile scoring: experience in 19,221 referred high-risk pregnancies, Am J Obstet Gynecol 157:880-884, 1987. Manning FA, Morrison I, Lange IR, et al: Fetal assessment based on fetal biophysical profile scoring: experience in 12,620 referred high-risk pregnancies: I. Perinatal mortality by frequency and etiology, Am J Obstet Gynecol 151: 343-350, 1985. Complete references and supplemental color images used in this text can be found online at www.expertconsult.com C H A P T E R 14 Prematurity: Causes and Prevention Tonse N.K. Raju and Caroline Signore BURDEN OF PRETERM BIRTH Despite major advances in perinatal medicine, preterm births (PTBs) in the United States have been increasing over the past two decades, reaching a high of 12.8% of live births in 2006 (Martin et al, 2009). This rate amounts to one PTB occurring at each minute of every day, throughout the year. From the perspective of a societal health care burden, these data are sobering, because infants born even 1 or 2 weeks before full term suffer higher rates of morbidity and mortality throughout life. Thus even a minimal increase in the PTB rate has a major effect on the societal burden of disease. Conversely, even a modest reduction in the PTB rate can have a major positive effect on the societal health care burden. In this chapter we focus on the etiology, prediction, and prevention of PTB. Preliminary data in the United States for 2007 showed that among the total number of live births, 2.04% were very preterm, 1.59% were moderate preterm, and 9.03% were late preterm (Hamilton et al, 2009). Thus, of all the PTBs the late PTB stratum was the largest at 71%, the very PTB stratum was intermediate at 16.1%, and the moderate PTB stratum was the smallest at 12.6%. The U.S. 2007 preliminary data also suggest a slight decline in the total PTB rate to 12.7%, from 12.8% in 2006. The 2007 decline was predominantly among late PTB, from 9.1% in 2006 to 9.0% in 2007 (Hamilton et al, 2009). However, trends over longer periods need to be examined to confirm these preliminary findings and to judge whether the PTB rates in the United States have ­leveled or begun to decline. DEFINITIONS ETIOLOGY AND RISK FACTORS The World Health Organization defines preterm birth as a birth occurring either before 37 completed weeks of gestation or on or before the 259th day, counting from the first day of the last menstrual period. Only completed weeks of gestation are reported; therefore an infant born 6 days after completing 35 weeks of gestation is noted as 35 weeks, not rounded up to 36 weeks. However, one can precisely denote the exact gestational age by using a superscript for the number of days after the completion of the gestational week. Thus, the infant in the example can be noted as being born at 356/7 weeks’ gestation. PTBs can be categorized into three major gestational age strata: those occurring at <32 weeks’ (on or before 224 days) gestation, usually referred to as very preterm births; those occurring between 320/7 and 336/7 weeks’ (between 225 and 238 days) gestation (a group with no specific name, but can be termed moderate preterm births); and those occurring between 340/7 and 366/7 weeks’ (239 to 259 days) gestation, referred to as late preterm births (Figure 14-1). It is worth noting that the phrase near term is no longer used to refer to the third group, because the phrase falsely conveys a message that such “borderline” preterm infants are almost as mature as term infants (Raju et al, 2006). Individual PTBs can be categorized into two major groups: (1) spontaneous PTB (accounting for up to 60% of all PTBs), which occur after the spontaneous onset of preterm labor, and in some cases, after spontaneous preterm rupture of the fetal membranes before the onset of labor; and (2) indicated PTB (accounting for about 40% of PTBs), which arise from interventions by the health care team to reduce poor outcomes in the presence of specific medical, surgical, or obstetric conditions and indications in the mother or fetus. EPIDEMIOLOGY In the United States, the PTB rate increased from 10.6% in 1990 to 12.8% in 2006—a 21% increase (Goldenberg et al, 2008; Martin et al, 2009). This increase was part of a general trend in the distribution of gestational age at delivery, which showed a dramatic, leftward shift by 1 week, such that the peak modal week (the gestational week at which most deliveries occurred) shifted from 40 weeks in 1992 to 39 weeks in 2002 (Davidoff et al, 2006). Whereas this shift has decreased births at 41 weeks’ gestation or later, it has also led to an increase in the number of all PTBs. 140 SPONTANEOUS PRETERM BIRTHS In most spontaneous PTBs, the precise cause for the onset of labor remains unknown; however, a variety of risk factors have been identified (Behrman and Butler, 2007) (Table 14-1). BEHAVIORAL AND PSYCHOLOGICAL CONTRIBUTORS Behavioral and psychological contributors, especially when occurring in constellation, tend to increase the risk for PTB. Some of these factors include excessive consumption of alcohol, smoking, use of cocaine, unfavorable diet, prolonged and stressful physical labor during pregnancy, shorter interval between pregnancies, and a variety of psychosocial stressors (Goldenberg et al, 2008). Sociodemographic and Community Contributors Increased risk for PTB has been associated with extremes of maternal age (<17 or >35 years), unmarried status, poverty, adverse neighborhood conditions, and lower educational attainment. The causal pathways for the CHAPTER 14 Prematurity: Causes and Prevention 141 CATEGORIZATION OF GESTATIONAL AGE AT BIRTH First day of LMP Day # 1 141 225 239 260 295 Week # 0 20 0/7 32 0/7 34 0/7 37 0/7 42 0/7 Very Preterm ( 224 days) Moderate Late Preterm Term Preterm (239 – 259 days) (260 – 294 days) Post Term ( 295 days) (225 – 238 days) FIGURE 14-1 Categorization of gestational age at birth. TABLE 14-1 Risk Factors for Preterm Birth* Category Specific Variable Effect Size and Comments Behavioral and psychological contributors Tobacco and alcohol Inconsistent effect; these variables may be confounders with lower socioeconomic status Other drugs Cocaine users experience a twofold increase in PTB; other drugs (e.g., ­marijuana) have inconsistent effects Nutrition Low prepregnancy weight and lower weight gain increase PTB risk; these ­factors may be confounders with other medical conditions Physical activity, employment Inconsistent effect of employment per se; but long work hours and stress have been associated with increased risk of PTBs Douching Frequent and prolonged vaginal douching during pregnancy can alter vaginal flora and enhance the risk for PTB Stress (chronic, acute, life events) Emotional stress, living conditions, major stressful life events, discrimination, and racism, and other forms of stress experiences may act through diverse mechanistic pathways, affect many body functions, such as the HPA, immune defense mechanisms, and autonomic nervous system, leading to PTB Maternal age U-shaped relationship with maternal age and PTB rate, with increased risk in women <17 and >35 years; in black women, risk seems to increase by age 30 Marital status Although unmarried mothers have higher rates of PTB, the protective effects of marital status on the PTB rate may vary across different ethnic groups Race and ethnicity Non-Hispanic black women have high rates of PTB; the causes are likely to multifactorial Socioeconomic conditions Disparities in PTB rates by SES (high rates with low SES) are well documented, although the etiologic pathways for such effects are unclear Community and neighborhood Adverse neighborhood conditions influence health outcomes, including PTB rates, through direct and indirect pathways Maternal medical illnesses and conditions Indicated PTBs may occur with chronic hypertension, hypertensive disorders of pregnancy, systemic lupus, hyperthyroidism, pregestational diabetes mellitus, cardiac disease, asthma, renal disorders, and gestational diabetes mellitus Underweight and low weight gain Low prepregnancy weight and lower-than-average increase in weight during pregnancy have been documented with higher rates of PTB Fetal conditions In vitro fertilization, assisted reproductive technology, multifetal pregnancy, congenital anomalies, umbilical cord accidents Pregnancy conditions Placenta previa, first-trimester vaginal bleeding, abruption of the placenta, ­uterine malformations Infections Maternal-fetal infections, bacterial vaginosis Other Interpregnancy interval of less than 6 months is estimated to increase PTB rate by 30% to 60%; family history of PTB is associated with higher rates of PTB; it is unclear if the effect is caused by genetic factors, environmental factors, or a combination of both Lead, air pollution Exposures linked to higher PTB rates include lead, tobacco smoke, air pollution (sulfur dioxide, particulates, carbon monoxide), and arsenic in drinking water; however, interactions between such exposures and confounding factors such as SES, race, and ethnicity need to be studied Sociodemographic ­contributors Medical and pregnancy conditions Environmental toxins Gene-environment interactions Genetic susceptibility and gene environmental interactions is a rapidly evolving field; according to some experts, epigenetic mechanisms are the final common pathway to explain diverse sociodemographic, racial, and ethnic factors and their effects on PTB rates HPA, Hypthalamopituitary axis; PTB, preterm birth; SES, socioeconomic status. *For a comprehensive discussion of various factors noted in this table, please see Behrman RE, Butler AS: Preterm birth, causes, consequences, and prevention, Washington, DC, 2007, The National Academies Press. 142 PART IV Labor and Delivery association between adverse sociodemographic and community contributors are complex and poorly understood. It should be noted, however, that these factors are interrelated and often coexist with other medical or obstetric conditions. Racial Disparity African American women are consistently found to have higher PTB rates than women of other racial and ethnic groups (16% to 18% versus 5% to 9%; Goldenberg et al, 2008). Black women are also more likely to experience preterm premature rupture of membranes (PPROM) and very PTB. The reason for this persistent racial disparity is unknown. Previous Preterm Birth The prior history of a PTB is consistently identified as one of the strongest predictors of preterm delivery, elevating the risk to more than twofold the background rate (Mercer et al, 1999). Risk is further elevated if there is a history of multiple PTBs, consecutive previous PTBs, and births at earlier gestational ages (Spong, 2007). Short Cervix Among asymptomatic women, cervical length less than 25 mm at 24 weeks’ gestation, as measured by transvaginal ultrasound, in singleton pregnancy is associated with a sixfold increase in the risk of delivery before 35 weeks’ gestation (Iams et al, 1996). Among women in suspected preterm labor, a cervical length of less than 15 mm may discriminate between women who are likely to deliver preterm and those whose pregnancies will continue (Tsoi et al, 2003). Genetic Factors The risk of PTB is increased in women with a positive family history of PTB. Single nucleotide polymorphisms in a number of maternal and fetal candidate genes involved in inflammatory pathways have been associated with PPROM and PTB (Varner and Esplin, 2005). Studies using genomic and proteomic approaches to further investigate molecular risk factors for PTB are in progress. Infections Overt or subclinical intrauterine infection, which might lead to rupture of the membranes before onset of labor, has been hypothesized to etiologically contribute to onset of labor, especially among very PTBs. The most common route of intrauterine infection is ascension of vaginal bacteria. Bacterial vaginosis (BV) is a common vaginal infection characterized by a shift in vaginal flora toward anaerobic species, accompanied by an abnormally high pH. Risk of PTB is increased up to twofold in women with BV (Hillier et al, 1995; Meis et al, 1995). Periodontal disease has also been linked to PTB (Jeffcoat et al, 2001); however, the biologic pathway is not well understood. Multifetal Gestation Twin and higher-order pregnancies account for 3% of all births. Multiple pregnancies are at increased risk of preterm delivery, with approximately 40% of twins delivered spontaneously before 37 weeks’ gestation. Uterine overdistention is believed to be the major etiologic factor in these births. An additional 20% of twins will develop maternal or fetal complications leading to indicated PTB. The mean gestational age at delivery is 35 weeks’ gestation for twins, 32 weeks’ gestation for triplets, and 29 weeks’ gestation for quadruplets (Elliott, 2005; Martin et al, 2003). Infertility Treatment The use of assisted reproductive technology (ART), in which gametes are manipulated outside the body, is steadily increasing. In 2006, 41,000 American women delivered after successful infertility treatment with ART, accounting for approximately 1% of all live births (Sunderam et al, 2009). The use of ART is strongly associated with multifetal pregnancies, dramatically increasing the risk for PTBs. However, even singleton pregnancies following ART have a twofold higher risk for PTB (Jackson et al, 2004), suggesting a complex relationship between a history of infertility, ART, and reproductive outcomes. Medical treatment of infertility (i.e., ovulation induction) is also associated with multiple gestation and PTB. In 2003, 21% of twin, 37% of triplet, and 62% of quadruplet and higher-order multiple births were attributable to non-ART ovulation induction (Dickey, 2007). Maternal Body Habitus: Underweight and Obesity Low prepregnancy body mass index (<18.9 kg/m2) is associated with a 1.5- to 2.5-fold increase in the risk of PTB and has previously been a target of PTB prevention efforts (Goldenberg et al, 1998; Siega-Riz et al, 1996). Currently there is an epidemic of overweight and obesity in all U.S. populations. Maternal obesity is not associated with an increased risk of spontaneous PTB (Hendler et al, 2005). However, maternal obesity is a risk factor for a variety of pregnancy complications, including congenital anomalies, preeclampsia, and macrosomia, which may result in the need for indicated PTB (Aly et al, 2010). Fetal Fibronectin Fetal fibronectin (FFN) is a glycoprotein involved in the adherence of the fetal membranes to the uterine wall. The presence of FFN in the cervicovaginal secretions has been considered a risk factor for PTB and has been evaluated for its predictor value of preterm delivery in large cohort studies. A positive test result for FFN at 22 to 24 weeks’ gestation in asymptomatic women is significantly associated with PTB (Goldenberg et al, 1998); however, because of its low positive predictive value, it is not a reliable universal screening tool in low-risk women. The most useful aspect of FFN assessment is its high negative predictive value: among women with signs and symptoms of preterm CHAPTER 14 Prematurity: Causes and Prevention 143 labor and a negative FFN, 99% will not deliver over the next 7 days (Lu et al, 2001). ischemia, uterine over distension, abnormal allograft reaction, allergy, cervical insufficiency, and hormonal dysregulation (Romero et al, 2006). INDICATED PRETERM BIRTHS Allostatic Load Medical and Pregnancy Conditions Well-recognized medical and pregnancy-related conditions are noted in Table 14-1. Many of these conditions are also encountered to a greater extent among women in lower socioeconomic strata and among those living under adverse conditions. Maternal and fetal conditions that can lead to indicated preterm deliveries include diabetes; hypertension; preeclampsia or eclampsia; maternal cardiovascular and pulmonary disorders; multifetal gestations; uterine structural disorders; polyhydramnios or oligohydramnios; placenta previa, abruption, and other uterine hemorrhagic events; chorioamnionitis and fetal infection; fetal distress; and congenital anomalies. Causes for Increasing Preterm Birth The precise reasons for increasing rates of PTB are unclear. The contribution of many factors is noted in Table 14-2 (Behrman and Butler, 2007; Raju, 2006; Raju et al, 2006). Etiologic Pathways Preterm labor and delivery arise from abnormal activation of parturition. No single mechanism describes all cases of PTB; rather, preterm parturition can be viewed as an obstetric syndrome with multiple etiologies, a long preclinical phase, frequent fetal involvement, and genetic factors that modify risk. Contributing pathologic processes may include intrauterine inflammation or infection, uterine TABLE 14-2 Possible Causes for Increasing Preterm Birth* Variable Comments Increasing maternal age at first pregnancy Number of pregnant women 35-39 years nearly doubled between 1993 and 2003 Increasing infertility treatment and multifetal gestations Number of women seeking infertility treatment increased twofold between 1996 and 2002 Change in practice parameters or guidelines Improved surveillance and increased medical interventions to prevent still births Increasing cesarean section rates and decreasing VBAC — Increasing maternal overweight and obesity Errors in estimating gestational age of macrosomic fetus Inaccuracy in estimating gestational age Lack of early ultrasound examination Iatrogenic or nonmedical reasons Maternal request for early delivery or other logistical considerations of the patient, family or health care team VBAC, Vaginal birth after cesarean. *For a review, please see Goldenberg et al (2008), Raju (2006), and Reddy et al (2007, 2009). Recent studies of the pathophysiology of and biologic responses to stress have uncovered a series of causal pathways for adverse outcomes caused by stresses at multiple levels and from different sources. The biologic responses to stressors have been collectively called allostatic load, a summary measure of major perturbations in the autonomic, endocrine, and immune systems. Allostatic load has been linked to accelerated aging, worsened immunologic tolerance, ineffective defense systems, and increased cardiovascular morbidity. Abnormal allostatic load during the entire life span of a woman can potentially modulate her biologic systems and worsen perinatal outcomes (Groer and Burns, 2009; Herring and Gawlik, 2007; Latendresse, 2009; Lu and Halfon, 2003; McEwen, 2006; Shannon et al, 2007). Studies are under way to understand the complex relationship between the effects of sociodemographic and biologic stressors on organ systems and their influence on perinatal outcomes. PREVENTION OF PRETERM BIRTH Despite intense research efforts, few widely effective strategies for preventing PTB have been identified. Because of the multifactorial nature of preterm labor and birth, it is unlikely that any single intervention could serve to prevent all cases of prematurity. In the United States, efforts have focused on addressing specific risk factors to prevent PTB (i.e., secondary prevention), such as treatment of infection, and interventions to mitigate the morbidity and mortality of PTB (i.e., tertiary prevention), such as administration of tocolytic drugs to women in preterm labor (Iams et al, 2008). A better understanding of the complex mechanisms leading to preterm labor and birth is needed to advance development of interventions to prevent PTB. In this section, a number of strategies—both encouraging and disappointing in terms of effectiveness—are described. Progesterone Progesterone is thought to have antiinflammatory effects and to be a smooth muscle relaxant that inhibits uterine contractions. Administration of progesterone has shown the most promise as an effective intervention for preventing PTB in women with certain risk factors for PTB. In a recent metaanalysis of 11 trials, progesterone significantly decreased PTB of singleton gestations less than 34 weeks in women with a prior spontaneous preterm birth (SPTB) by 85%, with a number needed to treat of seven (Dodd et al, 2005). However, progesterone does not prevent PTB in women carrying twins (Rouse et al, 2007) or triplets (Caritis et al, 2009). Petrini et al, (2005) estimated that if all eligible women with a history of prior SPTB were treated with progesterone, then 10,000 PTBs per year could be prevented. Progesterone treatment does not appear to pose a developmental risk to the fetus or infant. In a follow-up study of children exposed to 7-alpha hydroxyprogesterone caproate versus placebo in utero, there were 144 PART IV Labor and Delivery no differences in growth or ­neurodevelopmental measures between groups at age 4 years (Northen et al, 2007). Cerclage Cervical cerclage is a purse-string suture of the cervix intended to mechanically prevent cervical dilation. A number of investigators have evaluated the use of cerclage to prevent PTB in women with shortened cervical length, and they have not demonstrated a significant effect (Berghella et al, 2004; To et al, 2004). A more recent trial indicated a significant beneficial effect (adjusted odds ratio, 0.23; 95% confidence interval, 0.08 to 0.66) of cerclage in women with a history of previous SPTB and cervical length less than 15 mm (Berghella et al, 2008; Owen et al, 2009). Limited Embryo Transfer Preterm delivery of multifetal pregnancies arising from ART can be prevented in many cases by limiting the number of embryos transferred to the uterus. The proportion of in vitro fertilization cycles in which three or more fresh embryos were transferred declined significantly in the United States between 1996 and 2002 (92% to 54%; p <0.001); however, the proportion of single-embryo transfer cycles remained small (2.5% in 2002) (Reynolds and Schieve, 2006). Recent practice guidelines from the American Society for Reproductive Medicine call for wider consideration of single-embryo transfer in many clinical situations (American Society for Reproductive Medicine, 2008). Multifetal Pregnancy Reduction Reduction of multifetal pregnancy, or the selective termination of one or more fetuses in a high-order pregnancy, has been proposed as a means to reduce the risk of prematurity and other pregnancy complications. Although effective at preventing some very PTBs, the procedure carries an approximately 1% risk of loss of the entire pregnancy (Stone et al, 2002) and can create profound ethical dilemmas (American College of Obstetricians and Gynecologists, 2007). Treatment of Vaginal Infections Despite consistent associations between lower genital tract infection and PTB, multiple randomized trials have failed to demonstrate a benefit of screening and antibiotic treatment for vaginitis or genital tract colonization on the PTB rates (Carey et al, 2000; Iams et al, 2008; Klebanoff et al, 2001; McDonald et al, 2007). Treatment of Periodontal Disease PTB has been associated with maternal periodontal disease in a number of studies, and early clinical trials suggested that treatment of periodontitis reduces the risk of PTB (Jeffcoat et al, 2003; Lopez et al, 2002). On the other hand, a recent multicenter clinical trial in the United States found that, whereas periodontal disease improved in pregnant women randomized to treatment, there were no differences in rates of PTB, low birthweight, or fetal growth restriction in the treatment group compared with controls (Michalowicz et al, 2006). Tocolytics: Acute and Maintenance Preterm labor (PTL), is defined as the onset before 37 weeks’ gestation of regular uterine contractions that result in cervical change. Although numerous pharmacologic and nonpharmacologic treatments have been aimed at arresting PTL and preventing PTB, none have been shown to prolong pregnancy for more than 2 to 7 days (Iams et al, 2008). Nevertheless, tocolytic drugs are often used to allow for the administration of steroids and maternal transfer to a facility capable of caring for a preterm neonate. Drugs commonly used to treat PTL are summarized in Table 14-3. Maintenance tocolysis after an acute episode of PTL has not been shown to prevent PTB (Dodd et al, 2006; Gaunekar and Crowther, 2004; Nanda et al, 2002). MANAGEMENT OF PRETERM PREMATURE RUPTURE OF MEMBRANES Preterm premature rupture of membranes (PPROM) is defined as the prelabor rupture of membranes before 37 weeks’ gestation. The period between initial leakage of TABLE 14-3 Drugs Commonly Used to Treat Preterm Labor* Agent Class, Mechanism Maternal Side Effects Fetal and Neonatal Effects Magnesium sulfate Decreases intracellular calcium, inhibits contractile response Flushing, lethargy, nausea, muscle weakness Lethargy, decreased tone, respiratory depression Terbutaline β-Mimetic Tachycardia, hyperglycemia, myocardial ischemia, pulmonary edema Tachycardia, hyperinsulinemia, hyperglycemia Nifedipine Calcium channel blocker Hypotension, flushing None reported Indomethacin NSAID, prostaglandin synthase inhibitor Gastrointestinal upset Decreased renal function, oligohydramnios, closure of the ductus arteriosus Atosiban (not available in the United States) Oxytocin inhibitor Hypersensitivity, injection site reaction Increased risk of fetal or infant death if administered before 28 wk NSAID, Nonsteroidal antiinflammatory drug. *For a review, see American College of Obstetricians and Gynecologists (2007), Dodd et al (2006), Gaunekar and Crowther (2004), Nanda et al (2002), Simhan and Caritis (2007), and Stone et al (2002). CHAPTER 14 Prematurity: Causes and Prevention fluid and the onset of labor and delivery is known as the latency period. Without treatment, the latency period after PPROM is usually less than 1 week, with an inverse relationship between the gestational age at rupture and length of the latency period. In a large multicenter trial, a 7-day course of broad-spectrum antibiotics was shown to prolong the latency period for up to 3 weeks in women who had experienced PPROM at 24 to 32 weeks’ gestation. In addition, antepartum antibiotic treatment decreased the incidence of chorioamnionitis and major morbidity (respiratory distress syndrome, early sepsis, severe intra ventricular hemorrhage, or severe necrotizing enterocolitis) in the infants (Mercer et al, 1997). SUMMARY PTBs are occurring at epidemic proportions in the United States. With more than 4.3 million live births each year, a prematurity rate of 12.5% results in more than 537,000 PTBs annually. As described in other chapters, preterm infants are at greater risk for mortality, and short- and long-term morbidities compared with term infants. A recent study from Norway on the adult-age outcomes of preterm infants reported that even mild prematurity was associated with significantly elevated risk for long-term adverse medical, behavioral, psychological, and vocational outcomes (Moster et al, 2008). In addition to the families’ burden of illness, death, and disability, society bears a large burden of health care costs related to PTBs. The collective economic burden from prematurity was more than $26 billion in 2005 (Behrman and Butler, 2007). Given the complex nature of the risk factors and causal pathways for PTBs, and the difficulties in diagnosing, preventing, and treating them, there have been renewed calls in support of multidisciplinary research to address the problems of PTBs. The Institute of Medicine of the National Academies of Science issued a major publication providing specific recommendations and guidelines for research, surveillance, and education on this topic (Behrman and Butler, 2007). The U.S. Senate approved the PREEMIE 145 Act on August 1, 2006. This bill called for efforts to reduce preterm labor and delivery and the risk of pregnancyrelated deaths and complications, and to reduce infant mortality caused by prematurity. The U.S. House of Representatives passed the bill unanimously on December 9, 2006, and the president signed it into law on December 22, 2006 (March of Dimes Foundation, 2006). The March of Dimes Foundation initiated the Prematurity Campaign, designed to educate and inform health care professionals, industry, government agencies, and the general public about prematurity-related issues. These and other efforts have begun recently and will take time to yield positive results. Their collective effect on reducing the burden of PTBs needs to be monitored on a long-term basis. SUGGESTED READINGS Behrman RE, Butler AS: Preterm birth, causes, consequences, and prevention, Washington, DC, 2007, The National Academies Press. Dodd JM, Crowther CA, Cincotta R, et al: Progesterone supplementation for preventing preterm birth: a systematic review and meta-analysis, Acta Obstet Gynecol Scand 84:526-533, 2005. Goldenberg RL, Culhane JF, Iams JD, et al: Epidemiology and causes of preterm birth, Lancet 371:75-84, 2008. Goldenberg RL, Iams JD, Mercer BM, et al: The preterm prediction study: the value of new vs standard risk factors in predicting early and all spontaneous preterm births. NICHD MFMU Network, Am J Public Health 88:233-238, 1998. Iams JD, Romero R, Culhane JF, et al: Primary, secondary, and tertiary interventions to reduce the morbidity and mortality of preterm birth, Lancet 371:164175, 2008. Owen J, Hankins G, Iams JD, et al: Multicenter randomized trial of cerclage for preterm birth prevention in high-risk women with shortened midtrimester cervical length, Am J Obstet Gynecol 201:375, 2009. Raju TN, Higgins RD, Stark AR, et al: Optimizing care and outcome for latepreterm (near-term) infants: a summary of the workshop sponsored by the National Institute of Child Health and Human Development, Pediatrics 118:1207-1214, 2006. Romero R, Espinoza J, Kusanovic JP, et al: The preterm parturition syndrome, BJOG 113(Suppl 3):17-42, 2006. Simhan HN, Caritis SN: Prevention of preterm delivery, N Engl J Med 357: 477-487, 2007. Spong CY: Prediction and prevention of recurrent spontaneous preterm birth, Obstet Gynecol 110:405-415, 2007. Varner MW, Esplin MS: Current understanding of genetic factors in preterm birth, BJOG 112(Suppl 1):28-31, 2005 Complete references used in this text can be found online at www.expertconsult.com C H A P T E R 15 Complicated Deliveries: Overview Sarah A. Waller, Sameer Gopalani, and Thomas J. Benedetti Historically, child birth was often regarded as a perilous undertaking. However, over the past century in the United States, perinatal and maternal mortality have dramatically fallen with advances in modern obstetric care, such as widespread use of antibiotics, easy access to expedient cesarean delivery, and better understanding of the proper use of instruments such as forceps and vacuum extraction (Ali and Norwitz, 2009). Indeed, adverse outcomes are generally uncommon in modern obstetrics, and unlike in the past, most labor and delivery concludes with a healthy mother and neonate. Nevertheless, complicated deliveries still exist, and knowledge of their conduct and sequelae is still required for the administration of proper maternal and infant care. In this chapter we will first address the complicated vaginal delivery, with particular attention to neonatal outcomes. We will then discuss cesarean delivery and vaginal birth after a prior cesarean delivery (VBAC) and what neonatal implications these may have. Before discussing complicated labor and its neonatal effects, it is important to have a brief understanding of the conduct of normal labor and delivery. A comprehensive discussion of labor and delivery is beyond the scope of this chapter, and the interested reader is directed Williams Obstetrics, ed 23, chapter on normal labor. The first stage of labor begins with the onset of regular uterine contractions with concomitant cervical dilation and effacement, and it ends with complete cervical dilation. The first stage is further subdivided into a latent phase, the length of which is variable and can last for several hours, and an active phase, which usually begins when the cervix is dilated 3 to 4 cm and is marked by further rapid, progressive cervical dilation and effacement. Often the diagnosis of the transition from latent to active phase labor is retrospective, because the time of onset of active labor is variable by patient. The second stage begins with complete cervical dilation and terminates with the expulsion of the fetus from the birth canal. The third stage of labor concludes with the delivery of the placenta. Disorders of the conduct of labor are of either protraction, in which cervical dilation or fetal descent occurs but is at a rate much less than expected, or arrest. Both disorders are addressed by operative delivery if they are unresponsive to active medical management; this can be performed abdominally through cesarean section or vaginally by obstetric forceps or vacuum extraction if the cervix is fully dilated and specific criteria are fulfilled (see later discussion on operative vaginal delivery). All these modalities can have neonatal and maternal effects, and the choice of instrument or mode of delivery must always be selected taking these potential morbidities into account. CESAREAN SECTION Neonatal mortality within 7 days of birth was 17.8 per 1000 live births in 1950; it is currently 3.8 per 1000 (Ali and Norwitz, 2009). Certainly much of the improvement 146 in neonatal survival stems from advances in neonatal intensive care and resuscitative techniques. Along with improved technology available to the pediatricians, however, is the fact that the last 50 years have also seen a dramatic increase in the universal access to safe cesarean section, which affords quick and timely fetal delivery. However, a cause-and-effect relationship between cesarean delivery and improved neonatal outcome has never been demonstrated, and recent evidence suggests that some current obstetric cesarean delivery practices may actually cause harm (Bates, 2009 ). Currently almost one mother in three is giving birth by cesarean section, a record level for the United States. Recommendations from the U.S. Department of Health and Human Services Healthy People 2010 recommend a cesarean section rate of less than 15% for a first pregnancy and 63% for previous cesarean sections. In addition, there are wide variations in cesarean delivery rates among hospitals in a given state or region, suggesting that factors other than pregnancy risk factors may be responsible for the current cesarean delivery rate. This rise in cesarean delivery has been associated with a parallel drop in the vaginal operative delivery rate to less than 10%. These two trends have increased the cost of childbirth in the United States and are now threatening many state budgets, because almost 50% of obstetric care cost is paid with public funds (i.e., Medicaid). In response as the result of a budget crisis, the state of Washington reduced hospital payment for uncomplicated cesarean delivery (DRG 371: Cesarean section) by reimbursing only the equivalent of a complicated vaginal birth (DRG 372: Normal delivery with problems), a projected savings of $2 million per year. Many states are actively working to safely reduce the rate of cesarean delivery. Projects are under way to reduce elective delivery before 39 weeks’ gestation, safely promote VBAC and limit the occurrence of higher-order multiple gestations. Cesarean section is usually performed through either a Pfannenstiel or vertical skin incision. The uterine incision is often made transversely in the lower uterine segment, because it minimizes intraoperative blood loss and future risk of rupture during subsequent labor, compared with a vertical or classical incision. The risk of rupture in future labor is thought to be 0.5% to 1.0% for a low transverse incision, compared to 4% to 9% for a classical incision (Feingold et al, 1988). Cesarean delivery is also performed for disorders of protraction or arrest in the first stage of labor when conservative measures fail to augment delivery, such as oxytocin or amniotomy, or in the second stage when assisted or operative vaginal delivery is deemed unfeasible or unsafe. Accepted obstetric indications for operative delivery are as follows: 1. Fetal malpresentation (e.g., shoulder or breech) 2. Placenta previa 3. Prior classical uterine incision CHAPTER 15 Complicated Deliveries: Overview 4. Fetal status not reassuring, remote from vaginal delivery 5. Higher-order multiple gestation (triplet or greater) 6. Fetal contraindications to labor (alloimmune thrombocytopenia, neural tube defect) 7. Maternal contraindication to labor (e.g., history of rectal or perineal fistulas from inflammatory bowel disease, large lower-uterine segment, cervical leiomyoma preventing vaginal delivery) 8. Maternal choice after counseling regarding risks versus benefits COMPLICATED VAGINAL DELIVERY: OBSTETRIC FORCEPS AND VACUUM EXTRACTION DESCRIPTION OF THE OBSTETRIC FORCEPS Obstetric forceps have been used to facilitate vaginal deliveries since 1500 bc. Most commonly, the invention is credited to Peter Chamberlen and his brother of ­England. Designed originally as a means of extracting fetuses from women who were at high risk of dying during childbirth, forceps now are an alternative to cesarean delivery in women with a protracted second stage of labor. Originally, many of these instruments were furnished with hooks and other accessories of destruction, and they were intended to save the mother but not the fetus. Over the last 500 years, the modern instruments in current use have been through hundreds of modifications, safer techniques have been established, and the overriding goal now includes delivering an intact, living baby and a healthy mother (Meniru, 1996; O’Grady et al, 2002). Current obstetric forceps were first devised for practical use in the sixteenth and seventeenth centuries and were perfected over the past 300 years into the models in current use. Although there are many variations on the standard blueprint depending on the indication for its use, all obstetric forceps have a similar design. Forceps are made of stainless steel and consist of two blades (each approximately 37.5 cm long, crossing each other), a lock at the site of crossing, and a handle, whereby the instrument is grasped by the obstetrician. The long blades are parallel, divergent, or convergent, depending on the type of instrument; the lower 8 cm comprises the shank, which is the part between the blade proper and the handle, giving a length for the forceps that is sufficient to be locked easily outside the vagina. The four most common types of lock are the sliding lock, which can articulate anywhere along the shanks of the blade (Kielland and Barton’s forceps); the English, which is fixed, a double slot lock; the French, which is a screw lock; and the German, which is a combination of the English and French. The part of the forceps that grasps the fetal head is the blade; this is further divided into the heel, which is the part closest to the lock, and the toe, which is the most distal part of the blade. The blade can be either fenestrated—­meaning the body of the blade is hollow—or solid to prevent fetal head compression. A further modification is pseudofenestration, in which a solid blade has a ridged edge, combining the advantages of easier applicability and less fetal trauma that a solid blade affords with the ease of traction of a 147 FIGURE 15-1 Simpson forceps: a standard obstetric forceps with features common to all such instruments. TABLE 15-1 Types of Obstetric Forceps in Most Common Use Anatomic ­Modification General Use Tucker-McClane Solid blade Nonmolded vertex Simpson Parallel shanks Molded vertex or significant caput Elliot Convergent shanks Nonmolded vertex Laufe Pseudofenestrated blade; divergent shanks For preterm infants or EFW <2500 g English lock; absent pelvic curve For rotation of fetal vertex >45 degrees Long handles with no pelvic curvature For after-coming head in breech vaginal delivery Type Classic Rotational Kielland Breech Piper EFW, Estimated fetal weight fenestrated blade. Obstetric forceps also possess a rounded cephalic curve, which accommodates the fetal vertex, and a pelvic curve that mirrors the maternal pelvic curve (Figure 15-1). There are more than 60 different types of obstetric forceps described in the literature, but most of are not used currently. The forceps used most commonly today are described in Table 15-1 , along with their indications for use and the variations in anatomy, which distinguish one from the other. INDICATIONS FOR USE OF OBSTETRIC FORCEPS To an individual without obstetric training, the use of forceps can be a dangerous and difficult undertaking, fraught with potential trauma for both the mother and fetus. It is true that the use of this instrument, if not performed carefully or appropriately, can have serious consequences. Nevertheless, with properly trained hands, and a proper appreciation of its use, forceps can be lifesaving for both mother and fetus. The criteria for the safe application of obstetric forceps are as follows: 1. The cervix must be fully dilated. 2. The position of the fetal vertex must be known. Forceps should not be applied when the fetal presentation is in doubt. 148 PART IV Labor and Delivery 3. The fetal vertex must be engaged within the maternal pelvis. Often in difficult or challenging labors, significant caput can lead to the false impression that fetal station is lower than it actually is. For this reason, the obstetrician must be confident that the actual biparietal diameter has passed the pelvic inlet and that the leading part of the fetal skull is beyond the level of the ischial spines. In addition, when the presentation is occiput posterior, the leading point of the fetal skull may appear to be lower in the pelvis although the biparietal diameter has not yet passed through the pelvic inlet; this can also lead to an erroneous conclusion about fetal station. 4. When the forceps are properly applied, the sagittal suture must be exactly midway between the blades, and the lambdoidal sutures should be equidistant (usually one fingerbreadth) from the edge of the blade. If these conditions are not met, application of the forceps should be reconsidered. Furthermore, in 1988 the American College of Obstetricians and Gynecologists (ACOG) revised its classification of the type of forceps delivery, according to the station of the fetal vertex before forceps application (Ali and Norwitz, 2009), which is as follows: 1. Outlet forceps—the fetal vertex is visible at the labia without manually separating them, and the fetal skull has reached the pelvic floor 2. Low forceps—the leading point of the fetal skull is greater than 2 cm beyond the ischial spines 3. Mid forceps—the fetal head is engaged and is beyond the level of the ischial spines; the forceps should be applied only if cesarean delivery is not quickly or imminently possible, and the fetus is in distress; or there should be a high likelihood that the forceps operation will be successful 4. High forceps—the vertex is not engaged (i.e., not at the level of the ischial spines or beyond); under these circumstances, the forceps must never be applied. The forceps are further divided into whether they are rotational (sagittal suture is >45 degrees from the midline) or nonrotational (sagittal suture <45 degrees from the midline). The indications for use of the obstetric forceps are: 1. Maternal exhaustion or inability to push (endotracheal intubation with sedation or paralysis; neuromuscular disease) 2. Fetal heart tracing not reassuring or fetal distress 3. Maternal contraindications to pushing (cardiopulmonary disease, cerebrovascular aneurysm) 4. After-coming head in a vaginal breech delivery 5. Facilitate difficult extraction of the fetal vertex during cesarean delivery FORCEPS AND POTENTIAL NEONATAL MORBIDITY As stated previously, forceps were used for hundreds of years without regard to fetal survival and were primarily needed to facilitate or terminate difficult labors for maternal benefit. Today, with the widespread availability of cesarean delivery, considerations turn to providing the best neonatal outcome possible; therefore the difficult forceps deliveries of the past have been abandoned. Nevertheless, forceps still play a crucial role in modern obstetrics, and if judiciously used can provide a safer alternative to cesarean delivery for both mother and baby. The difficulty in interpreting the obstetric literature, in regard to neonatal morbidity incurred by forceps, is that the classification for type of forceps was revised by the ACOG in 1988; therefore prior studies do not use the same clinical criteria used currently to select appropriate candidates for forceps use. Furthermore, residency training in operative vaginal delivery has dramatically decreased over the past 30 years, potentially increasing fetal risk (Benedetto et al, 2007). Consequently, for proper interpretation of adverse outcomes, one must look to studies performed after the ACOG revised the classifications. The incidence of operative vaginal delivery in the United States is approximately 5% (1 in 20 deliveries), ranging from 1% to 23% with a 99% success rate ­(Menacker and Martin, 2008). The prevalence of forceps use varies widely by region (highest in the South), and recent estimates show that it accounts for approximately 25% (1:4 ratio with vacuum extractions) of all operative vaginal deliveries (Benedetto et al, 2006). National statistics show that of the 1,268,502 deliveries in 12 states (31% of U.S. deliveries) in 2005, 1% were forceps assisted and 3.7% were vacuum assisted, for a estimated total national operative vaginal delivery rate of 4.7% (Menacker and Martin, 2008). Unfortunately, the prevalence of low, outlet, and mid-forceps (the application of forceps when the head is engaged but the leading point of the skull is more than +2 cm station) deliveries nationally is not known, nor is the rate of rotational and nonrotational forceps use. Neonatal outcome by type of forceps application is also not generally known and is hindered by the fact that even with the universal ACOG classification scheme, the determination of fetal station and type of forceps can be subjective and dependent on the experience and examining skill of the obstetrician. Moreover the indication for forceps use varies widely by clinical situation, and the neonatal morbidity that can result from a “difficult pull” in a patient with a transverse arrest with marked fetal asynclitism may be different from the quick delivery of a 2600-g fetus whose mother is unable to push, even if both deliveries are by low-forceps. Nevertheless, what large studies show that long-term and short-term neonatal morbidity from outlet or low-­ forceps delivery is uncommon? In 2009, Prapas et al (2009) noted that the rate of vacuum- versus forceps-assisted deliveries had increased and that different maternal and neonatal outcomes have been proposed. The aim of their study was to compare the short-term maternal and neonatal outcomes between vacuum and forceps delivery. They conducted a medical record review of live born singleton, vacuum- and forceps-assisted deliveries. Of 7098 deliveries, 374 were instrument assisted, 324 were conducted by vacuum (86.7%), and 50 were assisted by forceps (13.3%). The incidence of third-degree lacerations and periurethral hematomas was similar between vacuum and forceps (3.4% vs. 2% and 0.3% vs. 0%, respectively), whereas perineal hematomas were more common in forceps compared CHAPTER 15 Complicated Deliveries: Overview with vacuum application (2% vs. 0.3%, respectively), albeit not significantly. The rate of neonates with Apgar scores ≤6 at 1 min was significantly higher after forceps compared with vacuum delivery (18% vs. 5.2%, respectively; p = 0.0003). The rate of neonatal trauma and respiratory distress syndrome did not differ significantly between the two groups. The conclusion was that both modes of instrumental vaginal delivery are safe in regard to maternal morbidity and neonatal trauma (Prapas et al, 2009). Alternatively, Benedetto et al, (2007) found that in healthy women with antenatally normal singleton pregnancies at term, instrument-assisted deliveries are associated with the highest rate of short-term maternal and neonatal complications. Of the 332 women who underwent an operative vaginal delivery, 201 met study criteria and were analyzed, with 54% forceps-assisted deliveries. Maternal complications were mostly associated with forceps-assisted and vacuum-assisted instrumental deliveries (odds ratio [OR], 6.9; 95% confidence interval [CI], 2.9 to 16.4; and OR, 3.0; 95% CI, 1.1 to 8.8, respectively, versus spontaneous deliveries). Neonatal complications were also mostly correlated with forceps-assisted and vacuumassisted instrumental deliveries (OR, 3.5; 95% CI, 1.9 to 6.7; and OR, 3.8; 95% CI, 2.0 to 7.4, respectively, versus spontaneous deliveries) (Table 15-2) (Bendetto et al, 2007). There are few randomized prospective trials specifically addressing the issue of neonatal morbidity arising from forceps operations, yet those that exist suggest no significantly increased risk from operative vaginal delivery when compared with spontaneous birth. Yancey et al (1991) randomized 364 full-term women at +2 station to elective outlet forceps delivery or unassisted birth. Women with suspected fetal macrosomia or chorioamnionitis were excluded. Neonates were examined at birth and 72 hours of age, and cranial ultrasound examinations were performed by neonatologists 24 to 72 hours after birth. The spontaneous and forceps-assisted births had no statistically significant differences in the incidence of scalp abrasions, facial bruising, cephalhematoma, subconjunctival hemorrhage, or abnormal cranial sonograms (Herbst and Källén, 2008). Several retrospective, population-based studies have further examined the issue of potential adverse neonatal sequelae arising from forceps procedures. Robertson et al TABLE 15-2 Risk of Intracranial Injury According to Type of Delivery Mode of Delivery Incidence of ­Intracranial Injury Vacuum 1 in 860 Forceps 1 in 664 Combined vacuum and forceps 1 in 256 Cesarean, in labor 1 in 907 Cesarean, not in labor 1 in 2750 Spontaneous vaginal delivery 1 in 1900 Adapted from Towner D, Castro MA, Eby-Wilkens E, Gilbert WM: Effect of mode of delivery in nulliparous women on intracranial injury, N Engl J Med 341:1709-1714, 1999. 149 (1991) reclassified forceps operations in accordance with the 1988 revised ACOG guidelines and examined neonatal outcomes by type of forceps operation, matched to cesarean section with a second stage of at least 30 ­minutes, from a similar station. They found that mid-forceps operations compared with a cesarean section from a comparable station resulted in greater neonatal resuscitation requirements, lower umbilical artery pH, and birth trauma, defined as nerve injuries or fractures. There was no increased risk of trauma from low forceps procedures compared with cesarean delivery, although the prevalence of arterial cord pH less than 7.10 was increased (Keith et al, 1988). It must be noted, however, that this study was not randomized and that an abdominal delivery group is not necessarily an appropriate control for operative vaginal delivery. Nevertheless, this large population-based study (from a database of 20,831) emphasizes the relative safety of outlet and low-forceps procedures, casting some doubt as to the safety of mid-forceps operations. The role of mid-forceps delivery in modern obstetrics, specifically in regard to rotation of 45 degrees or greater, has incited much controversy. The difficulty lies in the fact that no randomized trials exist comparing mid-forceps operations and other modes of delivery, and it is unlikely that any will be done in the near future given that training in this type of delivery has declined among obstetric residents in the United States (Kozak and Weeks, 2002; Learman, 1998). Hankins et al reported a case-control study of 113 rotational forceps of 90 degrees or greater matched to 167 controls with rotation 45 degrees or less. They found no significant differences in major injuries, defined as skull fracture, brachial plexus, facial or sixth nerve palsy, and subdural hemorrhage. There were also no differences in the prevalence of cephalhematoma, clavicular fracture, or superficial laceration. All deliveries were performed with Kielland forceps by resident staff members, with attending supervision as required. The prevalence of nerve injury in these forceps-assisted operations ranged from 2% to 3%; there were no skull fractures (Lipitz et al, 1989). The available literature supports the fact that neonatal morbidity from outlet or low forceps is exceedingly low, and comparable to spontaneous vaginal delivery. This finding comes from prospective data and large, population-based investigations. The evidence concerning neonatal safety from more advanced forceps operations (mid-forceps and rotations) shows that there is some degree of increased morbidity from these procedures, although whether this arises from the operative vaginal delivery itself or from difficult labor is not currently possible to discern (Meniru, 1996; O’Grady et al, 2002). In well-trained hands, the benefit of appropriately selected candidates for mid-­forceps or rotational forceps may be justifiable, although the number of physicians in the United States who are facile and comfortable with attempting these procedures is steadily declining. VACUUM DELIVERY: INDICATIONS, USES, AND COMPARISON WITH FORCEPS PROCEDURES Operative vaginal delivery for the indications listed previously can also be performed by the vacuum extractor. Vacuum extraction was first described in 1705 by Dr. James 150 PART IV Labor and Delivery Yonge, an English surgeon, several decades before the invention of the obstetric forceps. However, it did not gain widespread use until the 1950s, when it was popularized in a series of studies by the Swedish obstetrician Dr. Tage Malmström. By the 1970s, the vacuum extractor had almost completely replaced forceps for assisted vaginal deliveries in most northern European countries, but its popularity in many English-speaking countries, including the United States and the United Kingdom, was limited. By 1992, however, the number of vacuum-assisted deliveries surpassed the number of forceps-assisted deliveries in the United States, and by the year 2000 approximately 66% of operative vaginal deliveries were by vacuum (Ali and Norwitz, 2009; Hillier and Johanson, 1994). The situations that indicate the use of the vacuum and the requirements that must be fulfilled for its correct use are identical to those for the obstetric forceps. The device consists of a metal or plastic cup (flexible or semirigid) that is applied to the fetal vertex. Care is taken in its application to ensure that an adequate seal has been created, and that no maternal soft tissue is trapped between the suction device and the fetus. Traction is then applied to the fetal head in the line of the birth canal in an effort to assist delivery. It is also cautioned that rocking movements or torque to the cup are not used. The premature infant is a relative contraindication to vacuum application. It is generally advised that no more than three detachments occur before attempts at vacuum extraction are abandoned (Ali and Norwitz, 2009). In a laboratory experiment, Duchon et al (1988) compared the maximum force at suggested vacuum pressures (550-600 mm Hg) prior to detachments for different types of vacuum devices. They found that the average force of traction exerted before detachments ranged from 18 to 20 kg (Benedetto et al, 2007). This result is interesting to bear in mind when one considers the older data of Wylie who estimated the average tractive force required for delivery of infants weighing 15.95 kg for a primigravida and 11.33 kg for a multipara (Feingold et al, 1988) (Figure 15-2). The original vacuum device developed in the 1950s by the Swedish obstetrician Dr. Tage Malmström was a discshaped stainless steel cup attached to a metal chain for traction. Because of technical problems and lack of experience with this instrument, vacuum devices did not gain popularity in the United States until the introduction of the disposable cups in the 1980s. There are two main types FIGURE 15-2 A soft, bell-shaped vacuum extractor (top) and a rigid, mushroom-shaped vacuum extractor (bottom). (Courtesy Cooper S­urgical.) of disposable cups, which can be made of plastic, polyethylene, or silicone. The soft cup is a pliable, funnel- or bellshaped cup, which is the most common type used in the United States. The rigid cup is a firm mushroom-shaped cup (M cup) similar to the original metal disc-shaped cup and is available in three sizes. A metaanalysis of 1375 women in nine trials comparing soft and rigid vacuum extractor cups demonstrated that soft cups were more likely to fail to achieve a vaginal delivery, because of more frequent detachments (OR, 1.65; 95% CI, 1.19 to 2.29), but were associated with fewer scalp injuries (OR, 0.45; 95% CI, 0.15 to 0.60) and no increased risk of maternal perineal injury (Benedetto et al, 2007; Hillier and Johanson, 1994). For example, the risk of scalp laceration with the rigid Kiwi OmniCup (Clinical Innovations, Murray, Utah) was reported to be 14.1% compared with 4.5% using a standard vacuum device (p = 0.006). These and other authors concluded that handheld soft bell cups should be considered for more straightforward occiput-anterior deliveries, and that rigid M cups should be reserved for more complicated deliveries, such as those involving larger infants, significant caput succedaneum (scalp edema), occiput-posterior presentation, or asynclitism. The vacuum extractor is widely used in the United States, but is not free of preventing neonatal injury. Other than superficial scalp lacerations or abrasions, which usually heal without incident, as well as local soft tissue swelling or bruising, the use of the vacuum has been associated with cephalhematoma and subgaleal hemorrhages. Cephalhematoma occurs when the force created by the vacuum results in the rupture of diploic or emissary vessels between the periosteum and outer table of the skull; this fills the potential space that exists between the two with blood. Although cephalhematomas are often cosmetically alarming, they are limited to traveling along one cranial bone, because the firm periosteal attachments limit further extravasation of blood across suture lines. Thus large amounts of blood cannot usually collect in this space, and serious neonatal compromise from this bleeding is rare. In a randomized trial of continuous and intermittent vacuum application, Bofill examined factors associated with increasing the risk of cephalhematoma; they found that only asynclitism and traction time were independently related to this complication (Hartley and Hitti, 2005). There was a clear relationship between increasing time of vacuum application (up to 6 minutes) and cephalhematoma. Interestingly, Hartley and Hitti did not find a significant independent association of neonatal injury with continuous versus intermittent vacuum, or with decreasing gestational age or increasing birthweight. Furthermore, the number of detachments was not correlated with cephalhematoma. These results were further corroborated by Teng who conducted a prospective observational study of 134 vacuum extractions and found that only increasing total duration of vacuum application was associated with neonatal injury (Feingold et al, 1988). Metaanalysis of randomized trials comparing vacuum to forceps extractions showed that vacuums are more likely to fail to deliver the baby and lead to increased rates of cephalhematoma and retinal hemorrhage (Figure 15-3) (Vacca, 2007). Subgaleal hemorrhage poses much more of a potential risk for the neonate; it occurs when emissary veins above CHAPTER 15 Complicated Deliveries: Overview the skull and periosteum rupture, with blood dissecting through the loose tissue underlying the cranial aponeurosis, unimpeded by suture lines. A tremendous amount of blood, potentially the entire neonatal blood volume (approximately 250 mL) can fill this space, and thus can significantly compromise the neonate’s condition (Hillier and Johanson, 1994). Much of the literature about this rare complication of vacuum extraction was published in the 1970s and early 1980s, with few recent studies to detail associated risk factors. Plauche, in his classic paper on vacuum related neonatal injury, identified only 18 cases of subgaleal hematomas among 14,276 cases of vacuumassisted births, in contrast to a mean incidence of cephalhematoma of 6% (Keith et al, 1988). These morbidity 151 estimates are derived from data that are approximately 30 to 40 years old; nevertheless, Teng noted an incidence of cephalhematoma of 8%, and 0.7% for subgaleal hemorrhage in their more recent investigation, which agrees well with Plauche’s estimates (Herbst and Källén, 2008). A recent study from Australian investigators evaluated 37 cases of subgaleal hemorrhage at a single tertiary care center accrued over a period of 23 years, with an estimated prevalence of 1.54/10,000 total births. The finding was that this complication occurred most often in primigravidae, and that a large proportion of these infants (89.1% compared with 9.8% of the general control population) had an attempted vacuum extraction (Figure 15-4) (Hillier and Johanson, 1994; Kozak and Weeks, 2002). Subcutaneous edema Cephalhematoma Periosteum Bone FIGURE 15-3 A cephalhematoma is a hemorrhage that occurs under the periosteum of the skull and is thus confined to a defined space with limited capacity for expansion. (Adapted from Gilstrap LC, Cunningham FG, Hankins GDV, et al: Operative obstetrics, ed 2, Stamford, Conn, 2002, Appleton and Lange.) Skin Subgaleal hematoma Periosteum Bone FIGURE 15-4 A subgaleal hematoma spreads along subcutaneous soft tissue planes and has no immediate barrier to expansion, creating the potential for significant neonatal hemodynamic compromise. (Adapted from Gilstrap LC, Cunningham FG, Hankins GDV, et al: Operative obstetrics, ed 2, Stamford, Conn, 2002, Appleton and Lange.) 152 PART IV Labor and Delivery It must be kept in mind that the overall neonatal morbidity of the vacuum extractor is low, as ascertained from the large population-based study by Towner showed that the risk of intracranial injury is only 1 in 664 ­(Learman, 1998). Yet as outlined previously, there are definite neonatal risks associated with the use of vacuum extraction. The U.S. Food and Drug Administration has suggested that infants delivered with the vacuum have close monitoring for subgaleal or subaponeurotic hematoma, and that a high index of suspicion be maintained for this rare complication. It must be emphasized that the overall risk of adverse events attributable to the vacuum is extremely low, because the U.S. Food and Drug Administration estimated five serious complications per year of recent use during a time period when 228,354 vacuum deliveries were performed. However, it is likely that sequelae related to vacuum extraction often go underreported (Menacker and Martin, 2008). The choice of which instrument to use, forceps or vacuum, is usually determined by the obstetric care provider depending on the skill level and experience with either method. There have been several randomized trials exploring this issue (Menacker and Martin, 2008, Meniru, 1996; O’Grady et al, 2002; Prapas et al, 2009). The Cochrane Library has pooled the results from 10 randomized trials comparing neonatal morbidity and successful vaginal delivery between these two devices (Roberts et al, 2002). This analysis found that the vacuum was more likely than forceps to fail (OR, 1.69; 95% CI, 1.31 to 2.19), and was associated with a greater likelihood of Apgar scores less than 7 at 5 minutes (OR, 1.67; 95% CI, 0.99 to 2.81), cephalhematoma (OR, 2.38; 95% CI, 1.68 to 3.37), and retinal hemorrhage (OR, 1.99; 95% CI, 1.35 to 2.98). However the overall serious complication rate was low, and there was no difference in long-term morbidity between groups (Johnson and Menon, 2003). The greatest danger in the use of either vacuum or forceps comes in the combination of both instruments together. Towner showed that the use of one instrument after the other had failed and carried a neonatal intracranial injury risk of 1 per 256, significantly greater than that of either modality alone (Learman, 1998). This finding was further supported by the work of Gardella et al, (2001), who matched 11,223 women (a third of whom had combined instruments, vacuum alone, or forceps alone, respectively) to an equivalent number of spontaneous vaginal deliveries. These investigators found no statistically significant difference in intracranial hemorrhage when a single instrument was compared to spontaneous vaginal delivery; however, the combined use of both instruments markedly increased the risk of intracranial hemorrhage, seizures, and low 5-minute Apgar scores (Ron-El et al, 1981). The overall risk of nerve and scalp injury was greater when single-instrument delivery was compared with spontaneous delivery, but the overall incidence of each complication is rare. The vacuum extractor is an acceptable instrument if used judiciously and in the proper circumstances, carrying an overall minimal risk of serious neonatal complications. Its safety is comparable to the obstetric forceps, although it has a higher incidence of cephalhematoma, but a lower potential for facial nerve injury. The chance of failure is greater with the vacuum, which could then potentially tempt the health care provider to subsequently use the forceps. The use of both forceps and vacuum after one instrument has failed carries a higher risk of adverse outcomes, and it should be undertaken only with an understanding of the higher likelihood of neonatal morbidity. SHOULDER DYSTOCIA Shoulder dystocia is arguably the most dreaded complication in obstetrics. The problem posed by this entity is that although it is highly anticipated, it is unpredictable and can appear despite the most cautious measures taken to prevent it. Shoulder dystocia is defined as the delivery of the fetal head with an impaction of the fetal shoulder girdle or trunk against the pubic symphysis, making subsequent delivery either difficult or impossible without performing auxiliary delivery maneuvers. In some cases the posterior shoulder may be lodged behind the sacral promontory—a bilateral shoulder dystocia. Once shoulder dystocia occurs, a series of maneuvers—which have never been tested in a prospective fashion, because of the sporadic and unpredictable nature of this complication—are used to resolve it. The first step is usually the McRoberts maneuver, which consists of hyperflexing the maternal thighs onto the abdomen. This maneuver flattens the pubic symphysis and sacral promontory and facilitates delivery of both the anterior and posterior ­shoulders. If unsuccessful, this maneuver is usually followed by suprapubic pressure to remove the anterior shoulder from its impacted state behind the pubic symphysis. If these two maneuvers fail, either rotational maneuvers or extraction of the posterior fetal arm are usually tried. The Woods-Screw maneuver, or the Rubin’s rotational maneuver, is used in an attempt to rotate the infant’s shoulders in an effort to relieve the impaction of the shoulder against the pubic bone. Alternatively, delivery of the posterior arm can be accomplished by inserting the operator’s hand into the vagina and grasping the posterior fetal wrist and guiding it across the fetal chest and through the vaginal introitus. It is often necessary to perform an episiotomy to have sufficient room in the vagina to accomplish this maneuver. An alternative maneuver to fetal manipulation is the all-fours position, or Gaskin maneuver. With this maneuver, the mother is moved from the lithotomy position to a hands and knees position. Next the posterior fetal shoulder, which is now at the 12 o’clock position, is delivered with gentle downward traction. If the dystocia continues unresolved, the Zavanelli maneuver or cephalic replacement can be performed. After the fetal head is rotated from occiput transverse to occiput anterior, it is flexed and pushed back in the birth canal, and the child was delivered by emergent cesarean section. McRoberts maneuver, suprapubic pressure, or both will relieve greater than 50% of instances. Cephalic replacement should be necessary in rare cases. Recent data have focused on enhanced practitioner training for shoulder dystocia by means of simulation. It is estimated that almost 50% of currently practicing obstetric birth attendants have never successfully performed maneuvers other than the McRoberts maneuver and suprapubic pressure. Draycott et al (2008) have published data showing that the annual compulsory training of all attending obstetric physicians CHAPTER 15 Complicated Deliveries: Overview in Southmead Hospital in the United Kingdom reduced the incidence of fetal injury and brachial plexus injury within 3 years of introduction. These data await confirmation in the United States or other countries. The prevalence of shoulder dystocia varies depending on the population studied and the presence of various risk factors known to predispose women to this obstetric emergency. Estimates range from 0.2% to 1% in a low-risk population to 20% in higher-risk groups (Ali and Norwitz, 2009; Benedetto et al, 2007; Feingold et al, 1988; Hartley and Hittii, 2005; Herbst and Källén, 2008). Maternal obesity, fetal macrosomia, history of prior shoulder dystocia, and maternal diabetes mellitus are the most common associated variables, but are not of sufficient prognostic power to be clinically useful in predicting should dystocia (Ali and Norwitz, 2009; Benedetto et al, 2007; Herbst and Källén, 2008). Because shoulder dystocia has the potential to cause significant neonatal morbidity and mortality, efforts have been made to predict its occurrence; unfortunately, no clinical guidelines have been clinically tested or proved. Ultrasound examination is commonly used in patients with suspected fetal macrosomia or diabetes to detect large birthweight in infants who might be more likely to suffer shoulder dystocia. Third-trimester sonographic examination has an accuracy of 10% to 15% in the prediction of fetal weight, and is thus not highly reliable (Feingold et al, 1988; Hillier and Johanson, 1994; Keith et al, 1988). In addition, if ultrasound examination were completely reliable, the fetal weight cut-off that would prompt an elective cesarean section has not been determined. Lipscomb evaluated the deliveries of 227 mother-infant pairs at their institution with birthweights greater than 4500 g and found a shoulder dystocia rate of 18.5%; therefore the majority of those delivering vaginally did so without any adverse event occurring (Hartley and Hitti, 2005). In those with a shoulder dystocia, 51.7% infants had a neurologic injury, but at 2 months, all the infants had a normal neurologic examination. Therefore the occurrence of brachial plexus injury at birth does not necessarily guarantee permanent neurologic morbidity. This conclusion has been reinforced by several other studies. Gherman et al (2006) reviewed 285 cases of shoulder dystocia and found that 77 (24.9%) suffered fatal injury, most commonly of the brachial plexus (16.8%) or clavicular (9.5%) and humeral (4.2%) fractures. Almost half of the shoulder dystocias resolved with use of the McRoberts maneuver alone; the rest required Woods’ maneuver, posterior arm extraction, or the Zavanelli maneuver. The requirement of additional fetal manipulative procedures increased the risk of humeral fracture only and not clavicular or brachial plexus injury. The incidence of permanent musculoskeletal injury was only 1.4% (Kozak and Weeks, 2002). A prospective investigation evaluated the natural history of recovery following a birth-related brachial plexus injury of infants referred to a tertiary care, multidisciplinary neurological center. Enrollment required identification of injury in the newborn period, initial evaluation at the center between 1 and 2 months of age, and lack of antigravity movement in the shoulder or elbow persisting until 2 weeks of age. In this group of children subject to ascertainment bias (as those injuries resolving before 2 weeks of age would not have been included in the results), complete 153 neurologic recovery was documented in 66%, and only 14% had persistent, severe weakness (Learman, 1998). In the best systematic review of brachial plexus injury to date, the risk of permanent brachial plexus impairment, if recognizable at birth, was 15% to 20% (Pondaag et al, 2004). Rouse et al (1996) elaborated further on these concepts in their decision analysis, which showed that if one chose to perform an elective cesarean section for all women without diabetes and with sonographically predicted macrosomia (estimated fetal weight >4000 g), 2345 cesarean sections would need to be performed to prevent one permanent brachial plexus injury. If the 4500-g cutoff were selected, 50% more cesarean deliveries would be needed to prevent one permanent brachial plexus injury. In the mother with diabetes, if one chose a cut-off of 4500 g or greater, 443 cesareans would need to be done to prevent one permanent injury (Keith et al, 1988)—a tradeoff that most practitioners now believe is acceptable. The conclusions from this decision analysis have been borne out by several other investigators who have established that the risk of nerve injury certainly increases with rising birthweight, but the large number of macrosomic infants who have a normal, spontaneous vaginal delivery without sequelae does not justify a policy of elective cesarean for macrosomia alone in a nondiabetic population (Lipitz et al, Loucopoulos and Jewewicz, 1982; Menacker and Martin, 2008; Meniru, 1996; O’Grady et al, 2002; Pondaag et al, 2004; Roberts et al, 2002). Therefore at present there is no universally accepted method to prevent shoulder dystocia. Studies have shown that operative vaginal delivery, especially vacuum delivery, of a fetus suspected to have macrosomia either clinically or sonographically could increase the risk of shoulder dystocia (Ron-El et al, 1981). It seems wise to avoid difficult forceps or vacuum delivery if a patient is thought to have an infant weighing more than 4000 g, especially if she has diabetes or a past history of shoulder dystocia. In addition, the ACOG states that “for pregnant women with diabetes who are suspected of carrying macrosomic fetuses, a planned cesarean delivery may be a reasonable course of action, depending on the incidence of shoulder dystocia, the accuracy of predicting macrosomia, and the cesarean delivery rate within a specific population” (Ali and ­Norwitz, 2009; Benedetto et al, 2007). The documentation of the events surrounding shoulder dystocia are important, as is the discussion of current status and future status of an infant delivered with a birth injury after shoulder dystocia. Both obstetric and pediatric providers should debrief the shoulder dystocia event immediately after the delivery. In the case of fetal injury, it is optimal for both obstetric and pediatric providers to discuss the delivery events and subsequent newborn treatment plans with the mother before discharge. VAGINAL BREECH DELIVERY Three percent to 4% of all infants at term will be in the breech presentation at the time of delivery. There are three main types of breech presentations. The footling breech has one (single footling) or both (double footling) lower extremities presenting. The frank breech has both thighs flexed, but legs extended. The complete breech has both thighs and legs flexed. The vaginal delivery of a singleton 154 PART IV Labor and Delivery footling breech carries attendant risks of cord prolapse and head entrapment, and the consensus among obstetricians is that this presentation should be delivered operatively (unless the fetus is a second twin, see later discussion on Twin Delivery). The frank breech with buttocks presenting has a lower risk of these adverse events occurring, and thus could potentially deliver vaginally. The complete breech presentation will convert to frank or footling during labor, and the appropriate management scheme for delivery depends on which leading fetal part will descend. The mechanics of vaginal breech delivery are as follows. The frame of reference for the presenting part is the sacrum (i.e., sacrum anterior, posterior, or transverse). In the absence of urgent fetal indications, the singleton breech is allowed to deliver passively with maternal expulsive efforts until the infant has been delivered past the umbilicus. At this point the legs are gently reduced, and the trunk and body are gently rotated to bring the sacrum anteriorly. With the appearance of the scapula below the maternal symphysis, the arms are then delivered by gently sweeping them across the chest. Every effort is then made to keep the neck from extending during the delivery of the aftercoming head; this is accomplished during delivery of the body by an assistant exerting suprapubic pressure on the fetal head to keep it flexed. Once the body has delivered, the delivery of the head is accomplished by either the Mariceau-Smellie-Veit maneuver, in which one hand extends along the posterior neck and occiput and applies pressure to prevent hyperextension, while the other hand gently applies downward traction against the maxilla to flex the head forward as the head is delivered, or with Piper forceps directly applied to the fetal vertex. The feasibility of vaginal breech delivery and its safety have been the subject of much debate throughout the past half century. With the advent of safe, expedient cesarean delivery in the United States, many obstetricians have favored the operative approach as the method of choice for management of the breech presentation at term. The literature to support this point of view has produced conflicting conclusions, and its interpretation is consequently difficult. There has been an extensive body of literature over the past half century examining this issue. Unfortunately, there are only two randomized trials that have explored the question of which delivery route is best for the term singleton frank breech fetus (Ali and Norwitz, 2009; Benedetto et al, 2007), but there are several large retrospective series describing neonatal outcomes with the vaginal approach, most of which suggest that vaginal delivery in carefully selected patients carries a low risk of long-term neonatal morbidity and mortality. Diro et al (1999) evaluated 1021 term singleton breech deliveries occurring at their institution over a 4-year period. Infants with a clinically adequate pelvis and frank breech presentation less than 3750 g were allowed a trial of labor. They found an overall cesarean rate of 85.6%; however, for women allowed to deliver vaginally, the success rate, defined as vaginal delivery, was 50% (19 of 38 patients) for nulliparous women and 75.8% (116 of 153 patients) for multiparous women. The length of neonatal intensive care unit (NICU) stay was higher for the group delivered vaginally (17.4% vs. 12.1%; p = 0.036), but major morbidities between operative and vaginal delivery were not significantly different (Feingold et al, 1988). Long-term outcome was not evaluated. Of note, the women in this cohort had pelvic dimensions evaluated clinically, and not by x-ray or computed tomography pelvimetry, as has been performed in other studies. Norwegian investigators examined their similar policy of vaginal breech delivery and evaluated maternal and neonatal outcomes in a large cohort of women (Fang and Zelop, 2006). Patients were allowed a trial of vaginal delivery if they had adequate x-ray pelvimetry, and they were excluded if they had an estimated fetal weight greater than 4500 g or a footling presentation. Each vaginal breech was matched to a term vaginal vertex birth; each cesarean delivery was matched to the appropriate vertex control; 1212 breech deliveries were evaluated, 639 (52.7%) of which were vaginal, 172 (11.4%) were intrapartum cesarean section, and 138 (11.4%) were planned cesarean section. Once major or lethal anomalies and fetal disorders not related to delivery were excluded, there were no perinatal deaths attributable to mode of delivery. When births planned vaginally were compared with operative deliveries, there was an increased risk of 1-minute Apgar score less than 7 and traumatic morbidity in the vaginal group, but no significant differences in NICU admissions, 5-minute Apgar scores, or normal neonatal course. The conclusion reached was that short-term morbidity was worse in the group delivered vaginally, but long-term outcome was similar between groups. It must be remembered, however, that this study was not randomized, and the infants selected for vaginal delivery versus cesarean section were intrinsically different. In addition, all deliveries occurred in a tertiary care institution. The truly interesting point raised by these investigators is that, aside from the issue of cesarean versus trial of labor, singleton breech infants regardless of mode of delivery have an increased risk of morbidity compared with their vertex counterparts. Breech infants had higher incidences of NICU admissions, eventful courses, hip dislocation, and traumatic morbidity (soft tissue trauma, fracture, facial nerve paralysis, and brachial plexus palsy). Thus both the obstetrician and pediatrician must be aware that the infant in breech presentation requires careful attention upon birth for the presence of these potential factors. Christian evaluated their policy of offering a trial of vaginal breech delivery to women with an estimated fetal weight between 2000 and 4000 g and having computed tomographic pelvimetry documenting adequate pelvic dimensions (Herbst and Källén, 2008). Of 122 women evaluated, 85 were judged appropriate by these standards for vaginal delivery, of which 81.2% had a successful vaginal delivery. The only indices of neonatal outcome evaluated were the Apgar score, which was not different between groups, and neonatal cord pH, which was lower in a statistically but not clinically significant manner in the vaginal delivery group. The largest difficulty in interpreting the large number of retrospective studies examining the issue of vaginal breech delivery in the obstetric literature is that even the best designed reports have relatively small numbers of patients, and the possibility of a type II or beta error is high. In addition, the groups being compared (planned cesarean section and vaginal delivery) are different because the patients are not randomized (Herbst and Källén, 2008; Hillier CHAPTER 15 Complicated Deliveries: Overview and Johanson, 1994). Often the women chosen to have a cesarean section tend to have factors that would place their neonates at higher risk than those allowed a trial of labor. A metaanalysis evaluating seven cohort studies and two randomized trials compared 1825 trial of labor patients to 1231 elective cesarean patients and found a statistically significant, but clinically questionable, increased risk with vaginal breech delivery of 1.10% (Keith et al, 1988). There is a paucity of randomized trials to specifically address the role of the vaginal breech delivery in modern obstetrics. Collea et al randomized 208 women with a singleton frank breech presentation at term to vaginal delivery or cesarean section; they found a low overall risk of permanent birth injury or neonatal morbidity in the vaginal delivery group, although the incidence of neonatal morbidity was higher in the vaginal delivery group (Ali and Norwitz, 2009). Of note, a majority of conditions listed as morbidities (hyperbilirubinemia, meconium aspiration, mild brachial plexus injury) had resolved by hospital discharge. In addition, decreased neonatal morbidity with cesarean section was offset by a striking increase in higher maternal risk in the operative group. It must be remembered that this study was published in 1979, and standards of maternal and neonatal care have changed dramatically since then. Until 2001, no additional randomized investigations were available to settle the long raging debate of the optimal mode of delivery for term singleton breech. This controversy has recently been clarified by a large, multicenter, multinational trial that randomized 2088 women at 121 centers in 26 countries to planned vaginal birth or planned elective cesarean section (Benedetto et al, 2007). Criteria for enrollment were frank or complete term singleton breech with no evidence of fetal macrosomia. The investigation was halted when preliminary results showed that there was significantly increased neonatal mortality and severe morbidity in the vaginal breech arm compared with the cesarean arm (5.0% vs. 1.6%). This conclusion was not altered by the experience of the delivering obstetrician or maternal demographic factors such as parity and race. Maternal morbidity between both groups was comparable. The difference in outcome was even more striking in countries, such as the United States, with a low national perinatal mortality rate (5.7% vs. 0.4%). Criticisms of this study are that patients enrolled did not have computed tomography pelvimetry performed, which in some institutions is standard practice before considering a vaginal breech delivery. Furthermore, subjects did not have continuous fetal monitoring, but rather intermittent fetal auscultation every 15 minutes. In addition, the capability of various centers to perform emergent cesarean sections differs markedly, and this could have potentially affected the neonatal morbidity and mortality rate. Nevertheless, it is unlikely that another large study will ever be performed to examine this issue again, and the ultimate results are difficult to dispute given the excellent study design and adequate sample size. Indeed the ACOG recently issued a statement that “planned vaginal delivery of a term singleton breech may no longer be appropriate … patients with a persistent breech presentation at term in a singleton gestation should undergo a planned cesarean delivery” (Kozak and Weeks, 2002). As will be discussed later in Multifetal Delivery, this statement 155 does not apply to the vaginal delivery of a nonvertex second twin. The delivery of the vaginal breech is also an emotional issue; physicians trained in the art of the vaginal breech delivery maintain that for an appropriately selected candidate, vaginal breech delivery has acceptable neonatal risk and has the advantage of sparing the mother significant operative morbidity. Proponents of cesarean delivery further state that the level of resident training in the art of the singleton vaginal breech delivery has markedly diminished, with most graduating senior residents having performed few such births. Nonetheless, many practitioners will be required to assist in vaginal birth of a breech infant in unplanned situations. The acquisition of skills necessary to competently perform this procedure may need to be learned and practiced with simulation-based training, because the opportunities for training in most residency programs are few. MULTIFETAL DELIVERY With the advent of in vitro fertilization and the sophisticated assisted reproductive technologies (ARTs), the incidence of multifetal pregnancies has dramatically increased, particularly higher-order multiples. The incidence of twin gestations in patients undergoing in vitro fertilization is currently more than 30%, and it is 1% to 3% in higher-order multiples. As a result, the frequency of twin gestation in the United States has increased 65% since 1980; twins now account for 3% of births. Of these, 80% are dizygotic and 20% are monozygotic. This 3% of births accounts for 17% of preterm births and approximately 25% of infants of low birthweight and very low birthweight. The perinatal mortality rate of twins is sevenfold that of singletons, of which a small fraction is due to problems during labor and at delivery. The mode of delivery for twins is well delineated by several studies, and the issues surrounding the choice of vaginal birth versus cesarean is outlined in the following sections (Smith et al, 2005). TWIN DELIVERY Vertex-Vertex It is almost universally accepted that the appropriate method of delivery is vaginal if both twins are vertex. The first infant is delivered like a singleton infant. The second infant is delivered in a similar fashion, but care must be taken not to rupture membranes before the head is well engaged, because this may increase the risk of cord accident. Of note, the delivery of the second twin does not necessarily occur immediately after the first. Vertex-Nonvertex The first twin is usually delivered vaginally. The options for delivery of the second twin are as follows: cesarean section, breech extraction, or attempts at external cephalic version and vertex delivery of the second twin if successful. The subject of the optimal delivery choice for the second twin has been the subject of much controversy. Many obstetricians claim that cesarean section is the safest approach 156 PART IV Labor and Delivery to the nonvertex twin, whereas others claim that vaginal delivery affords equivalent neonatal outcome, sparing the mother from an unnecessary surgical procedure (Usta et al, 2005). Fortunately there is a large body of evidence in the literature addressing these issues. If the vaginal approach is chosen, once the first twin is delivered the obstetrician inserts a hand into the uterine cavity and, under sonographic guidance if necessary, finds the feet of the second twin. Once the feet are firmly grasped, they are brought down into the vagina, and the membranes are then ruptured. Traction is applied to the fetus along the pelvic curve; once the body has been delivered through the introitus, delivery of the arms, shoulders, and aftercoming head proceed in a fashion similar to that of a singleton breech. In 2003, Hogle et al performed a metaanalysis to determine whether a policy of planned cesarean or planned vaginal birth is preferable for twins. They found only four studies with a total of 1932 infants met their inclusion criteria. There were no significant differences in maternal morbidity, perinatal or neonatal mortality, or neonatal morbidity between the two groups. They did find significantly fewer low 5-minute Apgar scores in the planned cesarean group, principally because of a reduction among breech first twins. They concluded that, if twin A is vertex, “there is no evidence to support planned cesarean section for twins.” In contrast, Smith et al (2005) published a retrospective cohort study of 8073 twin births after 36 weeks of gestation in Scotland between 1985 and 2001. There was a death of either twin in two of 1472 (0.14%) deliveries by planned cesarean and in 34 of 6601 (0.52%) deliveries by other means (p = 0.05; OR for planned cesarean, 0.26; 95% CI, 0.03 to 1.03). They concluded that planned cesarean may reduce the risk of perinatal deaths of twins at term by 75% despite the lack of statistical significance in outcomes between the two groups. The data also suffer from the fact that 30 of the 36 deaths were in second twins, and there were no data regarding fetal presentation (Smith et al, 2004). There are several large cohort studies examining the issue of feasibility and safety of total breech extraction of the nonvertex second twin. These studies have almost unanimously reached the similar conclusion that the neonatal outcome for nonvertex second twins delivered vaginally is similar to the vertex first twin, but is not statistically different from those second twins delivered by cesarean section, regardless of birthweight or gestational age (Ali and Norwitz, 2009; Benedetto et al, 2007; Feingold et al, 1988; Hartley et al, 2005; Herbst and Källén, 2008; Hillier and Johanson, 1994; Keith et al, 1988). Hartley et al (2005) conducted a retrospective analysis of birth certificates and fetal and infant death certificates for 5138 twin pairs selected from those born in Washington State from 1989 to 2001. They concluded that if prompt vaginal delivery of twin B does not occur, the benefits of vaginal delivery for twin A might not outweigh the risks of distress and low Apgar scores in twin B and vaginal plus cesarean delivery for the mother (Hartley and Hitti, 2005). Although there are several retrospective studies evaluating the outcome of vaginally born nonvertex second twins, there is only one randomized trial (Hillier and Johanson, 1994). Rabinovici et al (1987) allocated 60 women with vertex-nonvertex twins to either operative or vaginal delivery. Maternal morbidity and hospital stay were increased in the surgical group, but there were no differences in neonatal outcome. In addition, Acker et al (1982) retrospectively reviewed 150 nonvertex second twins of all birthweights, 74 delivered by cesarean and 76 by breech extraction, and they found no mortality in either group and a 3.9% incidence of low Apgar scores in the breech extraction group, which was no different from that in the cesarean group. Chervenak et al (1986) reviewed 76 breech extractions for second twins, found no morbidity or mortality in the 60% weighing more than 1500 g, and concluded “for birthweights >1500 gm, routine cesarean for vertex/nonvertex twins may not be necessary.” The available body of evidence supports attempts at vaginal delivery of the nonvertex second twin. Of course the responsible obstetrician must choose a management plan most compatible with his or her experience and training. For those not versed in the techniques of successful vaginal breech extraction, cesarean delivery might be a more prudent plan. As in the case of the singleton vaginal breech, simulation training may play a role in the acquisition and maintenance of skills for safe vaginal breech birth. In addition to abdominal delivery and total breech extraction, there is the option of external cephalic version (i.e., attempting to turn a nonvertex fetus to vertex by abdominal manipulation). Studies have shown that this option is associated with a higher failure rate at successful vaginal delivery and other complications (such as cord accidents and malpresentations not amenable to vaginal delivery) when compared with primary breech extraction or cesarean section (Hillier and Johanson, 1994; Learman, 1998). Nonvertex-Nonvertex Because of to the theoretical risk of interlocking twins, as well as the recent data showing the greater morbidity for the singleton vaginal breech (see the preceding discussion), cesarean section is the recommended choice for delivery of the nonvertex-nonvertex presentation. Monochorionic, Monoamniotic Twins Monochorionic, monoamniotic twins share a single intra­ amniotic space, and thus have a higher risk of cord and extremity entanglement during the course of delivery. It is commonly accepted that the optimal mode of delivery is a planned cesarean section before labor ensues. HIGHER-ORDER MULTIPLE GESTATIONS Most perinatologists would suggest cesarean delivery for triplets and higher-order multiples (Lipitz et al, 1989). Although this practice is common, the data mandating operative delivery are far from conclusive. A Dutch study compared the outcomes of triplets delivered vaginally and abdominally at two institutions (Learman, 1998). One hospital favored cesarean section, whereas at another trial of labor was offered to all appropriate candidates. The success of vaginal delivery was relatively high (34 of 39 women [87%]). There was a higher incidence of neonatal mortality and postdelivery depression (as estimated by Apgar score) in the hospital favoring operative CHAPTER 15 Complicated Deliveries: Overview delivery compared with the vaginal delivery group. The biases inherent in this study are obvious, although the reported findings have been corroborated by several other reports from single institutions that offer trial of vaginal delivery to triplet gestations (Menacker and Martin, 2008; Meniru, 1996; O’Grady et al, 2002). Vintzileos et al (2005) attempted to estimate the risks of stillbirth and neonatal and infant deaths in triplets, according to mode of delivery; they used the “matched multiple birth” data file that was composed of triple births that were delivered in the United States during 1995 through 1998 and found that 95% of all triplets were delivered by cesarean delivery. Vaginal delivery (all vaginal) was associated with an increased risk for stillbirth (relative risk, 5.70; 95% CI, 3.83 to 8.49) and neonatal (relative risk, 2.83; 95% CI, 1.91 to 4.19) and infant (relative risk, 2.29; 95% CI, 1.61 to 3.25) deaths. They concluded that cesarean delivery of all three triplet fetuses is associated with the lowest neonatal and infant mortality rate and that vaginal delivery among triplet gestations should be avoided (Vintzileos et al, 2005). Most of the data on triplet births consists of small cohort studies and not randomized trials, and the possibility of type II or beta errors exists in the interpretation of many of these studies (Feingold et al, 1988; Keith et al, 1988). Thus delivery of triplet gestations vaginally while not an unreasonable approach has nearly disappeared from the practice of modern obstetrics in favor of routine cesarean delivery. Currently quadruplets and other higher-order multiples are usually delivered by cesarean section (Ron-El et al, 1981). VAGINAL BIRTH AFTER CESAREAN: NEONATAL ISSUES Cesarean section accounts nationally for one quarter of all deliveries (Ali and Norwitz, 2009). Surgery carries the maternal risks of increased blood loss, prolonged hospital stay, and longer recovery period compared with vaginal delivery. During the 1980s to 1990s, efforts were made to encourage women to attempt vaginal birth after a prior cesarean delivery, because success rates vary from 60% to 80% for vaginal delivery, dependent on the indications for the prior cesarean delivery (Benedetto et al, 2007). Whereas VBAC rates remain relatively high in the United Kingdom at 33% (range, 6% to 64%), the rates are decreasing rapidly in the United States from a high rate of 28.3% in 1996 to less than 10 per 1000 deliveries in 2006 (Caughey, 2009; Fang and Zelop, 2006). This decrease occurred primarily because VBAC can result in uterine dehiscence, in which the prior scar asymptomatically separates or, more seriously, uterine rupture occurs. This decrease has significantly affected the United States cesarean section rate of 31.1% (Caughey, 2009; Fang and Zelop, 2006). A full discussion of VBAC, studies supporting its safety, and controversies surrounding its feasibility is beyond the scope of this chapter, and the interested reader is urged to consult Williams Obstetrics, ed 23, for further details. This discussion instead focuses on neonatal risks from VBAC, particularly from its most dreaded complication, uterine rupture. Studies have uniformly shown a risk of uterine rupture with VBAC on the order of 0.5% to 1% (Benedetto 157 et al, 2007; Caughey, 2009; Feingold et al, 1988). A recent large, retrospective study evaluated 20,095 women with a history of prior cesarean delivery and found that rupture risk was 0.16% if the woman elected a repeated, elective operative delivery; 0.52% if VBAC occurred as a result of spontaneous labor; 0.77% if labor was induced without prostaglandins; and 2.5% if labor was induced with prostaglandins (Feingold et al, 1988). Thus VBAC carries the lowest risk if labor is spontaneous and not augmented. There are few large, well-designed studies specifically evaluating neonatal rather than maternal outcomes in VBAC. Most recently, Kamath et al (2009) performed a retrospective cohort study of 672 women with one prior cesarean section undergoing trial of labor. They found that infants born by cesarean delivery had higher rates of admission to the NICU (9.3% compared with 4.9%) and higher rates of oxygen supplementation for delivery room resuscitation (41.5% compared with 23.2). Yap retrospectively evaluated 38,027 deliveries occurring at a single tertiary care institution and found 21 cases of uterine rupture; 17 occurred after a history of a prior cesarean delivery (Herbst and Källén, 2008). The two neonatal deaths that occurred were a result of prematurity (23-week-old fetus) and multiple congenital anomalies; all live born infants were discharged from the hospital without neurologic sequelae. Thus the ultimate neonatal outcome despite uterine rupture was favorable. However, all deliveries occurred in a tertiary care institution with readily available obstetric anesthesiologists, neonatologists, and obstetricians. Most deliveries after diagnosis of rupture occurred within 26 minutes. A third group of investigators retrospectively identified 99 cases of uterine rupture occurring over a period including 159,456 births (Hillier and Johanson, 1994). Thirteen of these ruptures occurred before the onset of labor. There were six neonatal deaths, but four of these occurred in women with uterine rupture at admission, and thus were never given a trial of labor. There were five cases of perinatal asphyxia, but once again it is not detailed whether these occurred in women allowed a trial of labor or in those who had ruptured on presentation to the hospital. Moreover, many of these women had an undocumented prior scar, which in some institutions would warrant an elective repeated cesarean section. The aforementioned recent study evaluating 20,095 women with a prior cesarean delivery and their subsequent risk of rupture found a neonatal mortality of 5.5% (Feingold et al, 1988). However, because this was a population based study, it was not specified whether these deliveries occurred in tertiary care institutions with the capability of performing emergent operative rescue procedures in the event of uterine rupture. Finally Fang et al (2006) reviewed all of the literature to date in regard to adverse neonatal outcomes and found that the combined rates of intrapartum stillbirth and neonatal death were not statistically different between trial of labor and those who elected for repeated cesarean section. Thus the true neonatal risk of VBAC, especially in the event of uterine rupture, cannot be precisely estimated at the current time. There are no studies adequately evaluating long-term outcomes of surviving infants after uterine rupture (Fang and Zelop, 2006). 158 PART IV Labor and Delivery It appears that the risk of uterine rupture after a prior cesarean delivery is low, but this risk increases when labor is augmented with oxytocin or prostaglandins. It is appropriate to offer women VBAC, but they must be counseled carefully about the potential risk of uterine rupture. Careful documentation of the informed consent and labor management must be completed. Moreover, VBAC should ideally occur in a tertiary care institution or in facilities capable of rapidly performing an emergent cesarean section, because this improves the likelihood of minimizing adverse neonatal sequelae. CORD ACCIDENTS The term cord accident usually refers to adverse events affecting the fetus that occur as a result of a problem with the umbilical cord. This heterogeneous term encompasses umbilical cord prolapse, in which the cord delivers through the cervix and compression by a fetal part results in a significantly increased risk of asphyxia; it also includes such entities as cord entanglements or “true knots,” which can lead to fetal compromise. The incidence of such events is not clearly known, because the diagnosis is often one of exclusion. One large population-based study compared 709 cases of cord prolapse occurring among 313,000 deliveries to matched controls and found that low birthweight, male sex, multiple gestations, breech presentation, and congenital anomalies all increased the risk of umbilical cord prolapse (Ali and Norwitz, 2009). Not surprisingly, cord prolapse was associated with a high mortality rate (10%) that was reduced if cesarean rather than vaginal delivery was performed. The standard of care in cases of cord prolapse is to proceed immediately with cesarean section as quickly as possible while an assistant elevates the presenting fetal part with a vaginal hand to prevent compression of the umbilical cord. It is also of paramount importance to have appropriate pediatric support available at the time of delivery, because the newborn is likely to be depressed and require resuscitation. Cord accident, or in utero compromise, secondary to entanglement of the umbilical cord as a clinical entity is difficult to understand. Often in cases of in utero fetal demise (IUFD), a diagnosis or cause of fetal death is never found. It is tempting to attribute the demise to an event that compromises umbilical blood flow to the developing pregnancy. The literature on this subject is scarce. ­Hershkovitz et al (2001) identified 841 cases of true knots from a population of 69,139 deliveries (for a prevalence of 1.2%) and in a case-controlled study found that grand multiparity (>10 deliveries), chronic hypertension, history of genetic amniocentesis, male gender, and umbilical cord prolapse were all independently associated with true knots of the umbilical cord. The presence of a true knot was associated with both in utero fetal demise and greater FIGURE 15-5 (See also Color Plate 2.) A 28-year-old woman was admitted to the labor and delivery department with an intrauterine demise. Examination of the fetus shows the cord wrapped tightly around the torso, leg, and ankle, suggesting cord accident as a cause of death. No other pathologic abnormalities were found on autopsy. (Courtesy Thomas R. Easterling.) likelihood of cesarean delivery (Figure 15-5) (Benedetto et al, 1994). SUGGESTED READINGS American College of Obstetricians and Gynecologists: ACOG Committee Opinion No. 340: Mode of term singleton breech delivery, 108:235-237, 2006. American College of Obstetricians and Gynecologists: Fetal macrosomia. ACOG Practice Bulletin No. 22, Obstet Gynecol 96 , 2000. American College of Obstetricians and Gynecologists: Operative vaginal delivery. ACOG Practice Bulletin No. 17, Obstet Gynecol 95 , 2000. American College of Obstetricians and Gynecologists: Shoulder dystocia. ACOG Practice Bulletin No. 40, Obstet Gynecol 100:1045-1050, 2002. American College of Obstetricians and Gynecologists: Vaginal birth after cesarean section. ACOG Practice Bulletin No. 54, Obstet Gynecol 104:203-212, 2004. Draycott TJ, Crofts JF, Ash JP, et al: Improving neonatal outcome through practical shoulder dystocia training, Obstet Gynecol 112:14-20, 2008. Fang YM, Zelop CM: Vaginal birth after cesarean: assessing maternal and perinatal risks: contemporary management, Clin Obstet Gynecol 49:147-153, 2006. Hartley RS, Hitti J: Birth order and delivery interval: analysis of twin pair perinatal outcomes, J Matern Fetal Neonatal Med 17:375-380, 2005. Johanson RB, Menon BK: Vacuum extraction versus forceps for assisted vaginal delivery, Cochrane Database Syst Review (2):CD000224, 2000. Kamath BD, Todd JK, Glazner JE, et al: Neonatal outcomes after elective cesarean delivery, Obstet Gynecol 113:1231-1238, 2009. Prapas N, Kalogiannidis I, Masoura S, et al: Operative vaginal delivery in singleton term pregnancies: short-term maternal and neonatal outcomes, Hippokratia 13:41-45, 2009. Thorngren-Jerneck K, Herbst A: Low 5-minute Apgar score: a population-based register study of 1 million term births, Obstet Gynecol 98:65-70, 2001. Complete references and supplemental color images used in this text can be found online at www.expertconsult.com C H A P T E R 16 Obstetric Analgesia and Anesthesia Mark D. Rollins and Mark A. Rosen This chapter introduces some of the scientific background and clinical techniques used in providing obstetric analgesia and anesthesia. These practices provide substantial benefit to the patient in labor and are essential for operative delivery. Although the effects of obstetric analgesia and anesthesia on the fetus and neonate are typically benign, there is potential for significant neonatal effects. HISTORY OF OBSTETRIC ANESTHESIA Modern obstetric anesthesia began in Edinburgh, Scotland, on January 19, 1847, when the professor of obstetrics James Young Simpson used diethyl ether to facilitate child delivery by anesthetizing a woman with a contracted pelvis. Morton’s historic demonstration of the anesthetic properties of ether at the Massachusetts General Hospital in Boston had occurred only 3 months earlier. Fanny Longfellow, wife of Henry Wadsworth Longfellow, was the first American to receive anesthesia for childbirth, publicly proclaiming in 1847, “This is certainly the greatest blessing of this age” (Longfellow and Wagenknecht, 1956). Although anesthesia for surgery was rapidly and widely accepted, most of Simpson’s contemporaries in the United States, France, and England were critical of using anesthesia in obstetrics, presenting both medical and religious arguments in outspoken opposition. However, the public outcry in support of labor analgesia was vehement. In 1853, the negative reaction from Thomas Wolsley, editor of The Lancet, after John Snow administered ether for Queen Victoria’s eighth child was so strong that court physicians denied anesthesia had been used. The great debate was largely settled 4 years later when Victoria delivered her ninth and last child, and the use of a royal anesthetic was acknowledged. Although many physicians had remained opposed, public opinion had changed and women were requesting labor analgesia from their doctors. During the second half of the twentieth century, anesthesiologists made significant advances in techniques and improved safety for delivering labor analgesia. Hingson and Edwards (1943) developed the continuous caudal catheter that preceded development of the epidural catheter. Apgar (1953) initially proposed a simple neonatal scoring system as a guide for evaluating the effects of obstetric anesthesia and later as a guide for neonatal resuscitation. Other early pioneers in the emerging specialty of obstetric anesthesia were Gertie Marx (Marx and Orkin, 1958), John Bonica (1967), and Sol Shnider et al (1963). These pioneers helped to characterize the normal changes in maternal physiology related to pregnancy, confirm the safety and efficacy of obstetric analgesia, determine the effects of these techniques on uterine blood flow and placental transfer of anesthetic agents, and evaluate the effects of these techniques and agents on newborn well-being. ANATOMY OF LABOR PAIN Contraction of the uterus, dilatation of the cervix, and distention of the perineum cause pain during labor and delivery. Somatic and visceral afferent sensory fibers from the uterus and cervix travel with sympathetic nerve fibers to the spinal cord (Figure 16-1). These fibers pass through the paracervical tissue and course with the hypogastric nerves and the sympathetic chain to enter the spinal cord at T10 to L1. During the first stage of labor (cervical dilation), the majority of painful stimuli are the result of afferent nerve impulses from the lower uterine segment and cervix, as well as contributions from the uterine body causing visceral pain (poorly localized, diffuse, and usually described as “a dull but intense aching”). These nerve cell bodies are located in the dorsal root ganglia of levels T10 to L1. During the second stage of labor (pushing and expulsion), afferents innervating the vagina and perineum cause somatic pain (well localized and described as “sharp”). These somatic impulses travel primarily via the pudendal nerve to dorsal root ganglia of levels S2 to S4. Pain during this stage is caused by distention and tissue ischemia of the vagina, perineum, and pelvic floor muscles, associated with descent of the fetus into the pelvis and delivery. Neuraxial analgesic techniques that block levels T10 to L1 during the first stage of labor must be extended to include S2 to S4 for efficacy during the second stage of labor. Labor pain can have significant physiologic effects on the mother, fetus, and the course of labor. Pain stimulates the sympathetic nervous system, elevates plasma catecholamine levels, creates reflex maternal tachycardia and hypertension, and reduces uterine blood flow. In addition, changes in uterine activity can occur with the rapid decrease in plasma epinephrine concentrations associated with onset of neuraxial analgesia. Oscillations in epinephrine can result in a range of uterine effects from a transient period of uterine hyperstimulation (Clarke et al, 1994) to a transient period of uterine quiescence, or conversion of dysfunctional uterine activity patterns associated with poorly progressive cervical dilation to more regular ­patterns associated with normal cervical dilation (Leighton et al, 1999). CHANGES IN MATERNAL PHYSIOLOGY AND THE IMPLICATIONS During pregnancy, labor, and delivery, women undergo fundamental changes in anatomy and physiology. These alterations are caused by changing hormonal activity, biochemical shifts associated with increasing metabolic demands of a growing fetus, placenta, and uterus, and mechanical displacement by an enlarging uterus (Cheek and Gutsche, 2002; Parer et al, 2002). 159 160 PART IV Labor and Delivery T 10 11 12 L1 T-10 11 12 L-1 3 4 2 FIGURE 16-1 Parturition pain pathways. Nerves that accompany sympathetic fibers and enter the neuraxis at the T10, T11, T12, and L1 spinal levels carry afferent pain impulses from the cervix and uterus. Pain pathways from the perineum travel to S2, S3, and S4 via the pudendal nerve. (From Bonica JJ: Principles and practice of obstetric analgesia and anesthesia, Philadelphia, 1967, F.A. Davis Co.) MATERNAL CIRCULATORY SYSTEM Hypotension can occur when a pregnant woman is in the supine position because of compression of the vena cava by the gravid uterus. Significant aortoiliac artery compression occurs in 15% to 20% of parturients and vena caval compression is universal, often as early as 13 to 16 weeks’ gestation. Vena caval compression contributes to lower extremity venous stasis and can cause ankle edema and varices despite increased collateral circulation. Venous compression by the gravid uterus diverts some blood returning from the lower extremities through the internal vertebral venous plexus, the azygos, and the epidural veins. This increases the likelihood of entering an epidural vein with spinal or epidural anesthetic techniques. Anesthetic interventions that diminish sympathetic tone can further exacerbate the effects of vena caval compression induced by supine positioning, potentially causing profound hypotension. Therefore supine positioning is avoided during anesthetic administration in the second and third trimesters. Significant lateral tilt is used in all operative deliveries and frequently during labor analgesia to help preserve uterine blood flow and fetal circulation. Cardiac output increases during pregnancy, reaching an output 50% greater than the prepregnant state by the third trimester. During labor, maternal cardiac output increases during the first and second stages, reaching an additional 40% above prelabor values in the second stage ­(Robson et al, 1987). Each uterine contraction results in the autotransfusion of 300 to 500 mL of blood back into the maternal central circulation. The greatest increase in cardiac output occurs immediately after delivery, when values can increase as much as 75% above predelivery ­levels. This abrupt increase in cardiac output is secondary to the loss of aortocaval compression, autotransfusion from the contracted uterus, and decreased venous pressure in the lower extremities (Kjeldsen, 1979). Physiologic (dilutional) anemia of pregnancy occurs as a result of a greater increase in plasma volume (45%) than in red blood cell volume (20%) at term. Average blood loss at delivery—Approximately 500 mL for vaginal delivery and 1000 mL for cesarean section—is well tolerated because of this expanded blood volume and autotransfusion (normally in excess of 500 mL) from the contracted uterus after delivery (Cheek and Gutsche, 2002). MATERNAL AIRWAY AND RESPIRATORY SYSTEMS During pregnancy, the maternal airway has significantly increased mucosal edema and tissue friability throughout the pharynx, larynx, and trachea. These changes make laryngoscopy and intubation more challenging. In addition, the presence of comorbidities such as preeclampsia, upper respiratory tract infections, and the active pushing and increased venous pressure during the second stage further exacerbate airway tissue edema (Munnur et al, 2005). At term, minute ventilation is increased approximately 50%, oxygen consumption is increased by more than 20%, and functional residual capacity is decreased by 20%. The combination of these changes (increased oxygen consumption and decreased oxygen reserve) results in a state promoting rapid desaturation during periods of apnea. The changes in both airway and respiratory physiology during pregnancy make ventilation and intubation more difficult and increase the potential for complications. MATERNAL GASTROINTESTINAL SYSTEM The gravid uterus displaces the stomach cephalad and anterior and the pylorus cephalad and posterior. These changes reposition the intraabdominal portion of the esophagus into the thorax and result in decreased competence of the esophageal sphincter. Higher progesterone and estrogen levels further reduce esophageal sphincter tone. Gastric pressure is increased by the gravid uterus and by the lithotomy position used during vaginal delivery. Maternal gastric reflux symptoms increase with the gestational age of the pregnancy and ultimately affect the majority of parturients (Marrero et al, 1992). Beyond midgestation, women are at increased risk for pulmonary aspiration of acidic gastric contents; this is caused by decreased tone and competence of the lower esophageal sphincter as well as delayed gastric emptying with onset of labor or administration of opioids. This risk has important implications for induction of general anesthesia and airway management by the anesthesiologist, and it is discussed in detail under General Anesthesia. UTERINE AND FETAL CIRCULATION Uterine weight and blood flow increase throughout gestation from approximately 100 mL/min before pregnancy to approximately 700 mL/min (10% of cardiac output) at term gestation, with 50% to 80% of the uterine blood flow perfusing the intervillous space (placenta) and 20% to 50% supporting the myometrium. Uterine vasculature has limited autoregulation and remains (essentially) maximally dilated under normal conditions during pregnancy. CHAPTER 16 Obstetric Analgesia and Anesthesia Maternal uterine blood flow decreases as a result of either decreased uterine arterial perfusion pressure or increased arterial resistance. Decreased perfusion pressure can result from systemic hypotension secondary to hypovolemia, aortocaval compression, or significant decreases in vascular resistance from the initiation of neuraxial anesthesia or induction of general anesthesia. Uterine perfusion pressure can also decrease from increased uterine venous pressure associated with vena caval compression (e.g., supine position), uterine contractions (particularly uterine hypertonus by oxytocin hyperstimulation), or significant increase in intra-abdominal pressure (pushing during second stage or seizure activity). Stress-induced endogenous catecholamines and exogenous vasopressors can increase uterine artery resistance and decrease uterine blood flow. Despite these potential effects, phenylephrine (alpha-adrenergic) is useful for treating maternal hypotension secondary to neuraxial anesthesia, and it has been demonstrated to result in less fetal acidosis and base deficit compared to treatment with ephedrine (primarily ­beta-adrenergic) in many clinical trials (Lee et al, 2002a; Ngan Kee et al, 2009; Smiley, 2009). ANALGESIC OPTIONS FOR LABOR AND VAGINAL DELIVERY The pain of labor is highly variable and described by many women as severe. Factors influencing the patient’s perception of labor pain include duration of labor, maternal pelvic anatomy in relation to fetal size, use of oxytocin, parity, participation in childbirth preparation classes, fear and anxiety about childbirth, attitudes about and experience of pain, and coping mechanisms. Labor analgesia prevents autonomic reflex effects that can be deleterious for certain high-risk patients and their fetuses (e.g., patients with severe preeclampsia, valvular heart disease, myasthenia gravis). The American College of Obstetricians and the American Society of Anesthesiologists issued a joint statement indicating that a maternal request for pain relief is sufficient justification for administration of analgesics during labor (­American College of Obstetricians and Gynecologists, 2002). Although older observational studies associate increased cesarean delivery rates with early administration of epidural analgesia, more recent randomized studies found no difference in the rate of cesarean delivery or instrument-assisted vaginal delivery between women in whom analgesia was initiated early in labor versus later (Eltzschig et al, 2003; Wong et al, 2005). Consequently the American Society of Anesthesiology stated in its 2007 Practice Guidelines, “Neuraxial analgesia should not be withheld on the basis of achieving an arbitrary cervical dilation, and should be offered on an individualized basis. Patients may be reassured that the use of neuraxial analgesia does not increase the incidence of cesarean delivery” (American Society of Anesthesiologists Task Force on Obstetric Anesthesia, 2007). The choice of analgesic method resides primarily with the patient. The medical condition of the parturient, stage of labor, urgency of delivery, condition of the fetus, and availability of qualified personnel are also factors. Many different techniques are available to alleviate labor and delivery pain. Analgesia refers to pain relief without loss of consciousness; regional analgesia denotes partial sensory 161 BOX 16-1 Techniques for Labor Analgesia Nonpharmacologic analgesia Systemic medications Opioid analgesics Sedatives and anxiolytics Dissociative analgesics Inhalation analgesia Regional analgesia Epidural Spinal Combined spinal-epidural Paracervical block Pudendal block blockade in a specific area of the body, with or without partial motor blockade. Regional anesthesia is the loss of sensation, motor function, and reflex activity in a limited area of the body. General anesthesia results in the loss of consciousness and the goals for providing general anesthesia typically include hypnosis, amnesia, analgesia, and skeletal muscle relaxation. Techniques for labor analgesia must be safe for mother and fetus and individualized to satisfy the analgesic requirement and desires of the parturient; they also must accommodate the changing nature of labor pain and the evolving, varied course of labor and delivery (e.g., spontaneous vaginal delivery, instrumentally assisted vaginal delivery, and cesarean delivery). The current approaches to pain relief are outlined in Box 16-1. NONPHARMACOLOGIC ANALGESIA There is a variety of nonpharmacologic techniques for labor analgesia. These techniques include hypnosis, the breathing techniques described by Lamaze, acupuncture, acupressure, the LeBoyer technique (LeBoyer, 1975), transcutaneous nerve stimulation, massage, hydrotherapy, vertical positioning, presence of a support person, intradermal water injections, and biofeedback. A metaanalysis reviewing the effectiveness of a support individual (e.g., doula, family member) noted that parturients with a support individual used fewer pharmacologic analgesia methods, had a decreased length of labor, and had a lower incidence of operative deliveries (Hodnett et al, 2007). In a 2006 retrospective national survey of women’s childbearing experiences, although neuraxial methods of pain relief were rated as the most helpful and effective, nonpharmacologic methods of tub immersion and massage were rated more or equally helpful in relieving pain compared with the use of opioids (Declercq et al, 2006). Although many nonpharmacologic techniques seem to reduce labor pain perception, most studies lack the rigorous scientific methodology for the useful comparison of these techniques to pharmacologic methods. SYSTEMIC MEDICATIONS Systemic medications for labor and delivery are widely used, but are administered with limitations on both dose and timing because they readily cross the placenta and are 162 PART IV Labor and Delivery associated with a risk of neonatal respiratory depression in a dose-dependent fashion. Although the use of systemic opioid analgesics is common (e.g., fentanyl, meperidine, morphine, nalbuphine, butorphanol), the use of sedatives and anxiolytics (e.g., barbiturates, phenothiazine derivatives and benzodiazepines), and dissociative agents (e.g., ketamine, scopolamine) is rare. Opioid Analgesics Opioids are the most frequently used systemic analgesic, but for many patients opioids do not provide adequate analgesia during labor and delivery (Bricker and Lavender, 2002; Olofsson et al, 1996). Opioids are inexpensive, easy to administer, and do not require a trained anesthesia provider. However, they have a high rate of maternal side effects (sedation, respiratory depression, dysphoria, nausea, pruritus), can decrease fetal heart rate variability and fetal movements, and carry a potential risk of neonatal respiratory depression and changes in neurobehavior. Systemic administration of opioids at doses that are safe for mother and newborn provides some analgesia, but it is not a substitute for the analgesia provided by regional techniques. Systemic opioids are recommended for administration in the smallest doses possible with minimization of repeated dosing to reduce the accumulation of drug and metabolites in the fetus. Systemic opioids are most useful for patients with minimal pain, precipitous labor, or contraindications to neuraxial blockade, such as a coagulopathy. The opioid analgesics act by binding to opiate receptors located throughout the central nervous system. Although this is the main site of action, opioid receptors have also been identified in other peripheral tissues. Four main types of receptor have been identified: mu (μ), kappa (κ), delta (δ), and sigma (σ). The analgesic effects, side effects, and pharmacodynamic characteristics depend on the receptor affinity of each individual opioid. These receptors are also responsible for the associated respiratory depression, sedation, and dysphoria and may affect thermoregulation. Binding of opiate agonist agents (e.g., meperidine, morphine, and fentanyl) and subsequent receptor activation alters neural transmission of pain by inhibiting the presynaptic and postsynaptic release of neurotransmitters. The opioid antagonists (e.g., naloxone, naltrexone) are competitive antagonists and bind with high affinity to the μ receptor more so than the κ or δ opioid receptors. These drugs do not produce analgesia and are capable of displacing other agonist drugs from the receptor and reversing their effects. A third class of agents interacts with the receptors and results in both agonist and antagonist activity depending on the receptor type. These drugs, with high receptor affinity, are capable of analgesic effect (e.g., pentazocine, nalbuphine, and butorphanol). Opioids differ in pharmacokinetics, pharmacodynamics, method of elimination, and the presence of active metabolites, but all readily cross the placental barrier through passive diffusion. In order for systemic opioids to effectively alleviate labor pain, larger doses would be necessary. This dosing would risk excessive maternal sedation, maternal respiratory depression, loss of protective airway reflexes, newborn respiratory depression, and impairment of both early breastfeeding and neurobehavior. Consequently, large doses of opioids are avoided, and the doses used routinely for labor analgesia typically do not have significant adverse effects on either the mother or neonate. Meperidine remains the most widely used opioid worldwide. Maternal half-life of meperidine is 2 to 3 hours, with the half-life in the fetus and newborn being significantly greater and more variable at values between 13 and 23 hours (Kuhnert et al, 1979). In addition, meperidine is metabolized to an active metabolite (normeperidine) that can significantly accumulate after repeated doses. With increased dosing and shortened time interval between dose and delivery, there is greater neonatal risk of decreased Apgar scores, lowered oxygen saturation, prolonged time to sustained respiration, abnormal neurobehavior, and more difficulty initiating successful breastfeeding (Nissen et al, 1997). Morphine was used more frequently in the past, but currently is rarely used. Like meperidine it has an active metabolite (morphine-6-glucuronide) and a prolonged duration of analgesia (3 to 4 hours). The half-life is longer in neonates compared with adults, and it produces significant maternal sedation. Fentanyl is a synthetic opioid with a short duration of action (approximately 30 minutes), no active metabolites, and a ratio of fetal to maternal plasma concentrations of approximately 1:3. In small intravenous (IV) doses of 50 to 100 μg/hr there were no significant differences in Apgar scores, respiratory depression, or neurobehavior scoring compared with newborns of mothers who did not receive fentanyl (Rayburn et al, 1989a). In addition, a comparison of equianalgesic doses of IV fentanyl compared with IV meperidine (Rayburn et al, 1989b) demonstrated a decreased frequency of maternal nausea, vomiting, and prolonged sedation in the fentanyl group. In addition, neonates whose mothers received meperidine required naloxone more often compared with the fentanyl-exposed infants. There was no difference in the neuroadaptive testing scores of the two groups of infants. Opioids can be given by intramuscular (IM) or IV administration. IM injection of opioids is technically easy but leads to uneven analgesia, the possibility of late respiratory depression, and profound neonatal effects if not properly timed (Shnider and Moya, 1960). IM administration is associated with a high incidence of neonatal depression 2 to 4 hours after injection in some of the less lipid soluble opioids. IV administration is the most widely used technique to give opioids to a woman in labor, with effects that are more predictable and doses more easily timed. However, achievement of a steady blood level of opiate sufficient to provide analgesia is difficult, with the parturient frequently suffering underdosage (rarely overdosage). In an effort to improve pain relief and maternal satisfaction, continuous IV infusion of short-acting opiates (e.g., alfentanil, remifentanil) or self-administration of IV opiates is increasingly used for systemic opioid delivery ­during labor. Patient-controlled analgesia (PCA) has the implied advantage of allowing the patient to titrate her dose to the minimum required for analgesia with the lowest blood levels of opiates, resulting in considerably less placental transfer and fewer side effects with increased patient satisfaction (McIntosh and Rayburn, 1991). Remifentanil, an CHAPTER 16 Obstetric Analgesia and Anesthesia ultra-short–acting opioid rapidly metabolized by nonspecific serum esterases, is significantly metabolized by the fetus with umbilical artery–to–vein ratios of approximately 0.3 (Kan et al, 1998). It can be used effectively for labor PCA, but it appears difficult to achieve satisfactory analgesia without significant potential of maternal respiratory depression (Olufolabi et al, 2000; Thurlow and Waterhouse, 2000; Volikas, 2001). In a prospective randomized controlled trial comparing the effectiveness of epidural analgesia to a remifentanil PCA with optimized settings, epidural analgesia was significantly more effective than PCA in regard to labor pain, but no differences were noted in neonatal outcome measures (Volmanen et al, 2008). Although there are individual differences among opioids, all readily cross the placental barrier and exert neonatal effects in typical clinical doses, including decreased fetal heart rate variability and dose-related neonatal respiratory depression. All opioids can have maternal side effects, including nausea, vomiting, pruritus, and decreased gastrointestinal motility and stomach emptying. Sedatives and Anxiolytics Sedatives and anxiolytics are administered infrequently to pregnant patients because they increase risks of sedation and respiratory depression in both mother and newborn, especially when used with opioids. Sedatives and anxiolytics were used more frequently in the past to diminish the adverse motivational-affective component of labor pain. Examples of such drugs are barbiturates and benzodiazepines. Diazepam and midazolam are benzodiazepines used as anxiolytic agents in obstetrics. They rapidly cross the placenta, yielding approximately equal maternal and fetal blood levels within minutes of IV administration (Cree et al, 1973). In addition, the neonate has a limited ability to excrete diazepam, so the drug and its active metabolite may persist in significant amounts in the neonate for up to 1 week (Scher et al, 1972). Diazepam can result in neonatal hypotonia, lethargy, and hypothermia when used in large maternal doses (30 mg) (Cohen et al, 1993). However, when it is used in small doses (2.5 to 10 mg IV), minimal sedation and hypotonia have been observed (McAllister, 1980). Midazolam has a shorter duration of activity, but rapidly crosses the placenta and is associated with neonatal hypotonia in larger doses. The use of benzodiazepines remains somewhat controversial, but these agents can reduce maternal anxiety and are useful for treating convulsions associated with local anesthetic toxicity or eclampsia. However, all benzodiazepines are amnestics, and therefore may not be appropriate in many childbirth situations, depending on the desires of the parturient. Dissociative Analgesia The IM or IV administration of low-dose ketamine (0.25 mg/kg) produces a state called dissociative analgesia, which is characterized by analgesia and unreliable amnesia without loss of consciousness or protective airway reflexes (Galloon, 1976). This state is accompanied by a dreaming phenomenon, which may be unpleasant but can 163 be minimized by coadministration of benzodiazepines to improve amnesia. In divided doses totaling less than 1 mg/kg, ketamine provides adequate analgesia for vaginal delivery and episiotomy repair. Although these low doses are not associated with neonatal depression, higher doses are associated with decreased Apgar scores (Akamatsu et al, 1974). With larger doses, maternal airway protection cannot be guaranteed, increasing the risk of gastric content aspiration. Ketamine is best reserved for use as a supplement (in low doses) to other techniques or for situations in which (1) more reliable and safer agents or techniques are contraindicated or ineffective or (2) when rapid control is required because the mother’s pain is compromising the fetus (e.g., mother is moving uncontrollably, unable to effectively push, and jeopardizing delivery while the fetal head is presenting during a prolonged fetal heart rate deceleration). INHALATION ANALGESIA The use of nitrous oxide is widespread in Canada, ­Australia, the United Kingdom and other parts of the world, but its use in the United States as a labor analgesic is uncommon. It is usually administered as a 50% mixture with oxygen. At a 50% concentration (without coadministration of opioids), nitrous oxide is insufficient to cause unconsciousness or loss of protective airway reflexes. Appropriate equipment and trained personnel are essential to ensure safety (i.e., limiting the nitrous oxide concentration, avoiding administration of a hypoxic mixture, avoiding coadministration of other agents). Nitrous oxide is a weak analgesic but can provide satisfactory pain relief for some parturients. It can be used during the first, second, or third stage of labor either alone or as a supplement for a regional block or local infiltration (Rosen, 1971). The use of 50% nitrous oxide in a supervised fashion is safe and rapid acting, causes minimal maternal cardiovascular or respiratory depression, and does not affect uterine contractility. The effects of nitrous oxide are rapidly reversed with discontinuation, and it does not cause neonatal depression regardless of duration of administration (Rosen, 2002a). Although the use of inhaled halogenated agents such as sevoflurane in low-inspired concentrations was found to be safe and effective at reducing labor pain in a small study (Yeo et al, 2007), barriers to routine use include the need for specialized vaporizers, scavenging systems, and the lack of larger studies. NEURAXIAL (REGIONAL) ANALGESIA Neuraxial analgesia, including epidural, spinal, and combined spinal-epidural (CSE) techniques, has become the most widely used method for labor analgesia in the United States (Bucklin et al, 2005). Neuraxial techniques typically involve epidural and spinal administration of local anesthetic agents, and often the co-administration of epidural and spinal opioid analgesics. Other adjuvant agents, such as clonidine and neostigmine, can decrease the dose of local anesthetics or opioids required for effective analgesia; however, they are not routinely used and do not appear to offer a significant advantage when compared to local anesthetics with or without opioids (Eisenach, 2009; Parker et al, 2007; Roelants, 2006). 164 PART IV Labor and Delivery Neuraxial Local Anesthetics Local anesthetic agents consist of amine moieties linked by an intermediate chain containing an ester or amide. Local anesthetics reversibly block nerve impulse conduction via voltage-gated sodium channels. Their chemical structures are secondary or tertiary amines, which are weak bases, marketed as the hydrochloride salts to achieve aqueous solubility. The ester-linked local anesthetics (e.g., chloroprocaine, procaine, tetracaine) are rapidly metabolized by plasma cholinesterase, decreasing the risk of maternal toxicity and placental drug transfer (O’Brien et al, 1979). Amide-linked local anesthetics (e.g., lidocaine, bupivacaine, ropivacaine) are degraded by P-450 enzymes in the liver. Local anesthetics differ in their onset, peak plasma concentration, potency and duration based on their lipid solubility, protein binding, site of injection, and concentration. Vascular absorption of local anesthetics limits the safe dose that can be administered. Elevated plasma concentrations produce neurologic toxicity (seizures) or cardiovascular toxicity (myocardial depression, ventricular arrhythmia). Bupivacaine and ropivacaine are the most commonly used local anesthetics for labor analgesia. Ropivacaine is a pure amide-linked S-isomer, unlike bupivacaine, which is a racemic mixture. Because the R-isomer of bupivacaine can bind strongly to cardiac sodium channels, ropivacaine can decrease the possibility of severe cardiac toxicity associated with accidental intravascular injection of large doses of bupivacaine. However, ropivacaine is less soluble than bupivacaine and may be slightly less potent. Regardless of differences, bupivacaine and ropivacaine are extremely safe when appropriately used for epidural or intrathecal administration. An accidental, large intravascular dose of any local anesthetic can result in significant maternal morbidity or mortality. As with all drugs, placental transfer is determined by molecular size, lipid solubility, protein binding, and maternal drug concentration. Local anesthetics are weak bases with high lipid solubility and a low ionized fraction. However, the lower pH of the fetus has the potential to increase the fraction of ionized molecules, decrease lipid solubility, and result in ion trapping. Therefore in an acidotic fetus, higher concentrations of local anesthetic can accumulate (ion trapping). Increased concentrations of local anesthetics result in decreased neonatal neuromuscular tone similar to that seen with magnesium. If a direct intravascular or intrafetal injection of local anesthetics occurs, significant toxicity and depression can develop, signified by bradycardia, ventricular arrhythmia, and severe cardiac depression with acidosis. Neuraxial Opioids Although intraspinal (intrathecal) opiates were demonstrated in 1979 to be capable of producing profound analgesia in humans (Behar et al, 1979; Wang et al, 1979), the epidural injection of opioids as sole agents has proved to be of limited use for effective labor analgesia. In one study, high doses of epidural morphine (7.5 mg) provided satisfactory but not excellent analgesia for 6 hours in the first stage of labor, whereas 2 to 5 mg produced barely satisfactory analgesia in fewer than half the patients (Hughes et al, 1984). Besides the inadequate analgesia and the long onset time (approximately 1 hour), the side effect of pruritus was significant. In contrast, when lipid-soluble opioids (e.g., sufentanil, fentanyl) are administered in the epidural space as sole agents, analgesia is rapid and equivalent to that of systemic administration (Camann et al, 1992), but remains inferior to that of dilute concentrations of local anesthetics, and less effective for somatic pain associated with the second stage of labor. When lipid-soluble opioids are administered as an adjunctive agent with local anesthetics in the epidural space, they decrease the total local anesthetic dose required and lower the minimum local anesthetic concentration needed to achieve adequate labor analgesia (Buyse et al, 2007; Celleno and Capogna, 1988; Lyons et al, 1997). The most common maternal side effect of conventional doses of epidural fentanyl or sufentanil is pruritus. After maternal epidural administration, the lipid-soluble, poorly ionized opioid analgesics rapidly enter the fetal circulation. An epidural bolus injection of an opioid results in peak neonatal depression at 30 to 60 minutes (similar to an IV bolus dose). Although typical doses used for labor analgesia adversely affect the neonate, the potential for respiratory depression is a function of the amount and timing of drug administered. Subarachnoid (i.e., spinal, intrathecal) injections of fentanyl, sufentanil, meperidine, and morphine as sole agents are more promising for effective maternal labor analgesia. Analgesic effects of spinal opioids are more potent than epidural or systemic administration, but are of limited duration (2 hours) and are less effective than dilute epidural solutions of local anesthetics for analgesia in the second stage (Honet et al, 1992; Leighton et al, 1989). Spinal opioid administration is often performed as part of a CSE technique (discussed in the following section), with fentanyl or sufentanil being the most commonly used agents. The intrathecal opioid is often combined with a small dose of local anesthetic (e.g., 2.5 mg bupivacaine), decreasing the dose of opioid needed and incidence of pruritus (Wong et al, 2000). Reports of fetal heart rate changes after intrathecal administration of fentanyl or sufentanil (Cohen et al, 1993) may be caused by rapid onset of analgesia and rapid decrease in circulating catecholamines, with epinephrine decreasing faster than norepinephrine, resulting in an unopposed oxytocic effect on the uterus. This effect would increase uterine tone and decrease uterine blood flow. This mechanism is speculative, but it is suggested by observed cases and case reports (Friedlander et al, 1997). Some prospective randomized studies have found no difference in incidence of fetal bradycardia between epidural administration of local anesthetics and intrathecal opioids administered with the CSE technique (Fogel et al, 1999; Nageotte et al, 1997). A systematic review of studies comparing intrathecal opioids to other methods of labor analgesia noted an increase in fetal bradycardia (odds ratio [OR], 1.8; 95% CI, 1.0 to 3.1) and increased maternal pruritus (RR, 29.6; 95% CI, 13.6 to 64.6), but the risk of cesarean section because of FHR abnormalities was similar (Mardirosoff et al, 2002). NEURAXIAL TECHNIQUES FOR LABOR ANALGESIA Neuraxial techniques represent the most effective form of labor analgesia, and they achieve the highest rates of maternal satisfaction (Declercq et al, 2006). The patient CHAPTER 16 Obstetric Analgesia and Anesthesia remains awake and alert without sedative side effects, maternal catecholamine concentrations are reduced (Shnider et al, 1983), hyperventilation is avoided ­(Levinson et al, 1974), cooperation and capacity to participate actively during labor are facilitated, and predictable analgesia can be achieved, superior to the analgesia provided by all other techniques. Before initiating any neuraxial blockade, anesthesiologists assess the patient’s gestational and health history, perform a focused physical examination, discuss the risks, benefits, and alternatives, and obtain consent. In otherwise healthy parturients, routine laboratory tests are not required (American Society of Anesthesiologists Task Force on Obstetric Anesthesia, 2007). Resuscitation equipment and drugs must be immediately available to manage serious complications secondary to initiation of epidural or spinal blocks (discussed under Contraindications and Complications of Neuraxial Techniques). During initiation of the neuraxial blockade, mother and fetus are closely monitored. EPIDURAL ANALGESIA Epidural labor analgesia is a catheter-based technique to provide continuous analgesia during labor. The technique involves insertion of a specialized needle and catheter (Figure 16-2) between vertebral spinous processes in the back, into the epidural space. Most commonly, the needle is inserted at a lumbar space between L1 and L4. The needle traverses the skin and subcutaneous tissues, supraspinous ligament, interspinous ligament, and the ligamentum flavum, and it is advanced into the epidural space (Figure 16-3). The tip of the needle does not penetrate the dura, which forms the boundary between the intrathecal or subarachnoid space and the epidural space. To locate the epidural space, a tactile technique called loss of resistance is used. The tactile resistance noted with pressure on the plunger of an air- or saline-filled syringe dramatically decreases as the tip of the needle is advanced through the ligamentum flavum (dense resistance) into the epidural space (no resistance), which has an average depth of approximately 5 cm from the skin. Once the needle is properly positioned, a catheter is inserted through the needle. 165 The catheter remains in the epidural space and the needle is removed. The catheter is secured with adhesives and used for intermittent or continuous injections. Once the catheter is placed, analgesia is achieved by administration of local anesthetics, opioids, or both (see Neuraxial Local Anesthetics and Neuraxial Opioids, earlier) and maintained throughout the course of labor and delivery. The catheter can also be used for operative anesthesia (cesarean delivery) and postoperative analgesia, when necessary. After an incremental local anesthetic bolus to initiate ­reliable analgesia, local anesthetics are typically infused continuously with similar or lower blood drug levels compared with repetitive, intermittent boluses of local anesthetics (Hicks et al, 1988; Rosenblatt et al, 1983). Most importantly, the possibility of disastrous complications is reduced with continuous infusions, such as total spinal anesthesia or massive intravascular injections with cardiovascular collapse secondary to large bolus doses of local anesthetics (D’Athis et al, 1988). If the catheter enters the subarachnoid space instead, the level of sensory and motor blockade increases slowly without the sudden onset of complete subarachnoid blockade that can occur with large bolus techniques (Li et al, 1985). Patient-controlled epidural anesthesia is a delivery technique allowing the patient to self-administer small boluses of epidural analgesics with or without a background infusion. Studies comparing patient-controlled epidural anesthesia with continuous infusion technique have found decreased local anesthetic requirements, less anesthesia provider intervention, equivalent or improved patient satisfaction, equivalent or decreased motor blockade, and no significant differences in effects on the fetus or neonate (Halpern, 2005; van der Vyver et al, 2002). The choices for local anesthetics for epidural infusion include dilute solutions of bupivacaine, ropivacaine, lidocaine, or chloroprocaine (Lee et al, 2002b). The concentration and volume of the loading dose, which is administered before initiating the continuous infusion, and the volume and concentration of the infusion are highly variable. With higher concentrations of local anesthetics, the density of the motor blockade increases. With larger volumes, a greater dermatomal spread of analgesia is achieved. Most practitioners routinely use low concentrations of local A1 A B A1 C D C1 FIGURE 16-2 Photograph of typical needles and catheters used for neuraxial analgesia and anesthetic techniques. A, Epidural needle (18-gauge Tuohy) with magnification of tip shown at right (A1). B, Epidural needle (Tuohy) with catheter inserted through needle. C, Spinal needle (24-gauge Whitacre) with magnification of tip shown at right (C1). D, Spinal needle inserted through epidural needle for use in combined ­spinal-epidural technique. 166 PART IV Labor and Delivery Spinal cord L1 Filum terminate L2 L3 Needle in subarachnoid space Dura Ligamentum flavum Interspinous ligament L5 Needle in epidural space S1 S2 SPINAL ANALGESIA S3 Supraspinous ligament Sacrococcygeal ligament timing of neuraxial analgesia in labor on operative deliveries in nulliparous women (Marucci et al, 2007). Cesarean delivery rates (OR, 1.00; 95% CI, 0.82 to 1.23) and instrumental vaginal delivery rates (OR, 1.00; 95% CI, 0.83 to 1.21) were similar in the early neuraxial analgesia and later control groups. Neonates of patients with early neuraxial analgesia had a higher umbilical artery pH and received less naloxone than did patients when using other methods. The timing and use of neuraxial labor analgesia have differing effects on the length of labor. Compared with opioids for labor analgesia, the use of epidural analgesia minimally lengthens the first stage in some women and shortens it in others, but overall appears to lengthen the second stage by approximately 15 minutes (Anim-Somuah et al, 2005; Ohel et al, 2006; Sharma et al, 2004; Wong et al, 2005). S4 S5 Needle in caudal canal FIGURE 16-3 Schematic diagram of lumbosacral anatomy ­showing needle placement for subarachnoid, lumbar epidural, and caudal blocks. (From Rosen MA, Hughes SC, Levinson G: Regional anesthesia for labor and delivery. In Hughes SC, Levinson G, Rosen MA, editors: Shnider and ­Levinson’s anesthesia for obstetrics, ed 4, Baltimore, 2002, Lippincott ­Williams and Wilkins, p 125.) anesthetics and many coadminister an opioid with both the initial bolus and the infusion (e.g., 2 μg/mL fentanyl) for its synergistic effect. Dilute solutions of local anesthetics (0.0625% to 0.125% bupivacaine or ropivacaine) minimize the motor blockade and preserve the perception of pelvic pressure with descent of the fetus. Effects on the Progress of Labor and Rate of Operative Delivery The use of epidural analgesia has been associated with prolonged labor and increased rates of assisted delivery and cesarean section. Retrospective studies are difficult to interpret, because higher pain scores are predictive of increased labor duration (Wuitchik et al, 1989), and higher analgesic requirements are predictive of increased rates of cesarean section (Alexander et al, 2001). A Cochrane review of the effects of epidural analgesia, including CSE, compared with nonepidural analgesia or no analgesia during labor (Anim-Somuah et al, 2005) concluded that, although there was an increased risk of instrumental vaginal birth (RR, 1.38; 95% CI, 1.24 to 1.53; 17 trials including 6162 women), there was no evidence of a significant difference in the risk of caesarean delivery (RR, 1.07; 95% CI, 0.93 to 1.23; 20 trials including 6534 women). A recent metaanalysis summarized the results of studies addressing the Spinal analgesia is typically administered near the time of anticipated delivery. A small dose of a local anesthetic, opioid, or both is injected into the subarachnoid space. This dose of local anesthetic is far lower than that used for spinal anesthesia for cesarean section, and it has minimal effects on motor nerve function. Compared with epidural analgesia, it has the benefits of a more rapid onset, a lower failure rate, and being technically easier and quicker to perform. It has the significant disadvantage of a finite effective duration (approximately 90 minutes), but can be extremely useful for certain circumstances such as forcepsassisted delivery for a woman without epidural analgesia. COMBINED SPINAL-EPIDURAL ANALGESIA One variation of the lumbar epidural technique is a CSE analgesic. After placement of the epidural needle, but before insertion of the epidural catheter, a longer spinal needle is passed through the indwelling epidural needle (see Figure 16-2), puncturing the dura, and a small dose of local anesthetic or opioid is administered. If opioid alone is administered and the epidural catheter is placed but no local anesthetic is given, some analgesia can be achieved without motor blockade or sympathectomy. This method can allow the patient to walk safely (i.e., a “walking ­epidural”), but the analgesia is limited in efficacy and duration (approximately 2 hours in the first stage) and is rarely effective for the second stage of labor. If small doses of local anesthetics are administered through the spinal needle, segmental analgesia results more rapidly than with epidural administration of local anesthetics. Specifically, the onset of analgesia in the sacral root distribution is much more rapid with CSE compared with epidural alone. The epidural placement of the catheter allows continuation of the segmental analgesia initiated by the spinal technique. In a metaanalysis of 19 trials (2658 women) comparing CSE to epidural labor analgesia ­(Simmons et al, 2007), CSE had a faster onset of effective analgesia (especially in spread to sacral roots), was associated with more pruritus, and was associated with a clinically insignificant lower umbilical arterial pH. No differences were seen for maternal satisfaction, walking in labor, type of delivery, maternal hypotension, postdural puncture headache rate, or need for blood patch. CHAPTER 16 Obstetric Analgesia and Anesthesia CONTRAINDICATIONS AND COMPLICATIONS OF NEURAXIAL TECHNIQUES Certain conditions contraindicate neuraxial procedures; these include patient refusal, infection at the needle insertion site, significant coagulopathy, hypovolemic shock, increased intracranial pressure from mass lesion, and inadequate provider expertise. Other conditions such as systemic infection, neurologic disease, and mild coagulopathies should be evaluated on a case-by-case basis. Human immunodeficiency virus infection is not a contraindication to regional technique in the pregnant patient (Hughes et al, 1995). Infrequent but occasionally life-threatening complications can result from administration of regional anesthesia. The most serious complications are from accidental IV or intrathecal injections of local anesthetics. A prospective study of 145,550 epidurals in the United ­Kingdom noted unintended intravascular injection rates of 1 in 5000 and high spinal rates of 1 in 16,000 (Jenkins, 2005). An unintended bolus of IV local anesthetic causes dosedependent consequences ranging from minor side effects (e.g., tinnitus, perioral tingling, mild blood pressure, heart rate changes) to major complications (e.g., seizures, loss of consciousness, severe arrhythmias, cardiovascular collapse). The severity depends on the dose, type of local anesthetic, and preexisting condition of the parturient. Measures that minimize the likelihood of accidental intravascular injection include careful aspiration of the catheter before injection, test dosing, and incremental administration of therapeutic doses. If a local anesthetic overdose occurs, consider using a 20% IV lipid emulsion to bind the drug and decrease toxicity. Successful resuscitation and support of the mother will reestablish uterine blood flow; this will provide adequate fetal oxygenation and allow time for excretion of local anesthetic from the fetus. The neonate has an extremely limited ability to metabolize local anesthetics and may have prolonged convulsions if emergent delivery is required (­Morishima and Adamsons, 1967; Ralston and Shnider, 1978). A “high spinal” (or total spinal) can result from an unrecognized epidural catheter placed subdural, migration of the catheter during its use, or an overdose of local anesthetic in the epidural space (i.e., high epidural). Both high spinals and high epidurals can result in severe maternal hypotension, bradycardia, loss of consciousness, and blockade of the motor nerves to the respiratory muscles (Yentis and Dob, 2001). Treatment of complications resulting from both intravascular injection and high spinal is directed at restoring maternal and fetal oxygenation, ventilation, and circulation. Intubation, vasopressors, fluids, and advanced cardiac life support algorithms are often required. In any situation of maternal cardiac arrest with unsuccessful resuscitation, the fetus should be delivered by cesarean section if the mother is not resuscitated within 4 minutes of the arrest. This guideline for emergent operative delivery increases the chances of survival for both the mother and neonate (American Heart Association, 2005; Katz et al, 1986). In addition, a variety of less severe complications and side effects can occur with neuraxial blockade. The retrospective rates of inadequate epidural analgesia or 167 inadequate CSE analgesia requiring catheter replacement were 7% and 3%, respectively, at a U.S. academic center (Pan et al, 2004). The rate of accidental dural puncture during epidural catheter placement is 1.5%, and approximately half of these will result in a severe headache (Choi et al, 2003), which are typically managed with analgesics or a blood patch if necessary. Hypotension (decrease in systolic blood pressure greater than 20%) secondary to sympathetic blockade is the most common complication of neuraxial blockade for labor analgesia, with rates between 10% and 24% (Brizgys et al, 1987; Simmons et al, 2007). Prophylactic measures include left uterine displacement and hydration. Although standards for timing, amount, and hydration fluid remain controversial, dehydration should be avoided (Hofmeyr et al, 2004; Kinsella et al, 2000; Ko et al, 2007). Treatment of hypotension consists of further uterine displacement, IV fluids, and vasopressor administration. Small boluses of either phenylephrine or ephedrine can be used to treat hypotension. Although ephedrine (primarily β-adrenergic) was historically used, phenylephrine (primarily α-adrenergic) is associated with less fetal acidosis (Lee et al, 2002a; Ngan Kee et al, 2009; Smiley, 2009). If treated promptly, maternal hypotension does not lead to fetal depression or neonatal morbidity. A rise in core maternal body temperature is associated with labor epidural analgesia and may be influenced by several factors; these include duration, ambient temperature, administration of systemic opioids, and the presence of shivering. During the first 5 hours of epidural analgesia, there is no significant rise in body temperature (Mercier and Benhamou, 1997). In a retrospective study, temperature increases approximately 0.10° C per hour, and may reach 38° C in as many as 15% of parturients with a labor epidural compared with 1% without an epidural ­(Lieberman et al, 1997). Although Lieberman et al suggest a significant increase in the rate of newborn evaluation for sepsis (i.e., sepsis work-up [SWU]) with epidural analgesia, 67.3% of the noted SWUs were ordered in infants born to mothers without fever. Kaul et al (2001) found no association between epidural analgesia and SWU. However, patients who received epidural analgesia and their neonates had an increased body temperature at delivery, and approximately 6% of these women had temperatures equal to or exceeding 38° C. Although maternal temperature had no predictive value for the SWU, Kaul et al (2001) found other risk factors associated with the SWU, including low ­birthweight, gestational age, meconium, or respiratory distress requiring intubation at birth, hypothermia at birth, group-B β-hemolytic streptococcus colonization, and maternal preeclampsia or hypertension. Although the etiology of the maternal temperature rise remains uncertain, it need not affect the neonatal SWU. Epidural analgesia during labor does not increase the incidence of neonatal sepsis. Although it is suggested that the fever is associated with noninfectious inflammatory activation or altered thermoregulation, the fever is not associated with a change in white blood cell count or with an infectious process, and treatment is not necessary. Other potential side effects from neuraxial blockade include pruritus, nausea, shivering, urinary retention, motor weakness, low back pain, and a prolonged block. More serious complications of meningitis, epidural hematoma, and nerve or spinal cord injury are extremely rare. 168 PART IV Labor and Delivery A retrospective Swedish study of severe neurologic complications from neuraxial blockade included 200,000 obstetric epidurals and 50,000 obstetric spinals. Rates of serious neurologic events (i.e., neuraxial hematoma or abscess, nerve or cord damage) were 1:29,000 for obstetric epidural and 1:25,000 for obstetric spinal anesthetic procedures (Moen et al, 2004). regional anesthesia is contraindicated (e.g., coagulopathy, hemorrhage) or if a rapid onset is needed for emergent deliveries (e.g., fetal bradycardia, uterine rupture). Benefits of general anesthesia compared with regional anesthesia include a secure airway, controlled ventilation, rapid and dependable onset, and potential for less hemodynamic instability. PARACERVICAL BLOCK EPIDURAL ANESTHESIA A paracervical block is used to provide pain relief during the first stage of labor in up to 3% of patients (Bucklin et al, 2005). The technique consists of submucosal administration of local anesthetics immediately lateral and posterior to the uterocervical junction, which blocks transmission of pain impulses at the paracervical ganglion. Analgesia is not as profound as with epidural or spinal regional block, and the duration of analgesia is short (45 to 60 minutes), but the complications and side effects of epidural analgesia, such as hypotension, hypoventilation, and motor blockade, are avoided. Complications from systemic absorption or transfer of local anesthetic can occur. There is also the possibility of direct fetal trauma or injection. Paracervical block is associated with a 15% rate of fetal bradycardia based on a metaanalysis of studies (Rosen, 2002b). The mechanism of this phenomenon is unclear, and close fetal monitoring is warranted. The bradycardia may occur secondary to decreased uterine blood flow from the vasoconstrictor properties of local anesthetics, greater uterine activity, or transfer of the local anesthetic across the placenta to the fetus, causing direct toxic effects on the fetal heart. The bradycardia is usually limited to less than 15 minutes, and treatment is supportive, consisting of lateral positioning and administration of oxygen. Epidural anesthesia is an excellent choice for surgical anesthesia when an indwelling, functioning epidural catheter had been placed for labor analgesia. Epidural anesthesia provides the ability to titrate the desired level of anesthesia and extend the block time, if needed for a prolonged procedure. Epidural anesthesia is also ideal for patients in a nonemergent delivery who would not tolerate the abrupt onset of a sympathectomy from spinal anesthesia, such as some patients with cardiac disease. The volume and concentration of local anesthetic agents used for surgical anesthesia are larger than those used for labor analgesia. However, the technique of catheter placement, test dosing, and potential complications are similar. A typical dose regimen of epidural anesthesia for cesarean delivery could include 2% lidocaine (20 mL). Typically the anesthesiologist attempts to provide a dense block from the T4 level to the sacrum. This technique might not completely alleviate the visceral pain and pressure sensation associated with peritoneal manipulation, and adjuvant drugs are occasionally necessary. IV ondansetron and metoclopramide are frequently given to decrease nausea and vomiting associated with the operative delivery and hemodynamic effects from the dense neuraxial blockade (Lussos et al, 1992; Pan and Moore, 2001). Epidural morphine (3 to 5 mg) is typically given near the end of the procedure to decrease postoperative pain for up to 24 hours (Cohen et al, 1991). PUDENDAL BLOCK The obstetrician performs a pudendal block with a transvaginal technique, guiding a sheathed needle to the vaginal mucosa and sacrospinous ligament just medial and posterior to the ischial spine. The technique primarily blocks sensation of the lower vagina and perineum, and it is typically placed just before vaginal delivery. Although the technique provides analgesia for vaginal delivery or uncomplicated instrumental delivery, the rate of failure is high. In many centers, this technique is used when epidural or spinal techniques are unavailable. Complications in addition to failure include systemic local anesthetic toxicity, ischiorectal or vaginal hematoma, and, rarely, fetal injection of local anesthetic. ANESTHESIA FOR CESAREAN DELIVERY In the United States, the vast majority of cesarean deliveries are performed with neuraxial anesthesia. It offers the advantages of less anesthetic exposure to the neonate, has the benefit of an awake mother at the delivery, allows for placement of neuraxial opioids to decrease postoperative pain, and avoids the risks of maternal aspiration and difficult airway associated with general anesthesia. However, the use of general anesthesia is sometimes required if SPINAL ANESTHESIA For the patient without an epidural catheter, spinal anesthesia is the most common regional anesthetic technique used for cesarean delivery. The block is technically easier than epidural blockade, more rapid in onset, and more reliable in providing surgical anesthesia from the midthoracic level to the sacrum (Riley et al, 1995). The risk of profound hypotension is higher with spinal anesthesia than with epidural anesthesia, because the onset of the sympathectomy is more rapid. However, this risk can be nearly eliminated by avoidance of aortocaval compression, prehydration, and appropriate use of vasopressors. Colloid is significantly more effective than crystalloid (Ko et al, 2007). Historically, ephedrine was recommended as the vasopressor of choice, but more recent data confirm that a phenylephrine infusion at the time of spinal placement is effective at preventing hypotension and is associated with less fetal acidosis compared to ephedrine (Ngan Kee et al, 2004, 2009). Data also suggest that spinal anesthesia can be used safely for patients with preeclampsia (Hood and Curry, 1999; Wallace et al, 1995). A typical spinal anesthetic could consist of bupivacaine (12.5 mg) with morphine (200 μg) added to decrease postoperative pain. A large variety of other combinations of local anesthetics and opioids CHAPTER 16 Obstetric Analgesia and Anesthesia are used. A hyperbaric solution of local anesthetic is often used to facilitate anatomic and gravitational control of the block distribution. The medication will flow with the spinal curvature to a position near T4, and a head-down position can enhance the rostral spread of the block if needed. The duration of a single shot spinal anesthetic is variable, but normally provides adequate surgical anesthesia for greater than 90 minutes. In selected circumstances, the use of a CSE technique offers the advantage of a spinal anesthetic, with rapid onset of a dense block and the ability to administer additional local anesthetic through the epidural catheter if the procedure lasts for an extended time. A continuous spinal anesthetic technique with deliberate subdural catheter placement is a rarely used alternative, but is sometimes chosen in cases of accidental dural puncture during attempts to place an epidural catheter. This technique allows the advantage of a titratable, reliable, dense anesthetic, but carries the risks of high spinal if the intrathecal catheter is mistaken for an epidural catheter, or if the provider is unfamiliar with the technique. The rates of rare complications of meningitis or neurologic impairment from local anesthetic toxicity with the use of a spinal catheter may be somewhat higher than the other neuraxial techniques, but they remain unknown. Some data suggest that leaving the spinal catheter in place for 24 hours decreases the risk of postdural puncture headache (Ayad et al, 2003; Cohen et al, 1994). LOCAL ANESTHESIA Although cesarean delivery can be performed with local infiltration, it is accompanied with considerable discomfort to the woman and risks the possibility of local anesthetic overdose. Most obstetricians are not trained to perform the technique. However, local infiltration is useful in rare circumstances, such as acute fetal distress without an available anesthesia provider. GENERAL ANESTHESIA General anesthesia is used in obstetric practice for cesarean section typically when regional anesthesia is contraindicated or for emergencies, because of its rapid and predictable action. The major risks for maternal morbidity are pulmonary aspiration and failed intubation. Appropriate airway examination, preparation for unanticipated events, and familiarity with techniques and the algorithm for difficult intubation (American Society of Anesthesiologists Task Force on Management of the Difficult Airway, 2003) are critical for providing a safe general anesthetic. After denitrogenation of the lungs (i.e., preoxygenation), general anesthesia is induced by rapid-sequence administration of an IV induction agent, followed by a rapidly acting muscle relaxant. The trachea is intubated with a cuffed endotracheal tube, and a surgical incision is made after confirmation of tracheal intubation and adequate ventilation. Anesthesia is maintained by administering a combination of inhaled nitrous oxide and a potent inhaled halogenated agent (e.g., isoflurane), as well as sedative-hypnotics, opioid analgesics, and additional muscle relaxants if needed. During typical general anesthesia for cesarean delivery, opioids and benzodiazepines are 169 administered after the baby is delivered, to avoid placental transfer of these agents to the neonate. Before delivery of the baby, the primary anesthetic for the incision and delivery is the induction agent, because there is little time for uptake and distribution of the inhaled agents into the mother or fetus (Dwyer et al, 1995). If intubation attempts fail, the operative delivery may proceed if the anesthesiologist communicates that it is possible to reliably ventilate the mother’s lungs with either facemask or laryngeal mask airway (American Society of Anesthesiologists Task Force on Management of the Difficult Airway, 2003). Induction Agents Anesthesiologists use a variety of agents to rapidly induce unconsciousness. Among the most common are thiopental, propofol, etomidate, and ketamine. Each agent represents a different biochemical class, and each has specific advantages and cardiovascular effects. Sodium thiopental is a highly lipid-soluble, proteinbound barbiturate that can cause decreased cardiac output and hypotension and rapidly crosses the placenta. Historically, it was the most common induction agent for cesarean section under general anesthesia. In a study of healthy volunteers undergoing uncomplicated cesarean section, the mean umbilical artery (UA)–to–umbilical vein (UV) ratio of thiopental concentrations was 0.87 (Morgan et al, 1981). IV administration of an appropriate dose (4 to 6 mg/kg) renders the patient unconscious within a circulation time (30 seconds), peaks in the UV blood in 1 minute and in the UA blood in 2 to 3 minutes, and has no significant clinical effect on neonatal well-being. However, doses of 8 mg/kg can result in neonatal depression (Finster et al, 1972), and higher doses may require cardiorespiratory supportive techniques until the neonate can excrete the drug. This elimination may take up to 2 days (Fox et al, 1979). The lack of neonatal effects is unclear, but may be caused by first-pass metabolism by the neonatal liver, rapid redistribution into maternal vascular-rich tissue beds, additional dilution by the fetal circulation, and higher fetal brain water content. Propofol is a diisopropylphenol that is available as a 1% aqueous solution in an oil-in-water emulsion containing soybean oil, glycerol, and egg lecithin. Like thiopental, propofol is rapid in onset and can cause significant hypotension with a similar UA:UV ratio of 0.7 (Dailland et al, 1989). Unlike thiopental, propofol is preservative free and must be drawn up only hours before use. Other differences are that propofol decreases the incidence of nausea and vomiting, and it is currently not a controlled substance. Propofol has not been demonstrated to be superior to thiopental in maternal or neonatal outcome. Propofol administration has no significant effect on neonatal behavior scores with induction doses of 2.5 mg/kg, but larger doses (9 mg/kg) are associated with newborn depression (Gregory et al, 1990). Etomidate contains a carboxylated imidazole ring that provides water solubility in acidic solutions and lipid solubility at physiologic pH. Like thiopental, etomidate has a rapid onset of action because of its high lipid solubility, it rapidly crosses the placenta, and redistribution results in a relatively short duration of action. Unlike thiopental 170 PART IV Labor and Delivery and propofol, etomidate has minimal effects on the cardiovascular system, but it is painful on injection, can cause involuntary muscle tremors, has higher rates of nausea and vomiting, and can increase the risk of seizures in patients with decreased thresholds. At typical induction doses (0.3 mg/kg), etomidate administration can cause decreased neonatal cortisol production (<6 hours), but the clinical significance remains uncertain (Crozier et al, 1993). Ketamine, a structural analogue to phencyclidine, is more lipid soluble and less protein bound than thiopental. It is an analgesic, hypnotic, and amnestic with minimal respiratory depressive effects. Ketamine is biotransformed in the liver to active metabolites, such as norketamine. In contrast to thiopental, sympathomimetic characteristics of ketamine increase arterial pressure, heart rate, and cardiac output through central stimulation of the sympathetic nervous system, making it an ideal choice for a patient in hemodynamic compromise. Doses that are higher than those appropriate for induction of general anesthesia (1 mg/kg) can increase uterine tone, reducing uterine arterial perfusion. No neonatal depression is noted with typical induction doses (Little et al, 1972). In low doses (0.25 mg/kg), ketamine has profound analgesic effects, unlike barbiturates, but has been associated with undesirable psychomimetic side effects (e.g., illusions, bad dreams), which can be lessened by coadministration of benzodiazepines. Nitrous Oxide Inhaled nitrous oxide is often used as part of maintenance for general anesthesia, because of its minimal effects on maternal hemodynamics and uterine tone. As a sole agent, it is insufficient to provide an appropriate level of anesthesia for an operative procedure. It rapidly crosses the placenta with increasing UV-to–maternal artery ratios of 0.37 in the first 2 to 9 minutes increasing to 0.61 at 9 to 14 minutes (Karasawa et al, 2003). The effects of nitrous oxide on the neonate are not significant. Additional information about nitrous oxide is found in the previous section under Inhalation Analgesia. Inhaled Halogenated Anesthetics Isoflurane, sevoflurane, desflurane, and halothane are all halogenated hydrocarbons that differ in chemical composition, physical properties, biotransformation, potencies, and rates of uptake and elimination. In clinical use, specialized vaporizers deliver these volatile liquid agents, so that the inhaled concentrations can be carefully titrated by anesthesiologists because of the relatively profound cardiovascular effects and potential for uterine muscle relaxation. These agents are important components of general anesthesia for cesarean section, but readily cross the placenta. Without the use of these agents, the incidence of maternal recall of intraoperative events is unacceptably high (Schultetus et al, 1986; Tunstall, 1979). Placental transfer of inhalation agents is rapid because these are nonionized, highly lipid-soluble substances of low molecular weight. The concentrations of these agents in the fetus depend directly on the concentration and duration of anesthetic in the mother. Clinicians often confuse the use of general anesthesia and the terms fetal distress and depressed neonate. A depressed fetus will likely become a depressed neonate, and general anesthesia may be used because it is the most rapidly acting anesthetic to allow cesarean delivery. For a healthy fetus, the use of general anesthesia is not contraindicated. A Cochrane review of 16 studies comparing neuraxial blockade versus general anesthesia in otherwise uncomplicated cesarean deliveries found that “no significant difference was seen in terms of neonatal Apgar scores of six or less and of four or less at one and five minutes and need for neonatal resuscitation” (Afolabi et al, 2006). The authors concluded that there was no evidence to show that neuraxial anesthesia was superior to general anesthesia for neonatal outcome. Recent experimental animal studies have demonstrated neuronal apoptosis in the developing brain when a variety of agents are administered to induce and maintain general anesthesia (Istaphanous and Loepke, 2009; Loepke and Soriano, 2008). Implications for the fetus and neonate from brief anesthetic exposures are currently unknown because of a lack of human studies and difficulties extrapolating animal study methodology to humans. The induction–delivery interval is not as important in neonatal outcome as the uterine incision–delivery interval, during which uterine blood flow may be compromised and fetal asphyxia may occur. A long induction–delivery time may result in a neonate who is lightly anesthetized but not asphyxiated. If excessive concentrations of anesthetic are given for inordinately long times, neonatal anesthesia, evidenced by flaccidity, cardiorespiratory depression, and decreased tone, can be anticipated (Moya, 1966). It cannot be overemphasized that if the neonatal depression is caused by transfer of anesthetic drugs, the infant is merely lightly anesthetized and should respond easily to basic treatment measures. Treatment should include and focus on effective ventilation; cardiopulmonary resuscitation is rarely necessary. Ventilation will allow elimination of the inhalation anesthetic by the infant’s lungs. Rapid improvement of the infant should be expected. Otherwise, a search for other causes of depression is imperative. For these reasons, it is critical that clinicians experienced with neonatal ventilation are present at operative deliveries under general anesthesia in which the time from skin incision to delivery may be longer (i.e., known percreta, large fibroids), or maternal condition necessitates an atypical induction and maintenance of anesthesia (e.g., a patient with critical aortic stenosis undergoing an opioid-based induction). A discussion of the operative and anesthetic plan by the neonatologist, obstetrician, and anesthesiologist is crucial for optimizing the outcome of neonates in these situations. Neuromuscular Blocking Agents Succinylcholine remains the skeletal muscle relaxant of choice for obstetric anesthesia, because of its rapid onset and short duration of action. In doses of 1.5 mg/kg, appropriate intubating conditions are present within 45 seconds. Because it is highly ionized and poorly lipid soluble, only small amounts cross the placenta. Side effects include increased maternal potassium levels, myalgias, and succinylcholine is a known trigger agent for malignant hyperthermia in susceptible individuals. This depolarizing neuromuscular blocking agent is normally hydrolyzed CHAPTER 16 Obstetric Analgesia and Anesthesia in maternal plasma by pseudocholinesterase and usually does not interfere with fetal neuromuscular activity. If the hydrolytic enzyme is present either in low concentrations (Shnider, 1965) or in a genetically determined atypical form (Baraka, 1975), prolonged maternal or neonatal respiratory depression secondary to muscular paralysis can occur. Rocuronium is a rapid-acting, nondepolarizing neuromuscular blocker that is an acceptable alternative to succinylcholine. It provides adequate intubating conditions in approximately 90 seconds at doses of 0.6 mg/kg and in less than 60 seconds at doses of 1.2 mg/kg (Abouleish et al, 1994; Magorian et al, 1993). Unlike succinylcholine, it has a much longer duration of action, decreasing maternal safety in the event the anesthesiologist is unable to intubate or ventilate the patient. It has the benefit of not being a triggering agent of malignant hyperthermia or elevating serum potassium levels. During the operation, nondepolarizing neuromuscular blocking agents can be titrated to improve operating conditions. Under normal circumstances, the poorly lipidsoluble, highly ionized, nondepolarizing neuromuscular blockers (i.e., rocuronium, vecuronium, cis-atracurium, pancuronium) do not cross the placenta in amounts significant enough to cause neonatal muscle weakness (Kivalo and Saaroski, 1972). This placental impermeability is only relative, however, and neonatal neuromuscular blockade can occur when large doses are given (Older and Harris, 1968). The diagnosis of neonatal depression secondary to neuromuscular blockade can be made on the basis of the maternal history (e.g., prolonged administration of neuromuscular blockers, history of atypical pseudocholinesterase), the response of the mother to neuromuscular blocking drugs, and the physical examination of the newborn. The paralyzed neonate has normal cardiovascular function and good color, but lacks spontaneous ventilatory movements, has muscle flaccidity, and shows no reflex responses. The anesthesiologist can place a nerve stimulator on the neonate and demonstrate the classic signs of neuromuscular blockade (Ali and Savarese, 1976). Treatment consists of ventilatory support until the neonate excretes the drug, up to 48 hours. Reversal of nondepolarizing relaxants with cholinesterase inhibitors 171 may be attempted (e.g., neostigmine, 0.06 mg/kg), but adequate ventilatory support is the mainstay of treatment. Concomitant administration of an anticholinergic (e.g., atropine, glycopyrrolate) is normally necessary to prevent severe bradycardia from muscarinic side effects of the increased acetylcholine. SUMMARY This chapter serves as a general overview of the changes in maternal physiology during pregnancy and briefly discusses options and techniques for both labor analgesia and anesthesia for operative delivery. Its purpose is to allow the pediatrician and neonatologist a better understanding of the decisions and concerns of the anesthesiologist and the implications of his or her interventions. To provide the best care for mother and child, excellent communication is required between the obstetrician, pediatrician, anesthesiologist, and nurse. Only by facilitating these lines of communication and obtaining input from each discipline can patient care and safety become optimized. SUGGESTED READINGS Afolabi BB, Lesi FE, Merah NA: Regional versus general anaesthesia for caesarean section, Cochrane Database Syst Rev (4):CD004350, 2006. American Society of Anesthesiologists Task Force on Obstetric Anesthesia: Practice guidelines for obstetric anesthesia: an updated report by the American Society of Anesthesiologists Task Force on Obstetric Anesthesia, Anesthesiology 106:843-863, 2007. Anim-Somuah M, Smyth R, Howell C: Epidural versus non-epidural or no analgesia in labour, Cochrane Database Syst Rev (4):CD000331, 2005. Chestnut DH: Obstetric anesthesia: principles and practice, ed 4, Philadelphia, 2009, Mosby. Eltzschig HK, Lieberman ES, Camann WR: Regional anesthesia and analgesia for labor and delivery, N Engl J Med 348:319-332, 2003. Hughes SC, Levinson G, Rosen MA, et al: Shnider and Levinson’s anesthesia for obstetrics, ed 4, Philadelphia, 2002, Lippincott Williams & Wilkins. Marucci M, Cinnella G, Perchiazzi G, et al: Patient-requested neuraxial analgesia for labor: impact on rates of cesarean and instrumental vaginal delivery, Anesthesiology 106:1035-1045, 2007. Simmons SW, Cyna AM, Dennis AT, et al: Combined spinal-epidural versus epidural analgesia in labour, Cochrane Database Syst Rev (3):CD003401, 2007. Ngan Kee WD, Khaw KS, Tan PE, et al: Placental transfer and fetal metabolic effects of phenylephrine and ephedrine during spinal anesthesia for cesarean delivery, Anesthesiology 111:506-512, 2009. Wong CA, Scavone BM, Peaceman AM, et al: The risk of cesarean delivery with neuraxial analgesia given early versus late in labor, N Engl J Med 352:655-665, 2005. Complete references used in this text can be found online at www.expertconsult.com P A R T V Genetics C H A P T E R 17 Impact of the Human Genome Project on Neonatal Care Jeffrey C. Murray HISTORY OF THE HUMAN GENOME PROJECT Within 50 years of the discovery of the structure of DNA by Watson and Crick, the Human Genome Project achieved a major milestone with its description of an almost complete sequence of the approximately 3 billion nucleotides (A,C,G,T) contained within the human haploid genome. This effort was the result of an early vision by a group of scientists that with advances in technology and sequence analysis this task could be achieved (Watson and Cook-Deegan, 1991). The central goal was to provide a reference sequence of the human genome that could serve as a framework on which to better understand disease pathogenesis. Eventually, it was hoped, this would lead to improved treatments and prevention. There have been numerous spin-offs from the technology and analytic platforms developed. A wide range of other organisms have now had their genomes sequenced from pathogenic viruses to complex plants and animals. These sequences provide insights into disease pathogenesis that are infectious or immunologic and open new doors to therapies. Comparing DNA sequences across species to look for evolutionary conservation has also provided tremendous insights into the protein structure of genes and the regulatory elements that control the cell type, timing, and amount for the synthesis of any individual protein. The nature of how a gene is defined has been substantially altered by these insights, so that the gene is now recognized as a far more complex structure with elements dispersed sometimes 1 million nucleotides or more from the protein coding components classically thought of as “the gene” (Gerstein et al, 2007). Understanding the genetic relationships between individuals of different ancestral origins has led to the current effort to describe genetic variation across the human species as an essential aspect of understanding predisposition to or resistance from disease. Because the identification that carriers of sickle cell trait had increased resistance to malaria, or the identification of genetic factors conferring persistence of lactose tolerance into adult life as a mechanism associated with the advantages of dairy farming (Bersaglieri et al, 2004), it has been evident that individual genetic variation is an important contributor to health and disease. The hemoglobin system has provided an evolving paradigm for our understanding of the molecular nature of genetic disease beginning in the 1940s (Neel, 1949; Pauling et al, 1949) and onward to serving as a model for gene therapy in the modern era. The conception of the genome project led to wide ranging concerns about ethical, legal, and social aspects and resulted in the development of the Ethical, Legal and Social Issues (ELSI) Project to investigate these areas and to anticipate future concerns and problems. Finally, as part of the social and legal component, the work has led to legislation in the United States and abroad that is designed to protect individuals from discrimination based on their genetic background. The Human Genome Project has been an international effort from its beginnings and had critical predecessors in the human gene mapping (HGM) meetings that had an initial focus on identifying the chromosomal location of normal and disease causing genetic variants. The community established by the HGM meetings provided an infrastructure that enabled the more comprehensive sequence-based maps developed in the wake of the HGM meetings. One early outcome of these meetings and the recognized need to convert research findings to clinical utility was creation of Mendelian Inheritance in Man by ­Victor McKusicks, which has now evolved into Online Mendelian Inheritance in Man (www.ncbi.nlm.nih.gov/omim/)—a comprehensive catalogue of single-gene disorders that is an essential reference tool for learning about the genetic aspects of both rare and common disease. The international effort continues with many individual countries now focusing on genome efforts that are specifically relevant to their own high-risk medical conditions (e.g., malaria, HIV, hemoglobinopathies). These early collaborative international efforts also created the framework for one of the greatest successes of the Human Genome Project—that is, information generated relevant to the human DNA sequence and its variation should be held in public trust and there should be open access to DNA sequence and its annotation. This spirit of open access has, in turn, led to a stronger community ethic for the sharing of scientific and medical data that may be a legacy that will exceed the value of the sequence itself. There have been many technical and analytic advances that have enabled the sequence, and more recently its variation, to be understood and applied. Currently these advances extend to the characterization of hundreds of thousands of genetic variants on large populations of individuals. In the last few years this characterization has led to dramatic advances in the understanding of how common 173 174 PART V Genetics genetic variation can contribute to human genetic disease. This chapter will provide some detail on the various aspects of the Human Genome Project that are of current relevance. The interval of more than 50 years between the elucidation of DNA as the information-containing macromolecule and its structural description by Watson and Crick has also spanned the embedding of medical genetics as an important specialty in health care. This same time interval has seen the identification, treatment, and in some cases prevention of a wide range of disorders that have a strong genetic component (including Rh incompatibility and phenylketonuria) and a variety of other neonatal biochemical disorders and the hemoglobinopathies. Discussed in Chapters 3 and 27 are the details of the current state of our knowledge of these important advances in prenatal and neonatal screening. It now seems apparent that the past focus on adding one disease at a time to the armamentarium of those that can be studied and treated using genetic tools will soon give way to a comprehensive analysis available at the level of the individual genome. Much of the current effort in human genome work is devoted toward individualized medicine, in which physicians will have the ability to treat not just a disease but a specific person with a disease and to account for their genetic and environmental variation in response to therapy or risks associated with prevention. Motulsky (1957) recognized this pharmacogenetic (or pharmacogenomic when applied globally) approach in the early days of human genetic study as being essential to understanding the significant variation in response to therapeutic agents or toxins. Pharmacogenetics is already being applied clinically to identify patients at risk for toxic reactions to some chemotherapeutic agents as well as antiinfectious compounds. Individualization of diet and nutritional approaches is also on the immediate horizon, and it seems likely that over the next few years there will be in place high-throughput genotyping platforms that will provide comprehensive information to the pediatrician and neonatologist for providing direct alterations and care beginning in early infancy. These individualized genomewide profiles are already available to early adopters and are creating new models for how physicians will anticipate health care needs and outcomes for their patients. The history of the Human Genome Project, both its scientific and social basis, has been published widely in review articles and books. Its central conceit arose in the middle 1980s, when it became apparent that technology was advancing so that an undertaking could at least be considered (Watson and Cook-Deegan, 1991). By the early 1990s there were formal projects devoted to its implementation in academic settings (e.g., the U.S. Department of Energy, the National Institutes of Health [NIH] and the Medical Research Council in the United Kingdom) as well as several commercially based entities that had an interest in gene identification as a mechanism to developing therapeutic products. Besides the primary emphasis on DNA sequencing, preliminary steps that entailed the development of physical and genetic maps also proved critical to assembling the DNA sequence of humans into its linear order on each of the individual chromosomes. These steps were also necessary for the conversion of DNA sequence data into useful tools for gene identification and mutation discovery. The combination of academic and commercial interests led to a competition that in the end provided the stimulus to develop large-scale, high-throughput methodologies and the analysis tools that in 2001 enabled the project to reach one of its primary goals ahead of some of even the most aggressive predictions for when a complete human sequence might be in place (Lander et al, 2001; Venter et al, 2001). These successes and their resulting technologies and algorithmic advances were based on trials in less complex organisms (e.g., Escherichia coli, yeast, fruit fly) and were then quickly applied to model and experimental organisms and to identify the sequences of plants and animals of commercial and nutritional importance. At the time of this writing, there are 998 microorganisms (www.ncbi.nlm.nih.gov/genomes/lproks.cgi), as well as many higher organisms, whose sequences have been completed. SCIENCE AND THE HUMAN GENOME In the years since the description of the primary human sequence, a deeper understanding of the nature of the DNA sequence and its regulation has developed. Phenomena that are still relatively new in 2011 (e.g., copy number variants, micro-RNA, epigenetic regulation) are important aspects of how the human genome functions and interacts. There is doubtless much more to learn about our DNA sequence, but this chapter will review some current understandings. There are enormous numbers of highly conserved DNA sequences outside of traditional DNA-encoded protein sequences for which function is completely unknown. The next decades will be, in part, devoted to developing and understanding how these conserved sequences play a role in human health and disease. A few current features will be discussed, with many more evolving almost daily in the primary literature. The effects of these approaches have been reviewed recently for neonatology, and this chapter will expand on those themes (Cotten et al, 2006). SINGLE NUCLEOTIDE POLYMORPHISMS Single nucleotide polymorphisms (SNPs) are the workhorses of human genetic variation and indeed are only a more specific term and characterization of restriction ­fragment length polymorphisms (RFLPs), which were the original DNA variants studied in the human genome. The advent of the polymerase chain reaction eliminated the need for RFLP studies. SNPs are specific nucleotide sites in the human genome, where it is possible to have one of two different nucleotides, or polymorphisms, at that position on one of the DNA strands. For example, there might be either a T or a G at a specific site. These variant sites are common with up to 1% of the human DNA sequence being potentially variable between any two individuals, resulting in tens of millions of SNPs across the genome. Most variation is found across all human populations, although some variants appear to be highly population specific. These variants are usually normal, in the sense that they do not appear to be disease causing, although they may lie adjacent to DNA changes that do contribute to disease predisposition. Chip-based DNA sequence detection allows the assay of up to 1 million SNPs simultaneously on one individual at a cost of a few hundred dollars. CHAPTER 17 Impact of the Human Genome Project on Neonatal Care COPY NUMBER VARIANTS A relatively recent discovery in human genetic variation has been the importance of copy number variants as contributors to human inherited disorders. Although both small and large deletion and duplication events of the human DNA sequence have been known since the 1970s, based on cytogenetic banding, only recently has the role that these play in disease been recognized widely. Early copy number variants ranged in length from one to five nucleotides and proved to be useful in characterizing common genetic variation used in family linkage studies. In parallel with these were variants of which the central element might be 15 to 50 nucleotides in length—so-called mini satellites, which were also used in family-based linkage studies. The new class of variation is thousands to occasionally hundreds of thousands (or millions) of base pairs in length. These variations may contain one or multiple genes that can exist in two, three, or more copies arrayed in tandem at particular chromosomal positions. When these tandem arrays of largely identical sequences align themselves during meiosis, there is occasionally a misalignment that can result in the deletion or duplication of one or more of the copies. This event in turn can create a range of the number of copies present from zero to many (Zhang et al, 2009). When functional genes are contained within the copied element or functional regulatory elements, the amount of gene product made may be increased or decreased from a reference level. Early examples in which copy number variants contributed to disease include the Di George syndrome 22q- deletion and deletions associated with spinal muscular atrophy and Charcot-Marie-Tooth disease. New microdeletion syndromes are now being characterized with great precision using array-based DNA analysis that is rapidly replacing the traditional karyotype as the first line of chromosomal analysis. The recognizable syndromes that are found recurrently, and with an identifiable phenotype such as 22q-, are complemented by rare deletion or duplication events that result in congenital anomalies, developmental delay, or both, and where their etiologic nature can be inferred from the normal structure of the parental chromosomes. Finally, in areas such as autism, it is clear that these microdeletion duplication events are a major explanation for the sometimes sporadic, as well as familiar nature, of these disorders. Their contribution to human genetic disease is now suggested to be in the vicinity of 10% of all variant-contributed disease, and thus they form an important class for investigation. Despite their clear importance, there remains to be resolved many technical and clinical issues related to their identification and meaning (Aki-Khan et al, 2009). MITOCHONDRIAL DNA Mitochondria are the energy-producing organelles present in thousands of copies within each cell. Each mitochondrion has its own genome, distinct from the nuclear genome and thought to arise from incorporation of bacterial DNA by a eukaryotic cell. Although the mitochondrial genome is approximately 16,500 bp in length (compared with the 3 billion bp of the nuclear genome), there is a wide range of disorders associated with variation in 175 mitochondrial sequence. In addition, because mitochondria reside in the cytoplasm and are not found in sperm, they have a unique pattern of maternal-only inheritance, in which mothers pass their mitochondria to all of their offspring with their daughters in turn passing that on to subsequent generations and with no passing of mitochondrial DNA from males to their children. GENE IDENTIFICATION One of the primary benefits of the reference human DNA sequence is our ability to move quickly from finding the location of a gene on a chromosome to identifying the specifics of that gene and the disease-causing mutations. Gene discovery can provide immediate clinical benefit in the form of more accurate diagnoses and risk predictions and longer-term benefits when gene discovery leads to an understanding of gene function and physiology that can be converted into treatment and prevention. Gene mapping approaches to human gene identification have been in use since the late 1970s, and the first successes were the identification of single-gene disorders, sometimes termed monogenic or Mendelian because their inheritance patterns follow the traditional modes of autosomal dominant, autosomal recessive, and sex linked. Currently there are three primary methodologies under use for gene identification (Altshuler et al, 2008). The first involves linkage studies using large families with genetic disorders or many small families with the same disorder than can be studied and their data pooled. These linkage-based approaches can provide a relatively well-defined chromosomal location for single-gene conditions and has led to successful gene finding for cystic fibrosis, neurofibromatosis, and hundreds of additional, mostly rare, conditions. This approach can also be applied to common but genetically complex traits for which there are no simple inheritance pattens. This technique may require many hundreds of small families, and the resultant gene localization is imprecise. A second method for gene localization can make use of small chromosome rearrangements, such as balanced translocations or small deletions or duplications, that result in a phenotype for a known disorder. In these cases, the location of the chromosomal rearrangement immediately suggests that a gene at or near that rearrangement is etiologic and can be used to again directly search for evidence of a specific gene and mutation in that region. A third approach now in great favor is described next in some detail, and it is exerting a major influence on disease gene finding. HAPMAP AND GENOME-WIDE ASSOCIATION As noted previously, it is particularly challenging to find genes associated with complex traits that have multiple genetic and environmental contributors. This heterogeneity creates substantial difficulties in both finding and confirming that any particular gene plays a role in the disorder of interest. A new approach, enabled by advances in technology (i.e., DNA chips) and statistical analysis is proving to be remarkably successful. This approach takes the form of the genome-wide association (GWA) study in which a panel of genetic variants are densely arrayed on DNA chips and characterized in large case and control 176 PART V Genetics populations (Manolio and Collins, 2009). GWA can also be applied successfully to case-parent approaches, as will be described in more detail. The prerequisite to this comprehensive search for variation was the development of the human HapMap (http:// hapmap.ncbi.nlm.nih.gov/). The HapMap project developed a comprehensive understanding of the relationship of SNPs across multiple human ancestral groups including Europe, Africa, and Asia (International HapMap Consortium, 2007). HapMap is an essential component of a GWA study because it provides a comprehensive reference listing of the relationships of SNPs and copy number variants to enable their careful characterization in case and control populations. By looking for evidence of DNA sequence variation or allelic variation in which one allele is found significantly overrepresented in a case compared with a control population, there is a strong suggestion that the etiologic gene and variant lies in the vicinity of the surrogate or marker gene allele. In the last 2 years this approach has greatly extended the gene discovery process. Because these common complex traits have a far greater population impact, they are the ones that hold the greatest promise for providing information about the common contributors to pediatric disease. As of this writing approximately 100 associations have been identified (Manolio et al, 2009), but to date the emphasis has largely been on adult complex disorders (e.g., type 2 diabetes, cardiovascular disease, mental health disorders such as schizophrenia), with only a handful applied to pediatric traits. Autism (Wang et al, 2009), asthma (Himes et al, 2009), and cleft lip or palate (Birnbaum et al, 2009) have had new loci identified with the hope that this will provide insights into their pathogenesis. One challenge of the GWA study is that at present, whereas it can identify one or more loci associated with the disease, implementation of a GWA requires expensive technology and large (usually numbering in the thousands), well-phenotyped case and control populations. Thousands of cases and controls must be available to have sufficient power to detect the small effects seen. A second caveat is that the loci found usually have low relative risks or odds ratios, so that the clinical effect of any one identified locus is very small (Hardy and Singleton, 2009). However, collections of loci can have a combined substantial impact, and even a low odds ratio may identify a new biologic pathway that could provide great insights into etiology and treatment (Hirschhorn, 2009). As analytic approaches improve and as costs drop, it seems likely that the GWA study will become a common approach to a wide range of neonatal disorders in which either genetic risks or pharmacogenetic variation important in drug choice and dosing will be unraveled. Besides detecting disease-associated genetic risk factors, these studies also contribute to knowledge of the genetics of normal trait variation, such as height and skin color. Substantial advances in this area have also occurred with more than 50 genes found to have a role in factors such as height determination (Hirschhorn and Lettre, 2009). However, although identical twin studies tell us that height is almost entirely genetically determined across a broad range of environmental variation, the genetic findings to date explain only a small amount of the contributors to height. There is still much to learn in regard to normal trait variation and the ability to make predictions about a child’s future physical traits or cognitive behavioral range. New approaches beyond GWA are still critically needed to find this large amount of unexplained genetic contributors (Manolio et al, 2009). The final, and in some ways ultimate, strategy for identifying gene variation associated with disease will be to perform comprehensive genome-wide sequencing. In the last few years, completely new technologies have been applied to DNA sequencing, and the costs of DNA sequencing of a specific individual have dropped by four or five factors of 10 from the original cost of assembling the reference human genome. The reference genome was an amalgamation of many individuals and cost approximately $3 billion, or $1 per nucleotide of sequence. The goal is to eventually reduce this cost by $1000 per individual genome sequenced, which would place it well within the realm of current medical investigations, such as imaging, and well below the cost of many therapeutic approaches. At the time of this writing, several individuals have had their individual genome sequence reported (Wheeler et al, 2008). There is a newly initiated project under way to sequence the genome from approximately 1000 individuals around the world to provide a catalog of DNA sequence variation that can be used as a reference point for comparisons in the future for normal and disease trait variation studies. The high-throughput sequencing projects have an initial focus on the sequencing of all exons or conserved elements in the genome that are known to be at higher risk for containing disease-causing mutations. Nonetheless, mutations can also reside in the intergenic regions and outside of regions of conservation. Eventually wholegenome sequencing will be required for a comprehensive view of contributors to human inherited disorders. This high-throughput sequencing approach, when performed in sufficient numbers of individuals, can identify disorders in which there are common variants contributing to the disease process as well as individually rare variants that also compose a portion of the disease risk panel. Cystic fibrosis provides an example of this model in that of the single mutation (the ΔF508) consisting of a three-nucleotide deletion, which explains approximately 70% of the allelic variants that result in cystic fibrosis. The remaining 30%, however, are divided across an additional 20 to 50 that are relatively common, and then more than 1000 others that may be rare or family specific. Therefore to identify all the mutations associated with cystic fibrosis risk, one would need to perform comprehensive DNA sequencing. COMPARATIVE GENOMIC HYBRIDIZATION Comparative genomic hybridization (CGH), or arraybased hybridization, is another technical advance that is now providing, in many cases, a replacement for traditional karyotype-based chromosomal analysis. CGH has already had a substantial influence on prenatal and newborn testing for genetic disease and is discussed in more detail in Chapter 20. Although these array approaches cannot yet detect every form of chromosomal abnormality (e.g., balanced translocations are not detected), they are highly successful in detecting major structural whole chromosomal CHAPTER 17 Impact of the Human Genome Project on Neonatal Care aneuploidy such as Turner syndrome (45X) or trisomy 21, and they are also effective in identifying small deletions and duplications with a high degree of resolution. These approaches are increasingly becoming a first-line screen for chromosomal structural abnormalities, with the potential for replacing standard karyotypes almost entirely. The arrays also benefit from being able to use DNA directly, so that live cells are not required as they have been for traditional karyotypes. Thus material obtained from even a deceased infant or fetus can be used in the analysis. Because the amount of DNA required is also small, minimal DNA quantity further facilitates its technical application. There are a variety of competing approaches available for such CGH. The current generation of array-based tools use fragments of DNA that are far smaller than prior versions; therefore the resolution is in the range of thousands of nucleotides. This resolution provides detection of very small deletions or duplication events of potential etiologic importance and begins to approach the level of direct sequence comparison. At some point, direct genome sequencing may replace these tiled-array approaches for detection of deletions and duplications. In parallel with this improvement in resolution has been the recognition that there is no comprehensive catalog of the normal range of variation for rare deletion and duplication events. When new deletions or duplications are identified for which there is not a strong prior track record for their clinical importance (as would be the case for 22qfor example), then the interpretation as to whether they are causal for disease can be challenging. This challenge can sometimes be aided by examining parental samples in which the presence of a de novo deletion or duplication event may be more strongly indicative of contributing to a disease etiology than when the event is also identified in one or more other family members. However, caution needs to be exercised when such deletion-duplications are found in other family members, because there can be a range of penetrance for such abnormalities as well as other co-contributors that might be necessary for the full disease phenotype to be expressed (van Bon et al, 2009). Therefore caution must be exercised in family counseling, and it is essential to have the most accurate and up-to-date information on the nature and role of such variants when using them in a clinical setting. Fortunately these catalogs of normal variation are becoming available (Shaikh et al, 2009). With these caveats, it is still clear that there is an enormous amount to learn about how these structural variants can contribute to disease of the neonate. PHARMACOGENOMICS Pharmacogenomics, although originally described in the 1950s as a result of genetic variation associated with commonly recognized enzymes such as G6PD (Motulsky, 1957), is now a highly active area for the clinical application of genome studies. The U.S. Food and Drug Administration has already identified more than a dozen genes for which characterization of allelic variation of those genes may be an important component of therapeutic decision making, including several that have a pediatric application. Perhaps one of the best understood is the role of thiopurine methyltransferase and its genetic variation associated with 177 major adverse events associated with chemotherapeutic agents such as 6-mercaptopurine (Weinshiboum, 2006). It is now standard care to study pediatric patients preparing to begin chemotherapy for their risks associated with these allelic variants that affect thiopurine methyltransferase drug metabolism. These current pharmacogenetic variants are the beginnings of a much larger group of variants that will tie individual response to therapeutics and will eventually result in direct individualized medicine on an individual basis. In neonatology, many potentially toxic medications are used routinely; these include indomethacin or ibuprofen for treatment of patent ductus arteriosus, antibacterial and antifungal agents that have potential serious adverse consequences such as gentamicin, medications to treat pulmonary or systemic hypertension, antiarrhythmic agents, and others for which individual patient response may be based in part on enzyme allelic variation in those individuals. By identifying individual risk beforehand, one can either adjust medication doses or choose alternative medications to minimize complications and maximize therapy. In some cases it may be that larger doses of medication will be required based on pharmacogenetic variation as well. It is in the area of this pharmacogenetic variation that perhaps the greatest advances will come in the application of the genome project to neonatal care in the next decade. With the availability of large clinical trials, such as those overseen by the Neonatal Network and others, improved understanding of the role of genetic variation in drug response will almost certainly be obtained. Because the cost of studies of genetic variation are modest, and the consequences of using the wrong drug or the wrong dose potentially devastating, it will be critical for neonatologists and pediatricians to understand the evidence for which variants can contribute in a proven way to improved therapeutic approaches. Just as measuring drug levels has become a routine part of care in many settings, so will testing beforehand for genetics risk identification and dosage plans also become a part of neonatal care. INDIVIDUALIZED GENETIC INFORMATION One of the most dramatic recent advances has been the application of the SNP-based association technologies to commercial entities that are now providing direct consumer testing to interested parties. Currently there are multiple commercial organizations that, for a fee of between $100 and $1000, will extract DNA from a saliva sample and perform high throughout SNP analysis. The results are then provided to the individual with the whole array of GWA risks that have been published in peer-review journals. Although currently controversial (Ng et al, 2009), it seems likely that the use of these technologies will expand in the private or “recreational” setting, directly in the clinical care setting, and eventually into newborn screening. A challenge for clinicians is presented when families have obtained this information but have little context for its meaning, and anecdotal examples of its application to neonates in terms of risk are already arising. This example is yet another example of why clinicians need to remain current with the rapidly advancing genomic capabilities. 178 PART V Genetics MICROBIOMES AND THE METAGENOME As described previously, there have been approximately 1000 bacterial and viral genomes sequenced, as well as more complicated pathogenic eukaryotic species. These microbial and viral sequences can be used to establish the genetic framework of the many organisms that are found in or on an individual. Because the newborn infant is largely thought to be sterile at birth, but rapidly acquires its microbial flora through environmental exposures, an understanding of the relationships of these many microbial species and how they participate in maintaining health or contributing to disease will be critical to an understanding of neonatal health and disease. Humans have tens of thousands of microbial and viral species contained in or on their bodies in locations such as the gut, mouth, and skin (Turnbaugh et al, 2007). A few obvious disease states, such as necrotizing enterocolitis (Neu et al, 2008) or presumed sepsis, are already under active study for the role in which microbial communities might have a role in predisposition or disease prevention. It would seem likely that probiotic therapy at sometime in the future might be targeted to provide a milieu that would optimize the microbial flora or the neonatal gut to prevent or limit disease and to assist in nutrition. Our understanding of the transmittal of microbiomes to the infant from maternal and environmental sources is still limited (Palmer et al, 2007), but the capacity of the Human Genome Project and its DNA sequencing capabilities allows identification of the organisms involved with a high degree of precision. It is already apparent that there are many organisms present in and on newborns that have been previously unknown to science, because the identification of species has been limited by the ability to characterize them initially through culture or antigen characterization. For organisms for which appropriate culture media have not been identified, their existence was not known until the development of the ability to perform direct genome sequencing of material collected from the stool, urine, or elsewhere (DiGiulio et al, 2008). By bypassing the culture step and sequencing biologic samples with DNA sequenced directly, it is possible to establish the identification of new organisms that may be contributing directly to disease in ways that were not previously identifiable. This concept of the microbiome also suggests that in the future—whereas the core DNA sequence will need to be sequenced only once because it is relatively unchanging, except in the case of rare somatic events—the metagenome sequence is constantly evolving based on alterations in factors such as nutrition and antibiotic exposure. In the pediatric population, it may be that DNA sequences sampled from the metagenome will be performed repeatedly as a way of assaying responses to therapy or risks to disease. EPIGENETICS Epigenetic regulation or the modification of DNA sequences through environmental exposures, such as altered nutrition or drug exposure, is also an evolving part of the Human Genome Project’s contributions to our understanding of how genes function and work (Devaskar and Raychaudhuri, 2007). Many of these epigenetic changes are thought to arise in utero. As a result, a new field of inquiry into the developmental origins of adult disease (Joss-Moore and Lane, 2009; Simmons, 2009) is now expanding to help better understand the consequences of how DNA can be modified and gene function altered, whereas the linear DNA sequence remains the same. Alterations of DNA can include methylation of individual nucleotides or acetylation or other modifications that take place in the histones that help to establish the chromosomal and chromatin-based structure of DNA sequences. The developmental origins of disease hypothesis posits that in utero changes can be induced by altered exposure states such as malnutrition, and these exposure states can imprint on the DNA a modified form that is then transmitted cell to cell and established in the infant, with continuing effects on gene regulation into childhood and adulthood. For example, if this imprinting results in a more effective storage capacity for calories, then in adult life when calories are more readily accessible, adults may be predisposed to obesity and its consequences. Similarly, such changes might also regulate cell division and transfer and play a role in childhood leukemias. Because these effects may occur in utero and in the premature infant, it seems likely that in the future a better understanding of our nutritional and drug exposure to infants will be important, which could have direct consequences on the predisposition of neonates to later disease. Although there are some technologies in place to develop an understanding of epigenetic modifications and changes, their application in a clinical setting is still relatively modest. GENE THERAPY Although not a direct consequence of the Human Genome Project, the idea of gene therapy has developed in parallel with our understanding of the human genome. Early efforts to use direct gene replacement are continuing, and although many challenges still remain it seems likely that these forms of therapy will continue to expand and eventually be commonly applied in the pediatric setting. Enzyme replacement therapy for several biochemical disorders is now widely available and, for those disorders for which earlier treatment results in better outcomes, recognition of such disorders early in life becomes critical. In other settings direct gene replacement has begun to suggest it may be both safe and efficacious (Maguire et al, 2009), opening an important door to future therapies. In parallel with projects that examine gene replacement, there is also the development of a better understanding of how related phenomena such as micro RNAs or small interfering RNAs also function. These small interfering RNAs, for example, may prove to be highly effective therapeutic agents, and a wide range of pediatric disorders are now under study to determine whether these can be used in a therapeutic or preventive sense. The genome sequence is likely to contribute to an understanding of undiscovered elements that may have an important role in gene regulation that could be effective in therapeutic settings as well, and these advances are eagerly anticipated. Finally, advances in stem cell research and its therapeutic potential have now been enabled by new legislation CHAPTER 17 Impact of the Human Genome Project on Neonatal Care that is affording the NIH an opportunity to make stem cell lines available to investigators for studies using NIH funding (Collins, 2009). More readily available stem cells will likely increase the pace of research into the use of these potentially powerful therapeutic agents. ETHICAL ASPECTS OF HUMAN GENETICS The ethical, legal, and social aspects of the Human Genome Project have been of at least equal importance to the technology developments that have provided the advances in basic science and translational science described elsewhere in this chapter. It was Watson’s insight early on in the process of establishing the Human Genome Project to allot 3% to 5% of the NIH Genome Institute budget directly for ELSI Project activities. The goal was to anticipate the controversies that might develop and to address them proactively whenever possible. In addition, the ELSI Project has also been involved in a wide range of concerns about genetic discrimination. ELSI Project investigators and their colleagues worked hard to develop model legislation that could eliminate genetic discrimination in the job and for health insurance purposes. The successful passage of the Genetic Information Nondiscrimination Act (Tan, 2009) in the summer of 2008 was a major milestone at the federal level in providing protection for all citizens from workplace and health insurance discrimination. While many states had already addressed these issues, this federal assurance provides a greater level of national protection that is likely to facilitate the ease with which individuals enroll in both clinical genetics studies and research-oriented studies. The second critical aspect of the ELSI Project is addressing the need to have common forms and formats for enrolling research subjects in what are now the highly complex and large genetic investigations required to perform GWA studies described elsewhere in this chapter. It was recognized early that obtaining informed consent for participation in a genetics study was a constantly moving target, because technologies often outstripped the ability of social scientists to be prepared for the scale and possibilities of what might come next. These issues have been dramatically highlighted in recent years during which the scale of genetic characterization is being performed on the SNP ChIP assays, and the recognition that even aggregate data might be able to result in individual identification (Homer et al, 2008) has dramatically altered the clinical and research landscape for genetics. GWA studies in particular have been scrutinized in great depth, and the NIH has detailed guidelines in place for 179 how consent documents should be worded and the information that must be provided to research subjects. Similar oversight is provided for external users of the GWA data, which, to the benefit of the research community, have been deposited in readily accessible databases after appropriate scientific and ethical review. Another area of influence of the genome project has been on forensic DNA analysis. Although such a discussion is beyond the scope of this chapter, the ability to identify family relationships or ancestral origins has now made DNA analysis a widely used legal investigative and as a commercial tool. It has had spectacular successes in freeing convicts whose DNA evidence has exonerated them from being a participant in major crimes. DNA analysis is also applied widely to assess paternity and to help suggest the ancestral origin of an individual when it might be unknown, as with adoptees or descendants of slaves. SUMMARY The advances in the genome project have now begun to be evident at the bedside. Individualized or personalized medicine already has applications in pharmacogenetic variation, and risk profiles can also be generated for common diseases. These risk profiles may one day assist in anticipatory guidance for common disorders such as asthma, diabetes, or preterm birth. The microbiome will also represent an opportunity to better identify disease etiology and to improve therapeutic specificity. As long as protections against discrimination continue to be aggressively implemented, detailed genetic information will greatly improve patient care. SUGGESTED READINGS Altshuler D, Daly MJ, Lander ES: Genetic mapping in human disease, Science 322:881-888, 2008. Christensen K, Murray JC: What genome-wide association studies can do for medicine, N Engl J Med 356:1094-1097, 2007. Devaskar SU, Raychaudhuri S: Epigenetics: a science of heritable biological c adaptation, Pediatr Res 61:1R-4R, 2007. Hardy J, Singleton A: Genome-wide association studies and human disease, N Engl J Med 360:1759-1768, 2009. Hirschhorn JN: Genome-wide association studies: illuminating biologic pathways, N Engl J Med 360:1699-1701, 2009. Joss-Moore LA, Lane RH: The developmental origins of adult disease, Curr Opin Pediatr 21:230-234, 2009. Manolio TA, Collins FS, Cox NJ, et al: Finding the missing heritability of complex diseases, Nature 461:747-753, 2009. Complete references used in this text can be found online at www.expertconsult.com C H A P T E R 18 Prenatal Genetic Diagnosis Diana W. Bianchi As a result of the expanding number of prenatal diagnostic tests that are performed on pregnant women, clinicians know a lot about their patients long before they even touch them. To date, prenatal genetic diagnosis has focused broadly on detection of fetal structural abnormalities. Whereas in the past the major indication for prenatal diagnosis was advanced maternal age or a family history of a single-gene disorder, currently most pregnant women receive an invasive prenatal diagnosis because of abnormal screening results. This chapter discusses the common methods of prenatal genetic diagnosis, the information they convey, and the implications for the newborn. NONINVASIVE TECHNIQUES MATERNAL SERUM SCREENING Maternal serum screening is part of routine obstetric care worldwide. It is used to identify a high-risk pregnancy in a low-risk population of pregnant women. First-­ trimester maternal serum screening consists of measuring pregnancy-associated plasma protein A and the free beta subunit of human chorionic gonadotropin (hCG). Secondtrimester maternal serum screening consists of measurement of alpha-fetoprotein (AFP), hCG, unconjugated estriol (uE3), and inhibin A—proteins that are made by the fetus or placenta. Biochemical Screening for Neural Tube Defects AFP is one of the major proteins in fetal serum. Its precise physiologic role is unknown. It can be detected as early as 4 weeks’ gestation when it is synthesized by the yolk sac (Bergstrand, 1986). Subsequently, it is produced in the fetal liver and peaks in the fetal serum between 10 and 13 weeks’ gestation. AFP is then excreted into the fetal urine or leaks into the amniotic fluid through the skin before keratinization at 20 weeks’ gestation. It is also present in cerebrospinal fluid. AFP in maternal serum is exclusively fetal in origin (Crandall, 1981). Maternal serum AFP peaks at 32 weeks’ gestation because of greater placental permeability for the protein (Ferguson-Smith, 1983). For interpretation of results, accurate gestational dating and knowledge of maternal race, weight, and presence or absence of diabetes are critical. Brock and Sutcliffe (1972) observed that there were markedly increased levels of AFP in the amniotic fluid of fetuses with anencephaly and open neural tube defects. Subsequently, it was shown that elevated amniotic fluid AFP levels were associated with increased maternal serum AFP (Ferguson-Smith, 1983). The possibility of a screening test for open neural tube defects became apparent. In the initial collaborative efforts aimed at studying maternal serum AFP, results were expressed as multiples of 180 the median (MoM) to allow comparisons between laboratories. It has become a convention to describe results greater than 2.5 MoM as abnormally high and less than 0.6 MoM as abnormally low. Both findings require further investigation. If the AFP is elevated, the patient is offered an ultrasonographic examination to verify gestational age, determine fetal viability, and diagnose many of the structural abnormalities that can be associated with an elevated AFP. Although the AFP test was developed to screen for neural tube defects, abnormally high results are not specific for this condition (Box 18-1). If the ultrasonographic examination is unrevealing, the patient undergoes amniocentesis to assay the amniotic fluid for the presence of AFP and acetylcholinesterase, which are elevated in open spina bifida (Crandall et al, 1983). Although an elevated AFP is compatible with a normal diagnosis, a study of 277 infants with elevations of maternal serum AFP and normal levels of amniotic fluid AFP revealed a higher incidence of intrauterine growth restriction and non–neural tube anomalies (Burton and Dillard, 1986). Aneuploidy Screening Maternal serum AFP screening has also been used to detect chromosomally abnormal fetuses since the observation was made that a low AFP value was more likely in a fetus with trisomy 18 or 21 than in a normal fetus (Merkatz et al, 1984). Several prospective studies have demonstrated that expressing risk for Down syndrome as a combined function of maternal age and AFP value, and offering amniocentesis to all women with a risk of 1 in 270 or greater (the equivalent risk in a 35-year-old woman based on age alone), makes it possible to detect approximately one third of otherwise unexpected cases of Down syndrome in fetuses (Dimaio et al, 1987; Palomaki and Haddow, 1987). Low AFP values are probably caused by decreased hepatic production in the affected fetus. Although it would make sense to ascribe this phenomenon to the small size of the liver, one study found no association between fetal weight and low AFP values in chromosomally abnormal fetuses (Librach et al, 1988). The differential diagnosis of a decreased AFP level is shown in Box 18-2. Because a low AFP value detects only one third of fetuses with Down syndrome, a normal AFP value does not rule out trisomy 21. Experience with using low maternal AFP levels as a screen for fetal chromosome abnormalities has led to the evaluation of many other proteins produced by both the fetus and placenta. Three of these—uE3, hCG, and inhibin A—were incorporated into second-trimester maternal serum screening panels (Wenstrom et al, 1999). Measurements of all four can be combined to improve the sensitivity and specificity of Down syndrome detection. AFP, uE3, and hCG are only weakly correlated with one another, and CHAPTER 18 Prenatal Genetic Diagnosis BOX 18-1 D  ifferential Diagnosis of ­Abnormally High Maternal Serum Alpha-Fetoprotein Levels ll ll ll ll ll ll ll ll ll ll ll ll ll ll ll ll Incorrect gestational dating Multiple pregnancies Threatened pregnancy loss Fetomaternal hemorrhage Anencephaly Open spina bifida Anterior abdominal wall defects Congenital nephrosis Acardia Lesions of the placenta and umbilical cord Turner syndrome Cystic hygroma Renal agenesis Polycystic kidney disease Epidermolysis bullosa Hereditary persistence (autosomal dominant trait) BOX 18-2 D  ifferential Diagnosis of ­Abnormally Low Maternal Serum Alpha-Fetoprotein Levels ll ll ll ll Incorrect gestational dating Trisomy 18 Trisomy 21 Intrauterine growth restriction their values are all independent of maternal age (Norton, 1994). Elevations of hCG are the most specific markers for fetal trisomy 21 (Bogart et al, 1987; Rose and ­Mennuti, 1993), whereas estriol levels are approximately 25% less than normal. Pregnancies affected by fetal trisomy 18 also have reduced levels of uE3 and hCG. Each measurement is compared with population-specific normal values (i.e., MoM), which are then converted to a likelihood ratio that is expressed as a numeric risk for Down syndrome. Women whose serum screening results indicate a fetal Down syndrome risk of greater than 1 in 270 are offered amniocentesis. Approximately 5% of all serum screen values are calculated to be false-positive results to achieve a sensitivity of detection of at least 70% for cases of Down syndrome. In 1996 the National Health Technology ­Assessment Program in the United Kingdom launched a study comparing the performance of first-trimester to second-trimester screening—the Serum, Urine, and Ultrasound ­Screening Study (SURUSS) (Wald et al, 2003). In addition to the well-validated second-trimester markers, the first-­ trimester serum markers (pregnancy-associated plasma protein A and free beta hCG) and an ultrasound marker, nuchal translucency (NT), were measured. The results suggested that combining all measurements in an integrated panel improved the sensitivity of detection of trisomy 21 to 93% at a fixed false-positive rate of 5%; this was followed by the largest obstetric clinical trial ever performed in the United Sates and funded by the National Institutes of Health—the First and Second Trimester Evaluation of Risk (FASTER) study (Malone 181 et al, 2005a). In this prospective study of 38,189 pregnant women, 117 had a fetus affected with Down syndrome. All study subjects were required to have a first-trimester NT measurement, as well as first- and second-trimester serum screening. The results of the study showed that first-trimester combined serum and sonographic screening was better than conventional second-trimester screening, and that the highest sensitivity was achieved by the integrated test that combined results from both trimesters. The excellent results achieved by both of these large-scale studies resulted in the American College of Obstetricians and Gynecologists recommendation that all pregnant women be offered screening for Down syndrome as part of their routine prenatal care (American College of Obstetricians and Gynecologists, 2007). ULTRASONOGRAPHIC EXAMINATION OF THE FETUS Ultrasonographic examination of the fetus is the best noninvasive method for gestational dating, definition of fetal anatomy, serial measurements of fetal growth, and evaluation of dynamic parameters such as cardiac contractility, fetal urine production, and fetal movement. In addition, it has been suggested that antenatal visualization of the fetus promotes maternal-infant bonding (Fletcher and Evans, 1983). The advent of antenatal ultrasonography has had a significant influence on the types of patients who are admitted to the neonatal intensive care unit. First Trimester With the advent of transvaginal scanning and improved imaging, coupled with advances in first-trimester biochemical screening, fetal ultrasonography during the first trimester has become routine. A major emphasis of firsttrimester sonography is screening for aneuploidy, using the following markers: nuchal translucency measurement, absent nasal bone, cystic hygroma, and abnormal ductus venosus blood flow. NUCHAL TRANSLUCENCY MEASUREMENT Ultrasonographic visualization and quantification of the normal subcutaneous fluid-filled space at the back of the fetal neck is known as the nuchal translucency measurement (Said and Malone, 2008). This particular fluid-filled space is especially well seen in the first-trimester fetus (10 to 14 weeks’ gestation; Figure 18-1). In normal fetuses, the maximal thickness of the subcutaneous translucency between the skin and soft tissue overlying the fetal spine increases as a function of crown-rump length. Normal standards have been established for each gestational week. An enlarged NT measurement is thought to be caused by abnormal or delayed development of the lymphatic system, fetal cardiovascular abnormalities, or defects in the intracellular matrix (Said and Malone, 2008). An increased NT thickness is associated with a higher risk of fetal trisomy 21 (Snijders et al, 1998), cardiac defects in chromosomally normal fetuses (Hyett et al, 1996), and anomalies such as diaphragmatic hernia, omphalocele, skeletal defects, Smith-Lemli-Opitz syndrome, and spinal 182 PART V Genetics chromosomally normal fetuses (Cicero et al, 2001). The prevalence of absence of the nasal bone is affected by ethnicity (most common in individuals of African ancestry), gestational age, and NT thickness. Using standardized imaging techniques, the nasal bone is absent in 70% of fetuses with trisomy 21, 55% of fetuses with trisomy 18, and 34% of fetuses with trisomy 13 (Cicero et al, 2004). ABNORMAL DUCTUS VENOSUS BLOOD FLOW FIGURE 18-1 (See also Color Plate 3.) Sagittal ultrasonographic image of a first-trimester fetus. The nuchal translucency measurement is the distance between the two crosses. A measurement larger than normal standards for gestational age indicates that the fetus is at high risk for Down syndrome, congenital heart disease, or both. (Courtesy Dr. Fergal Malone.) muscular atrophy (Sonek, 2007). In the largest prospective study (96,127 pregnant women) to examine the association of NT and aneuploidy, Snijders et al (1998) detected 82.2% of the cases of trisomy 21 at a false-positive rate of 8.3%. The implementation of large-scale, first-trimester sonographic screening protocols into clinical practice has emphasized the need for quality assessment, training, and standardized measurements. Assessment of the risk of fetal chromosomal abnormalities in the first trimester allows pregnant women the option of earlier (invasive) diagnostic testing. The disadvantage of this approach is that earlier diagnosis of aneuploidy may preferentially identify fetuses already destined for miscarriage. CYSTIC HYGROMA Cystic hygroma manifests sonographically in the first trimester as an enlarged echolucent space that extends along the entire fetal body (Said and Malone, 2008). What distinguished it from the NT is the presence of characteristic septations. Cystic hygroma in the first trimester is the single most powerful sonographic marker that is associated with aneuploidy (Malone et al, 2005b). In the FASTER trial there were 134 cases of cystic hygroma among 38,167 study subjects (incidence of 1 in 285). In the 134 cases, 67 were associated with chromosome abnormalities, (51%), 22 were associated with major structural anomalies (primarily skeletal and cardiac) (34%), and five were associated with fetal death (8%) (Malone et al, 2005b). Therefore pediatricians and neonatologists need to have a high level of suspicion for anomalies in live born infants with an in utero history of cystic hygroma. ABSENT NASAL BONE In the original description of his eponymous syndrome, Dr. Langdon Down noted that affected individuals had a small nose (Down, 1995). In one study, absence of the nasal bone in the fetal facial profile was noted in 43 of 59 (73%) fetuses with trisomy 21 and in 3 of 603 (0.5%) of The use of first-trimester Doppler assessment has identified abnormal blood flow velocity patterns across the ductus venosus (Said and Malone, 2008). Blood flows in the ductus during ventricular systole (the S wave) and diastole (the D wave) have characteristic forms with high velocity (Sonek, 2007). This velocity decreases during atrial contraction (the A wave), but in normal fetuses forward blood flow is maintained. When there is complete cessation or reversal of forward flow, the A wave is considered to be abnormal. Abnormal ductal flow is associated with chromosomal abnormalities, cardiac defects, and adverse pregnancy outcomes (Sonek, 2007). Second Trimester Within the context of prenatal genetic diagnosis, ultrasonography can be used to detect congenital anomalies. A malformation is present in 2% to 3% of live births (Nelson and Holmes, 1989). This risk is doubled in twins. Fetal structures that are normally filled with fluid are especially well visualized by ultrasonography. In approximately 10% of infants with anomalies, the central nervous system is involved (Hill et al, 1983). Ultrasonography is particularly useful in the diagnosis of anencephaly, microcephaly, encephalocele, and hydrocephalus. By 20 weeks’ gestation, the fetal facial structures can be examined with this method for cyclopia, cleft lip, and micrognathia. Nuchal thickening is suggestive of Down syndrome, familial pterygium coli, and other chromosome abnormalities (Benacerraf et al, 1987; Chervenak et al, 1983). Fetal cardiovascular structures may be reliably examined at 20 weeks’ gestation. The presence or absence of four cardiac chambers, the dynamic relationships between the cardiac valves and the locations of the vessels allow diagnoses such as hypoplastic left heart, double-outlet right ventricle, tricuspid atresia, tetralogy of Fallot, and Ebstein anomaly. Pericardial effusion and arrhythmias may be similarly observed. Gastrointestinal anomalies occur in approximately 0.6% of live births, and one third of them are associated with chromosome abnormalities (Barss et al, 1985). The decrease in fetal swallowing seen in some cases of bowel obstruction (from atresia, stenosis, annular pancreas, or diaphragmatic hernia) may lead to polyhydramnios that results in a uterine size greater than expected for gestational dates. Although gastroschisis and omphalocele are readily diagnosed on ultrasonography, they may be confused with each other, and their differing prognoses may cause considerable parental anxiety (Griffiths and Gough, 1985). Gastroschisis usually occurs as an isolated anomaly; infants generally do well after surgical repair. The kidneys are identifiable by 14 weeks’ gestation, but the presence of perirenal fat and large adrenal glands may obscure the CHAPTER 18 Prenatal Genetic Diagnosis diagnosis of renal agenesis (Hill et al, 1983). Renal cysts, hydronephrosis, and obstructive uropathy are easily visualized. Oligohydramnios is indicative of poor renal function. Multiple standard curves have been developed for fetal anthropometric measurements (Elejalde and Elejalde, 1986; Saul et al, 1988). These instruments are particularly helpful in diagnosing skeletal dysplasias and evaluating growth restriction. Fetal genitalia may be determined reliably in the second trimester. In addition, ultrasonographic examination is of benefit in the diagnosis and management of multiple pregnancies. Although there have been no documented adverse outcomes related to ultrasound exposure during human pregnancy, the reported experimental biologic effects—altered immune response, cell death, change in cell membrane functions, formation of free radicals, and reduced cell reproductive potential—necessitate judicious use of this technology. Another concern is the appropriate pediatric follow-up for prenatally observed conditions with unclear clinical significance, such as minimal hydronephrosis and echogenic bowel. SECOND TRIMESTER GENETIC SONOGRAM The genetic sonogram is a term that is used to describe the use of a second-trimester sonographic examination to adjust the risk of fetal aneuploidy (Breathnach et al, 2007). The genetic sonogram relies on the identification of major structural anomalies in association with soft markers—that is, findings that are normal variants that often resolve in the third trimester. An example of a soft marker for trisomy 18 is the presence of choroid plexus cysts. Therefore the presence of choroid plexus cysts, with other soft markers such as echogenic bowel, single umbilical artery, or both, in association with an omphalocele or a neural tube defect, would strongly increase the suspicion of fetal trisomy 18. NEW IMAGING MODALITIES New fetal imaging modalites include three-dimentional ultrasound examination and magnetic resonance imaging (Lee and Simpson, 2007). Three-dimensional sonography improves detection of cleft palate, and magnetic resonance imaging improves assessment of fetal brain anomalies. CELL-FREE NUCLEIC ACIDS IN MATERNAL BLOOD Lo et al (1997) first described the circulation of large amounts of cell-free DNA in maternal plasma and serum samples. Since then, knowledge regarding the biology of fetal cell-free DNA and mRNA in the maternal circulation has greatly expanded (Maron and Bianchi, 2007). The placenta is the predominant source of the circulating nucleic acids. To date, two main clinical applications have transitioned from bench to bedside: fetal rhesus D genotyping (Bianchi et al, 2005) and fetal gender determination (Rijnders et al, 2001). Both have been routinely available in the United Kingdom and Europe for a number of years and are just beginning to be available in the United States (Hill et al, 2010). Fetal gender determination is particularly useful in guiding maternal steroid administration in 183 patients at risk of congenital adrenal hyperplasia, as well as aiding further management of fetuses at risk for X-linked disorders. Extensive research is ongoing toward using circulating cell-free fetal nucleic acids for the noninvasive prenatal diagnosis of aneuploidy and in developing assays of normal fetal developmental gene expression at different gestational ages (Maron et al, 2007). INVASIVE TECHNIQUES AMNIOCENTESIS Amniocentesis refers to the removal of up to 20 mL of amniotic fluid from the pregnant uterus. Contained within this fluid are cellular components (desquamated fetal epithelial and bladder cells) that serve as sources of chromosomes, DNA, or enzymes. Most of the cellular elements are nonviable; therefore amniocytes generally require tissue culture under specific conditions to provide enough material for diagnosis (Gosden, 1983). Herein lies one of the major disadvantages of the procedure, in that results are received late in the second trimester after fetal movement has been perceived by the pregnant woman. In contrast, the amniotic fluid itself may be assayed biochemically immediately after being removed for the presence of AFP, acetylcholinesterase, bilirubin, lecithin, sphingomyelin, or phosphatidylcholine. The indications for genetic amniocentesis are (1) an abnormal maternal serum screen result indicating an increased risk of aneuploidy or neural tube defect, (2) maternal age of 35 years or older at the time of delivery, because there is an increased risk for fetal chromosome abnormalities, (3) a previous pregnancy that resulted in a fetus or an infant with chromosome abnormalities, (4) one parent with a balanced chromosome translocation, (5) a family history of a child with a neural tube defect, (6) a family history of a metabolic disorder for which the enzyme defect is known, (7) maternal history of an X-linked disorder, (8) a family history of a disorder for which DNA diagnosis is available, and (9) detection of a sonographic abnormality that is associated with aneuploidy. Extensive clinical experience with amniocentesis has accrued over the past 30 years (Simpson et al, 1976; U.S. National Institutes of Health, 1976). Historically, institutions in the United States currently quoted a 1% to 2% incidence of minor complications, such as amniotic fluid leakage, uterine cramping, and vaginal spotting after the procedure. The incidence of more serious complications, such as chorioamnionitis and miscarriage, is 0.25% to 0.5% (Centers for Disease Control and Prevention, 1995). More recently, however, the rate of fetal loss after secondtrimester amniocentesis was calculated using data from the FASTER trial (Eddleman et al, 2006). The spontaneous fetal loss rate at less than 24 weeks’ gestation did not differ statistically significantly between those that did and did not undergo the procedure. The procedure-related loss rate after amniocentesis was 0.06%. Because results of amniocentesis are received relatively late in the pregnancy, the Canadian early and Mid-­ Trimester Amniocentesis Trial focused on evaluation of the procedure when performed between 12 and 15 weeks’ gestation. This study showed a higher fetal loss rate and a 1.3% risk of fetal clubfoot when amniocentesis was 184 PART V Genetics performed between 11 and 13 weeks’ gestation (Canadian Early and Mid-Trimester Amniocentesis Trial Group, 1998). The rate of clubfoot in the early amniocentesis group was tenfold the risk in the general population. The cause of clubfoot is thought to be a disruption of normal foot development secondary to transient oligohydramnios (Farrell et al, 1999). For these reasons, early amniocentesis is not generally recommended. Decidua Basalis Amniotic Cavity CHORIONIC VILLUS SAMPLING The increased use of first-trimester biochemical and ultrasound screening has increased uptake of chorionic villus sampling as the diagnostic follow-up procedure of choice. Chorionic villus sampling (CVS) involves the aspiration of the chorion frondosum between 10 and 11 weeks’ gestation (Figure 18-2). The fact that the procedure is performed early is advantageous, because most women at this point do not have external manifestations of pregnancy and have not yet perceived fetal movement. The chorionic villi are composed of syncytiotrophoblast and mesenchymal core cells that are actively growing and dividing. In contrast to the dying epithelial cells shed into the amniotic fluid, chorionic villus cells do not require prolonged culture to provide enough mitoses for a cytogenetic diagnosis. Karyotype results are generally available within 1 week of the procedure. Initially, direct preparations derived from syncytiotrophoblast were used for analysis, but the number of apparently mosaic abnormal results proved unacceptable. Cultured preparations derived from the cell of the mesenchymal core are more closely related in embryonic origin to the actual fetus (Bianchi et al, 1993). It is currently recommended that both direct and cultured preparations be used for cytogenetic analysis. Mosaicism, defined as the presence of two or more cell lines carrying different chromosomal constitutions, is a true biologic (not technical) problem in CVS. In several large studies, 0.8% to 1.7% of 1000 cases demonstrated a chromosome abnormality that was present in the villus but not in the fetus (Hogge et al, 1986; Ledbetter et al, 1992) This finding has led to the observation that postzygotic nondisjunction is more common than was previously suggested. The indications for CVS are the same as those for amniocentesis, with two exceptions. First, neural tube defects cannot be diagnosed by this procedure, and AFP or other serum screening is not routinely offered at this early point in gestation. Second, evidence shows that the fetomaternal hemorrhage associated with placental biopsy results in elevated maternal serum AFP immediately after the procedure (Brambati et al, 1988). CVS is performed transcervically or transabdominally. With the transcervical technique, the inherent risks of fetal and maternal infection appear to be greater, because it is impossible to sterilize the cervix. Under ultrasonographic guidance, a flexible catheter is passed through the endocervix and placed into the chorion frondosum. A small segment of placenta is then aspirated into sterile tissue culture medium, and the catheter is withdrawn (Jackson, 1985). In contrast, the transabdominal technique uses a needle to obtain villus material; sterilization of the skin surface is straightforward (Brambati et al, 1988). With either Chorion Laeve Coelomic Space Chorion Frondosum FIGURE 18-2 A pregnant uterus containing a fetus at approximately 9 weeks’ gestation. The chorion frondosum, if sampled for biopsy, can provide fetal cells for chromosome, enzyme, or DNA analysis. (From Jackson LG: First trimester diagnosis of fetal genetic disorders, Hosp Pract (Off Ed) 20:40, 1985.) method, approximately 10 to 50 mg of tissue is obtained. Subsequently, the villi are dissected from maternal decidua and processed for tissue culture or DNA extraction. The safety and accuracy of these techniques have been extensively monitored. A randomized National Institutes of Health clinical trial compared amniocentesis with CVS (Rhoads et al, 1989). The Centers for Disease Control and Prevention and the National Institutes of Health held a consensus conference to summarize worldwide experience. After adjustment for confounding factors such as gestational age, the risk of miscarriage after CVS was found to be on the order of 0.5% to 1.0% (Centers for Disease Control and Prevention, 1995). The advantages and disadvantages of CVS are summarized in Box 18-3. For patients at high risk for single-gene disorders amenable to DNA diagnosis (e.g., cystic fibrosis, sickle cell anemia, Duchenne muscular dystrophy), CVS is probably the preferred prenatal diagnostic method. Alternatively, for patients with relatively low risk (e.g., a 35-year-old woman being tested for chromosome abnormalities), an amniocentesis may be more appropriate. There have been a few reports of serious maternal sepsis and transient bacteremia in association with transcervical CVS (Barela et al, 1986; Silverman et al, 1994). The 1% incidence of mosaicism in villus samples may necessitate a further invasive technique, such as amniocentesis or cordocentesis, to confirm or refute diagnoses. However, the detection of mosaicism in a CVS sample may identify a fetus at risk for uniparental disomy. Uniparental disomy refers to the inheritance of both copies of a chromosome pair from a single parent. This inheritance occurs when a trisomic fetus undergoes rescue and loses CHAPTER 18 Prenatal Genetic Diagnosis BOX 18-3 A  dvantages and Disadvantages of Chorionic Villus Sampling ADVANTAGES Performed in first trimester; results available quickly ll Cells obtained are mitotically active ll Amount of tissue obtained is preferable for DNA analysis ll Placental mosaicism is detected ll DISADVANTAGES ll Miscarriage rate slightly higher ll Increased fetomaternal hemorrhage after procedure ll Risk of serious maternal infection ll Risk of fetal limb and jaw malformations the extra copy of the chromosome. One third of the time, the fetus will be left with two chromosomes that originated in a single parent. Uniparental disomy is an important mechanism in conditions such as Prader-Willi syndrome. Initially, there was concern regarding the association between CVS and the risk of limb deficiencies in infants whose mothers underwent the procedure. The risk of limb malformations was first suggested by a number of studies describing an increased incidence of transverse limb anomalies and the hypoglossia-hypodactyly syndrome in infants whose mothers had undergone CVS (Burton et al, 1992; Firth et al, 1991; Firth et al, 1994; Hsieh et al, 1995). The overall rate of non–syndrome-related transverse limb deficiency from 65 centers performing CVS is 7.4 per 10,000 procedures (Centers for Disease Control and Prevention, 1995). This number is reasonably similar to that in population-based registries that monitor infants with all limb deficiencies, which give a rate of 5 to 6 per 10,000 live births (Centers for Disease Control and Prevention, 1995). Factors likely to influence the rate of limb malformations include gestational age (risk of limb deficiency is greatest at or before 9 weeks’ gestation and decreased at or after 10 weeks’ gestation), type of catheter used, and operator experience. The overall risk of limb deficiency appears to be on the order of 1 per 3000 procedures (Centers for Disease Control and Prevention, 1995). A higher incidence of hemangioma has also been suggested in infants born after CVS (Burton et al, 1995). The most important issue at present, however, is the fact that relatively few maternal-fetal medicine and obstetric clinics have personnel who are adequately trained to perform the procedure (Cleary-Goldman et al, 2006). CORDOCENTESIS Percutaneous umbilical blood sampling, or cordocentesis, was first described as a means of obtaining fetal immunoglobulin M (IgM) measurements in the prenatal diagnosis of congenital toxoplasmosis (Daffos et al, 1985). Under continuous ultrasonographic imaging, the insertion site of the umbilical cord into the placenta is identified. The umbilical vein is punctured with a 20-gauge needle, the sample is withdrawn, and the umbilical cord is observed for signs of hemorrhage. The technique has been used diagnostically in many clinical settings (Forestier et al, 185 BOX 18-4 Indications for Fetal Blood ­Sampling ll ll ll ll ll ll ll ll Thrombocytopenia Rapid third-trimester diagnosis of chromosome abnormalities* Immunodeficiency Congenital infection Acid-base abnormalities Metabolic disorders* Fetal blood type incompatibility* Hemoglobinopathy* *Diagnosis can also be made by a DNA-based test performed on any nucleated fetal material, such as amniocytes. 1988) (Box 18-4). Regarding genetic diagnosis, the lymphocytes are a source of cells for a rapid karyotype; this is helpful in two situations: (1) when anomalies have been noted on ultrasonographic examination, but it is too late in gestation to perform an amniocentesis (antenatal diagnosis of trisomy 13 or 18 influences delivery room management), and (2) for confirmation of a fetal karyotype when amniocentesis or CVS has shown mosaicism (Gosden et al, 1988). SUMMARY There has been a major shift away from using maternal age as an indication for prenatal genetic diagnosis. Instead, the results of biochemical and sonographic screening tests, which are increasingly being performed in the first trimester, are used to guide the need for amniocentesis and CVS. Safety considerations and patient preferences have influenced the emphasis on noninvasive prenatal diagnosis. The results of large-scale clinical trials, such as the SURUSS and the FASTER study, have shown a high sensitivity and specificity for fetal Down syndrome detection using noninvasive techniques. Circulating cell-free fetal nucleic acids in maternal blood are already being used for noninvasive diagnosis of fetal gender and rhesus D genotype, and many potential opportunities exist to use this material for future clinical applications. SUGGESTED READINGS American College of Obstetrician and Gynecologists: Screening for fetal chromosomal abnormalities, ACOG Practice Bulletin No.77, Obstet Gynecol 109: 217-227, 2007. Breathnach FM, Fleming A, Malone FD: The second trimester genetic sonogram, Am J Med Genet C Semin Med Genet 145C:62-72, 2007. Eddleman KA, Malone FD, Sullivan L, et al: Pregnancy loss rates after midtrimester amniocentesis, Obstet Gynecol 108:1067-1072, 2006. Malone FD, Canick JA, Ball RH, et al: First-trimester or second-trimester screening, or both, for Down’s syndrome, N Engl J Med 353:2001-2011, 2005a. Malone FD, Ball RH, Nyberg DA: First-trimester septated cystic hygroma: prevalence, natural history, and pediatric outcome, Obstet Gynecol 106:288-294, 2005b. Maron JL, Bianchi DW: Prenatal diagnosis using cell-free nucleic acids in maternal body fluids: a decade of progress, Am J Med Genet C Semin Med Genet 145C: 5-17, 2007. Said S, Malone FD: The use of nuchal translucency in contemporary obstetric practice, Clin Obstet Gynecol 51:37-47, 2008. Complete references and supplemental color images used in this text can be found online at www.expertconsult.com C H A P T E R 19 Evaluation of the Dysmorphic Infant Chad R. Haldeman-Englert, Sulagna C. Saitta, and Elaine H. Zackai Genetic disorders have a major impact on public health, as indicated by several large epidemiologic studies (Table 19-1) (Hall et al, 1978; McCandless et al, 2004; Scriver et al, 1973). The latest data indicate that genetic factors contribute to more than two thirds of the conditions prompting admission to a children’s hospital (McCandless et al, 2004). Early identification of the genetic nature of a given condition may then help to appropriately focus resources for providing better care to these individuals. It is therefore critical to implement a systematic approach to evaluating a dysmorphic or malformed infant. This chapter outlines such a general approach. The clinical geneticist incorporates the following five essential tools in the evaluation of a child suspected of having a primary genetic disorder: ll History: prenatal, birth, and medical ll Pedigree analysis or family history ll Specialized clinical evaluation that includes a detailed dysmorphology examination ll Comprehensive literature search ll Focused genetic laboratory analyses (e.g., chromosomes, fluorescence in situ hybridization, DNA microarray, sequencing) HISTORY PRENATAL A complete gestational history should be generated, including results of prenatal testing such as maternal serum screening, ultrasonography, chorionic villus sampling (CVS), and amniocentesis (Box 19-1). The maternal age at conception should be documented, because the risk of chromosomal anomalies such as nondisjunction rises with maternal age. It is important to identify prenatal exposures to infection and medications, maternal habits such as alcohol and drug use, maternal chronic illnesses such as maternal diabetes, and pregnancy-related complications. An additional significant historical component involves the presence of abnormal levels of amniotic fluid. Oligohydramnios (decreased amniotic fluid volume) can be associated with either a fluid leak or a genitourinary abnormality, whereas polyhydramnios (excess amniotic fluid volume) can be seen in fetuses with neuromuscular disease or gastrointestinal malformations. It is also important to identify exposure to environmental agents that might act as teratogens. Teratogens are environmental agents that may cause structural and functional diseases in an exposed fetus. Each teratogen may have a characteristic expression pattern, with a specific range of associated structural anomalies and dysmorphic features. Effects and the extent of the effects depend on the time of exposure, duration, and dosage as well as interactions 186 TABLE 19-1 Genetic Disorders in Pediatric Hospital Admissions Genetic disorders Montreal (1973) Chromosome, single gene (%) 7.3 Seattle (1978) Cleveland (2004) 4.5 11 Polygenic (%) 29 49 60 Nongenetic disorders (%) 64 47 29 12,801 4115 5747 Total number of admissions Data from Hall JG, Powers EK, McIlvane RT, et al: The frequency and financial burden of genetic disease in a pediatric hospital, Am J Med Genet 1:417-436, 1978; McCandless SE, Brunger JW, Cassidy SB: The burden of genetic disease on inpatient care in a children’s hospital, Am J Hum Genet 4:121-127, 2004; and Scriver CR, Neal JL, Saginur R, et al: The frequency of genetic disease and congenital malformation among patients in a pediatric hospital, Can Med Assoc J 108:1111-1115, 1973. BOX 19-1 Elements of Prenatal History for the Dysmorphic Infant MATERNAL HEALTH Age Disease: diabetes, hypertension, seizure disorder MODE OF CONCEPTION Natural Assisted reproductive technologies Fertility medications In vitro fertilizatioin Intracytoplasmic sperm injection Gamete intrafallopian transfer Artificial insemination EXPOSURES Medications Alcohol Environmental agents Infections (gestational age at exposure) PRENATAL TESTING Ultrasonography (gestational age performed) Triple screen Chorionic villus sampling, amniocentesis, and indications with maternal and genetic susceptibility factors. In general, more severe effects are typically correlated with exposure early in the pregnancy and with more extensive (i.e., higher dose) exposure. The list of well-documented human teratogens is short and includes such substances as alcohol, thalidomide, warfarin, trimethadione, valproate, and hydantoin. If history of an exposure is documented, an effort should be made to identify the developmental time and level of exposure. This information is critical, because CHAPTER 19 Evaluation of the Dysmorphic Infant Male Marriage or union Female Divorced Sex unspecified Consanguinity Number of children of sex indicated Monozygotic twins 187 TABLE 19-2 Dermatoglyphic Patterns Associated With Specific Dysmorphic Disorders Dermatoglyphic Pattern Associated Disorders Excess arches Trisomy 13, trisomy 18, Klinefelter syndrome (47,XXY), deletion 5p (cri du chat), fetal phenytoin exposure Excess ulnar loops Trisomy 21 Excess whorls Smith-Lemli-Opitz syndrome, Turner syndrome (45,X), 18q deletion 3 2 Affected Dizygotic twins Nonpenetrant carrier, may manifest disease Obligate carrier, will not manifest disease ? BOX 19-2 Adjunct Studies in the Evaluation of the Dysmorphic Infant ll ll ll ll ll ll ll ll ll Investigation of internal malformation Assessment of neurologic function Identification of organ systems involved Ultrasonography magnetic resonance imaging Brain imaging Electroencephalography as indicated Electromyography as indicated Prognosis Treatment and intervention the counseling and calculation of recurrence risk for a given malformation are vastly different if environmental exposures are involved. BIRTH Another important component of the gestational history is obtaining information on fetal activity, size, and position. Often the mother’s subjective impressions can be further confirmed by examining obstetric records of the perinatal period. A history of hypotonia may be further supplemented by reports of poor fetal movements and breech presentation. Perinatal information including gestational age, fetal position at delivery, the length of labor, type of delivery, and any evidence of fetal distress, such as passage of meconium, are all relevant data (Table 19-2). Apgar scores, the need for resuscitation, birth parameters (weight, length, and head circumference), any malformations seen at birth, and all abnormal test results should be noted. MEDICAL A full review of the medical issues of the child should include the baby’s general health, test results, identification of any chronic medical issues, and need for hospitalization. Evaluation of growth, review of systems, developmental assessment, and notation of unusual behaviors can also provide important clues to a diagnosis. PEDIGREE ANALYSIS AND FAMILY HISTORY A critical part of any genetic evaluation is to obtain the family history (Box 19-2); this is best accomplished by creating a three-generation pedigree, which is a schematic II 2 1 I Proband Twins of unknown zygosity 1 2 3 Pedigree with generations and individuals numbered Diseased individual Miscarriage Sillbirth No offspring Adopted into family Adopted out of family 1 2 Multiple unions FIGURE 19-1 Symbols commonly used for pedigree notation. (From Nussbaum RL, McInnes RR, Willard HF, eds: Thompson and Thompson’s genetics in medicine, ed 7, Philadelphia, 2007, WB Saunders, p 117.) diagram depicting familial relationships using standard accepted symbols (Figure 19-1). This formal record can also be used to summarize positive responses elicited during the interview. Special attention should be paid to ethnic origins of both sides of the family, consanguinity, and any first-degree relatives with similar malformations to those of the patient being evaluated, also known as the index case, proband, or propositus. An extended family history should be used to identify relatives with congenital anomalies, developmental abnormalities, or physical differences. Often photographs can provide clear objective evidence of a descriptive history. Reproductive histories, especially of the parents, should be elicited. Specifically, questions should be asked about infertility, miscarriages, and stillbirths. The occurrence of more than two first-trimester miscarriages increases the probability of finding a balanced translocation in one parent (Campana et al, 1986; Castle and Bernstein, 1988). A balanced translocation is a rearrangement of genetic material such that two chromosomes have an equal exchange without loss or gain of material. There are typically no associated clinical features with such a rearrangement. However, when chromosomes align to recombine for meiosis in the sperm or egg, this exchange produces a risk of unequal distribution and an unbalanced translocation in the resulting fetus. In this case, there would be aneuploidy for part of a chromosome. It has been estimated that 25% of stillbirths exhibit single or multiple malformations, and in at least half of these cases there is a genetic etiology for the malformations. Couples with two or more pregnancy losses should undergo routine chromosome analysis or karyotyping. When possible, such analysis should be performed on the stillborn fetus or on products of conception. 188 PART V Genetics Obtaining a formal family history is helpful in discovering information that is often critical to making a diagnosis. Positive responses may help to discern a Mendelian pattern of inheritance for a given genetic disorder. For example, a disease affecting every generation, with both males and females involved, such as Marfan syndrome, would most likely be autosomal dominant. A pattern of X-linked recessive disease, such as hemophilia, would show affected males related through unaffected or minimally affected females; transmission in this pattern should not occur from father to son. A congenital malformation can be described as a “morphologic defect of an organ, part of an organ, or larger region of the body resulting from an intrinsically abnormal developmental process” (Jones, 2006). The term dysmorphology was introduced by Dr. David Smith in the 1960s to describe the study of human congenital malformations (Aase, 1990). This study of “abnormal form” emphasizes a focus on structural errors in development with an attempt to identify the underlying genetic etiology and pathogenesis of the disorder. In a landmark study, Feingold and Bossert (1974) examined more than 2000 children to define normal values for a number of physical features. These standards were devised as screening tools to objectively identify children with differences possibly attributable to a genetic disorder. Important measurements include head circumference, inner and outer canthal distances, interpupillary distances, ear length, ear placement, internipple distances, chest circumference, and hand and foot lengths. Other graphs and measurements using age-appropriate standards can be found in compendia such as the Handbook of Physical Measurements (Hall et al, 2007). The assessment should begin with newborn growth parameters that can reflect the degree of any prenatal insult. Measurements such as height, weight (usually reflecting nutrition), and head circumference should be plotted on newborn graphs. Gestational age–appropriate graphs should be used for premature infants. It is often helpful to express values that are outside the normal range as 50th percentile for a different gestational age. For example, a full-term baby with microcephaly may have a head circumference of less than the 5th percentile for 38 weeks. This can be expressed as a measurement at the 50th percentile for 33 weeks, which imparts the degree of microcephaly more clearly. A complete physical examination should include assessment of patient anatomy for features varying from usual or normal standards. This assessment can often provide clues to embryologic mechanisms. The data obtained should then be interpreted in regard to normal standards using comprehensive standard tables that are available for these purposes. Special attention to familial variants should be given. The shape and size of the head and fontanels should be noted as well as the cranial sutures, with assessment for evidence of craniosynostosis or an underlying brain malformation. Any scalp defects should also be noted. The shape Interpupillary distance PHYSICAL EXAMINATION FOR DYSMORPHOLOGY in 2.8 cm 7.0 2.6 6.5 2.4 6.0 2.2 5.5 2 5.0 1.8 4.5 1.6 4.0 1.4 3.5 1.2 3.0 1 2.5 ) .5 (1 4) 0 . 4. (1 2) 5 . 3. (1 .0) 0 3. 5 (1 9) 2. (0. ) 0 .6 2. (0 .4) 5 1. 0 (0 1. Inner canthal distance 4 5 6 7 8 9 10 cm 1.6 2 2.4 2.6 3.2 3.6 4 in Outer canthal distance FIGURE 19-2 Various eye measurements are depicted (top). A indicates the outer canthal distance, B indicates the inner canthal distance, and C indicates the interpupillary distance (IPD), which is difficult to measure directly. The IPD can be determined using the graph at the bottom or with the Pryor formula: IPD = (A − B) 2 + B. (From Feingold M, Bossert WH: Normal values for selected physical parameters: an aid to syndrome delineation, Birth Defects 10:1-16, 1974.) of the forehead, appearance of the eyebrows (noting synophrys), and the texture and distribution of hair should be noted. The spacing of the eyes, or canthal measurements (Figure 19-2), the interpupillary distances (Figure 19-3; see also Figure 19-2), palpebral fissure lengths (Figure 19-4), presence or absence of colobomata and epicanthal folds, and noting whether the palpebral fissures are turned upward or downward are components of the dysmorphology examination. Examination of the ears should include a search for preauricular and postauricular pits, tags, and assessment of the placement (Figure 19-5), length (Figure 19-6), and folding of the ear is important. Ear development occurs in a temporal frame similar to that of the kidneys, and external ear anomalies can be associated with renal anomalies. Evaluation of the nose should cover the shape of nasal tip, the alae nasi, presence of anteverted nares, the length of the columella, and patency of the choanae. 189 CHAPTER 19 Evaluation of the Dysmorphic Infant The mouth and throat are examined for the presence of a cleft lip or palate; the shape of the palate and uvula are noted, and the presence of unusual features, such as tongue deformities, lip pits, frenula, and natal teeth, are recorded. A small retrognathic or receding chin, which can be a part of several syndromes or an isolated finding, should be noted. The neck is inspected for excess nuchal folds or skin and evidence of webbing. Any bony abnormalities in the neck should prompt an evaluation of the cervical vertebrae to confirm cervical and airway stability. Evaluation of the chest and thorax involves lung auscultation and cardiac examination. Abnormal findings should prompt a consultation with a cardiologist and appropriate echocardiographic or invasive studies as needed. External measurements include determining the internipple distance and its ratio in regard to the chest circumference (Figure 19-7). The abdominal examination is focused on determining whether organomegaly is present, a finding typically associated with an inborn error of metabolism. The umbilicus should also be examined, with any hernias and the number of vessels present in the newborn cord being noted. A two-vessel cord, in which only a single artery is present, can be associated with renal anomalies. The genitourinary examination concentrates on determining whether anomalies such as hypospadias, chordee, cryptorchidism, microphallus, and ambiguous genitalia are present. These external anomalies may be associated with internal anomalies involving the upper urinary tract as 34 250 1.3 32 150 Mean Interpupillary distance 6 75 50 25 2.25 2.00 5 3 1.75 1.50 28 1.2 150 250 1.1 26 1.0 24 0.9 22 0.8 20 4 18 3 0 3 6 9 12 15 18 21 2 3 4 5 6 7 8 9 10 11 12 13 14 1.25 Months Years Age FIGURE 19-3 A nomogram for interpupillary distance at different ages for both sexes. (From Feingold M, Bossert WH: Normal values for selected physical parameters: an aid to syndrome delineation, Birth Defects 10:1-16, 1974.) Inches % 97 in cm Palpebral fissure length (mm) 30 16 0.7 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Age from birth (years) FIGURE 19-4 A graph of palpebral fissure length from birth to age 16 years for both sexes. (From Hall JG, Allanson JE, Gripp KW, et al: Handbook of physical measurements, Oxford, 2007, Oxford University Press.) FIGURE 19-5 Ear placement. Using the medial canthi (A and B) as landmarks, one draws a central horizontal line and extends it to a point (C ) on the side of the face. Ears attached below this line are considered low set. 190 PART V Genetics well. The anus is examined for evidence of tags, its placement, and its patency. The back should be assessed, especially for the shape of the spine and any associated defects, such as myelomeningocele. These defects prompt further radiologic evaluation to assess for potential functional limitations. In addition, a sacral dimple or hair tuft at the base of the spine should be noted, because either could signify developmental abnormalities in the underlying neural tissue. Minor anomalies are often manifested in the extremities. Gross differences in the hands and feet include polydactyly (more than five digits), whether the extra digits are located in a preaxial or postaxial position should be noted, syndactyly (fusion of the digits), clinodactyly (incurving of the digits), and extremity length, which should be expressed as a percentile measured on age-appropriate in cm 7 97 Total ear length 2.5 % 75 50 6 2.25 25 2 5 3 1.75 4 1.50 0 3 6 9 12 15 18 21 2 3 4 5 6 7 8 9 10 11 12 13 14 3 Months Years Age FIGURE 19-6 Graph showing various percentiles for ear length plotted against age. (From Feingold M, Bossert WH: Normal values for selected physical parameters: an aid to syndrome delineation, Birth Defects 10:1-16, 1974.) graphs (Figures 19-8 and 19-9). Often these data can provide important clues to a unifying syndrome. Dermal ridge patterns, or dermatoglyphics, are formed on the palms and soles early in embryonic life, and they vary considerably among individuals. This variation can be inherited and can be influenced by disturbances to the development of the peripheral limb buds. Environmental exposures and chromosomal aberrations can greatly affect the formation of these structures and are reflected by the dermatoglyphic pattern of an individual. Each of the distal phalanges has one of three basic dermal ridge patterns: arches, whorls, or loops (Figure 19-10). The predominance of a single pattern can be an associated feature of a genetic disorder. For example, the occurrence of arches on eight or more digits is a rare event, but is frequently encountered in children with trisomy 18 (Box 19-3). Deltas, or triradii, form at the convergence of three sets of ridges on the palm. This junction is where the hypothenar, thenar, and distal palmar patterns converge. There are typically no triradii in the hypothenar area of the palm, but when patterning is present or is large, a distal triradius arises, which is found in only 4% of normal Caucasian individuals but in 85% of patients with trisomy 21 (Down syndrome). A single transverse palmar crease is found in 4% of controls, but in more than half of patients with trisomy 21 (Figure 19-11) and in even greater proportions in patients with other trisomies. The hallucal area of the foot, located at the base of the big toe, also has a dermal ridge pattern, usually a loop or whorl. A simple pattern or open field in this region is found in less than 1% of controls but in more than 50% of patients with Down syndrome (Figure 19-12). This unusual dermal pattern is also associated with hypoplasia of the hallucal pad and a wide space between the great and second toes in these patients. An examination of the skin is also important, to look for phakomatoses or skin manifestations that herald the presence of an underlying disorder. Examples are café-au-lait spots (associated with neurofibromatosis type I) and ash leaf spots (associated with tuberous sclerosis and detected 40 38 36 34 32 30 28 26 24 22 20 18 16 14 12 10 8 cm % 20 97 75 50 25 3 19 Years Age FIGURE 19-7 The internipple distance as a percentage of the chest circumference plotted against age for both sexes. (From Feingold M, Bossert WH: Normal values for selected physical parameters: an aid to syndrome delineation, Birth Defects 10:1-16, 1974.) 17 6 15 5 13 4 11 9 3 7 2 5 0 3 6 9 12 15 18 21 2 3 4 5 6 7 8 9 10 11 12 13 14 Months Total hand length 7 % 97 75 50 25 3 0 3 6 9 12 15 18 21 2 3 4 5 6 7 8 9 10 11 12 13 14 Internipple distance as % of chest circumference in Months Years Age FIGURE 19-8 The total hand length plotted against age for both sexes. (From Feingold M, Bossert WH: Normal values for selected physical parameters: an aid to syndrome delineation, Birth Defects 10:1-16, 1974.) 97 75 50 25 A 11 10 3 7 6 5 Inches 8 Foot length, females (cm) 9 4 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 97 75 50 25 10 9 3 8 7 6 5 4 3 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 3 Age (years) 191 Inches 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Foot length, males (cm) CHAPTER 19 Evaluation of the Dysmorphic Infant B Age (years) FIGURE 19-9 Total foot lengths plotted against age for boys (A) and girls (B). (From Hall JG, Allanson JE, Gripp KW, et al: Handbook of physical measurements, Oxford, 2007, Oxford University Press.) BOX 19-3 Underlying Mechanisms of Malformation SYNDROME Pathogenetically related pattern of anomalies SEQUENCE Pattern of anomalies derived from a presumed or known prior anomaly or mechanical disturbance Simple Arch ASSOCIATION Nonrandom occurrence of multiple anomalies FIELD DEFECT Disturbance of a developmental field leading to a pattern of anomalies Loop with the use of a Wood’s lamp). Irregular pigmentation, such as hypomelanosis of Ito, can be suggestive of chromosomal mosaicism, in which the different skin pigmentation patterns represent a different, mixed chromosomal composition. Hemangiomas and other skin diseases are also noteworthy. Finally a careful neurologic examination with input from a specialist is often warranted in the child with multiple anomalies, because the neurologic status is often the most reliable prognostic indicator. Evaluation of tone, feeding, unusual movements, and the presence of seizure activity are critical pieces of diagnostic information. ADJUNCT STUDIES Whorl (Spiral) FIGURE 19-10 Basic fingerprint patterns (dermatoglyphics). (From Holt SB: The genetics of dermal ridges, Springfield, Ill, 1968, Charles C Thomas.) An exhaustive physical examination often reveals differences that require further evaluation for diagnostic, prognostic, and treatment purposes (Box 19-4). Poor feeding or cyanosis may lead to detection of internal organ malformations on echocardiogram or abdominal ultrasonography. Differences in head shape suggest the need for skull radiographs, three-dimensional computed tomography, 192 PART V Genetics BOX 19-4 P  rocesses Leading to Altered Form or Structure DEFORMATION Abnormal form resulting from mechanical forces DISRUPTION Morphologic defect caused by interference with a previously normal developmental process DYSPLASIA Altered morphology because of abnormal organization of cells into a given tissue BOX 19-5 Examples of Morphologic Differences ll ll ll ll ll FIGURE 19-11 The patterns on the hand of a patient with Down syndrome depicting the palmar crease (D). (From Holt SB: The genetics of dermal ridges, Springfield, Ill, 1968, Charles C Thomas.) ll ll ll ll Malformation Deformation Disruption Dysplasia Cardiac septal defects, cleft lip Club foot Amniotic bands Localized: hemangioma Generalized (skeletal): achondroplasia electroretinograms are needed to predict visual prognosis. Often there are well-characterized genetic disorders that have a specific pattern of abnormal findings in these highly specialized studies. Even when a unifying diagnosis is reached, there is often variation in the clinical phenotype, and determining the patient’s prognosis on a system-bysystem basis is typically the most appropriate and accurate way to proceed. LITERATURE REVIEW FIGURE 19-12 The distal sole of the foot of a patient with Down syndrome, depicting the characteristic “open field” pattern. (From Holt SB: The genetics of dermal ridges, Springfield, Ill, 1968, Charles C Thomas.) or magnetic resonance imaging of the brain. A disproportionality of the limbs prompts a skeletal survey and bone age measurement. It is prudent in children with anomalies involving multiple systems to obtain the input of relevant specialists. This step is often essential to medical decision making and the planning of interventions. Abnormal neurologic findings should prompt a consultation with a trained specialist and interpretation of studies such as head ultrasonography, brain magnetic resonance imaging, brainstem auditory evoked responses, and electroencephalogram for seizure activity. Muscle dysfunction might result in the ordering of electromyography, muscle ultrasound, or nerve conduction studies. Visual involvement requires a funduscopic examination by an experienced pediatric ophthalmologist, and sometimes studies of visual evoked responses or The occurrence of malformations can fit into one of several categories (Box 19-5). The next step toward reaching a diagnosis is to analyze the data generated from the evaluation and attempt to categorize the findings. A syndrome is a “collection of anomalies involving more than one developmental region or organ system,” (Aase, 1990). The word itself means a “running together” or “pattern of multiple anomalies thought to be pathogenically related.” Therefore a given congenital anomaly may be an isolated defect in an otherwise normal individual or part of a multiple malformation syndrome. Furthermore, the primary malformation itself can determine additional defects through an interrelated cascade of physical and functional processes; if ensuing malformations are related to one primary defect, factor, or event, a pathogenetic sequence has occurred. A classic example is the Pierre-Robin sequence or constellation, consisting of a small recessed jaw, midline U-shaped cleft palate, and relatively large and protruding tongue. The primary anomaly is the small jaw, which does not allow adequate room for the tongue and displaces it superiorly. The displaced tongue prevents closure of the palatine shelves, causing the cleft palate. The recurrence risk with this isolated occurrence is negligible. However, CHAPTER 19 Evaluation of the Dysmorphic Infant BOX 19-6 E  lements of Perinatal and Birth History for the Dysmorphic Infant ll ll ll ll ll ll ll ll Fetal activity Delivery Type (e.g., indication for cesarean section) Gestational age Fetal presentation Apgar scores, history of distress, or resuscitation Growth parameters Malformations noted one should also remember that such sequences can also be part of a larger constellation of findings that does fit into a syndrome, as Pierre-Robin sequence can when it is part of velocardiofacial syndrome (caused by a deletion of chromosome 22q11) or Stickler syndrome (associated with mutations in the type II collagen gene). The recurrence risk for an affected individual passing on these particular syndromes to their children is 50%. In addition, a cluster of several malformations that are not developmentally related can occur in a nonrandom fashion called an association that may appear without characteristic dysmorphic features. One such statistically nonrandom association of defects consists of vertebral defects, anal atresia, cardiac defects, tracheoesophageal fistula, esophageal atresia, renal dysplasia, and limb anomalies (VACTERL). It should be noted that not all features need to be present and that the extent of involvement of each system is widely variable. Recently, the CHARGE syndrome (coloboma, heart disease, atresia choanae, retarded growth and development, genital anomalies, and ear anomalies/deafness) was considered an association until the CHD7 gene was identified to cause approximately 60% of the cases (Lalani et al, 2006). These associations often manifest as sporadic rather than familial occurrences. Because they are not clearly related by a common etiology or pathogenesis, they are not considered syndromes and do not technically constitute a diagnosis. Instead, they are a recognition of a statistically significant association of features. It is important to remember that many of these same anomalies can occur as features of chromosomal aneuploidy or other syndromes. Syndromic malformations tend to occur in more than one developmental field. A field defect or complex is a set of primary malformations in a developmental field that originates from a single or primary abnormality in embryonic development (see Box 19-5). When generating a differential diagnosis of malformations that might occur together, the evaluator must also consider structures that may appear abnormally formed, but in fact are structures that underwent normal development and then received some insult that distorted their true form (Box 19-6). For example, a deformation describes the abnormal form, shape, or position of a part of the body that was caused by mechanical forces. Examples are clubfoot, hip dislocation, and craniofacial asymmetry; they can result from intrinsic (embryonic) or extrinsic (intrauterine) mechanical forces that alter the shape or position of an organ or part that had already undergone normal differentiation. Deformations are estimated to occur in 2% of births, 193 BOX 19-7 E  lements of Pedigree ­Analysis and Family History for the ­Dysmorphic Infant Identification of relatives with: Congenital anomalies (especially those similar to proband’s) ll Mental retardation Photographs (objective evidence) Parental reproductive history ll Pregnancy losses (gestational ages) ll Infertility Medical histories of primary relatives Ethnic origin Consanguinity ll and such factors as fetal crowding from the presence of multiple fetuses and uterine malformations, as well as oligohydramnios, and a face presentation during delivery can cause them. Along similar lines, a disruption describes a “morphologic defect of an organ, part of an organ, or larger region of the body resulting from the extrinsic breakdown of, or an interference with, an originally normal developmental process,” (Aase, 1990). The classic example of a disruption is entanglement of the fetus in amniotic bands. Amniotic bands are ribbons of amnion that have ruptured in utero and cause disruptions of normal developmental processes in the fetus, either through physical blockage or interruption of the blood supply or by entangling and tearing of developing structures. This effect is seen most often with digits and limbs, and remnants of the bands, or constriction marks, can frequently be seen at birth. If the fetus should swallow a band, a cleft palate might result; the etiology is a very different etiology from that of cleft palate occurring as a primary malformation. Recurrence risk counseling of the parent would be very different in these two scenarios. Dysplasias occur when there is “an abnormal organization of cells into tissue(s) and its morphologic results,” (Aase, 1990). Dysplasia tends to be tissue specific rather than organ specific (e.g., skeletal dysplasia) and can be localized or generalized. In summary, structural or morphologic changes identified at birth can occur during intrauterine development as a result of malformations, deformations, disruptions, or dysplasia. However, approximately 90% of deformations undergo spontaneous correction. Malformations and disruptions often require surgical intervention when possible. Dysplasias are typically not correctable, and the affected individual experiences the clinical effects of the underlying cell or tissue abnormality for life (see Box 19-7). Examples of these entities are listed in Box 19-7. After the history and physical evaluation are complete, a cross-reference of two or more anomalies is useful to generate a differential diagnosis. When the rest of the neonate’s physical and history findings are added, the possibilities can often be narrowed down to a few entities that may be amenable to diagnostic testing. If multiple anomalies are present, it is usually best to start with the least common. As Aase (1990) has stated, “The best clues are the rarest. The physical features that will be the most 194 PART V Genetics helpful on differential diagnosis are those infrequently seen either in isolation or as part of syndromes. Quite often, these are not the most obvious anomalies or even the ones that have the greatest significance for the patient’s health.” Cross-referencing is usually best accomplished by using published compendia of malformation syndromes. These compendia have been supplemented by databases that are accessible online (i.e., GeneReviews, Online Mendelian Inheritance in Man, and PubMed). The availability of such tools allows the cross-referenced features to be compared easily with those of other described syndromes that may include similar malformations. This systematic review produces a differential diagnosis for the constellation of features described and identifies references to pertinent literature. The recognition of patterns of genetic entities involves the comparison of the proband with the examiner’s personal experience of known cases and a search of the literature. Multiple anomalies may be causally related, occur together in a statistically associated basis, or occur together merely by chance. Diagnosis of a genetic disorder relies heavily on the ability of the clinician to suspect, detect, and correctly interpret physical and developmental findings and to recognize specific patterns. Accurate diagnosis of a syndrome in a child is important to the identification of major complications and their treatment if possible. It is also crucial for long-term management of patients and for parental counseling about recurrence in future offspring. SPECIALIZED LABORATORY TESTS In sorting through the array of possibilities listed, the geneticist uses one other important tool—the availability of highly specialized cytogenetic and molecular genetic testing, including: ll ll ll ll Karyotype Fluorescence in situ hybridization DNA microarray (see Chapter 20) ll Comparative genomic hybridization ll Single nucleotide polymorphism or oligonucleotide arrays Molecular analysis The standard karyotype, or analysis of stretched and stained chromosome preparations usually taken from a peripheral blood sample, can often confirm a suggested diagnosis or explain a set of major malformations not classically encountered together. Further description of specific chromosomal abnormalities is addressed in Chapter 20; it is sufficient to note that multiple malformation syndromes can result from large visible chromosome rearrangements that lead to deletion or addition of material (aneuploidy). These rearrangements can involve an entire arm of a chromosome or can be submicroscopic, requiring further special testing. Such small deletions can often be detected by fluorescence in situ hybridization analysis, which is performed using a probe specific for the deleted region. It has become the standard of care in several centers to offer more specialized molecular testing, such as a DNA microarray or individual gene sequencing, as an adjunct to or instead of karyotype analysis. In general, microarraybased methods are currently focused on detecting copy number changes (smaller deletions or duplications not detectable by a karyotype) and can be performed in a targeted or genome-wide fashion (Emanuel and Saitta, 2007). As more information is made available about the role of individual gene mutations in newborns with congenital malformations, testing for these mutations is becoming available in diagnostic laboratories. It can be useful to check with a geneticist or genetics counselor for the availability of gene mutation testing that may be clinically available or performed on a research basis. GeneTests is an internet database of laboratories worldwide that provides such services. DIAGNOSIS There are cases in which, after a detailed examination, exhaustive literature search, and genetic testing, no unifying diagnosis is evident. Aase (1990), a dysmorphologist, advises, “Don’t panic! The absence of a diagnosis may be distressing to the diagnostician and the family, but it is much less dangerous than the possibility of assigning the wrong diagnosis with the risk of erroneous genetic and prognostic counseling and possibly hazardous treatment.” Therefore, in cases in which there is no clear diagnosis, prognosis and treatment should be determined according to the organ systems involved and the extent of their impairment. In addition, when the infant has a severe, untreatable impairment or the patient’s condition is critical, it may be prudent to offer and obtain consent for a full postmortem examination by an experienced pathologist. A skin sample, and sometimes blood, can be taken from the expired fetus for establishing a cell line or for extracting DNA for future testing. Information gained from such investigations may often become relevant for family members, including the parents, allowing one to provide accurate recurrence risk counseling and perhaps offer prenatal testing of a new pregnancy. Such information can often help to provide closure for the family as well. When should a genetics evaluation be considered? The following clinical situations prompt a further genetic evaluation and counseling by a specialist: ll ll ll ll ll Multiple anatomic anomalies History of maternal exposure to teratogens Familial disorders Increased carrier frequency or ethnic risk Multiple pregnancy losses As described previously, if a birth defect is identified in the presenting patient or proband, especially if the defect is associated with other anatomic anomalies, short stature, or developmental delays, the features of a specific genetic syndrome may be present. A known history of maternal exposure to a potential teratogen would also be an indication for consultation. Conditions appearing to be familial, a family history of hereditary disorders involving malformation of a major organ, or major physical differences such as unusual body proportions, short stature, or irregular skin pigmentation, would warrant genetic investigation. Mental retardation, blindness, hearing loss, or neurologic deterioration in multiple family members suggests a genetic etiology. Likewise, a strong family history of cancer or a defined ethnic risk such as Ashkenazi Jewish heritage and CHAPTER 19 Evaluation of the Dysmorphic Infant its association with a higher carrier frequency for TaySachs disease would be an indication for genetic evaluation. The occurrence of multiple pregnancy losses would also raise the suspicion of a genetically influenced cause and indicate the need for further investigation and counseling. SUMMARY Diseases with underlying genetic bases have significant effects on health care and its delivery. An appreciation of these entities, coupled with an organized, systematic evaluation, can help to define the nature of a given disorder and aid in the development of the optimal plan of treatment and care for the patient. SUGGESTED READINGS Aase JM: Diagnostic dysmorphology, New York, 1990, Plenum. Carey JC, editor: Elements of morphology: standard terminology, Am J Med Genet 149A:1-127, 2009. Gehlerter TD, Collins FS, Ginsburg D, editors: Principles of medical genetics, ed 2, Baltimore, Md, 1998, Williams & Wilkins. 195 GeneTests. Available online at www.ncbi.nlm.nih.gov/sites/GeneTests/?db= GeneTests. Hall JG, Allanson JE, Gripp KW, et al: Handbook of physical measurements, Oxford, 2007, Oxford University Press. Hennekam RCM, Krantz ID, Allanson JE, editors: Gorlin’s syndromes of the head and neck, ed 5, New York, 2010, Oxford University Press. Jones KL: Smith’s recognizable patterns of human malformation, ed 6, Philadelphia, 2006, WB Saunders. Nussbaum RL, McInnes RR, Willard HF, editors: Thompson and Thompson’s genetics in medicine, ed 7, Philadelphia, 2007, WB Saunders. Online Mendelian Inheritance in Man (OMIM). Available online at www.ncbi.nlm.nih.gov/omim/. Rimoin DL, Connor JM, Pyeritz RE, et al: Emery and Rimoin’s principles and practice of medical genetics, ed 5, New York, 2006, Churchill Livingstone. Schinzel A: Catalogue of unbalanced chromosome aberrations in man, Berlin, 2001, Walter de Gruyter. Spranger J, Benirschke K, Hall JG, et al: Errors of morphogenesis: concepts and terms, J Pediatr 100:160-165, 1982. Stevenson RE, Hall JG, editors: Human malformations and related anomalies, ed 2, New York 2006, Oxford University Press. Complete references used in this text can be found online at www.expertconsult.com C H A P T E R 20 Specific Chromosome Disorders in Newborns Chad R. Haldeman-Englert, Sulagna C. Saitta, and Elaine H. Zackai It has been estimated that 3% of newborns have a major structural anomaly that will affect their quality of life. Although most of these patients have a single malformation, 0.7% of infants have multiple major malformations. In an additional 2%, a major anomaly is discovered by the age of 5 years. Early identification of the genetic nature of a given condition can aid in treatment and help to identify resources for providing better health care to these individuals. This chapter will focus on genetic disorders and syndromes with underlying chromosomal abnormalities that typically manifest in the newborn period, in addition to the shift in genetic evaluation and diagnosis with the advent of newer array-based diagnostic techniques. HUMAN KARYOTYPE Chromosomes consist of tightly compacted DNA whose structure is maintained by association with histones and other proteins. When treated and stretched, chromosomes from dividing cells can be visualized under the light microscope as linear structures with two arms joined by a centromere. The short arm is designated p (petit) and the long arm is designated q. The ends of the p and q arms are known as telomeres. Human chromosomes were first visualized in 1956 (Lejeune and Turpin, 1960), and each pair shows a distinctive size, centromeric position, and staining or banding pattern after treatment with special dyes, allowing it to be identified and classified. Each chromosome is identified by a number, in general from largest to smallest, in standard international cytogenetic nomenclature. This presentation, or karyotype (Figure 20-1), normally consists of 46 chromosomes, with 22 pairs of autosomes and one set of sex chromosomes—two X chromosomes for females (46,XX), and an X and a Y for males (46,XY). Karyotype analysis is performed in cells undergoing mitosis, or cell division, in which the chromosomes condense and can be stained and visualized. Thus, cells that can be stimulated to divide and grow in culture, such as peripheral blood lymphocytes, skin fibroblasts, and amniocytes, are typically used. Cells from bone marrow and chorionic villi are normally undergoing rapid cell division and can also be karyotyped successfully. Historically, several different staining methods have been described. However, G-banding (Giemsa staining) is the standard cytogenetic method used. This technique permits a resolution of at least 400 bands among all the chromosomes and can be adapted to allow for high-resolution analysis of up to 800 bands to analyze structural rearrangements as small as 5 to 10 million or megabase pairs (Mb). Gamete formation, either spermatogenesis or oogenesis, is accomplished by a process known as meiosis. In the first part of meiosis (meiosis I), homologous chromosomes align as pairs and cross over, exchanging genetic material, 196 also known as recombination. In this stage, reduction division, the recombined pairs separate and the typical diploid content (46 chromosomes) of the cell is reduced by half to a haploid complement of 23 chromosomes. In the next stage, meiosis II, the chromosomes separate, similar to mitosis (Nussbaum et al, 2007). The full diploid state of the cell will be restored at the time of fertilization. An imbalance of genetic material, or aneuploidy, occurs from a net loss or gain of genetic material during sperm or egg formation or, less commonly, during the initial divisions of the embryo. This missing or extra genetic material can be small pieces or parts of chromosomes or an entire chromosome itself. The classic recognizable aneuploidy syndromes involve trisomy (three copies of a given chromosome) such as those of chromosomes 13, 18, and 21, or monosomy (only a single copy) of a complete chromosome, such as X. Trisomies in particular can occur from nondisjunction, a failure of normal chromosome separation. In such cases, a pair of homologues does not separate in meiosis; one daughter cell receives both homologs of that pair, and the other cell receives none. This event can occur in either part of gamete division, meiosis I or meiosis II. Most human meiotic nondisjunction arises during oocyte formation, specifically in maternal meiosis I. Nondisjunction of meiosis I is especially pronounced in trisomies of the acrocentric chromosomes (13, 14, 15, 21, and 22) and in XXX trisomy (MacDonald et al, 1994; Zaragoza et al, 1994). The occurrence of meiotic nondisjunction increases significantly with maternal age. Therefore prenatal karyotyping from amniocentesis or chorionic villus sampling is offered to women aged 35 years and older (Hook and Cross, 1982). Nondisjunction can also occur in mitosis, with uneven division of genetic material during early embryonic cell division. This can result in two cell lines, one trisomic lineage that is potentially viable and one monosomic line. If this event occurs after the first postzygotic division, cells with a normal chromosome complement may also exist with cells containing an aneuploid complement as a mosaic chromosome constitution. Partial aneuploidy may result from several mechanisms, such as rearrangements of material between nonhomologous chromosomes that can occur in the gametes of a balanced translocation carrier. The carrier parent who has no net loss or gain of genetic material is usually phenotypically normal; however, the offspring are at increased risk for an unbalanced rearrangement and its phenotypic consequences. Attention has also been focused on deletion syndromes caused by the loss of genetic material from several chromosomes (e.g., 1p-, 4p-, 5p-), with a resulting, often recognizable phenotype. Other microdeletions have been associated with segmental duplications, or large blocks of DNA that contain chromosome-specific repetitive CHAPTER 20 Specific Chromosome Disorders in Newborns 197 FIGURE 20-1 G-banded human male karyotype. The 46 chromosomes are arranged into 23 pairs, each with a specific banding pattern. sequences (Emanuel and Saikh, 2001). It is thought that the repeats can mediate misalignment between two homologs and have been shown to be present in regions of the genome prone to instability, such as the pericentromeric regions of chromosomes 7q11, 15q11, 17q11, and 22q11, leading to the phenotypes seen in WilliamsBeuren syndrome, Prader-Willi or Angelman syndrome, Charcot-Marie-Tooth disease or hereditary neuropathy with liability to pressure palsies, and DiGeorge or velocardiofacial syndrome, respectively (Emanuel and Saitta, 2007). The use of fluorescent in situ hybridization (FISH) and chromosome-specific subtelomeric FISH and the advent of DNA microarrays has allowed the identification of smaller chromosomal rearrangements that were largely unrecognized until recently. The use of high-resolution microarrays in infants with multiple congenital anomalies has in many cases led to the identification of a specific genotype, with clinical investigations then further defining the associated phenotype (Bejjani and Shaffer, 2008). TRISOMIES DOWN SYNDROME (TRISOMY 21) Lejeune and Turpin (1960) demonstrated that trisomy of human chromosome 21 caused the constellation of findings recognized as Down syndrome (Figure 20-2). This chromosome disorder was the first to be described and is the most common viable autosomal trisomy, occurring in approximately 1 in 700 to 800 live births (Hook, 1992). The vast majority (>90%) occurs secondary to meiotic nondisjunction, and a pronounced maternal age effect is encountered. Approximately 3% to 5% of cases are caused by a translocation, that could be either de novo or passed from a balanced translocation-carrier parent, that becomes unbalanced and trisomic in the baby. Typically, the translocated chromosome 21 rearranges with another acrocentric chromosome, often chromosome 14, resulting in a Robertsonian translocation. Mitotic nondisjunction, or mosaic Down syndrome, has been demonstrated in approximately 3% of cases as well, with variable features ranging from normal to a typical Down syndrome phenotype. Clinical Features It is the more common occurrence of Down syndrome in babies of older mothers that led to the use of prenatal karyotyping for advanced maternal age (>35 years) at the time of conception, with samples typically obtained by amniocentesis after 15 weeks’ gestation, or chorionic villus sampling at 10 to 12 weeks’ gestation. Maternal serum analyte testing is used for prenatal screening purposes, with results showing low alpha-fetoprotein, low unconjugated estriol, and elevated total human chorionic gonadotropin. 198 PART V Genetics FIGURE 20-2 Newborn with Down syndrome (trisomy 21) illustrating some of the characteristic facial features, including upward-slanting ­palpebral fissures and a flat facial profile. Associated ultrasonographic findings for Down syndrome, including a cardiac defect, shortened long bones, underdeveloped fetal nasal bone, nuchal translucency or thickening, echogenic small bowel, and duodenal atresia (“double-bubble” sign), may be seen in 50% to 60% of fetuses. Most patients with Down syndrome, if not diagnosed prenatally, are usually recognized at birth because of the typical phenotypic features, which then prompt karyotype analysis. The constellation of physical findings associated with Down syndrome consists of brachycephaly, presence of a third fontanel, upward-slanted palpebral fissures, epicanthal folds, Brushfield spots in the irises, flattened nasal root, small posteriorly rotated ears with over-folded superior helices, prominent tongue, short neck with excess nuchal skin, single palmar creases, brachydactyly, fifth-finger clinodactyly, exaggerated gap between the first and second toes, open field hallucal pattern, and hypotonia (see Figure 20-2). Often the physical features conform to an easily distinguishable phenotype, but in some cases, prematurity or ethnic variations can make a clinical diagnosis less straightforward. An immediate karyotype is indicated to confirm the diagnosis and as preparation for counseling the family. Malformations involving many organ systems have been described in Down syndrome, and whether the diagnosis is known prenatally or determined in the newborn period, several clinical investigations are warranted when this diagnosis is suggested. The most common malformation is congenital heart disease (seen in approximately 50% of cases), which may require surgical intervention. Atrioventricular canal defects are often encountered (mean of 40%), although ventricular septal defects (VSDs), atrial septal defects (ASDs), tetralogy of Fallot, and patent ductus arteriosus (PDA) are all described in the disorder. An echocardiogram is indicated in all cases, and medical and surgical interventions of cardiac lesions are routine. Gastrointestinal malformations, especially duodenal atresia (2% to 5%), in addition to Hirschsprung disease and less frequently encountered conditions, such as esophageal atresias, fistulas, and webs throughout the tract, have been described. It is critical to carefully monitor the baby’s feeding and bowel function before considering discharge from the nursery. Although growth parameters can be in the range of 10% to 25% at birth, significantly decreased postnatal growth velocity is encountered in these patients. Separate growth curves have been devised for patients with Down syndrome (Fernandes et al, 2001), because growth retardation involving height, weight, and head circumference has been well documented. An initial ophthalmologic evaluation is also indicated in the first few months of life and then annually, because strabismus, cataracts, myopia, and glaucoma have been shown to be more common in children with Down syndrome. In addition, hearing loss of heterogenous origin is present in approximately half of patients, with middle ear disease contributing to this problem. Spinal cord compression caused by atlantoaxial subluxation from ligamentous laxity and subsequent neurologic sequelae can be a complication of the disorder. Screening radiographs are typically performed at approximately 3 years of age. Physicians should be vigilant in evaluating the cervical spine, especially before administration of anesthesia and before an older child’s participation in sports. Other associated disorders that merit screening are hypothyroidism in approximately 5% of patients, often with the presence of thyroid autoantibodies. Initial evaluation occurs with newborn screening programs, followed by additional thyroid-stimulating hormone and free thyroxine levels at 6 months, 12 months, and then yearly thereafter. Bone marrow dyscrasias, such as neonatal thrombocytopenia, and transient self-resolving myeloproliferative disorders, such as leukemoid reaction, have been observed in the first year of life. An elevated rate of leukemia with a relative risk of 10 to 18 times normal up to age 16 years has been described. Acute nonlymphoblastic leukemia is seen at higher rates in congenital or newborn cases, but the distribution becomes similar to that of non–Down syndrome patients after age 3 years. Survival of patients with Down syndrome is shorter after a diagnosis of acute lymphoblastic leukemia than in diploid patients (Epstein, 2001). Patients with Down syndrome demonstrate a wide range of developmental abilities, with highly variable personalities and behavioral phenotypes as well (Pueschel et al, 1991). Central hypotonia with concomitant motor delay is most pronounced in the first 3 years of life, as are language delays. Therefore immediate and intensive early intervention and developmental therapy are critical for maximizing the developmental outcome. A wide range of intelligence has been described, with conflicting data on genetic and environmental modifiers of outcome (Epstein, 2001). Seizure disorders occur in 5% to 10% of patients, often manifesting in infancy. The most common causes of death in patients with Down syndrome are related to congenital heart disease, to infection (e.g., pneumonia) that is thought to be associated with defects in T-cell maturation and function, and to malignancy (Fong and Brodeur, 1987). Once medical and surgical interventions for the correction of associated congenital malformations are complete and successful, the long-term survival rate is good. However, fewer than half of patients survive to 60 years, and fewer than 15% survive CHAPTER 20 Specific Chromosome Disorders in Newborns 199 past 68 years. Neurodegenerative disease with features of Alzheimer’s disease is encountered in most patients who are older than 40 years, although frank dementia is not typical. Men with Down syndrome are almost always infertile, whereas small numbers of affected women have reproduced (Epstein, 2001). In counseling the family of a newborn diagnosed with Down syndrome, it is important to include the organ systems affected in their baby and the severity of each malformation when defining a prognosis. Above all, the wide variability of the phenotype should be emphasized, with a care plan tailored to the needs of the individual patient. Genetic Counseling If a complete (full chromosome) or mosaic trisomy 21 is found, parental karyotypes are generally not analyzed, because the karyotypes are normal in virtually all cases. After having one child with Down syndrome, a mother’s recurrence risk for another affected child is approximately 1% higher than her age-specific risk (Hook, 1992). This fact is especially significant in younger mothers, whose age-specific risks are low. If a de novo translocation resulting in Down syndrome is found, the recurrence risk is less than 1%. If the mother is found to carry a constitutional balanced Robertsonian translocation, the risk for another translocation Down syndrome fetus is approximately 15% at the gestational age when amniocentesis is offered, and 10% at birth. However, if the father is the translocation carrier, the recurrence risk is significantly smaller, approximately 1% to 2% (Epstein, 2001). Whereas newer arraybased diagnostic techniques will identify the copy number change associated with the trisomy, structural rearrangements such as Robertsonian translocations are not readily detected. In this situation, a karyotype will provide information regarding the mechanism of the copy number change, which is needed for accurate recurrence risk counseling. EDWARDS SYNDROME (TRISOMY 18) Trisomy 18 is encountered in 1 in 6000 live births and is associated with a high rate of intrauterine demise. It is estimated that only 5% of conceptuses with trisomy 18 survive to birth and that 30% of fetuses diagnosed by second-­ trimester amniocentesis die before the end of the pregnancy (Hook, 1992). Findings on prenatal ultrasonography can raise suspicion for the disorder—growth retardation, oligohydramnios or polyhydramnios, heart defects, myelomeningocele, clenched fists, and limb anomalies. Maternal serum screening can show low values for alphafetoprotein, unconjugated estradiol, and total human chorionic gonadotropin, indicating the need for subsequent karyotype analysis and fetal ultrasonographic monitoring. Clinical Features Phenotypic features present at birth consist of intrauterine growth restriction (1500 to 2500 g at term), small narrow cranium with prominent occiput, open metopic suture, low-set posteriorly rotated ears, and micrognathia with small mouth. Characteristic clenched hands with FIGURE 20-3 Newborn with trisomy 18, showing prominent occiput, characteristic facial appearance, and clenched hands. overlapping digits, excess of arches on dermatoglyphic examination, hypoplastic nails, and “rocker-bottom” feet or prominent heels with convex soles (Figure 20-3) are also described. Additional malformations encountered in this syndrome are congenital heart disease (ASD, VSD, PDA, pulmonic stenosis, aortic coarctation), cleft palate, clubfoot deformity, renal malformations, brain anomalies, choanal atresia, eye malformations, vertebral anomalies, hypospadias, cryptorchidism, and limb defects, especially of the radial rays. The prognosis in this disorder is extremely poor, with more than 90% of babies succumbing in the first 6 months of life and only 5% alive at 1 year old. Death is caused by central apnea, infection, and congestive heart failure. The newborn period is characterized by poor feeding and growth, typically requiring tube feedings. A few patients have been described who have survived into childhood and beyond. Universal poor growth and profound mental retardation with developmental progress stopping at that of a 6-month-old infant (Baty et al, 1994) have been documented. Malignant tumors such as hepatoblastoma and Wilms’ tumor have been described in some of the survivors. In the few patients in whom cardiac surgery was performed, outcome was not shown to be improved. Genetic Counseling The typical estimate of a recurrence risk for trisomy 18 in a future pregnancy is a 1% risk over the maternal age– ­specific risk for any viable autosomal trisomy (Hook, 1992). Trisomy occurring from a structural rearrangement, such as a translocation, warrants parental karyotype analysis before the recurrence risk can be assessed. PATAU SYNDROME (TRISOMY 13) It has been estimated that approximately 2% to 3% of fetuses with trisomy 13 survive to birth, with a frequency of 1 in 12,500 to 21,000 live births (Hook, 1992). As with 200 PART V Genetics also seen. The fingers can be flexed or overlapped and can show camptodactyly. An increased frequency of nuclear projections in neutrophils, giving a drumstick appearance similar to that of Barr bodies, can also be found. This finding would be especially striking in males, in whom Barr bodies would not be expected. As with trisomy 18, prognosis for the fetus with trisomy 13 is extremely poor, with 80% mortality in the neonatal period, and less than 5% of patients surviving to 6 months old. Mental retardation is profound, and many patients are blind and deaf as well. Feeding difficulties are typical. Genetic Counseling Recurrence risk data suggest that, as with trisomy 18, the chance that a woman will have a child with any trisomy after a pregnancy affected by trisomy 13 is rare. The estimated risk is 1% higher than the maternal age–related risk for the recurrence of any viable autosomal trisomy in a subsequent pregnancy. TURNER SYNDROME (45,X) FIGURE 20-4 Stillborn with trisomy 13. The facial appearance is that of cebocephaly, which is associated with holoprosencephaly. There is an extra digit on the ulnar border of the right hand. other trisomies, amniocentesis performed for advanced maternal age or indicated by fetal ultrasonographic findings may lead to a prenatal diagnosis of trisomy 13. Clinical Features Trisomy 13–associated malformations include congenital heart disease, cleft palate, holoprosencephaly, renal anomalies, and postaxial polydactyly (Figure 20-4). In addition, microcephaly, eye anomalies, and scalp defects can suggest the diagnosis. Brain malformations such as holoprosencephaly are found in more than half the patients, with concomitant seizure disorders. Microcephaly, split sutures, and open fontanels are encountered. A scalp defect (cutis aplasia) that can sometimes be mistakenly attributed to a fetal scalp monitor is specific to the disorder, being found in 50% of cases. Eye malformations, including iris colobomas and hamartomatous cartilage “islands,” can be seen on funduscopic examination. Congenital heart disease is present in approximately 80% of patients, usually VSD, ASD, PDA, or dextrocardia. Limb anomalies, such as postaxial polydactyly, single palmar creases, and hyperconvex narrow fingernails, are In early embryogenesis, two active X chromosomes are required for normal development. Turner syndrome, a phenotype associated with loss of all or part of one copy of the X chromosome in a female conceptus, occurs in approximately 1 in 2500 female newborns. The 45,X karyotype or loss of one entire X chromosome accounts for approximately half of the cases. A variety of X chromosome anomalies—including deletions, isochromosomes, ring chromosomes, and translocations—account for the remainder of the causes. It is important to note that approximately 0.1% of fetuses with a 45,X complement survive to term; the vast majority (>99%) is spontaneously aborted. This fact underscores the requirement for both X chromosomes during embryonic development. Additional studies indicate that in approximately 80% of cases, it is the paternally derived X chromosome that is lost (Willard, 2001). Clinical Features There is wide phenotypic variability in patients with Turner syndrome. Features present at birth include short stature, webbed neck, craniofacial differences (epicanthal folds and high arched palate), hearing loss, shield chest, renal anomalies, lymphedema of the hands and feet with nail hypoplasia, and congenital heart disease. Typical cardiac defects are bicuspid aortic valve, coarctation of the aorta, valvular aortic stenosis, and mitral valve prolapse. Growth issues, especially short stature, are the predominant concern in childhood and adolescence; the mean adult height of patients with Turner syndrome is 135 to 150 cm without treatment. Growth hormone therapy, which is routinely offered starting at approximately 4 to 5 years old, can lead to an average gain of 6 cm or more in final adult height (Willard, 2001). Primary ovarian failure caused by gonadal dysplasia (streak gonads) can result in delay of secondary sexual characteristics and primary amenorrhea. Cyclic hormonal therapy is initiated at the age of puberty to aid the development of secondary sex characteristics CHAPTER 20 Specific Chromosome Disorders in Newborns 201 and menses as well as to help bone mass. Infertility, related to gonadal dysplasia, is typical and has been successfully treated with assisted reproduction techniques and donor oocytes. It is important to evaluate for structural cardiovascular defects in the patient before pregnancy. In terms of intellectual development, specific difficulties with spatial and perceptual thinking lead to a lower performance intelligence quotient; however, this syndrome is not characterized by mental retardation. TRIPLOIDY (69,XXX OR 69,XXY) As its name implies, triploidy is a karyotype containing three copies of each chromosome. Mechanisms that lead to this state include fertilization of the egg by two different sperm (dispermy) and complete failure of normal chromosome separation in maternal meiosis. The vast majority of triploid fetuses are spontaneously aborted, accounting for up to 15% of chromosomally abnormal pregnancy losses. Live births of affected fetuses are rare, and reports of survival beyond infancy are only anecdotal. Mosaicism with combinations of diploid and triploid cells (mixoploid) has also been documented. Malformations, including hydrocephalus, neural tube defects, ocular and auricular malformations, cardiac defects, and 3-4 syndactyly of the fingers, are associated findings. In addition, the placenta is often abnormal, typically large, and cystic. DELETION SYNDROMES In addition to the aneuploid conditions described previously, partial monosomy of a chromosome can lead to a recognizable pattern of malformations. Three known syndromes that are associated with the deletion or loss of genetic material from the short, or p arms of chromosomes 1, 4, and 5 are described. All these syndromes are associated with deletions that involve the loss of many genes located in a specific region. CHROMOSOME 1p DELETION SYNDROME (1p-) Monosomy for the distal short arm of chromosome 1, or deletion of 1p36, has been associated with a constellation of clinical findings. A characteristic facies consisting of frontal bossing, large anterior fontanel, flattened midface with deep set eyes, and developmental delay has been described (Figure 20-5). Orofacial clefting, hypotonia, seizures, deafness, and cardiomyopathy are also noted. This deletion syndrome is estimated to occur in approximately 1 in 10,000 live births, and it is the most frequently occurring subtelomeric deletion. Greater recognition of the phenotype, the availability of a specific FISH test, and current widespread use of DNA microarrays will likely lead to improved diagnosis of this condition. The majority of deletions arise de novo in the patient, with approximately 3% being attributable to malsegregation of a balanced parental translocation. The size of the deletion varies, from submicroscopic (<5 Mb) to large, cytogenetically visible deletions greater than 30 Mb. A correlation between the size of the deletion and the severity of clinical features is suggested. FIGURE 20-5 Child with deletion of chromosome 1p. WOLF-HIRSCHHORN SYNDROME (4p-) Distal deletions of the short arm of chromosome 4 are associated with a recognizable pattern of malformation. This syndrome is estimated to occur in 1 in 50,000 births and has features such as intrauterine growth restriction, microcephaly, midline structural defects such as cleft lip and cleft palate, cardiac septal defects, and hypospadias. The characteristic facial features are described as the Greek helmet facies, as evidenced by hypertelorism with epicanthi, a high forehead with a prominent glabella, and a beaked nose. Prominent, low-set ears are also seen. Hypotonia, failure to thrive, and developmental delay are common, with one third of infants dying in the first year of life. 202 PART V Genetics Many patients have lived well into childhood and even into adulthood, although profound growth and mental retardation is typical and often accompanied by seizures. Although most deletions are cytogenetically visible on karyotype analysis, small submicroscopic deletions have also been described. In cases in which the clinical features are suggestive but the karyotype is not revealing, further cytogenetic analysis using specific 4p telomere probes or a DNA microarray can be diagnostic. More than 80% of 4p deletions arise de novo in the patient, with minimal risk of recurrence. In the 10% to 15% of cases resulting from a translocation, analyzing parental samples is clinically indicated for appropriate recurrence risk counseling. CRI DU CHAT SYNDROME (5p-) Partial monosomy of chromosome 5p is seen in approximately 1 in 50,000 live births and is associated with a multiple congenital anomaly syndrome named for the unusual cry of the affected babies, described as similar to that of a cat, or cri du chat. The constellation of features associated with this disorder includes low birthweight, microcephaly, round face, hypertelorism or telecanthus, downward-slanting palpebral fissures, epicanthi, and broad nasal bridge. Hypotonia and cardiac defects are also seen, including ASD, VSD, and tetralogy of Fallot. Early issues include failure to thrive and pronounced developmental delay. The unusual cry usually resolves during infancy, and survival into adulthood is possible but is typically associated with severe mental retardation. Intensive therapy appears to provide some benefit, and more sensitive measures of cognition demonstrate clearly better receptive language skills than expressive language ability. Therefore children may understand more complex verbal language than their expressive skills might demonstrate (Cornish et al, 1999). It is estimated that almost 100 genes are lost when the putative critical region from 5p15.2 to p15.33 is deleted (Shaffer et al, 2001; Zhang et al, 2005). Close to 90% of 5p deletions arise de novo in the affected child, incurring a minimal risk of recurrence (<1%). The remainder arise from malsegregation of a balanced translocation in a carrier parent, which would be associated with a 10% to 15% risk of recurrence of an unbalanced karyotype in a future live born infant. Parental chromosome analysis is indicated for proper recurrence risk counseling. DNA MICROARRAYS Recently, emphasis has been placed on characterizing variations of the genome that fall in the range between the single nucleotide and visible chromosomal changes—­ submicroscopic structural variants (less than 5 Mb) that cannot be seen by standard karyotype. An emerging technology rapidly gaining acceptance in various clinical settings involves the use of DNA microarrays (Bejjani and Shaffer, 2008; Miller et al, 2010). In general, microarraybased methods are focused on detecting copy number changes, and rearrangements such as balanced translocations or inversions are not detectable using this methodology. Microarray testing can be performed in either a targeted or genome-wide fashion. Two types of DNA microarray platforms are used currently: array comparative genomic hybridization, which can utilize bacterial artificial chromosomes containing large DNA segments as probes, or small oligonucleotides as DNA probes. Other approaches use single nucleotide polymorphism arrays that include probes based on known polymorphisms present in the human genome. Single nucleotide polymorphism– based arrays or dense oligonucleotide arrays have the advantage of detecting gains or losses of shorter stretches of the genome, because the probes for a given region can be densely arrayed, and the sensitivity for detecting alterations of that region is greatly enhanced. With a single test, microarrays can detect genomic errors associated with disorders that are usually identified by cytogenetic analysis and multiple FISH studies. Whereas microarray analysis is proving robust and providing an exceptional level of resolution from a diagnostic perspective, the major difficulty with the current interpretation of the results lies in assigning causality and clinical significance to the multiple alterations that are detected in each individual. Toward this end, the availability of databases with information on normal variation in multiple ethnic populations and testing of unaffected parents remain standard approaches to discerning whether a copy number change is likely to cause disease. SEGMENTAL DUPLICATIONS AND MICRODELETION SYNDROMES A greater appreciation of the complexity of the human genome and its structure has now been afforded with the completion of the human genome sequence. This work has focused attention on regions of the genome that are prone to rearrangement. The presence of such unstable regions appears to have a significant role in the etiology of several genetic disorders (Emanuel and Shaikh, 2001; Emanuel and Saitta, 2007). These disorders result from inappropriate dosage of crucial genes in a given genomic segment via either structural mechanisms (deletion or duplication) or functional mechanisms (imprinting or uniparental disomy). It has also been demonstrated that many of these regions of genomic instability have a common element: the presence of large, chromosome-specific low copy repeats that most likely mediate misalignment and unequal crossover during recombination, leading to rearrangements such as a deletion and duplication (Figure 20-6). Many of these large repeat structures are localized to a single chromosome or within a single chromosomal band. Examples of such genomic disorders are hemophilia A (inversion of Xq28), Sotos syndrome (in which a number of patients have a deletion of 5q35), Smith-Magenis syndrome (deletion of 17p11.2), Charcot-Marie-Tooth disease (interstitial duplication on 17p12), and the reciprocal deletion of this same region of 17p12, leading to hereditary neuropathy with liability to pressure palsies and a small percentage of patients with neurofibromatosis type I (deletion involving 17q11.2). In this section, we focus on several deletion syndromes that occur on chromosomes whose underlying genomic structure contains segmental duplications such as ­Williams-Beuren syndrome (involving chromosome 7q11.2), Prader-Willi syndrome or Angelman syndrome (involving an imprinted region of chromosome 15q11 CHAPTER 20 Specific Chromosome Disorders in Newborns Normal Recombination Event A B C D A B C D Crossover at LCR-D A B C D A B C D A Misalignment followed by Recombination A B C D A B C D Crossover between A & D A B A B C DA B C D AD Duplication on A Deletion on B (a case of 22q11 Deletion Syndrome) B FIGURE 20-6 A, Alignment of low copy repeats (LCRs) before exchange. B, Misalignment of LCRs before exchange can result in rearrangement. through 15q13), and DiGeorge or velocardiofacial syndrome (DGS/VCFS), the most commonly occurring microdeletion syndrome in humans, involving chromosome 22q11.2. It is important to note that in several of these microdeletion syndromes, there is the possibility that a reciprocal duplication of the exact same region may also occur (see Figure 20-6). Typically, the duplication syndromes cause fewer abnormalities and have wide phenotypic variability, but they are often characterized by developmental delays or behavioral abnormalities. In addition, there are newly recognized microdeletion syndromes (e.g., 1q21.1, 3q29, 15q13.3, 16p11.2, and 17q21) that are rapidly being identified because of the increased use of DNA microarrays in the clinical and research settings. These regions are similarly flanked by segmental duplications, likely predisposing to their rearrangements. Many of the patients with these newly recognized rearrangements are not clinically diagnosed in the newborn period, because they may not have significant anatomic malformations or dysmorphic features that prompt a genetic evaluation; however, they can present during childhood with variable clinical findings. Typically there is developmental delay, often autistic spectrum disorder, or an intellectual disability. These syndromes are complicated because they have been reported in unaffected family members, which proves to be a challenge when ascribing causality to the rearrangement. 203 When an abnormality is identified by microarray analysis, the genes located in the affected region can be identified. It is then possible to decide what role, if any, the affected genes have in the observed phenotype. For example, deleted genes could be identified that may predispose a patient to cancer because of a germline loss of one copy of a tumor suppressor gene (Adams et al, 2009), resulting in careful surveillance for tumor formation. Therefore a new genetic paradigm is developing where patients are initially genotyped (genotype-first), followed by evaluation of the altered genes in a given region to determine their possible effect on the future phenotype. WILLIAMS-BEUREN SYNDROME (7Q11.2 DELETION) The estimated incidence of Williams-Beuren syndrome is 10,000 live births. The phenotype has a variable spectrum, but usually consists of distinctive facies, growth and developmental retardation, cardiovascular anomalies, and occasionally infantile hypercalcemia (Figure 20-7). Babies with Williams-Beuren syndrome usually show some degree of intrauterine growth restriction with mild microcephaly. Facial features include epicanthal folds with periorbital fullness of subcutaneous tissues, flat midface, anteverted nostrils, long philtrum, thick lips, large open mouth, and stellate irises that may not be discernible at birth. Most infants have a cardiovascular abnormality; supravalvular aortic stenosis (SVAS) is the most commonly associated defect, seen in more than 50% of cases. Pulmonary artery stenosis is also often encountered. It is interesting to note that isolated SVAS can also exist as a separate autosomal dominant trait and has been shown to occur from mutations of the elastin gene that is located within the deletion region on 7q11.2. Patients with ­Williams-Beuren syndrome are typically missing one copy of the elastin gene. Hypercalcemia, which is manifested in approximately 10% of patients with this disorder, is severe and persists through infancy. Umbilical and inguinal hernias are also associated features. Issues in infancy include feeding and growth problems, with pronounced irritability and colicky behavior. Hoarse voice, strabismus, hypertension, and joint mobility restrictions may develop later in childhood. In terms of development, the typical mild to moderate mental retardation can be masked by advanced language skills, although gross motor and visual-motor integration skills are especially affected. Attention-deficit disorders are common, and a characteristic outgoing personality is often described in affected children. Many of the classic features of Williams-Beuren syndrome are not clearly discernible in the newborn period, but the diagnosis should be suggested in any child with SVAS, hypercalcemia, and facial features consistent with the disorder. The diagnosis can be confirmed quickly by performing a DNA microarray or FISH using probes specific for the deleted region of chromosome 7q11.2. Because the condition is typically sporadic and most deletions arise de novo, the risk of recurrence in subsequent pregnancies is minimal. An affected adult, however, would pass on the condition in an autosomal dominant manner, with a 50% risk of the disorder in his or her child. 204 PART V Genetics A B C D FIGURE 20-7 Williams syndrome. A, Neonate with a coarse face, periorbital fullness, wide mouth, and thick lips with decreased Cupid’s bow. B, Neonate profile showing periorbital fullness, flat nasal bridge with full tip, and prominent cheeks. C, This infant has periorbital fullness, flat nasal bridge, thick lips with decreased Cupid’s bow, pouty lower lip, and low-set, full cheeks. D, Infant profile showing dolichocephaly (increased anteroposterior diameter of head), a higher nasal bridge than in the neonate, full nasal tip, pouty lower lip, long neck, sloping shoulders, and part of pectus excavatum. 22q11.2 DELETION SYNDROME A deletion of chromosome 22q11.2 has been identified in the majority of patients with DiGeorge syndrome, VCFS, and conotruncal anomaly face syndrome, leading to the realization that these clinical entities all reflect features of the same genomic disorder (McDonald-McGinn et al, 1997). The list of findings associated with 22q11.2 deletion is extensive and varies by patient. Estimates indicate that 22q11 deletion occurs in approximately 1 in 3000 live births (Burn and Goodship, 1996). This disorder is the most common microdeletion syndrome occurring in humans and is a significant health concern in the general population. The phenotype is characterized by a conotruncal cardiac anomaly and often aplasia or hypoplasia of the thymus and parathyroid glands. The majority of patients with a deletion can receive a diagnosis as newborns or infants with significant cardiovascular malformations, including interrupted aortic arch type B, truncus arteriosus, or tetralogy of Fallot, along with functional T-cell abnormalities and hypocalcemia. In addition, facial dysmorphia may be present (Figure 20-8), including hooded eyelids, hypertelorism, overfolded ears, bulbous nasal tip, a small mouth, and micrognathia. Since the initial report by DiGeorge in 1968, the spectrum of associated clinical features has been expanded to include anomalies such as palatal anomalies, CHAPTER 20 Specific Chromosome Disorders in Newborns A 205 B FIGURE 20-8 Facial differences associated with 22q deletion. vascular rings, feeding and swallowing problems, gastroesophageal reflux, renal agenesis, and hypospadias (McDonald-McGinn et al, 1997). Before advances in the medical and surgical management of children with complex congenital cardiac disease and immune deficiencies, this disorder was associated with significant morbidity and mortality. Developmental delays or learning disabilities have been reported in most patients with 22q11.2 deletion, and a wide range of developmental and behavioral findings have been observed in young children (Emanuel et al, 2001). In the preschool years, children with a 22q11.2 deletion were most commonly found to be hypotonic and developmentally delayed with language and speech difficulties. Severe or profound retardation was not seen, and one third of patients functioned within the average range. The vast majority of patients (80% to 90%) have the same large deletion, approximately 2.4 to 3 Mb, that is detected by FISH (Figure 20-9) or DNA microarray. The deletion remains unchanged when inherited from an affected parent. However, the phenotype can be widely variable, even within a family. Although smaller recurrent deletions that are half the size of the common deletion occur (1.5 Mb), a smaller size does not indicate milder symptoms, making genotype-phenotype correlations difficult. To date, no explanation for the great phenotypic variability has been forthcoming, and it is currently an area of active investigation (Emanuel and Shaikh, 2001). Most 22q11 deletions occur as de novo events, with less than 10% of them being inherited from an affected parent. The prevalence of these de novo 22q11.2 deletions indicates an extremely high mutation or rearrangement rate within this genomic region that is probably related to the presence of recombination-permissive duplicated DNA sequences, segmental duplications, or low copy chromosome 22–specific repeats in 22q11 FIGURE 20-9 (See also Color Plate 4.) Fluorescence in situ hybridization study of a 22q deletion. (Courtesy Beverly S. Emanuel.) (Emanuel and Shaikh, 2001; Emanuel and Saitta, 2007; Shaikh et al, 2007). NEW MICRODELETION AND MICRODUPLICATION SYNDROMES As mentioned previously, several newly identified genetic syndromes have been described with the increased use of DNA microarrays (Slavotinek, 2008). These regions are frequently flanked by segmental duplications, which is the likely reason for their prevalence in diverse patient populations, as well as the fact that there are reciprocal 206 PART V Genetics rearrangements (deletion or duplication). Almost all of the patients with these deletions or duplications were not initially recognized based on their clinical features, but were instead ascertained by microarray analysis. Deletions of 1q21.1 of 1.35 Mb have been seen in patients with variable presentations including developmental and behavioral abnormalities, mild facial dysmorphia, and microcephaly (Brunetti-Pierri et al, 2008; Mefford et al, 2008). The 3q29 microdeletion syndrome is approximately 1.6 Mb and was initially discovered (­Willatt et al, 2005) in patients with mild-to-moderate mental retardation, microcephaly, and nonspecific facial dysmorphia. Duplications of this region have also been described in patients with developmental delays, mental retardation, and microcephaly (Lisi et al, 2008). The 15q13.3 region is distal to the more commonly known 15q11-q13 locus associated with Prader-Willi and Angelman syndromes. Patients with the 1.5-Mb deletion of 15q13.3 have variable phenotypes, but typically manifest with seizures or abnormal electroencephalograms (Sharp et al, 2008). A smaller 680-kb deletion has also been described in a range of patients with neurobehavioral phenotypes (Shinawi et al, 2009). Interestingly, this deletion is often maternally inherited, suggesting that there might be an imprinting mechanism involved. A recurrent 550-kb deletion of 16p11.2 has been discovered in approximately 1% of patients with autism (Weiss et al, 2008), and the reciprocal duplication associated with attention deficithyperactivity disorder and schizophrenia (McCarthy et al, 2009; Shinawi et al, 2010). 17q21.31 deletions involving approximately 500 to 650 kb were also found in patients with developmental delays, learning disabilities, and variable facial dysmorphia (Koolen et al, 2008; Shaw-Smith et al, 2006). DISORDERS OF IMPRINTED CHROMOSOMES A growing recognition of mechanisms regulating gene expression has emerged in the last decade. Two of the most exciting concepts with important clinical correlates are imprinting and uniparental disomy. The term genomic imprinting implies that a whole region of a chromosome or a group of genes in a given region are subject to a difference in their expression that depends on whether they reside on the maternally inherited or paternally inherited chromosome. In these cases, a genetic disorder might manifest according to the parent from which the genomic region was inherited. The genes in an imprinted region are not necessarily mutated, but they are marked such that the cell can distinguish between the maternal and paternal copies and coordinate expression on the basis of that distinction. At the molecular level, it appears that differences in the methylation of the DNA, and its replication and regulation at the transcriptional level, appear to be involved in this mechanism. This mechanism has become an area of important research advancements, and there are now more than 100 known or predicted genes and chromosomal regions thought to be important in human disease (www.geneimprint.com/site/genes-by-species). PRADER-WILLI SYNDROME It has been demonstrated that occasionally, instead of inheriting one copy of each chromosome from each parent, both copies of a given chromosome can come from the same parent. This phenomenon, known as uniparental disomy (UPD), is associated with advanced maternal age. It becomes a significant issue when the involved chromosome is imprinted or has regions on it that are imprinted. Prader-Willi syndrome (PWS) involves the loss of activity from the paternally derived proximal long arm of chromosome 15 (15q11). This loss can occur through deletion or disruption of this region or through maternal uniparental disomy such that no paternal chromosome 15 is present (Nicholls et al, 1989). Newborns have pronounced central hypotonia, hyporeflexia, and a weak cry. The poor tone manifests as sucking and swallowing difficulties that can lead to failure to thrive and the need for feeding tubes in infancy. Facial differences that have been described include bifrontal narrowing, almond-shaped eyes, and a small, downturned mouth. Genitalia are often hypoplastic, with cryptorchidism being common in boys with this syndrome. The commonly reported small hands and feet are not usually demonstrated in the newborn. Strabismus and hypopigmentation relative to the family are also common. A history of poor fetal activity during the pregnancy can often be elicited, especially if the mother has had prior pregnancies. Consistent with the hypotonia, breech presentation and perinatal insults are found more frequently than usual. The extreme hypotonia begins to improve in the first year of life, as does motor development, although developmental delay is the rule, especially for gross motor skills and speech. The feeding improves in the first few years of life and gives way to often uncontrollable hyperphagia and obesity. This issue and other behavioral problems, including severe temper tantrums and obsessive-compulsive disorder, are encountered throughout life. The majority of patients manifests mild to moderate mental retardation. Early diagnosis and preemptive implementation of behavioral therapy are essential components of the optimal management of these issues. Deletions of the region critical in Prader-Willi syndrome have been demonstrated in up to 70% of patients. The deletion can be detected by DNA microarray or FISH analysis using a probe specific for this region of chromosome 15. A small number of patients have a disruption of this area as the result of a chromosomal translocation. To date, no single gene in this region has been implicated as the cause, but expression of the genes in this region is under intense investigation. It has been noted, however, that patients who have Prader-Willi syndrome as a result of a deletion of the region are more likely to be hypopigmented. This feature has been attributed to deletion of a gene involved in pigmentation, the so-called P gene (­Rinchik et al, 1993). Recurrence risks are negligible in cases in which de novo deletions are found and sporadic occurrence is usually encountered. Approximately 20% to 25% of patients with PraderWilli syndrome show maternal UPD that can be detected by means of a molecular assay designed to assess specific CHAPTER 20 Specific Chromosome Disorders in Newborns 207 FIGURE 20-10 Macrosomic infant with macroglossia and lax abdominal musculature. These findings are typical of Beckwith-Wiedemann syndrome. (From Viljoen DL, Jaquire Z, Woods DL: Prenatal diagnosis in autosomal dominant Beckwith-Wiedemann syndrome, Prenat Diagn 11:167-175, 1991.) methylation differences between maternal and paternal alleles. Methylation analysis is abnormal in more than 99% of affected individuals, but will not determine whether the cause is attributed to a deletion or maternal UPD. Further study is required if an abnormal result is obtained. A maternal age effect has been demonstrated in UPD cases, and recurrence risks in families without deletions are estimated at 1 in 1000. In addition, this region of the genome is subject to regulation by imprinting. Large chromosome-specific segmental duplications are found in 15q11 and have been implicated in mediating the recurrent deletion of this genomic region (Emanuel and Shaikh, 2001; Emanuel and Saitta, 2007). ANGELMAN SYNDROME Loss of genetic material from the 15q11 region from the maternal copy of chromosome 15 is associated with Angelman syndrome. Clinical features are not evident in the newborn period and infancy, but include significant mental retardation, seizures, ataxic gait, tongue thrusting, inappropriate bursts of laughter, and facial differences including protruding jaw, wide mouth, thin upper lip, and widely spaced teeth. The mental retardation and hypopigmentation overlap with the features of PraderWilli syndrome, but Angelman syndrome is a distinct entity. Seventy percent to 75% of patients have a deletion of 15q11 that is detectable by FISH analysis. A small percentage (3% to 5%) have evidence of paternal isodisomy (two paternal copies) of the entire chromosome 15, with no apparent maternal chromosome. Unlike in PraderWilli syndrome, Angelman syndrome has been associated with mutations in a single gene, UBE3A—an enzyme involved in the ubiquitin pathway of protein degradation, detected in up to 10% of patients. In addition, mutations of an imprinting center locus on chromosome 15 are thought to be associated with 1% to 2% of Angelman phenotypes (Shaffer et al, 2001). Methylation analysis can be performed and will detect abnormalities in approximately 75% to 80% of patients, because of a deletion or UPD. If methylation analysis is normal but Angelman syndrome is still suspected, UBE3A sequence analysis should be considered. The vast majority of cases result from a sporadic event, and the risk of recurrence can be best evaluated once the genetic mechanism has been determined for a given patient. BECKWITH-WIEDEMANN SYNDROME Beckwith-Wiedemann syndrome affects approximately 1 in 14,000 newborns and manifests as an overgrowth syndrome in the neonatal period. The characteristic findings are macrosomia, abdominal wall defect, and macroglossia (Figure 20-10). Affected babies are large for gestational age with proportionate length and weight. Infants of mothers with diabetes also manifest with macrosomia, but are more likely to have a weight disproportionately greater than length. Advanced bone age is also noted in Beckwith-Wiedemann syndrome. Hemihypertrophy caused by asymmetric growth is common, as is visceromegaly of various organs, including the spleen, kidneys, liver, pancreas, and adrenal glands. Other characteristic features of the syndrome are macroglossia, linear creases of the earlobe with indentations on the posterior helix, and severe hypoglycemia. Although the hypoglycemia responds quickly to therapy, it can be present for several months; therefore recognition of the condition and immediate therapeutic intervention are critical in these cases. The hypoglycemia resolves spontaneously with age, and the physical diagnostic features also become less prominent with age, making the diagnosis more difficult to ascertain. Equally important is the establishment of routine ultrasonographic surveillance at regular intervals, because children with Beckwith-Wiedemann syndrome are at increased risk of malignant tumors, especially Wilms’ tumor. The estimated risk is as high as 8% for patients with hemihypertrophy. Many centers currently perform ultrasonography at 3-month intervals until the school-age years (approximately 8 years old). Monitoring of serum alphafetoprotein levels at the same intervals until age 4 years has proved valuable, as several cases of hepatoblastoma have also been reported and detected with this adjunct study. Although most cases of Beckwith-Wiedemann appear to arise de novo, up to 15% may be familial. In familial cases, 208 PART V Genetics the transmission is autosomal dominant, because of mutations of the CDKN1C gene in 40% of familial cases, but only 5% to 10% of de novo cases. In addition, this region of the genome (11p15.5) appears to be imprinted such that the maternal allele is not usually expressed. The insulinlike growth factor type 2 (IGF2) gene is located in this region and encodes an important factor involved in fetal growth. Mutations causing overexpression of the paternal allele or underexpression of the maternal allele can result in an imbalance of expression leading to the overgrowth and tumor formation encountered in these patients. Paternal UPD has proved to be a mechanism involved in 10% to 20% of sporadic cases of Beckwith-Wiedemann syndrome. A method for detecting methylation abnormalities at two distinct genetic loci within 11p15 accounts for approximately 60% of patients and is related to the overexpression of the IGF2 gene. Therefore all the available testing methods combined can detect the cause in approximately 85% of the patients with Beckwith-Weidemann syndrome. The recurrence risk for future affected siblings or offspring of the proband depends on the specific genetic abnormality causing the disorder and can range from low (UPD, methylation abnormality) to as high as 50% (CDKN1C mutation). RUSSELL-SILVER SYNDROME Russell-Silver syndrome presents in neonates with intrauterine growth retardation followed by postnatal growth deficiency. The head size is usually normal, causing a relative macrocephaly that may have the appearance of hydrocephalus. Facial features can include a broad and prominent forehead, triangular-shaped face with a small chin, and downturned corners of the mouth. The fingers can show brachydactyly, camptodactyly, or more commonly fifth-finger clinodactyly. Other concerns involve limb-length discrepancy, delayed bone age, café au lait macules, hypospadias in males, developmental delays, diaphoresis, and hypoglycemia during the first 3 years of life. These patients are often examined by a geneticist as a toddler with growth retardation, proportionate short stature, and normal head circumference. When evaluating patients with growth retardation, it becomes important to know the prenatal and postnatal growth parameters, because they might provide a clue to the diagnosis of Russell-Silver syndrome. The molecular mechanisms underlying the pathogenesis show that Russell-Silver syndrome is likely caused by abnormalities of imprinted genes. Maternal UPD of chromosome 7 is present in 7% to 10% of patients, and the symptoms are likely caused by overexpression of the maternal GRB10 gene that suppresses activity of various growth factor receptors. In approximately 35% of the patients, imprinting abnormalities of 11p15.5 occur because of a loss of the paternally expressed IGF2 gene, leading to decreased prenatal and postnatal growth. This finding contrasts with some patients with BeckwithWiedemann syndrome, in whom IGF2 is overexpressed causing increased growth. To that end, patients with Russell-Silver syndrome do not have a significantly increased risk for neoplasia compared to patients with BeckwithWiedemann syndrome, and routine cancer surveillance protocols are typically not recommended. SUMMARY This chapter has summarized the rapidly expanding field of chromosomal and genomic disorders, concentrating on those that commonly manifest in the newborn. The widespread development and clinical implementation of molecular cytogenetic techniques have allowed the identification of subtle rearrangements that were previously undetectable. These advances now enable the discernment of new syndromes in which the chromosomal anomaly may be defined before a characteristic phenotype is recognized. In addition, a greater understanding of the role of segmental duplications and their effects on human genetic disorders as well as of the influence of mechanisms that regulate gene expression, such as imprinting, is emerging. The tremendous advances in genomics led by the completion of the Human Genome Project and the development of new molecular diagnostic tools present new challenges for clinicians to better diagnose, understand, and care for patients with genetic disorders and their families. SUGGESTED READINGS GeneReviews. Available on line at http://www.ncbi.nlm.nih.gov/sites/GeneTests/? db=GeneTests Jones KL: Smith’s Recognizable Patterns of Human Malformation, ed 6, Philadelphia, PA, 2006, WB Saunders. Online Mendelian Inheritance in Man (OMIM): Available on line at http://www. ncbi.nlm.nih.gov/omim/ Rimoin DL, Connor JM, Pyeritz RE, Korf BR, editors: Emery and Rimoin’s Principles and Practice of Medical Genetics, ed 5, New York, NY, 2006, Churchill Livingstone. Schinzel A: Catalogue of Unbalanced Chromosome Aberrations in Man, Berlin, 2001, Walter de Gruyter. Shaffer LG, Slovak ML, Campbell LJ, editors: An International System for Human Cytogenetic Nomenclature (ISCN) 2009: Recommendations of the International Standing Committee on Human Cytogenetic Nomenclature, Basel, 2009, S Karger. Slavotinek AM: Novel microdeletion syndromes detected by chromosome microarrays, Hum Genet 124:1-17, 2008. Complete references used in this text can be found online at www.expertconsult.com P A R T V I Metabolic and Endocrine Disorders of the Newborn C H A P T E R 21 Introduction to Metabolic and Biochemical Genetic Disease Stephen Cederbaum Inborn errors of metabolism or biochemical genetic disorders are one type of genetic disease that may be encountered in the neonatal period as an acute or more indolent illness. In these disorders, a mutation in a gene leads to an absent or defective gene product or enzyme and results in the accumulation of the precursor of the enzyme or a byproduct of it, a shortage of the product of the enzymatic reaction, or a combination of both. In reality the effects of many inborn errors on normal physiology are much more profound, causing many changes in gene expression and normal biochemical function. One example is a case of propionic acidemia in which interference with the mitochondrial respiratory apparatus is rarely measured except by the ascertainment of an elevated blood lactate level, but which may be much more frequent and represent only the most obvious and easily measured alteration. Inborn errors of metabolism may be inherited by any genetic mechanism—autosomal dominant, autosomal recessive, sex-linked recessive, or through a mutation in the independently inherited mitochondrial genome (mtDNA), which leads to a circumstance in which the mother alone passes the abnormal DNA to all of her children, but the affected or carrier father to none of his offspring. Most inborn errors are inherited as autosomal recessive conditions, with the carrier parents rarely expressing any obvious metabolic phenotype. A small minority is inherited in a sex-linked recessive or codominant manner, and they will be discussed in the context of their particular disease. Examples would be ornithine transcarbamylase deficiency (a urea cycle disorder) and Fabry disease (a lysosomal storage disorder), the latter not appearing in the neonatal period. When viewed from the perspective of disease mechanism, most genetic disorders, whether single gene or involving imbalance of chromosomal materials, could be considered to be inborn errors of metabolism. One or more changes in the DNA result in either altered gene expression or expression of a mutated gene, which then lead secondary to an altered product of the reaction or reactions. We will not use this expansive and grandiose interpretation of inborn errors, but rather confine ourselves to the more traditional definition described in the first paragraph. Thus disorders such as cystic fibrosis and spinal muscle atrophy, considered inborn errors that require broader interpretation, will not be considered. The advent of newborn screening in the early part of the 1960s for phenylketonuria (PKU) established a new paradigm for approaching inborn errors of metabolism and making the diagnosis and treatment prior to the symptomatic presentation, and hence preventing rather than treating the condition. This approach has proved to be remarkably successful for PKU and congenital hypothyroidism, with few patients becoming mentally retarded. In subsequent years the menu of tests expanded gradually, but it has greatly expanded in the last decade with the implementation of expanded newborn screening using tandem mass spectrometry (MS/MS) technology, which allows for ascertainment of a constellation of disorders and should alter the probability of, and the diagnostic testing for, inborn errors when incorporated into the diagnostic algorithms of the ill newborn. Moreover, the advancing technology permits some of the newborn dried blood spot to be used for a wider palette of tests, some of which may already be available. The consequences of this expansion have a downside for the neonatologist and the neonatal intensive care unit staff members. For a variety of reasons the sick and premature newborn, usually receiving intravenous alimentation and having immature or damaged organs, is far more likely to have a false-positive newborn screening test and require follow-up testing. It is important to recognize that a number of traditionally tested diseases are outside of this class of disorders, such as hypothyroidism, congenital adrenal hyperplasia, cystic fibrosis, and hemoglobinopathies. When considering inborn errors of metabolism, it is important to consider the molecular basis of mutation. Large deletions of a gene are certain to eliminate enzymatic function and any residual gene product. Smaller deletions, especially if they remove one or more of the in-phase coding triplets, may permit a stable protein to be made and to function to some extent. Most small deletions, however, cause the synthesis of unstable and out-of-phase proteins that do not function and have a short half-life within the cell. Some single-based change mutations can introduce a stop codon, causing the synthesis of the polypeptide to halt abruptly and leave a nonfunctional enzyme. Singlebase changes introducing a new amino acid vary in their effects from complete loss of activity to a lesser impact, and finally to having no effect whatsoever. When considering that the mutation test is then all modulated through the unique genetic background of the individual, there is no single final phenotype, severity, or time of onset for any genetic disorder. This variation must be considered when any diagnosis, genetic or not genetic, is considered. 209 210 PART VI Metabolic and Endocrine Disorders of the Newborn CLASSIFICATION OF INBORN ERRORS OF METABOLISM Each professional uses classification systems to permit effective reasoning as to possible causes of a symptom complex. A system for understanding inborn errors of metabolism is shown in Box 21-1. Each group has common characteristics, modes of presentation, types of molecules involved, and tests that would be applied. Because the demarcation between the groups is not sharp, other systems can see them differently. This discussion is restricted to those disorders that may be symptomatic in the newborn period or in early infancy, whereas many severe disorders would be unlikely to be associated with neonatal disease and will be given less emphasis. The disorders more commonly seen in the neonatal period are listed in Box 21-2. The first group consists of newborns with progressive lethargy, poor suck, neurologic deterioration, and often death. They have inborn errors of amino acids, the urea cycle, organic acids, or sugar metabolism. This group of BOX 21-1 C  lassification of Inborn Errors of Metabolism, 2007 Small molecule disorders ll Amino acids ll Organic acids ll Sugars Lysosomal storage disorders ll Mucopolysaccharides ll Sphingolipids ll glycolipids Energy metabolism disorders ll Oxidation disorders ll Fatty acid mobilization and metabolism disorders ll Glycogen storage diseases Peroxisomal and membrane biogenesis disorders Carbohydrate-deficient glycoprotein disorders Cholesterol biosynthetic disorders Disorders of biogenic amines, folate, and pyridoxine Transport disorders Purine and pyrimidine metabolism disorders Receptor disorders patients is the product of normal pregnancies and deliveries and becomes ill after 36 hours of life, when the maternal circulation no longer cleanses the accumulating small molecules from the fetal or infant blood and the offending metabolites accumulate in intoxicating amounts (Box 21-3). Examples include maple syrup urine disease, methylmalonic and propionic acidemias, galactosemia, and ornithine transcarbamylase deficiency. The general characteristics are given in Box 21-4; they are the disorders for which expanded newborn screening may lead to earlier detection and a more rapid diagnosis. These patients’ condition is most likely to resemble sepsis, and they should be treated with antibiotics. Most disorders manifesting acutely in the newborn period will be detected by the newborn screen, with only some urea cycle disorders and lactic acidoses likely to be missed by this screening panel. When diagnosed, these conditions are BOX 21-3 M  etabolic Diseases With Newborn Coma Secondary to Toxic Metabolite Accumulation or Mitochondrial Failure ll ll ll ll ll ll ll ll ll ll ll ll ll ll BOX 21-4 C  haracteristics of Small Molecule Disorders High levels of metabolites in body fluids Normal physical phenotype ll Neonatal presentation ll Periods of stability and instability ll Considered to be intoxication disorders ll Can often be treated by external manipulation Lysosomal storage disorders (format) ll Usually born normally ll Course is progressive, relentless, and indolent ll Deposition of material seen clinically and microscopically ll May be deforming ll Cannot be addressed exogenously Disorders of energy metabolism ll Mixed presentation between the first two categories ll Can be catastrophic at presentation ll May be present at birth or develop later ll Usually progressive ll May cause malformations ll May have episodes of deterioration ll Usually not treatable by exogenous means ll May be tissue specific or preferential ll ll BOX 21-2 C  ommon Types of Inborn Errors of Metabolism With Newborn Presentation ll ll ll ll ll ll ll ll ll ll ll ll Amino acid disorders Organic acid disorders Disorders of ammonia metabolism Disorders of carbohydrate metabolism Disorders of gluconeogenesis or hypoglycemia Disorders of fatty acid oxidation Primary lactic acidoses (respiratory chain defects) Disorders of vitamin or metal metabolism Storage diseases (infrequently) Peroxisomal disorders Disorders of sterol metabolism Congenital defects in glycosylation Galactosemia Inborn errors of ammonia metabolism Maple syrup urine disease Nonketotic hyperglycinemia Methylmalonic acidemia with or without homocystinuria Propionic acidemia Isovaleric acidemia Multiple carboxylase deficiency Glutaric aciduria type 2 Fatty acid oxidation defects Primary lactic acidosis Pyruvate dehydrogenase deficiency Pyruvate carboxylase deficiency Mitochondrial respiratory chain or electron transport chain defects CHAPTER 21 Introduction to Metabolic and Biochemical Genetic Disease treated with dialysis, limitation of protein intake except for galactosemia, fluid, and caloric support, and some specific interventions. The association of identifying physical and laboratory characteristics and various disorders is listed in Tables 21-1 and 21-2. The individual disorders are discussed in Chapters 22 and 23. When a diagnosis of a metabolic disorder appears likely, plasma amino acids, urine organic acids, plasma acylcarnitine, and plasma carnitine tests should be repeated, and ammonia and lactate should be determined. The second major category of inborn errors is the lysosomal storage diseases. This group of disorders results from defective function of a catabolic hydrolase located in the lysosome that is generally responsible for breaking down complex glycosaminoglycans and sphingolipids that are products of normal cellular turnover (see Box 21-4). Unlike the small molecule disorders in which the metabolites are found freely circulating in the body fluid compartments, these compounds 211 accumulate intracellularly, are not removed by the maternal circulation, and are present in limited amounts in the body fluids. They most often cause no apparent symptoms in the newborn period or early infancy, because the pathologic metabolites accumulate slowly with time. Exceptions to this finding are the severe form of α-glucosidase deficiency or Pompe disease, the neonatal form of α-galactosidase deficiency, or Krabbe disease and galactosialidosis. These findings are discussed in Chapter 23, and some are listed in Table 21-1. The disorders of mucopolysaccharides and glycolipids lead to the characteristic features pejoratively and inappropriately referred to as gargoylism, which consist of an exaggerated eyebrow, coarse-appearing facies, thick skin, hirsutism, and multiple abnormalities of the bones and joints seen on a radiograph. Attention to the disorder is often drawn by the hepatosplenomegaly. The metabolites are synthesized in the body and are not influenced by dietary intake. TABLE 21-1 Unique or Characteristic Physical Findings in Inborn Errors* (Major Examples) Finding Error Finding Error Hepatomegaly Galactosemia Glycogen storage diseases Gluconeogenic defects Disorders of fatty acid oxidation and ­transport Mitochondrial respiratory or electron ­transport chain defects Hereditary tyrosinemia type 1 Urea cycle defects Peroxisomal defects Niemann-Pick disease type C Congenital defects in glycosylation Retinitis pigmentosa Mitochondrial respiratory or electron transport chain defects Sjögren-Larsson syndrome Peroxisomal disorders Abetalipoproteinemia Optic atrophy or hypoplasia Pyruvate dehydrogenase complex deficiency Mitochondrial disorders Leigh disease Peroxisomal disorders Corneal clouding or opacities Mucolipidoses Mucopolysaccharidoses Steroid sulfatase deficiency Cataracts Galactosemia Lowe syndrome Mitochondrial respiratory or electron transport chain defects Peroxisomal disorders Congenital defects in glycosylation Dislocated lens Methionine synthetase deficiency Sulfite oxidase deficiency Bone or limb deformities or contractures Storage, peroxisomal, or connective tissue disorders Inborn errors of cholesterol biosynthesis Thick skin Mucolipidoses Gangliosidoses Mucopolysaccharidoses Desquamating, eczematous, or vesiculobullous skin lesions Acrodermatitis enteropathica Organic acidemias Early-onset forms of porphyria Ichthyosis Gaucher disease type 2 Steroid sulfatase deficiency Alopecia Multiple carboxylase deficiency Steely or kinky hair Menkes disease Persistent diarrhea Glucose galactose malabsorption Congenital lactase deficiency Congenital chloride diarrhea Sucrase isomaltase deficiency Acrodermatitis enteropathica Congenital folate malabsorption Wolman disease Galactosemia Hepatosplenomegaly Gangliosidoses Niemann-Pick disease type C Mucopolysaccharidoses Wolman disease Ceramidase deficiency Macrocephaly Glutaric acidemia type 1 Canavan disease Microcephaly Mitochondrial respiratory or electron ­transport chain defects Leigh disease Methylmalonic acidemia with ­homocystinuria Coarse facial features Macroglossia Dystonia or extrapyramidal signs Macular “cherry red spot” *For Gangliosidosis Mucolipidoses Mucopolysaccharidosis type VII Sialidosis Galactosialidosis Pompe disease Gangliosidoses Mucopolysaccharidoses Mucolipidoses Gaucher disease type 2 Glutaric acidemia type 1 Krabbe disease Crigler-Najjar syndrome Biopterin defects GM1 gangliosidosis Galactosialidosis Niemann-Pick disease type A Tay-Sachs disease (GM2 gangliosidosis) discussion of specific disorders, see Chapters 22 and 23. 212 PART VI Metabolic and Endocrine Disorders of the Newborn TABLE 21-2 Characteristic or Unique Laboratory or Diagnostic Testing Outcomes in Inborn Errors (Major Examples) Outcome Error Outcome Error Metabolic acidosis with or without increased anion gap Organic acidemias Maple syrup urine disease Fatty acid oxidation defects Ketothiolase deficiency Ketogenesis defects Disorders of pyruvate metabolism Mitochondrial respirator or electron transport chain defects, including Leigh disease Galactosemia Glycogen storage disease type 1 Gluconeogenesis defects Thrombocytopenia Organic acidemias Pearson syndrome Anemia Organic acid disorders Wolman disease Pearson syndrome Severe liver failure Galactosemia Vacuolated lymphocytes or neutrophils Lysosomal storage disorders Respiratory alkalosis Urea cycle disorders Cardiomegaly Hyperammonemia Urea cycle disorders Methylmalonic acidemia Organic acidemias Fatty acid oxidation disorders Ketosis Organic acidemias Maple syrup urine disease Glutaric acidemia type 2 Ketogenesis defects Glycogen storage disease type 1 Gluconeogenesis disorders Pompe disease Barth syndrome Fatty acid oxidation defects Mitochondrial respiratory or electron ­transport chain defects Carbohydrate-deficient glycoprotein ­syndrome Electrocardiographic abnormalities Pompe disease (short PR interval, large QRS interval) Fatty acid oxidation disorders Mitochondrial respiratory or electron ­transport chain defects Lactic acidosis Mitochondrial respiratory or electron transport chain defects, including Leigh disease Pyruvate dehydrogenase complex deficiency Pyruvate carboxylase deficiency Organic acidemias Glutaric acidemia type 2 Fatty acid oxidation defects Ketogenesis defects Glycogen storage disease type 1 Gluconeogenesis disorders Ventricular hypertrophy Pompe disease Organic acidemias Glutaric acidemia type 2 Fatty acid oxidation defects Mitochondrial respiratory or electron ­transport chain defects, including Leigh disease Dysostosis multiplex Gangliosidoses Mucopolysaccharidoses Mucolipidoses Sialidosis Stippled calcifications of patellae Peroxisomal disorders Cholesterol biosynthetic defects Adrenal calcifications Wolman disease Rhizomelica Rhizomelic chondrodysplasia punctata Hair abnormalities Menkes disease Argininosuccinicaciduria Basal ganglia lesions on MRI Organic acidemias (later in life) Pyruvate dehydrogenase complex deficiency Mitochondrial respiratory or electron ­transport chain defects, including Leigh disease Cerebellar atrophy or hypoplasia Pyruvate dehydrogenase complex deficiency Mitochondrial respiratory or electron transport chain defects, including Leigh disease Carbohydrate-deficient glycoprotein syndrome Agenesis of corpus callosum Pyruvate dehydrogenase complex deficiency Pyruvate carboxylase deficiency Mitochondrial respiratory or electron transport chain defects Hypoglycemia Hyperinsulinism Glycogen storage disease type 1 Gluconeogenesis disorders Maple syrup urine disease Glutaric acidemias Fatty acid oxidation defects Ketogenesis defects Galactosemia Severe liver failure Mitochondrial respiratory or electron transport chain defects Lipemia Glycogen storage disease type 1 Lipoprotein lipase deficiency Positive urinaryreducing substances Galactosemia Hereditary fructose intolerance Lowe syndrome Discolored urine Alkaptonuria Tryptophan malabsorption Leukopenia Organic acidemias Glycogen storage disease type 1B Barth syndrome Pearson syndrome MRI, Magnetic resonance imaging. The third important category of metabolic disorders is insufficient generation of energy by the mitochondrial machinery. These disorders can be caused by the inability to provide substrates such as glucose in glycogenoses; the inability to deliver substrate to the site of oxidation, such as the fatty acid and carnitine transport disorders; the inability to break down fatty acids in a stepwise fashion to provide reduced flavin adenine dinucleotide to be oxidized; or the deficient function of the mitochondrial respiratory pathway and energy-generating system itself. These disorders have characteristics in between those of the acute, small molecule disorders and the storage disorders CHAPTER 21 Introduction to Metabolic and Biochemical Genetic Disease Box 21-5 S  igns and Symptoms of Inborn Errors in the Newborn ll ll ll ll ll ll ll ll ll Neonatal catastrophe (life threatening) Poor suck and feeding Gastrointestinal problems, vomiting Respiratory distress Cardiac failure Neurologic abnormalities: alertness, tone, seizures Organomegaly Ocular abnormalities Cutaneous changes (see Box 21-4). They differ from the small molecule disorders in the possible onset immediately at birth or before, and from storage disorders in the generally normal physical features with hepatomegaly alone, a regular feature of glycogen storage disorders. The small molecule and energygenerating disorders are discussed in Chapter 22 and the lysosomal storage disorders are discussed in Chapter 23. Of the remaining groups that are encountered less frequently, the peroxisomal biogenesis disorders, carbohydrate-deficient glycoprotein disorders, and SmithLemli-Opitz syndrome (a cholesterol biosynthetic disorder) are discussed in Chapter 23. Other disorders are too infrequent to be considered in a general neonatology textbook. Disorders of biogenic amines are discussed in Chapter 22 and in greater depth in Chapter 63 on neonatal seizures, along with consideration of pyridoxine and folate disorders. SIGNS AND SYMPTOMS OF INBORN ERRORS The limited symptomatic repertoire of the sick newborn is well established, but it is worth repeating. For this reason, the first thought when confronting a newborn in deteriorating condition, with lethargy, poor suck, temperature instability, and neurologic abnormalities, is sepsis (Box 21-5). Most metabolic specialists have never confronted a sick newborn who has not had a “septic workup” and who is not receiving standard antibiotics. The issue then becomes when to perform a metabolic workup. The standard answer is that it should be performed when the neonatologist is concerned that the newborn in extremis does not fit the pattern that they expect from a child with sepsis or hypoxia. That threshold would vary by individual. Negative results of tests for infectious agents, a nonconfirmatory white count, hypoglycemia, unexpectedly severe acidosis, or hyperammonemia could be important triggers. Although dysmorphic features are not characteristic of inborn errors, there are some that may have subtle or occasionally pronounced abnormality on a physical examination; they are listed in Box 21-6 and in Table 21-2. Modern neonatology has one tool that was previously unavailable: the expanded newborn screen. This screening will diminish the probability of many disorders, and the newborn screening follow-up hotline should be on the speed dial of every neonatal intensive care unit. The only acutely presenting disorders not ascertained by these studies are most hyperammonemias and lactic acidoses. These test results are available immediately in any tertiary or 213 Box 21-6 M  etabolic Diseases With Congenital Malformations or Dysmorphic Features ll ll ll ll ll ll ll Cholesterol biosynthetic disorders Peroxisomal disorders Glutaric aciduria type 2 Primary lactic acidoses Congenital defects in glycosylation Lysosomal storage disorders Menkes disease secondary care hospital. With a high level of concern and near normal levels of lactate and ammonia, the standard battery of metabolic studies should be performed only when the index of suspicion for a metabolic disorder is particularly high and no alternative explanation for the poor condition of the patient is likely. When deemed necessary, these studies should include plasma amino acids, plasma acylcarnitine levels, and urinary organic acids. The abnormalities associated with individual disorders are discussed in Chapter 22. EMERGENCY TREATMENT An acutely ill child with an inborn error of metabolism is an emergency, and rapid rescue treatment is mandatory. When considering a differential diagnosis, special emphasis should be placed on disorders for which there is treatment, as opposed to those for which there is no treatment. As with all acutely ill patients, supportive care, including cardiorespiratory, hemodynamic status, fluids, and electrolytes, is the mainstay of treatment. Transfer from institution to institution should not be performed unless the patient’s condition is stable and there is adequate vascular access for emergency treatment. Transfer to a tertiary care center with experience in caring for these children is desirable and should be performed as quickly as possible. Virtually all disorders require a maximum source of calories (150 calories/kg is desirable, but 120 calories/kg is a minimum target) to prevent or diminish catabolism, and glucose is the most important part of this, especially in the absence of primary lactic acidemia or acidosis. As much lipid as is safe should be added to compensate for the caloric deficit. Hemodialysis, and especially extracorporeal membrane oxygenation, will remove most metabolites rapidly, but these are extreme interventions in a newborn. These interventions should be performed routinely in extreme and symptomatic hyperammonemia (ammonia levels of 400 mol/L or more), and many would consider dialysis in severe acidosis caused by acids other than lactate or ketones. Once a diagnosis is established, specific therapy can be initiated. It is important to emphasize that during this period of generic therapy, prolonged deprivation of exogenous protein or amino acids will cause endogenous protein breakdown and exacerbate the metabolic process. As a result, protein is added to the intravenous support fluids after 36 to 48 hours, beginning with 0.25 to 0.5 g/kg/day and increasing gradually to maintenance levels of 1.2 to 1.5 g/kg/day, pending a definitive diagnosis and assuming that it is tolerated. 214 PART VI Metabolic and Endocrine Disorders of the Newborn SUGGESTED READINGS Fernandes J, Saudubray J-M, van den Berghe G, et al, editors: Inborn metabolic diseases: diagnosis and treatmented, ed 4, Germany, 2006, Springer-Verlag. Online Mendelian Inheritance in Man, OMIM: McKusick-Nathans Institute of Genetic Medicine, Johns Hopkins University (Baltimore, Md.) and National Center for Biotechnology Information, National Library of Medicine (Bethesda, Md.), Accessed {date of download}. Available at www.ncbi.nlm.nih.gov/omim/. Saudubray JM, Charpentier C: Clinical phenotypes: diagnosis/algorithms. In Valle D, et al, editors: The metabolic and molecular bases of inherited disease. Available online at www.ommbid.com, see Chapter 66. Saudubray J-M, Desguerre I, Sedel F, et al: Classification of inborn errors of metabolism. In Fernandes JM, Saudubray J-M, van den Berghe G, Walter JH, editors: Inborn metabolic diseases: diagnosis and treatment, Heidelberg, 2006, Springer-verlag, pp 1-47. Valle D, et al, editors: The metabolic and molecular bases of inherited disease. Available online at www.ommbid.com. C H A P T E R 22 Inborn Errors of Carbohydrate, Ammonia, Amino Acid, and Organic Acid Metabolism Stephen Cederbaum and Gerard T. Berry The inborn errors of carbohydrate, ammonia, amino acid, and organic acid metabolism have one factor in common: all can be associated with acute, life-threatening disease during the newborn period. The most notable exception in this broad group of small-molecule disorders is phenylketonuria (PKU). There are often few signs secondary to classic PKU in the first 6 months of life, underscoring the importance of newborn screening in establishing the diagnosis of this disease. Although this chapter will present the most common phenotype for these disorders, special emphasis will be placed on the neonatal presentations. The fact that a number of the more mild forms can manifest later merely emphasizes the importance of newborn screening in their mitigation. The disorders that constitute each group are listed in Boxes 22-1 to 22-4. The interrelationships among the major metabolites, biochemical cycles, and organelle pathways in the most critical facets of intermediary metabolism are simplified and depicted in Figure 22-1. The primary lactic acidosis and mitochondrial respiratory chain disorders, as well as the defects in fatty acid oxidation, are included in the section on inborn errors of organic acid metabolism. INBORN ERRORS OF CARBOHYDRATE METABOLISM syndrome was a much more common occurrence, associated with unlimited intake of lactose in the proprietary formula or breast milk. Because death from Escherichia coli sepsis can occur with only 1 to 2 weeks of exposure to galactose, the incidence of the disorder in the prescreening era was estimated to be less than 1 in 200,000 births. Today, even when picked up later after screening, the disorder is rarely fatal because the liver dysfunction is reversible and sepsis is largely prevented. If patients survive the neonatal period and there is no diagnosis, the most common initial clinical signs of GALT deficiency is poor growth, irritability, lethargy, vomiting, and poor feeding. Jaundice may be present in the first few weeks of life and can persist. Initially the hyperbilirubinemia may be indirect and is only later associated with an elevation of the direct component as well. The current tendency to change formulas can mask the disease if a non–lactose-containing formula is substituted fortuitously. With continual lactose ingestion, multiorgan toxicity syndrome ensues, which is associated with liver disease that can progress to cirrhosis with portal hypertension, splenomegaly, ascites, renal tubular dysfunction, and sometimes full-blown renal Fanconi syndrome. Anemia, primarily caused by decreased red blood cell (RBC) survival, and lethargy; brain edema associated with a bulging fontanel HEREDITARY GALACTOSEMIA BOX 22-1 Inborn Errors of Carbohydrate Metabolism Galactose-1-Phosphate-Uridyltransferase Deficiency ll The three enzymes of the galactose metabolic pathway that are responsible for the rapid conversion of galactose to glucose in the liver after ingestion of dietary lactose or the breakdown of endogenous galactose-containing compounds are galactokinase, galactose-1-phosphate uridyl transferase (GALT), and uridine diphosphate (UDP) galactose-4-epimerase (Figure 22-2). All three enzymes have been associated with inborn errors of galactose metabolism (Berry et al, 2006; Fridovich-Keil and Walter, 2008). Although a deficiency of any of the three enzymes can lead to galactose accumulation in plasma, the term galactosemia refers to GALT deficiency, the most common of the three enzyme deficiencies in the newborn period. The frequency of clinically significant galactosemia is estimated at 1 in 60,000 births in the United States, but would be higher if the more frequent partial deficiencies are considered. The clinical syndrome of transferase-deficiency galactosemia has changed since the advent of newborn screening. In instances of the rapid availability of newborn screening results (3 to 4 days of life) patients rarely require hospitalization. In the past, a severe multiorgan toxicity ll ll ll Hereditary galactosemia Glycogen storage diseases Hereditary fructose intolerance Fructose-1,6-bisphosphatase deficiency BOX 22-2 Inborn Errors of Ammonia Metabolism ll ll ll ll ll Ornithine transcarbamylase deficiency Argininosuccinicaciduria Citrullinemia Carbamylphosphate synthetase deficiency Transient hyperammonemia of the newborn BOX 22-3 Inborn Errors of Amino Acid Metabolism ll ll ll ll ll Maple syrup urine disease Hereditary tyrosinemia type 1 Nonketotic hyperglycinemia Methionine synthetase deficiency Phenylketonuria 215 216 PART VI Metabolic and Endocrine Disorders of the Newborn BOX 22-4 Inborn Errors of Organic Acid Metabolism ll ll ll ll ll ll ll ll Methylmalonic acidemia Propionic acidemia Isovaleric acidemia Multiple carboxylase deficiency Glutaric acidemia type 1 Fatty acid oxidation disorders ll Glutaric acidemia type 2 ll Very-long-chain acyl-CoA dehydrogenase deficiency ll Medium-chain acyl-CoA dehydrogenase deficiency ll Short-chain acyl-CoA dehydrogenase deficiency ll Long-chain 3-hydroxy acyl-CoA dehydrogenase deficiency ll Carnitine transporter defect ll Carnitine palmitoyltransferase type I deficiency ll Carnitine palmitoyltransferase type II deficiency ll Acylcarnitine translocase deficiency Defects in ketone metabolism ll Ketothiolase deficiency ll Succinyl-CoA: 3-ketoacid-CoA transferase deficiency ll 3-Hydroxy-3-methylglutaryl-CoA lyase deficiency Primary lactic acidoses ll Pyruvate dehydrogenase complex deficiency ll Pyruvate carboxylase deficiency ll Phosphoenolpyruvate carboxykinase deficiency ll Mitochondrial respiratory/electron transport chain defects ll Barth syndrome ll Pearson syndrome ll Leigh disease CoA, Coenzyme A. can also occur. Cataracts may be evident in the first few weeks of life. However, some infants are born with congenital cataracts that are associated with abnormalities of the embryonal lens; they are central in nature and require slit-lamp examination for documentation. After initiation of a lactose-free diet in the newborn period, the problems related to liver and kidney disease, anemia, and brain edema usually disappear, unless there has been severe organ damage such as hepatic cirrhosis. Most infants begin to grow and develop at a normal rate. However, patients treated prospectively can manifest long-term complications related to speech defects, delay in language acquisition, learning problems, frank mental retardation, autistic features, and hypergonadotropic hypogonadism in most of the females. The cause of these so-called diet-independent complications is unknown. Patients with galactosemia continue a lactose-restricted diet for their entire lives, although many lapse as they get older and may not suffer from detectable consequences. A minority of patients with GALT deficiency develop a neurodegenerative condition. One of the first abnormalities to be detected—­albuminuria—reflects a poorly understood renal glomerular component. This component develops within 24 to 48 hours of ingestion of lactose and disappears as quickly after elimination of lactose from the diet. In addition to hyperbilirubinemia, there may be mild to severe elevations of serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels and various abnormalities related to renal tubular dysfunction, such as hyperchloremic metabolic acidosis, hypophosphatemia, glucosuria, and generalized aminoaciduria. Vitreous hemorrhages are newly recognized complications in the newborn period. FIGURE 22-1 Intermediary metabolism interactions among the glycolytic, citric acid cycle, mitochondrial respiratory–electron transport chain, amino acid, organic acid, and urea cycle pathways. Defects in these primarily catabolic pathways are the chief source of inborn errors of metabolism that involve the small, simple—not the large—complex molecules. ATP, Adenosine triphosphate. (Adapted from Cowett RM, editor: Principles of perinatal-neonatal metabolism, ed 2, New York, 1998, Springer-Verlag, p 800.) CHAPTER 22 Inborn Errors of Carbohydrate, Ammonia, Amino Acid, and Organic Acid Metabolism 217 FIGURE 22-2 The important reactions in the galactose metabolic pathway are shown in relation to exogenous, via lactose primarily, and endogenous de novo synthesis of galactose. The reactions catalyzed by enzymes that have not been well delineated or are purported to exist are shown by broken lines. Carbon skeletons exit the galactose pool via galactokinase (GALK)–mediated conversion to galactose-1-phosphate (galactose-1-P), aldose reductase–mediated conversion to galactitol, and as galactonate. In patients with severe galactose-1-P-uridyltransferase (GALT) deficiency, there is little or no conversion of galactose-1-P to uridine diphosphate galactose (UDPgalactose). The epimerization of UDPglucose to UDPgalactose by UDPgalactose-4-epimerase, the utilization of UDPgalactose in the synthesis of glycoconjugates such as glycoproteins, and their subsequent degradation may constitute the pathways of de novo synthesis of galactose. (Adapted from Berry GT, Nissim I, Gibson JB, et al: Quantitative assessment of whole body galactose metabolism in galactosemic patients, Eur J Pediatr 156:S44, 1997.) When an infant’s condition is initially diagnosed, either through the newborn screening program or because of the recognition of clinical signs, blood galactose levels may be as high as 5 to 20 µmol/L, and the RBC galactose1-phosphate level is significantly elevated, as are urine galactitol levels. During this phase of severe hypergalactosemia, positive reducing substances are present in the urine. After the patient starts a lactose-free diet, the RBC galactose-1-phosphate levels in patients with classic galactosemia fall, but they never return to the normal range, remaining mildly elevated for the lifetime of the patient. As noted in the newborn screening chapter, a patient’s condition can be ascertained by elevated plasma galactose levels, reduced levels of GALT, or both. Because newborn blood spots can be obtained before significant lactose ingestion occurs, screening by galactose in plasma alone is unsatisfactory. Many programs screen for GALT deficiency alone and sacrifice diagnosis of the far less frequent and less lethal kinase and epimerase deficiencies. Many modern screening follow-up programs eschew the measurement of RBC galactose-1-phosphate levels and instead confirm the enzymatic deficiency with a quantitative assay and perform mutation analysis. Mutation analysis can identify those with a poorer prognosis and those who are unlikely to exhibit clinical symptoms. The mutations also provide information for genetic counseling and prenatal diagnosis, with the latter also performed effectively by enzymatic means. A GALT variant labeled the Duarte variant, after its place of discovery, is far more frequent than the more severe mutations and gives approximately three fourths of the activity of the normal allele. Compound heterozygotes for this allele (D2) and a null allele (G) have approximately 25% of normal activity in red blood cells and are not associated with clinical symptoms. GLYCOGEN STORAGE DISEASES The glycogen storage diseases (GSDs) can be divided into those types that primarily affect the liver and those that affect striated muscle (Kishnani et al, 2001; Smit et al, 2006). With some forms, such as GSD type 3 or debrancher deficiency, both striated muscle and the liver may be affected. According to European prevalence data, the overall frequency of GSD is 1 in 20,000 to 25,000 ­(Kishnani et al, 2001). With the exceptions of GSD type 2 and or Pompe disease, a lysosomal defect, most of the patients with glycogenosis do not come to clinical attention in the newborn period. With the exception of a form of phosphorylase kinase deficiency, which is labeled as type IX and is inherited in a sex-linked recessive manner, all are inherited in an autosomal recessive manner. Patients with the three most common forms of GSD—types 1, 3, and 6—have a phenotype that mimics a small-molecule disorder, because glucose homeostasis is affected. Glucose-6-Phosphatase Deficiency GSD type 1 is caused by decreased activity of glucose-6phosphatase, the enzyme that is perched at the terminus of both glycogenolysis and gluconeogenesis. Several different biochemical abnormalities can result in this phenotype, now classified as GSD types 1a, 1b, and 1c. The enzyme that resides on the anticytoplasmic side of internal membrane spaces of the hepatocyte catalyzes the hydrolysis of glucose-6-phosphate to glucose and phosphate. Impairments in the transport of either glucose-6-phosphate (type 1b–much less frequent than type 1a) or phosphate (type 1c–extremely rare) may cause decreased function of this enzyme. The frequency of GSD type 1 is estimated to be 1 in 100,000 births, with most cases being type 1a. This condition infrequently manifests in the neonatal period, because “demand” feeding or feeding at short 218 PART VI Metabolic and Endocrine Disorders of the Newborn intervals prevent symptomatic hypoglycemia and hepatomegaly has usually not yet occurred. The major clinical findings are poor growth and enlarged abdominal girth as a result of hepatomegaly, and any of the signs that may be related to hypoglycemia. The major laboratory findings are fasting hypoglycemia, ketosis, lactic acidosis, hyperlipidemia (i.e., hypertriglyceridemia), and hyperuricemia. In patients with type 1b disease caused by a defect in the microsomal transporter of glucose-6-phosphate, there may be a history of recurrent infections because of neutropenia and defective neutrophil function and inflammatory bowel disease that may be present in the first year of life. Diagnosis used to be based on hepatic enzyme analyses, but molecular diagnostic testing is currently the first choice. The most important aspect of therapy is the prevention of brain damage from hypoglycemia and growth failure. The mainstay of therapy is frequent feedings and restriction of lactose and sucrose (Kishnani et al, 2001; Smit et al, 2006). The use of continuous nasogastric feedings or uncooked cornstarch, particularly during the night, has significantly improved the care of affected children, and although it does not correct all the biochemical perturbations, it does improve growth and can prevent hypoglycemic spells. Intercurrent illness with increased glucose utilization is particularly hazardous and a careful plan for emergencies is essential. The leukopenia of type 1b is helped by a regimen of granulocyte colony stimulating factor. Lysosomal α-1,4-Glucosidase Deficiency GSD type 2, or Pompe disease, is a deficiency of the lysosomal enzyme α-glucosidase. The clinical presentation of patients is most often in the newborn or early infancy period with symptoms of heart failure. The frequency is estimated at 1 in 100,000 births. There is usually marked cardiomegaly with massive biventricular cardiac wall thickening and a typical abnormal electrocardiogram, confirming biventricular hypertrophy and with a short PR interval. Decreased cardiac output can lead to heart failure and passive congestion. Infants may also have generalized hypotonia because of a skeletal myopathy. There is increased deposition of glycogen within the lysosomes of striated muscle. Except in the instance of passive congestion, the liver is not usually enlarged. Diagnosis is based on enzyme analysis in dried blood spots or leukocytes (Goldstein et al, 2009; Pompe Disease Diagnostic Working Group et al, 2008; Zhang et al, 2008). Cardiac transplantation has been performed to prevent death in infancy. Onset of symptoms as late as adulthood occurs when the mutations in one or both of the alleles cause less complete enzyme loss. In the last several years enzyme replacement therapy for this disorder has been developed and has had a favorable effect on the outcome. More than half of the patients with infantile onset, for whom the disease had previously been fatal within the first year, can now be rescued with varying degrees of residual disability. Patients with adolescent and adult onset may do better depending on the speed with which the diagnosis was made, therapy initiated, and the degree of irreversible muscle damage that has occurred. HEREDITARY FRUCTOSE INTOLERANCE Fructose-1,6-Bisphosphate Aldolase B Deficiency Hereditary fructose intolerance is a rare disorder caused by a deficiency of the enzyme fructose-1,6-bisphosphate aldolase in liver (Steinmann et al, 2001). The enzyme deficiency results in an impairment in the conversion of fructose-1-phosphate to glyceraldehyde and dihydroxyacetone phosphate and therefore in the effective metabolism of fructose. The disorder is inherited as an autosomal recessive trait. Symptoms are triggered by eating fructose or sucrose, which produces fructose. The signs begin when juices and fruit are added to the infant diet and are mitigated by the aversion that the infant develops to these foods and drinks. The major clinical findings are pallor, lethargy, poor feeding, vomiting, loose stools, poor growth, hepatomegaly, and any sign that could be related to hypoglycemia. The major laboratory findings consist of hypoglycemia; hypophosphatemia; elevations of serum ALT and AST, including any of the findings that may be associated with hepatocellular disease; and the presence of reducing substances in the urine. The liver disease may be severe. Patients may be jaundiced with hyperbilirubinemia. There may be bleeding diathesis. In addition to liver disease, renal tubular dysfunction can lead to the renal Fanconi syndrome. The intuition of the physician is crucial in establishing the diagnosis. Previously an intravenous fructose tolerance test was performed under controlled circumstances, to determine whether serum phosphate and glucose levels decrease and serum AST and ALT values rise after 15 to 30 minutes of fructose administration. In the past, confirmation depended on enzyme analysis, but molecular diagnostic testing is more widely available now and avoids both the risk of provocative testing and the laborious enzymatic analysis. The treatment consists of elimination of dietary fructose and sucrose. Many patients enter adulthood with excellent teeth and an aversion to foods containing sucrose or fructose. FRUCTOSE-1,6-BISPHOSPHATASE DEFICIENCY Deficiency of the enzyme fructose-1,6-bisphosphatase results in an inability to hydrolyze fructose-1,6-bisphosphate to fructose-6-phosphate (Steinmann et al, 2001). It is a rare disorder. This enzyme is indispensable in gluconeogenesis. The main clinical features are hypoglycemia and signs related to glucose deprivation in the central nervous system (CNS). The symptoms are due primarily to inadequate food intake and not fructose ingestion, although fructose may exacerbate the abnormalities induced by fasting adaptation. Enlargement of the liver because of diffuse steatosis may be present only during periods of fasting and enhanced gluconeogenesis. The laboratory findings consist of hypoglycemia, ketosis, and lactic acidosis. The acidosis caused by accumulation of lactic, 3-hydroxybutyric, and acetoacetic acids can be severe in this disease. Diagnosis depends on enzymatic or mutation analysis. The therapy consists primarily of avoidance of fasting and the need for gluconeogenesis. CHAPTER 22 Inborn Errors of Carbohydrate, Ammonia, Amino Acid, and Organic Acid Metabolism Inborn Errors of Ammonia Metabolism Inborn errors of the urea cycle, because they manifest in infancy, exemplify the small-molecule weight disorders in which symptoms occur episodically, are exacerbated by protein catabolism, and can be treated by the control of catabolic stimuli, dietary modification, and the enhanced removal of offending metabolites. Four disorders in particular cause the majority of cases that occur in the newborn period and in infancy; they are carbamyl phosphate synthetase I deficiency, ornithine transcarbamylase (OTC) deficiency, argininosuccinate (ASA) synthetase deficiency (citrullinemia), and ASA lyase deficiency (argininosuccinic acidemia). The other four genetic disorders of the urea cycle, N-acetylglutamate synthase deficiency, arginase deficiency, ornithine transporter deficiency, and citrin deficiency, are rare or do not manifest with neonatal hyperammonemia, or both. In aggregate, disorders of the urea cycle may be as frequent as 1 in 25,000 births or more. The entire cycle is outlined in Figure 22-3 and will be important in understanding the therapeutic modalities. One round through the urea cycle condenses two molecules of toxic ammonia and one molecule of bicarbonate to form a molecule of urea that is nontoxic and is readily excreted in the urine. N-Acetylglutamate, the product of the first enzyme in the cycle, is an obligatory activator of carbamyl phosphate synthetase. These two reactions and OTC, which is responsible for citrulline synthesis, occur in the mitochondrion. The next three reactions occur in the cytoplasm after citrulline is transported out of the mitochondrion. The transport of ornithine back into the mitochondrion and the shuttling of aspartate to the cytoplasm are clearly imperative and account for the two other genes that cause urea cycle defects. With the exception of arginase deficiency, each of these enzyme deficiencies has been associated with disease in the newborn period. Clinical presentation in the newborn period is similar for all these defects (Brusilow and Horwich, 1995; Leonard, 2006). Almost all the infants are well in the first 12 to 24 hours of life until they begin to feed poorly, vomit, hyperventilate, become irritable and lethargic, and become comatose, usually with seizures. When these diseases are not treated aggressively, they are almost always fatal. The treatment requires specific therapy to lower the waste nitrogen burden, including the toxic substance ammonia, and address the increased intracranial pressure. The severe encephalopathic and life-­threatening features may be related in large part to brain edema. Chronic hepatomegaly has been reported in patients with argininosuccinic aciduria, whereas hepatomegaly is evident only during hyperammonemic episodes in the other urea cycle disorders. Acute hyperammonemia is associated with transaminase elevation and liver synthetic dysfunction, but these are rarely more than transient. Histologic examination of the liver shows modest fatty infiltration and fibrosis. Children with argininosuccinic aciduria can also manifest a specific abnormality of the hair known as trichorrhexis nodosa. The main laboratory finding in the urea cycle defects (UCDs) is a plasma ammonium elevation. Plasma ammonium values may vary in different laboratories. In general, however, with automated chemistry testing for ammonia, 219 the normal plasma values in older infants, children, and adults range between 10 and 35 μmol/L. However, the normal plasma ammonium value in newborns may occasionally be as high as 110 μmol/L but is usually somewhat lower. In patients with newborn-onset UCDs who are acutely ill, the plasma ammonium levels are often higher than 1000 or 2000 μmol/L. Patients with UCDs usually do not have metabolic acidosis unless they are in a terminal state with vascular collapse or respiratory failure. Instead, the characteristic acid-base abnormality associated with hyperammonemia is respiratory alkalosis caused by the effect of ammonia on the respiratory control centers in the brainstem. The various UCDs can usually be distinguished on the basis of the pattern and levels of plasma amino acids. Because citrulline is the product of the carbamyl phosphate synthetase type 1 (CPS-I) and OTC reactions and the substrate for ASA synthetase, its value is critical. In newbornonset N-acetylglutamate synthase (NAGS), CPS-I and OTC deficiencies, plasma citrulline concentrations may be undetectable and are always low. With OTC deficiency, there is increased urinary orotate excretion secondary to carbamyl phosphate accumulation and pyrimidine synthesis. With NAGS or CPS-I deficiency, carbamyl phosphate production is decreased or absent, and orotate excretion is decreased. In citrullinemia, the eponymous amino acid citrulline has markedly elevated concentrations. With argininosuccinic aciduria, plasma citrulline concentration is moderately elevated, in the range of 100 to 300 μmol/L, and can be readily detected during a study of plasma by amino acid analysis. Because the ability of infants with these disorders to excrete waste nitrogen as urea is impaired, therapy is initially focused on the reduction of nitrogen intake by decreasing dietary protein and providing essential amino acids or the ketoacid analogues. This approach theoretically permits adequate growth without an excessive nitrogen load. Excessive protein leads to hyperammonemia, but too much restriction of protein during long-term therapy leads to poor growth and can provoke catabolism to maintain essential amino acid levels. Actually this approach fails when the patient is in a catabolic state and in negative nitrogen balance, as occurs in the catastrophically ill infant presenting in the first week of life with massive hyperammonemia. For such an infant with hyperammonemia and coma, the mainstay of therapy is dialysis treatment. Hemodialysis (or extracorporeal membrane oxygenation) is the most effective way of reducing plasma ammonium levels, because it affords the greatest clearance of ammonia (Rutledge et al, 1990). Continuous arteriovenous hemofiltration (CAVH) provides a lower clearance rate, but has the added benefit of continuous use and a lesser likelihood of major swings in intravascular volume that can exacerbate an already fragile state and cerebral edema. Ammonia clearance with peritoneal dialysis is only approximately one tenth that of CAVH and is not recommended for specific UCD therapy in the newborn period. Ammonia is not cleared effectively by exchange transfusion. While the intensive care personnel are waiting for dialysis to be started, alternative waste nitrogen therapy using intravenous sodium benzoate, sodium phenylacetate (together as Ammonul), and, for patients with 220 PART VI Metabolic and Endocrine Disorders of the Newborn A B FIGURE 22-3 A, Depicted is the nonhomogeneous distribution of enzymes involved in ammonia metabolism in hepatocytes of an acinar sinusoid as they are linearly distributed from the portal triad to the region of the central vein or terminal hepatic venule. The specific enzymatic reactions are shown for an individual periportal and perivenous hepatocyte. The glutamine synthetase (GS) and ornithine aminotransferase (OAT) enzyme activities are expressed exclusively in the one- to three-cell layers surrounding the central vein—that is, the region of zone 3 of the liver lobule—whereas the urea cycle enzymes are concentrated within the periportal hepatocytes. However, the urea cycle enzyme activities are higher in zone 1 immediately surrounding the portal triad than in the middle zone 2. The hepatocytes are shown as squares, the hepatic mitochondria as shaded circles, and the lining of the space of Disse as the broken lines on either side of the linear array of hepatocytes. Arg, Arginine; ASA, argininosuccinate; Asp, aspartate; ATP, adenosine triphosphate; Cit, citrulline; CP, carbamylphosphate; CPS-I, carbamylphosphate synthetase type 1; Gln, glutamine; Glu, glutamate; α-KG, α-ketoglutarate; NH3, ammonia; CO2, CO2 or bicarbonate; Orn, ornithine; P5C, pyrroline 5-carboxylate; P5CDH, pyrroline-5-carboxylate dehydrogenase. B, The medications sodium phenylacetate and phenylbutyrate and sodium benzoate promote alternative waste nitrogen disposal by participating in these two mitochondrial reactions. (Adapted from Tuchman M, et al: Hepatic glutamine synthetase deficiency in fatal hyperammonemia after lung transplantation, Ann Intern Med 127:447, 1997.) ASA synthetase and ASA lyase deficiencies especially, arginine hydrochloride should be given (Brusilow, 1991; Brusilow and Batshaw, 1979; Brusilow et al, 1979). The medications and the proper dose for a bolus and 24-hour sustaining infusions of Ammonul are available from Ucyclyd Pharma (Scottsdale, Arizona). This product is rarely stocked except in pharmacies of tertiary care metabolic centers. Arginine hydrochloride is thought to be helpful in NAGS, CPS-I, and OTC deficiencies and is given at a continuous dose of 250 mg/kg per 24 hours. A dose of 600 mg/ kg/day is recommended for ASA lyase particularly and for citrullinemia. Arginine supplementation the body arginine CHAPTER 22 Inborn Errors of Carbohydrate, Ammonia, Amino Acid, and Organic Acid Metabolism pool, and in citrullinemia and argininosuccinic acidemia it increases ornithine to promote synthesis of citrulline and ASA. ASA contains both waste nitrogen atoms destined for excretion as urea and has a renal clearance rate equal to the glomerular filtration rate, provided that it is continuously synthesized and excreted. Accordingly, ASA should serve as an effective substitute for urea as a waste nitrogen product. The excretion of citrulline, which contains only one molecule of ammonia, also exceeds that of ammonia, although it does not appear to be eliminated as efficiently as ASA. Sodium benzoate promotes ammonia excretion when it is conjugated with glycine to form hippurate, which is cleared by the kidney at fivefold the glomerular filtration rate (see Figure 22-3).Theoretically, 1 mole of waste nitrogen is synthesized and excreted as hippurate for each mole of benzoate administered. The hippurate synthetic mechanism resides primarily in the hepatic mitochondria and depends on an intact mitochondrial energy system for adenosine triphosphate (ATP) synthesis. The glycine consumed in this reaction can be replaced by either serine or the glycine cleavage pathway. Sodium phenylacetate, as well as sodium phenylbutyrate, which is used for long-term therapy in the absence of sodium benzoate, conjugates with glutamine to form phenylacetylglutamine, which is excreted by the kidney (see Figure 22-3). Sodium phenylbutyrate is converted to phenylacetate in the liver. Glutamine contains two nitrogen atoms, whereas glycine contains one. Two moles of waste nitrogen are removed for each mole of phenylacetate administered. This acetylation reaction occurs in the kidney and the liver. The outcome for patients with severe newborn-onset CPS-I and OTC deficiencies is poor. Sometimes dialysis therapy cannot rescue severely affected boys with X-linked OTC deficiency in the first few days of life (Enns et al, 2008). Prospectively administered alternative pathway therapy in conjunction with high-calorie fluids usually prevents death and severe hyperammonemia in patients known from family studies or prenatal diagnosis to be at risk. This therapy is done best in collaboration with an expert in the treatment of urea cycle disorders. Even after institution of successful therapy, the morbidity and mortality rates are high in these severely affected patients and mental retardation is common in survivors (Brusilow and Horwich, 1995). There is a significant correlation between the duration of newborn hyperammonemic coma and the developmental quotient score at 12 months of age (Msall et al, 1984). Four of five reported children in whom duration of coma was 2 days or less had normal intelligence quotient scores, whereas all seven children in whom coma lasted 5 days or longer were severely mentally retarded. Currently liver transplantation is recommended for patients with CPS-I and OTC deficiencies in the newborn period. The patient should have almost no residual enzyme activity and grow to a sufficient size for the treatment to be feasible and safe. Early diagnosis and treatment is important because of the devastating effects of prolonged newborn hyperammonemic coma. Mutational analysis of DNA is available for all of these disorders at www.genetests.org. If the lesion in a particular family is known, prenatal diagnosis by means of direct DNA analysis is also feasible. With the exception 221 of OTC deficiency, all the UCDs are inherited as autosomal recessive traits. TRANSIENT HYPERAMMONEMIA OF THE NEWBORN Transient hyperammonemia of the newborn is a distinct clinical syndrome that was first identified by Ballard et al (1978). The disease usually develops in premature infants during the course of treatment for respiratory distress syndrome. The plasma ammonium level may be enormously elevated, as high as that found in any of the patients with the most severe type of UCD. Its onset is usually in the first 24 hours after birth, when the infant is undergoing mechanical ventilatory support. Affected babies can manifest all the signs associated with hyperammonemic coma. The diagnosis may not be obvious, however, because many of these same infants are receiving sedatives and muscle relaxants to optimize therapy of their life-threatening pulmonary disease and an ammonia level is not determined. Important clues are the absence of deep tendon reflexes, the absence of normal newborn reflexes, and a decrease or absence of response to painful stimuli. As with hyperammonemic coma in UCD, this medical emergency requires dialysis therapy. The cause of this disease is unknown. The plasma amino acid levels are similar to those found in CPS-I or OTC deficiency. Investigators have hypothesized that the disorder may be caused by impairment of hepatic mitochondrial energy production or shunting of portal blood away from the liver, such as in patent ductus venosus. The mortality rate in transient hyperammonemia of the newborn is high. A patient who can be treated early and aggressively may survive the episode. There is no evidence that any of the survivors have suffered any further episodes of hyperammonemia, nor has there been any further evidence of impaired ammonia metabolism. INBORN ERRORS OF AMINO ACID METABOLISM MAPLE SYRUP URINE DISEASE Branched-Chain 2-Keto Dehydrogenase Complex Deficiency Maple syrup urine disease (MSUD) is a rare inborn error of amino acid metabolism (Chuang et al, 2001; Wendel and de Baulny, 2006). It is inherited as an autosomal recessive trait and is caused by a deficiency of the enzyme branched-chain 2-keto dehydrogenase (BCKAD) complex. This enzyme catalyzes the conversion of each of the 3-ketoacid derivatives of the branched-chain amino acids (BCAAs)—leucine, isoleucine, and valine—into their decarboxylated coenzyme metabolites within the mitochondria (Figure 22-4). The disease occurs in 1 in 200,000 newborn infants around the world, but in the Mennonite communities of the United States, the frequency is 1 in 358 because of a founder effect for a point mutation in the E1α gene. The most common presentation of the disorder occurs in the newborn period. Other forms with later onset caused by less complete enzyme deficiencies occur, but they are not the subject of this chapter. The newborn 222 PART VI Metabolic and Endocrine Disorders of the Newborn FIGURE 22-4 The branched-chain amino acids leucine, isoleucine, and valine are reversibly transaminated to their corresponding 2-keto analogues, which are substrates for the single decarboxylase (branched-chain 2-keto dehydrogenase) enzyme deficient in maple syrup urine disease. Reduced activity of isovaleryl coenzyme A (CoA) dehydrogenase, the next enzyme in the leucine degradative pathway, is the cause of isovaleric acidemia. The immediate precursor of ketones, 3-hydroxy-3-methylglutaryl CoA (HMG-CoA), is the final product of the leucine catabolic pathway. However, its production more strongly depends on the oxidation of fatty acids as in ketogenesis and in cholesterol biosynthesis. Propionyl CoA, which accumulates in both propionic and methylmalonic acidemia, can be synthesized from isoleucine, valine, odd-chain fatty acids, cholesterol, methionine, threonine, and thymine. The adenosyl form of vitamin B12 is the important cofactor in the l-methylmalonyl CoA mutase–catalyzed conversion of l-methylmalonyl-CoA to the citric acid cycle intermediate, succinyl CoA. presentation is associated with a severe and catastrophic illness and usually results in death without specific medical intervention. Typically the infants are well at birth; only after 2 or 3 days of ingestion of breast milk or formula do they begin to manifest poor feeding and regurgitation. Lethargy becomes evident; the cry may be shrill and high pitched. There may be hypotonia alternating with hypertonia and opisthotonic posturing. The odor of maple syrup may be detected in saliva, on the breath, in urine and feces, and in cerumen obtained from the ear. The babies become more and more obtunded and eventually lapse into a deep coma. The anterior fontanel may be bulging. Seizures may develop. The life-threatening encephalopathic features may simply be related to brain edema (Chuang et al, 2001). The clinical biochemical laboratory hallmark is metabolic acidosis. The anion gap may be raised, but not necessarily. There is almost always ketonuria. The plasma ammonium values are usually normal. The complete blood count is usually normal. The levels of the plasma BCAAs leucine, isoleucine, and valine are elevated, with striking elevation in leucine, the probable toxic metabolite. The normal ranges for plasma leucine, isoleucine, and valine in the newborn period are 29 to 152 μmol/L, 11 to 87 μmol/L, and 71 to 236 μmol/L, respectively. In patients who are critically ill, the leucine levels may range between 1900 and 6900 μmol/L. In the past, almost every newborn infant who was recognized to have MSUD and was severely ill was treated with peritoneal dialysis in a tertiary care center. The treatment was clinically successful in most instances, but it did not allow for as rapid a rate of reduction in plasma BCAA levels as did hemodialysis or CAVH (Rutledge et al, 1990). It is now clear that a nutritional approach works as well as peritoneal dialysis in newborns with MSUD (Berry et al, 1991; Chuang et al, 2001; Morton et al, 2002; Parini et al, 1993; Wendel and de Baulny, 2006). It is best performed with the use of a BCAA-free modified parenteral nutrition solution for infants and older children with acute metabolic decompensation, but this is rarely available in any center that does not have a larger Mennonite population. Protein-free intravenous alimentation augmented with a low-volume, enteral solution of the 17 nonbranched chain amino acids for 24 to 48 hours is satisfactory. An insulin drip may also be necessary to curtail the effects of the catabolic stimulus (Berry et al, 1991; Wendel et al, 1982). On the basis of the rate of plasma leucine decline, the staff at Children’s Hospital of Philadelphia have found that this nutritional therapy is comparable to peritoneal dialysis when plasma leucine levels are as high as 1908 to 3053 μmol/L. However, CAVH or hemodialysis may achieve more rapid normalization of the plasma BCAAs and their corresponding branched-chain ketoacids when the levels of leucine are in the range of 4580 to 6870 μmol/L. Newborn infants with an odor of maple syrup, a metabolic CHAPTER 22 Inborn Errors of Carbohydrate, Ammonia, Amino Acid, and Organic Acid Metabolism acidosis, and ketonuria require immediate metabolic consultation assistance and a confirmatory plasma amino acid quantitation. This is truly a medical emergency with a high risk of death and permanent brain damage. Most patients who are successfully treated by 7 days of age do not have mental retardation. Patients whose diagnosis and therapy are delayed—and perhaps even in those whose disorder is diagnosed in the first week of life, but who have suffered such severe damage because of increased intracranial ­pressure—often have substantial decreases in developmental or intelligence quotient, as well as signs compatible with spastic diplegia or quadriplegia. Expanded newborn screening is going to lead to a changed picture for this disorder. In jurisdictions in which the turnaround time for the program are as fast as 3 to 4 days of life, the patient will likely be less ill and the diagnosis will be known. The outcome will be better because less risk of permanent damage will have occured. The mainstay of long-term therapy for patients with MSUD who survive the newborn period is a special formula devoid of the BCAAs. The amount of BCAAs necessary to sustain growth but maintain plasma leucine, isoleucine, and valine levels in the normal range is supplied with a regular proprietary formula in limited amounts. In the first week of life, this amount is approximately 100 mg of leucine per kilogram of body weight daily and 50 mg/kg daily of both isoleucine and valine. The BCAA requirements, and thus the rates of utilization of the BCAAs for protein accretion, drop rapidly in the first year of life in conjunction with the decline in growth velocity in young infants. As with PKU, it is imperative that the administration of the special amino acid formula be carefully monitored by means of frequent plasma amino acid quantitations. One of the most common errors made in the treatment of MSUD is the failure to administer adequate amounts of supplemental isoleucine and valine solutions to maintain normal plasma levels, because proprietary formulas alone do not always meet the needs of each growing baby (i.e., adequate amounts of each BCAA are not supplied by a formula alone). Deficiency of BCAAs can result in a severe exfoliative rash and anemia. The rash may mimic that of severe acrodermatitis enteropathica (Giacoia and Berry, 1993). As with many of the inborn errors of amino acid, organic acid, and ammonia metabolism, protein administration in the infant whose disease is controlled and who is in metabolic balance with diet therapy must be discontinued during periods of intercurrent infections because catabolism, possibly driven by counterregulatory hormones or cytokines, triggers metabolic decompensation through the release of branched-chain ketoacid from skeletal muscle. Metabolic decompensation is further exacerbated by poor nutritional intake. It is imperative that during these times, adequate calories and fluids be given to prevent the crisis from escalating into a medical emergency. In selected instances, a molecular diagnosis of MSUD may be undertaken; this diagnosis is especially useful in targeted populations such as the Mennonites. Rarely patients may have a defect in the enzyme 2(E2) component (Danner et al, 1985) or, as discussed in the Primary Lactic Acidosis section, in the enzyme 3(E3) component. The rare patient with MSUD may respond favorably to high-dose thiamine therapy with a reduction in BCAAs and a lower need for protein restriction. 223 HEREDITARY TYROSINEMIA TYPE 1 OR HEPATORENAL TYROSINEMIA Fumarylacetoacetate Hydrolase Deficiency Tyrosinemia type 1 is caused by a deficiency of the enzyme fumarylacetoacetate hydrolase (Tanguay et al, 1995). This enzymatic reaction is distal in the phenylalanine and tyrosine pathway, and the disorder is actually an inborn error of organic acid metabolism, with hypertyrosinemia being a secondary and variable biochemical effect. Tyrosinemia type 1 is a rare disease inherited in an autosomal recessive manner. The highest incidence is found in those of French-Canadian ancestry as a result of a founder effect (De Bracketeer and Larochelle, 1990). The phenotype is variable. At one extreme, the patient is in early infancy and has severe, usually fatal disease in which liver disease dominates the clinical picture. At the other end of the spectrum is a more chronic phenotype, and patients exhibit hypophosphatemic rickets related to renal Fanconi syndrome. These patients usually also have evidence of liver disease, although milder. Most of the patients with severe liver disease phenotype do not come to clinical attention in the newborn period. Careful search, however, may reveal laboratory findings compatible with this disease, even in newborn infants. In the phenotype seen in early infancy, the clinical findings are hepatomegaly, a bleeding diathesis, and jaundice. Ascites is not uncommon. The laboratory findings consist of abnormal liver function and dysfunction tests with increases in serum AST, ALT, and direct and indirect bilirubin; prolongations of PT and PTT; and findings related to renal Fanconi syndrome, such as glycosuria, hypophosphatemia, hypouricemia, proteinuria caused by β2-microglobulin hyperexcretion, and generalized amino aciduria and organic aciduria. More specific laboratory abnormalities consist of an abnormally high serum alpha-fetoprotein level. There may be an elevation of plasma tyrosine; however, the hypertyrosinemia and a more prominent hypermethioninemia are caused by a secondary impairment in liver function. The most important metabolites are those related to the substrate, fumarylacetylacetic acid, the handling of which is defective. Increased levels of the diagnostic metabolite succinylacetone can be detected on urine organic acid quantitation by gas chromatography–mass spectrometry (GC-MS) in most instances and even more sensitively by an isotope dilution method, and in RBCs by the δ-aminolevulinic acid dehydratase inhibition assay. As in all metabolic disorders, the metabolite abnormalities may not be present or detectable at all times. Newborn screening methods are likely to alter the phenotype of this disorder. Increasingly, analysis of blood spot succinylacetone is being added to the expanded newborn screening panels and will likely lead to preemptive diagnosis and treatment in the majority of cases. The falsenegative rate for this analysis is not known at this time. Newborn screening by older methods has been performed in Quebec for many years. In the past, most of the infants with the hepatic form of tyrosinemia type 1 died in early to late infancy. However, a medical therapy has been developed that uses the agent 2-(2-nitro4-trifluoromethylbenzoyl)-1,3-cyclohexanedione 224 PART VI Metabolic and Endocrine Disorders of the Newborn (NTBC), which inhibits p-hydroxyphenylpyruvate dioxygenase (Lindstedt et al, 1992), thus blocking the conversion of p-hydroxyphenylpyruvate, the transamination product of tyrosine, to fumarylacetoacetate and thereby retarding the synthesis and accumulation of succinyl acetate and succinyl acetone. This therapy has been used successfully to improve liver and renal function as well as to normalize or improve the various laboratory abnormalities. It is unclear, however, whether NTBC therapy will eliminate the need for liver transplantation, which has been the mainstay of long-term therapy until recently. One of the most devastating complications for older infants with tyrosinemia type 1 who have been stabilized clinically and biochemically is the development of hepatocellular carcinoma. This development is prevalent in tyrosinemia treated with diet, and most patients who survive infancy or who have a chronic phenotype succumb to liver cancer. It is recommended that NTBC therapy be started in infants who are acutely ill while they are waiting for a liver transplantation, which may be deferred until the third year of life, while monitoring the liver for development of a hepatoma. Early treatment with NTBC surely retards the development of tumors, but patients receiving therapy develop hepatomas and preemptive transplantation, although at a later age. Liver transplantation is still a highly prevalent therapy. Most liver transplantations have been successful, and the patients have done well with immunosuppressive therapy with only minimal persistent evidence of renal tubular dysfunction while eating a normal diet. Hepatorenal tyrosinemia has no direct effect on the nervous system, and most patients are free of neurologic findings and mental retardation. Because succinylacetone can inhibit delta-amino levulinic acid dehydratase, the enzyme deficient in a rare type of acute intermittent porphyria, porphyria symptoms have frequently occurred in the preNTBC era and have proved fatal in some instances. Hepatorenal tyrosinemia is not to be confused with transient tyrosinemia of the newborn, which is prevalent in premature infants and is probably the most common disturbance of amino acid metabolism in man; it is secondary to a delayed maturation of p-hydroxyphenylpyruvate dioxygenase activity (Levine et al, 1939; Tanguay et al, 1995) or with the hypertyrosinemia secondary to liver disease. Current screening methods are unlikely to detect this condition that was found frequently in the past. NONKETOTIC HYPERGLYCINEMIA Glycine Cleavage Complex Deficiency Nonketotic hyperglycinemia is a rare defect of glycine metabolism (Dulac and Rolland, 2006; Hamosh and Johnston, 1995.) and is inherited as an autosomal recessive trait. The disorder is caused by deficient activity of the glycine cleavage enzyme complex (GCC), which consists of the products of four genes. Its frequency is fewer than 1 in 200,000 newborn infants. Although several variant forms exist, most infants who come to clinical attention in the newborn period—presumably most patients with this ­disease—have a severe, catastrophic illness that mimics the most acute forms of ammonia, amino acids, or organic acid metabolism. The infants may be healthy at birth but begin to show severe hypotonia and seizures after 12 to 36 hours. They quickly become comatose, and there is a loss of all the newborn and deep tendon reflexes. The clinical findings are predominantly those of an acute encephalopathy. The electroencephalogram (EEG) usually shows a characteristic pattern of spike and slow waves. The babies may have hiccups from diaphragmatic spasms. The main laboratory finding is an elevation of plasma glycine and a proportionally higher elevation of glycine in the cerebrospinal fluid (CSF). Unfortunately, plasma glycine levels may be normal in the newborn period. The urine glycine is usually also elevated. The most common diagnostic criterion is the raised CSF–plasma glycine ratio. Values greater than 0.08 are considered to be diagnostic, whereas those between 0.04 and 0.08 are highly suspicious of the diagnosis. The amino acid serine, which is also a product of the defective enzyme reaction, is depressed in plasma, and there is a corresponding increase in the glycine-to-serine ratio in body fluids. Prenatal CNS lesions have been reported (Dobyns, 1989). The pathophysiology of this brain disease is not well understood; it is believed that glycine, an active neurotransmitter, may interfere with the activity of specific chloride channels, thus perturbing the membrane potential and depolarization of neurons. There also appears to be an impairment in alpha-motor neuron outflow tract activity, producing a clinical state that mimics Werdnig-Hoffmann disease. No acidosis or ketosis is seen in this disease—thus the name, nonketotic hyperglycinemia. Many of the disorders of organic acid metabolism, such as methylmalonic acidemia and propionic acidemia, are also associated with elevations of plasma glycine, presumably because of a secondary impairment in the GCC, and have been referred to in the past as the ketotic hyperglycinemia syndromes. Although many therapies have been tried in this ­disease— such as protein restriction, benzoate to trap glycine as the byproduct hippurate, strychnine to affect the lower motor neuron function, and dextromethorphan to block N-methyl-d-aspartate receptors—there is no generally effective treatment. The seizures are variably controlled and life may be prolonged, but no meaningful improvement occurs in most instances. Many affected babies die in the newborn period despite medical support with assisted ventilation, and most others die within 2 to 4 years. A transient form of nonketotic hyperglycinemia has also been reported in newborns (Luder et al, 1989; Schiffmann et al, 1989). Valproate may result in hyperglycinemia because of secondary inhibition of the GCC, as is postulated to explain secondary hyperglycinemia in disorders of organic acid metabolism, such as propionic acidemia. Milder forms exist and milder symptoms appear at a later age. This disorder cannot be diagnosed by newborn screening. METHIONINE SYNTHETASE DEFICIENCY, OTHER DISORDERS OF HOMOCYSTEINE REMETHYLATION AND CYSTATHIONINE SYNTHASE DEFICIENCY In humans, the essential amino acid methionine is converted to homocysteine, and in the process a methyl group is transferred to an acceptor molecule from CHAPTER 22 Inborn Errors of Carbohydrate, Ammonia, Amino Acid, and Organic Acid Metabolism S-­adenosylmethionine that serves as a donor for methyl groups in many different reactions (see Figure 22-5) (Mudd et al, 1995). Subsequently the homocysteine may either be completely metabolized through the cysteine pathway to sulfate or it may be remethylated back to methionine to maintain the critical levels of methionine. Defective methionine remethylation or a deficiency in methionine synthetase leads to a form of homocystinemia (the name homocystinuria is anachronistic and does not reflect the current reality of ascertainment or testing as well as does homocystinemia). Patients with classic homocystinemia, caused by a deficiency of the cystathionine beta-synthase enzyme, rarely manifest signs in early infancy (Mudd et al, 1995). In contrast, patients with methionine synthetase deficiency or methionine remethylation defect may come to clinical attention in the newborn period (Rosenblatt and Fenton, 2001). The patients may have either an abnormality in vitamin B12 metabolism, which can also produce methylmalonic acidemia and homocystinuria, or an isolated defect in folate metabolism or the methionine synthetase enzyme (Rosenblatt and Fenton, 2001). In this reaction the methyl group from 5-methyltetrahydrofolate, which is derived from 5,10-methylene tetrahydrofolate, is transferred to methylcobalamin and subsequently to homocysteine. The clinical findings associated with the methionine synthetase deficiencies are poor growth and development. There may be severe cortical atrophy and possible brain lesions caused by thromboses of the arteries or veins, as in classic homocystinuria. Infrequently they can manifest as an acute intoxicating disorder in the newborn period. The laboratory findings consist of an elevation in plasma homocysteine values and a normal or decreased methionine value. Often the homocysteine values are not as elevated as in classic cystathionine β-synthetase deficiency. If there is a defect in folate or vitamin B12 metabolism that produces secondary impairment in methionine synthetase activity, there may also be megaloblastic anemia. Some patients with a defect in cobalamin metabolism respond to treatment with high doses of intramuscular hydroxycobalamin. The treatment of methionine synthesis deficiency is a normal protein intake or methionine supplementation to restore methionine levels in plasma to normal and to restore CNS pools of methionine, which may be critical in one-carbon transfer reactions. It is also possible to retard homocysteine accumulation and restore methionine levels by administering betaine, which can enhance remethylation of homocysteine to methionine through the alternate betaine methyltransferase pathway. Some investigators have also used pharmacologic doses of cobalamin, folate, and pyridoxine to stimulate flux through either the methionine synthetase or the classic homocysteine pathway. Unfortunately, if infants with this type of disorder are not treated early, there is usually a permanent and devastating effect on cognitive and motor function. Milder and occasionally asymptomatic forms are known. Classic homocystinemia caused by cystathionine synthase deficiency can be ascertained by expanded newborn screening in a minority of cases. The elevated methionine that is the critical detected metabolite may require several days to become apparent; that is unfortunate because cystathionine synthetase deficiency is a treatable disorder. 225 Expanded newborn screening has altered the landscape for several B12 processing defects. Ascertainment of elevated methylmalonic acid with the C3 acylcarnitine, the C4 dicarboxylacyl carnitine, or both usually allows for earlier diagnosis and treatment, with apparent mitigation of long-term effects. These disorders are more common than classical homocystinemia, the latter having a frequency in most populations of 1 in 200,000 births or less. PHENYLKETONURIA Phenylalanine Hydroxylase Deficiency PKU is the most common inborn error of amino acid metabolism that can result in mental retardation (Scriver et al, 2001; Walter et al, 2006). Its frequency is between 1 in 10,000 to 1 in 25,000 births, depending on where in the United States it is studied. There are areas of the world such as Ireland, Turkey, and Iran in which it is more frequent, and others such as Finland, Japan, and among Ashkenazi Jews in which it is less prevalent. PKU is caused by a defect in the activity of the enzyme phenylalanine hydroxylase, which converts phenylalanine to tyrosine, a reaction that resides primarily in the liver. Thus in PKU it is the deficiency of this liver enzyme that results in brain disease. Because of the paucity of findings in the newborn period or early infancy, PKU usually was not diagnosed until late infancy or later until newborn screening was instituted. It is important for physicians caring for newborns to be aware of the pitfalls of screening, which are summarized in Chapter 27. This disease, inherited as an autosomal recessive trait, exemplifies the interaction of a gene and the manipulatable environment (i.e., diet) in the expression of a disease (Scriver et al, 2001). After birth, the baby with PKU who is undergoing normal postnatal catabolism, or ingesting adequate amounts of breast milk or a proprietary formula, will experience a gradual and persistent increase in plasma phenylalanine levels. The cutoff value for newborn screening differs by jurisdiction, but both phenylalanine value and the phenylalanine:tyrosine ratio are considered when calling a presumptive positive. The diagnosis is rarely missed if testing occurs after 24 hours of age. By 5 to 7 days of age in the untreated patient, the eponymous phenylketone is found in urine, and another side product of phenylalanine accumulation, phenylacetic acid, may impart the characteristic “mousey” odor to urine and the patient. This odor may be detected when levels of phenylalanine exceed 600 to 900 μmol/L, but it is not detected in every patient. With modern newborn screening programs, the neonatologist will rarely see a case of PKU in the hospital and will usually be dealing with a patient who has a preemptive diagnosis and is treated as an outpatient. Greater detail about the disorder can be obtained in Scriver et al (2001), but a synopsis will be given here. In the first 6 months of life, the affected babies may have difficulty with feeding and vomiting. In some instances, persistent vomiting has been associated with the diagnosis of pyloric stenosis, for which corrective surgery has been performed, perhaps inappropriately. Developmental delay is usually evident in the second 6 months of life. Patients may have seizures, sometimes infantile spasms in early 226 PART VI Metabolic and Endocrine Disorders of the Newborn infancy associated with a hypsarrhythmic EEG pattern. Persistent elevation of plasma phenylalanine levels greater than 600 μmol/L may be sufficient to result in mental retardation. The mechanism of brain disease in PKU is still unknown. It may be related to the effect of high phenylalanine levels on the transport of amino acids across the blood-brain barrier and then into brain neurons or glial elements. The plasma levels of tyrosine may be decreased, especially in patients who are not receiving tyrosine supplementation in the special amino acid powder used as a daily nutrient. Investigators have speculated for many years that hypotyrosinemia has a role in the CNS deficits, because tyrosine (only a liver enzyme) cannot be synthesized within the CNS by phenylalanine hydroxylase (a liver enzyme). Older infants often exhibit long tract findings, such as spastic quadriparesis and spastic quadriplegia. The untreated infant demonstrates microcephaly acquired postnatally and may also have severe behavioral problems in addition to a diagnosis of autism. Typical findings consist of elevated serum plasma phenylalanine levels, normal or subnormal plasma tyrosine levels, and increased urinary excretion of phenylpyruvic acid, phenyllactic acid, and phenylacetic acid (rarely measured routinely, but easily seen if urine organic acids are studied). The mainstay of therapy is a low-protein diet and the use of a special amino acid–containing formula that does not include phenylalanine. The patient must receive an adequate amount of phenylalanine from protein in proprietary formulas and later from table foods, which is tracked by means of a phenylalanine exchange system, to allow for the normal daily utilization of phenylalanine for protein synthesis while maintaining plasma phenylalanine levels in a range as close as possible to normal, but less than 360 μmol/L. As the patient ages, care-givers may have to be content with levels less than 600 μmol/L. Deviations from normal plasma levels are believed to be associated with chronic, perhaps acute, effects on brain function and testing performance; therefore the diet should be maintained for life. More recently, other forms of therapy have been used. Preparations of the large neutral amino acids that compete for the same transporter as phenylalanine may be given, especially to patients who cannot comply with the low–natural protein diet, and more importantly the mandatory daily ingestion of the medical food devoid of phenylalanine, but replete with unpalatable amino acids. Positive effects have been frequently claimed, but few carefully controlled studies are available. The positive effects of this supplement have variously been attributed to competition with phenylalanine absorption in the gastrointestinal tract, competition for transport into the brain, and repletion of neurotransmitters. A small fraction of patients with full-blown PKU respond favorably to high doses of tetrahydrobiopterin, the natural cofactor for phenylalanine hydroxylase, with a substantial fall in plasma phenylalanine levels and an increased tolerance for natural protein. The more mild forms of hyperphenylalaninemia are even more responsive. The therapy is expensive. Injections or implantations of a reservoir in which there is an enzyme that breaks down phenylalanine in the body are in the early phases of testing. The phenylalanine hydroxylase gene has been cloned and sequenced, and the mutations responsible for most abnormalities in humans are known. DNA sequencing and mutational analysis can be used to identify carriers in families and to provide a scientific rationale during family counseling. In rare cases, patients with hyperphenylalaninemia have severe disease not because of deficiency of the phenylalanine hydroxylase apoenzyme, but because of deficiency of the active cofactor of this enzyme, tetrahydrobiopterin. Several defects in the metabolism of biopterin can produce this type of hyperphenylalaninemia. Patients with these uncommon types of PKU usually come to attention in the newborn period because of severe seizure activity. Patients with biopterin deficiency can have evidence of severe brain damage despite treatment with a low-phenylalanine diet. Brain damage may be related to deficiencies of other neurotransmitters whose synthesis also depends on adequate levels of tetrahydrobiopterin. These various other defects, such as the dihydropteridine reductase, 6-pyruvoyl tetrahydropterin synthetase, and the guanosine triphosphate cyclohydrolase deficiencies, can be ascertained by urinary measurements of neopterin and biopterin in urine. They are responsive to therapy with neurotransmitter replacement and tetrahydrobiopterin. If women with PKU are poorly treated during pregnancy, their children may be born with microcephaly and congenital heart defects. Mental retardation is common. Adequate treatment before 8 weeks’ gestation is essential, and control of phenylalanine levels before pregnancy is most desirable. INBORN ERRORS OF ORGANIC ACID METABOLISM Defects in the catabolism of the BCAAs are responsible for most of the disorders of organic acid metabolism (Figure 22-5). Typical examples are methylmalonic, propionic, and isovaleric acidemias. An organic acid is any organic compound that contains a carboxy functional group but no α-amino group as in an amino acid. In this section, we consider the disorders of fatty acid oxidation, ketone body metabolism, and lactic acid metabolism as well as the more classic inherited defects in organic acid metabolism. Until recently, these disorders were diagnosed only after symptoms have appeared. It is now more common to be confronted with an asymptomatic or early symptomatic patient and a good idea of the diagnosis. As discussed earlier, acidosis and encephalopathy are usually the hallmarks of these syndromes when they manifest in the newborn period. METHYLMALONIC ACIDEMIA l-Methylmalonyl–Coenzyme Mutase Deficiency A Methylmalonic acidemia, along with propionic acidemia, is thought to be the most common of disorders of organic acid metabolism (Fenton et al, 2001; Wendel and de Baulny, 2006). Although more than one enzyme defect may result in methylmalonic acidemia, all are inherited as autosomal recessive traits. In California, defects in cobalamin metabolism are more frequent than apoenzyme deficiency, specifically l-methylmalonyl–coenzyme a (CoA) CHAPTER 22 Inborn Errors of Carbohydrate, Ammonia, Amino Acid, and Organic Acid Metabolism 227 FIGURE 22-5 The pathways, metabolites, and vitamin cofactors important in the interconversion of methionine and homocysteine are shown in this abbreviated scheme of the trans-sulfuration pathway. In a methyl-transfer reaction, homocysteine is converted to methionine; the reaction is catalyzed by a cobalamin-containing enzyme, 5-methyl-tetrahydrofolate-homocysteine methyltransferase or methionine synthase. An alternate reaction involves betaine-homocysteine methyltransferase. Therefore vitamin B12 and folate are important vitamins in homocysteine and methionine metabolism. The 5,10-methylene tetrahydrofolate (THF) derived from dietary folate is converted to 5-methyl THF by the enzyme 5,10-methylene THF reductase. In classic homocystinuria, the deficient enzyme is the cystathionine β-synthetase, which catalyzes the conversion of homocysteine to cystathionase, a precursor of cysteine. This pathway is readily operative in the adult, so that cysteine is a nonessential amino acid. In newborn infants, however, synthesis of cysteine does not occur at the same rate as in adults. Adenosylmethionine-dependent methyl transfer may be extremely important in the central nervous system. Most of the patients who come to clinical attention in the newborn period have defects in the remethylation of homocysteine to methionine; therefore these patients may have a severe deficiency of adenosylmethionine in the brain and spinal cord. mutase (see Figure 22-4). This enzyme is present in mitochondria, and it depends on adenosyl-cobalamin for activity. Impaired function can result from either a mutation of the l-methylmalonyl-CoA mutase apoenzyme or deficient availability of the adenosyl form of vitamin B12. The latter may result from impaired cellular metabolism of vitamin B12, including defective activity of the enzyme adenosylcobalamin synthetase. Some patients, but not usually newborn infants with methylmalonic acidemia caused by apoenzyme deficiency, are responsive to pharmacologic therapy with vitamin B12. Methylmalonic acidemia, therefore, is one of the important disorders that can be considered a vitaminresponsive inborn error of metabolism. Because of deficient activity of ­l-methylmalonyl-CoA mutase, the substrate l-methylmalonyl-CoA accumulates in mitochondria and is subsequently hydrolyzed to methylmalonic acid. Methylmalonic acid is capable of diffusing out of cells in which it is being produced, so it may be detected in excess in blood, CSF, and urine of patients with these various forms of the disease. The precursors of l-methylmalonyl-CoA are the BCAAs isoleucine and valine, in addition to methionine, threonine, thymine, and odd-chain fatty acids (see Figure 22-4). There are many different phenotypes of methylmalonic acidemia that range from severe, catastrophic newbornonset disease in the first week of life to an almost benign form that has been detected in adults with partial l-methylmalonyl-CoA mutase deficiency. Most patients in the newborn period have a severe phenotype. Dialysis therapy is not effective in some of the patients in the first week of life with catastrophic illness, possibly because of a delay in diagnosis and treatment or the breakdown of other systems, including the heart and circulation. The most striking presentation is in the second or third day of life. The baby is usually well at birth, as with UCDs, but then gradually begins to manifest problems with feeding, vomiting, lethargy, and perhaps seizures. There may be respiratory distress as a symptom of metabolic acidosis. The important laboratory findings include metabolic acidosis usually associated with an increased anion gap, ketosis, and hyperammonemia. The elevation of plasma ammonium may be as high as in severe newborn-onset hyperammonemic syndromes. Because ketonuria is relatively uncommon in newborn infants, even in stressed infants with hypoglycemia caused by poor feeding, and because diabetes mellitus is so uncommon in the newborn period, the physician caring for newborn infants must always consider an inborn error of organic acid metabolism when confronted with an acutely ill baby with ketosis. Other laboratory findings are thrombocytopenia, leukopenia, and anemia caused by effects of the metabolite on hematopoietic elements in bone marrow. Plasma amino acid analysis may reveal elevations of several amino acids, such as glycine and alanine. Secondary carnitine deficiency with elevations of carnitine esters is expected (Chalmers et al, 1984). Methylmalonic acid is detected at high levels in urine with GC-MS. Older patients with methylmalonic acidemia during crisis have suffered acute pancreatitis and devastating lesions to the basal ganglia. The diagnosis can be confirmed with an assay of the activity of l-methylmalonyl-CoA mutase enzyme in cultured skin fibroblasts. In addition, various other disturbances in vitamin B12 metabolism can be studied by analyzing skin fibroblasts in culture. It is more common 228 PART VI Metabolic and Endocrine Disorders of the Newborn to dissect the condition clinically, ascertain the response to vitamin B12, and use mutation analysis to confirm the exact site of the genetic lesion. The treatment of acute disease consists of protein restriction, empirical therapy with vitamin B12 (1 mg/day intramuscularly), intravenous fluids with 10% glucose and sodium bicarbonate to correct dehydration, electrolyte imbalance and acidosis, high-calorie feeds via a nasogastric tube, and often dialysis. The use of carnitine (25 to 200 mg/kg/day intravenously or orally) is often recommended, but its usefulness as an acute treatment has not been proved. The treatment of chronic state centers on the judicious use of a low-protein diet, an amino acid supplement low in the MMA precursors, carnitine to alleviate free carnitine deficiency, and appropriate calories and fluid. Vitamin B12 is used only when a specific and reproducible response is noted. Most diagnoses are ascertained by expanded newborn screening. If the results are rapidly available, the outlook for the neonatal period may be improved. Episodes of decompensation will still occur. Patients who are particularly brittle or who develop renal failure receive either liver, kidney, or combined liver-kidney transplants, but their efficacy at curing the primary disease process has not been proved. PROPIONIC ACIDEMIA Propionyl Coenzyme A Carboxylase Deficiency Propionic acidemia was the first classic defect in organic acid metabolism to be described in humans (Fenton et al, 2001; Wendel and de Baulny, 2006). The first patient, described by Childs et al in 1961, was sick on the first day of life with severe metabolic acidosis and ketosis. He responded to massive alkali therapy and survived the newborn period. He subsequently suffered multiple episodes of metabolic decompensation with ketoacidosis, usually precipitated by infections or protein ingestion. He also had developmental delay, a seizure disorder, and episodic neutropenia and thrombocytopenia. He died at 7 years old. His sister also demonstrated ketosis and metabolic acidosis in the first week of life, but because of better control of her disease, there were few severe episodes of decompensation in the first and second decades of life. Propionic acidemia, which is caused by a selective deficiency of propionyl-CoA carboxylase, is inherited as an autosomal recessive trait. This disorder was originally called ketotic hyperglycinemia, because of elevations in plasma glycine along with ketosis. The precursors of propionyl-CoA include the amino acids isoleucine, valine, methionine and threonine, thymine, and odd-chain fatty acids (see Figure 22-4). Of the several hundred patients described with this disease, most clinical presentations are in the newborn period with poor feeding, vomiting, lethargy, and hypotonia. The patients commonly have seizures and hepatomegaly. The metabolic acidosis may be severe with or without an increase in the anion gap. Ketosis is usually present, but not always. Patients who survive may often show choreoathetosis, because of persistent damage to the basal ganglia. Usually episodes of metabolic decompensation characterized by acidosis and ketosis can be precipitated by excessive protein intake or more commonly infection, with the attendant catabolism. Such episodes can result in permanent neurologic damage. Subsequent findings therefore include developmental delay, seizures, cerebral atrophy, and EEG abnormalities. During the acute attacks, leukopenia, thrombocytopenia, and less rarely anemia, probably caused by suppression of maturation of bone marrow hematopoietic precursors, may be exacerbated. The diagnosis can be confirmed with GC-MS analysis. The urine has excess concentrations of various propionate metabolites, such as methyl citrate, propionylglycine, 2-methyl-3-hydroxybutyrate, 2-methylacetoacetate, and several other, rarer compounds. The plasma glycine value may be elevated, and during acute attacks the plasma ammonium value is frequently increased. The enzyme activity of propionyl-CoA carboxylase can be assayed in white blood cells or extracts of cultured skin fibroblasts for definitive diagnosis. Numerous mutations have been described in the two genes that encode the subunits of this multimeric enzyme. Therapy consists of a low-protein diet and adequate calories. As in MMA, there is secondary carnitine deficiency with elevations of propionylcarnitine. The use of l-carnitine to relieve a deficiency of free carnitine and promote greater urinary excretion of propionylcarnitine to lower mitochondrial propionyl-CoA levels is not proved. Because intestinal bacteria can also contribute to propionate production, antimicrobial therapy with metronidazole has been used during an acute attack and for longterm therapy. In the acutely ill newborn, the immediate treatment consists of elimination of protein, total parenteral nutrition, administration of an adequate amount of calories (10% glucose intravenously or a nonprotein formula, or both, via nasogastric tube infusion), administration of alkali to eliminate metabolic acidosis, and platelet transfusion if warranted by thrombocytopenia. Patients frequently require hemodialysis, because of severe acidosis and coma. It is unclear whether administration of sodium benzoate and sodium phenylacetate, as in the treatment of hyperammonemia associated with UCDs, is of benefit to the patient with secondary hyperammonemia. Several patients with propionic acidemia have undergone liver transplantation with mixed success. Episodes of acidosis are often recurrent between 1 and 3 years of age, after which some ability to modulate the effects of catabolism occurs. Most patients with this condition are now ascertained by expanded newborn screening. If the results are rapidly available, the outlook for the neonatal period may be improved. Episodes of decompensation will still occur. ISOVALERIC ACIDEMIA Isovaleryl–Coenzyme A Dehydrogenase Deficiency Isovaleric acidemia is caused by a selective deficiency of the enzyme isovaleryl-CoA dehydrogenase (Sweetman and Williams, 2001; Wendel and de Baulny, 2006). It is inherited as an autosomal recessive trait. There are two major phenotypes; an acute form manifests as catastrophic disease in the newborn period, and the late-onset type is characterized by CHAPTER 22 Inborn Errors of Carbohydrate, Ammonia, Amino Acid, and Organic Acid Metabolism chronic, intermittent episodes of metabolic decompensation. The degree of enzyme deficiency and the mutations differ between the two extreme presentations. In the acute form, the infants become extremely sick in the first week of life. There is usually a history of poor feeding, vomiting, lethargy, and seizures. The characteristic sweaty feet or rancid cheese odor caused by isovaleric acid is noted on the body or in urine, especially if it is acidic. Metabolic acidosis is present, usually with an elevated anion gap and ketosis. There may be secondary hyperammonemia, thrombocytopenia, neutropenia, and sometimes anemia, resulting in pancytopenia. The babies usually lapse into a coma. Dialysis therapy may be necessary. As with other organic acid disorders for which an amino acid determines organic acid production (see Figure 22-4), treatment also consists of protein restriction, intravenous fluids with glucose, and perhaps sodium bicarbonate, protein-free formula with calories via nasogastric tube, glycine (250 mg/kg/day), and carnitine supplementation (Sweetman and Williams, 2001). Intravenous l-carnitine may be beneficial. The excretion of isovaleric acid as the glycine conjugate is highly efficient, and symptomatic relief can occur rapidly. Glycine is often available in compounding pharmacies that make their own IV alimentation solutions. In the chronic, intermittent form of isovaleric acidemia, patients have repeated episodes of metabolic decompensation precipitated by infections, primarily or excessive protein intake. Some of these episodes may mimic Reye syndrome. The same therapeutic principles are applied as for the treatment of the neonatal disorder. The mainstay of long-term therapy is a diet with limited natural protein, a valine-free amino acid supplement, and long-term administration of glycine (Berry et al, 1988), which enhances the production of the nontoxic compound isovalerylglycine and serves to reduce the free levels of isovaleric acid in body fluids. In addition, carnitine administration can augment the excretion of isovaleryl carnitine (Berry et al, 1988; Mayatepek et al, 1991). The benefit of carnitine treatment in the chronic state, however, remains unproved. Some patients who remain largely asymptomatic are ascertained through the expanded newborn screening programs. Newborn screening and the rapid treatment protocols can alter the outlook for this condition. In the prescreening era, many patients with neonatal onset could not be saved, and a number of those with the more indolent forms have been retarded. This situation should change with preemptive treatment. The diagnosis can be made from marked elevations of isovalerylglycine in urine. There is usually also increased excretion of 3-hydroxyisovaleric acid. The enzyme ­isovaleryl-CoA dehydrogenase can be assayed in extracts of cultured skin fibroblasts. As with many other conditions, the typical metabolic profile should allow confirmation directly by mutation analysis bypassing the enzyme assay. MULTIPLE CARBOXYLASE DEFICIENCY There are two enzymatic defects leading to deficiency of the same suite of carboxylase enzymes, holocarboxylase synthetase deficiency and biotinidase deficiency (Wolf, 2001). All the carboxylase enzymes—propionyl-CoA carboxylase, pyruvate carboxylase, 3-methylcrotonyl-CoA 229 carboxylase, and acetyl-CoA carboxylase—require covalent linkage with biotin for normal activity. In holocarboxylase synthetase deficiency, the enzyme that catalyzes the covalent linkage of biotin to a lysine residue of these various carboxylases is deficient. In the second form of the disorder, the absence of the enzyme biotinidase does not allow for biotin recycling after a carboxylase enzyme is degraded, and hydrolyzed biotinidase deficiency does not usually manifest in the newborn period. Patients with severe deficiency in holocarboxylase synthetase, however, characteristically are catastrophically ill in the newborn period. Patients have severe metabolic acidosis with lactic acidosis and become comatose. As with biotinidase deficiency, administration of biotin is life saving. Holocarboxylase synthetase deficiency is one of the few disorders of organic acid metabolism for which the administration of a vitamin in megadoses produces a dramatic turnabout in clinical and laboratory findings. Diagnosis depends on GC-MS analysis of urine and the demonstration of markedly increased levels of lactate, 3-methylcrotonylglycine, and propionate metabolites. In most patients the affinity of the holocarboxylase synthetase enzyme for biotin is diminished, and for the biochemical disturbances to normalize or improve, patients need between 10 and 60 mg of biotin daily. The enzyme deficiency can be confirmed in cultured skin fibroblasts or by mutation analysis. The disorder is inherited as an autosomal recessive trait. Both biotinidase deficiency, ascertained by direct enzyme assay, and holocarboxylase synthetase deficiency are detected in the expanded newborn screening panels. The outcomes should alter because of this earlier diagnosis. GLUTARIC ACIDEMIA TYPE 1 Glutaryl–Coenzyme A Dehydrogenase Deficiency An isolated deficiency of glutaryl-CoA dehydrogenase causes glutaric acidemia type 1 (GA-1) (Goodman and Frerman, 2001). Multiple phenotypes are known. In the most dramatic presentation, which accounts for fewer than half of the known patients with GA-1, the illness develops acutely in the first year of life, usually after an infection or other catabolic event. Acute encephalopathy is followed by the development of what appears to be a severe form of extrapyramidal cerebral palsy (Hoffmann et al, 1991). The affected infants have incurred bilateral damage to the caudate and putamen in the basal ganglia, resulting in an incapacitating dystonic syndrome. Some patients have a slowly progressive course with developmental delay, hypotonia, dystonia, and dyskinesia in the first 2 years of life. Other patients are relatively asymptomatic. In general, GA-1 is not a disorder that is associated with acute disease in the newborn period. However, macrocephaly at birth is common (Goodman and Frerman, 2001). The etiology of any of the CNS lesions is unknown, but toxicity caused by dicarboxylic acid accumulation has been postulated. Magnetic resonance imaging (MRI) of the head typically shows bilateral widening of the Sylvian fissures associated with hypoopercularization, resulting in the “bat-wing” appearance. Sometimes the fluid accumulation mimics subdural hygromas, and a few patients have 230 PART VI Metabolic and Endocrine Disorders of the Newborn been noted to have subdural hematomas. The congenital nature of these findings suggests that GA-1 has its onset in utero and affects CNS development. Usually infants do not demonstrate bilateral damage to the basal ganglia in the newborn period. Diagnosis depends on the demonstration of glutaric acid and 3-hydroxyglutaric acid in urine and is confirmed by the demonstration of deficiency of glutaryl-CoA dehydrogenase activity or protein levels on Western blot analysis in cultured fibroblasts, or more commonly by mutation analysis. Newborn screening may be complicated in this disorder. Whereas the majority of patients have positive findings of elevated C5DC in the newborn acylcarnitine profile, a minority with severe disease may have normal or near normal results. These same patients might not show the characteristic elevation of 3-hydroxyglutarate in urine. For this reason an initial elevation of C5DC in urine can and should elicit the close attention of a metabolic specialist who then must help to interpret results of the follow-up testing. A rare disorder, GA-1 is especially common in SaulteauxOjibway Indians of Canada (Greenberg et al, 1995) and in the Old-Order Amish of Lancaster County, Pennsylvania (Morton et al, 1991). A low-protein diet can help in the treatment of these patients; acute illnesses, usually viral in nature, should be treated vigorously with fluids containing glucose and adequate amounts of calories to prevent or diminish catabolism. Bicarbonate may be necessary to correct acid-base imbalance. Carnitine has also been used to reduce mitochondrial glutaryl-CoA levels. FATTY ACID OXIDATION DISORDERS Glutaric Acidemia Type 2 or Multiple Acyl Coenzyme A Dehydrogenase Deficiency Glutaric acidemia type 2 (GA-2) is characterized by deficiency of multiple acyl-CoA dehydrogenase enzymes (Goodman and Frerman, 2001; Stanley et al, 2006). All these dehydrogenase proteins have in common the binding of a protein called electron transfer flavoprotein (ETF). This protein is responsible for accepting electrons in any of these oxidative dehydrogenation reactions from the flavin adenine dinucleotide (FAD) cofactor. A mutation in any of the three genes that encode the protein subunits of either ETF or the ETF dehydrogenase, which is responsible for further transferring the electrons from ETF to coenzyme Q10 within the mitochondria leads to this condition. Deficiency of riboflavin, a component of FAD, or its transport may also lead to this condition. Among the compounds accumulating in this condition are short-, medium-, and long-chain fatty acids, glutaric and isovaleric acids, dimethylglycine, and sarcosine. The three major phenotypes of GA-2 are (1) a newbornonset type with congenital anomalies, (2) a newborn-onset type without anomalies, and (3) a milder or later-onset type, sometimes called mild acyl-CoA dehydrogenase deficiency or ethylmalonic adipic aciduria. Severely affected newborn patients with GA-2 often have multiple malformations, may be premature, and are usually ill within the first week of life (Frerman and Goodman, 2001). The patients demonstrate hypotonia, encephalopathy, hepatomegaly, hypoglycemia, and metabolic acidosis; often the odor of isovaleric acid is present. There may be facial dysmorphism consisting of a high forehead, low-set ears, hypertelorism, and a hypoplastic midface. The kidneys may be palpably enlarged, associated with large renal cysts; they may also have rocker-bottom feet, muscular defects of the inferior abdominal wall, and anomalies of the external genitalia, including hypospadias and chordee. Most of the patients with GA-2 and multiple malformations do not survive the first weeks of life. In some, the malformations are not so noticeable, and only renal cysts are identified at autopsy. Some of these patients have cardiomyopathy. In contrast, some of these infants can survive the newborn period. The phenotype in the third form of GA-2 is highly variable. Some patients are relatively free of disease and have intermittent episodes of vomiting, dehydration, hypoglycemia, and acidosis during childhood or adult life. In some, there may be hepatomegaly and muscle disease. Laboratory studies in patients with the newborn-onset type usually show severe metabolic acidosis with lactate elevation and increased anion gap, mild to moderate hyperammonemia, and hypoglycemia without a moderate or large ketonuria. The liver function and dysfunction test results may be abnormal, with increases in serum transaminases and prolongations of PT and PTT. A chest radiograph may show heart enlargement because of hypertrophic cardiomyopathy. Abdominal ultrasonography or computed tomography may reveal renal cysts. The diagnosis is made from the characteristic pattern of organic acid metabolites on urine GC-MS analysis. It consists of glutarate; ethylmalonate; 3-hydroxyisovalerate; 2-hydroxyglutarate; 5-hydroxyhexanoate; adipic, suberic, sebacic, and dodecanedioic acids; isovalerylglycine; isobutyrylglycine; and 2-methylbutyrylglycine. The ketones, acetoacetic acid, and 3-hydroxybutyric acid are usually not present or are only minimally elevated if present, being inappropriate for the degree of fatty acid metabolites that indicate free fatty acid mobilization and the potential for enhanced ketogenesis. The renal cystic disease may also be associated with evidence of impaired renal tubular function. Generalized aminoaciduria may be present. The amino-containing compound sarcosine may be elevated in serum as well as in urine. There may be secondary carnitine deficiency, and abnormal carnitine esters such as isovalerylcarnitine and butyrylcarnitine can be detected in blood. The ETF and ETF dehydrogenase deficiency can be detected through the use of specific antibodies on Western blot analysis of cultured skin fibroblasts or of functional assays in skin cells. Mutation analysis in this condition as in others may be superseding the more cumbersome enzyme studies. Prenatal diagnosis can be achieved by demonstration of glutarate in amniotic fluid, from results of dehydrogenase assays in cultured amniocytes, or by mutation analysis. Despite much debate, criteria for newborn screening have not been well established, and many false positive tests have occurred for this rare disorder. Although treatment with intravenous glucose, riboflavin, carnitine, and diets low in protein and fat generally have not been successful for catastrophically ill newborns, there has been some success in patients with milder or lateronset disease. Riboflavin is suggested to be administered to CHAPTER 22 Inborn Errors of Carbohydrate, Ammonia, Amino Acid, and Organic Acid Metabolism the newborn with severe disease, but riboflavin at a dose of 100 to 300 mg/day has been effective in only a few older patients (Gregersen et al, 1982; Harpey et al, 1983). The rationale for this therapy is that riboflavin, being a precursor of FAD, increases the concentrations of FAD and allows for better interaction with mutated and defective ETF or ETF dehydrogenase proteins. Finally, some artificial electron acceptors such as methylene blue have been used in the newborn period without success. VERY-LONG-CHAIN, MEDIUM-CHAIN, AND SHORT-CHAIN ACYLFATTY ACID OXIDATION DEFECTS Fatty acids are oxidized in a complex process by which they are taken up into the cell and transported into the mitochondrion for beta oxidation and energy production. There are at least 31 enzymes involved in the process and defects in the majority have been described. Of these many disorders, the most frequent are those of carnitine metabolism and transport, and various steps in the beta oxidation pathways. The presentations are sufficiently different to bear independent description, but in general they all have hypoketotic hypoglycemia and energy deficits. Most spare the CNS. When severe, the defects have been associated with coma, hypoglycemia, liver disease, cardiac, and skeletal myopathy (Roe and Ding, 2001; Stanley et al, 2006). All of these defects involve abnormalities in enzymes that participate in or facilitate the mitochondrial beta oxidation of fatty acids. In general the pathophysiology associated with these disorders has the potential to put the patient in a life-threatening condition in which there is a state of catabolism and enhanced liberation of free fatty acids by adipose stores. The most prevalent disorders will be discussed here. Medium-Chain Acyl-Coenzyme A Dehydrogenase Deficiency The most common of these disorders is the medium-chain acyl-CoA dehydrogenase (MCAD) deficiency, with a frequency of approximately 1 in 20,000 births in Northern European populations (Roe and Ding, 2001; Stanley et al, 2006). Most patients’ disease does not manifest until late infancy, but some with the most severe enzyme deficiencies manifest in the first days of life and may die without explanation, something recently appreciated with the advent of expanded newborn screening. It is also a well-known cause of sudden infant death syndrome having been found in 2% to 3% of cases in which it has been investigated. The typical patient is an older infant who, after an infection, experiences anorexia, vomiting, dehydration, lethargy, and hypoglycemia that may be associated with seizures. Similarly, older patients have features that mimic those of Reye syndrome and can die because of brain edema. In MCAD deficiency, the initial episode is associated with a high mortality rate. If not diagnosed by newborn screening, the laboratory studies usually show hypoglycemia and an absence of moderate to large ketones in urine that would be expected to accompany hypoglycemia. The plasma ammonium values may be mildly elevated, the liver 231 may be enlarged, and serum ALT and AST levels may be slightly increased. During an acute episode, urine GC-MS analysis of organic acids characteristically shows increased levels of adipic, suberic, and sebacic acids as well as their unsaturated and hydroxylated analogues. However, this result is not pathognomonic of a disorder of fat metabolism, because infants who are fed medium-chain triglyceride–enriched formulas also show dicarboxylic aciduria. Acylcarnitine studies of plasma from patients with MCAD deficiency demonstrates increased levels of C8, C6, and C10 species. Urine also contains glycine conjugates such as suberylglycine and hexanoylglycine. A secondary carnitine deficiency can be present. The MCAD enzyme can be assayed in cultured skin fibroblasts, but mutation analysis is the more common means of confirming the diagnosis. After the MCAD gene was cloned and sequenced, a large number of more common and private mutations were identified. The most conspicuous gene is R329E, a highly prevalent mutation in people of Northern European ancestry and known to cause a severe enzyme deficiency. The major treatment is avoidance of fasting for more than 12 hours, especially in association with an intercurrent illness. Some practitioners are more conservative and advocate fasting of no more than 4 to 6 hours in the first year of life, whether sick or well. Although not of proven efficacy, many physicians recommend a moderately fat-restricted diet for these patients. It is thought that once the disease is diagnosed and preventive programs are instituted, the disease should not be life threatening or fatal. With the advent of newborn screening, a larger group of patients has been found and some more common mutations in these patients have never been seen in a case ascertained because of symptoms. Finding these individuals has given rise to the notion that many patients found by newborn screening are destined never to become symptomatic. Whereas ascertainment of such mutations can be reassuring, too little is known about them to allow complete reassurance and full relaxation of preventive precautions. Newborn screening also finds individuals heterozygous for a severe mutation when no other mutation is found. These patients are considered to be biochemically manifesting heterozygotes and are thought to be at no medical risk from this condition. Very-Long-Chain Acyl-Coenzyme A Dehydrogenase Deficiency Most of the patients with a previous diagnosis of long-chain acyl-CoA dehydrogenase (LCAD) deficiency (Roe and Ding, 2001; Stanley et al, 2006) actually had very-long-chain acyl-CoA dehydrogenase (VLCAD) deficiency (Yamaguchi et al, 1993). Patients with this disease may be ill in the newborn period because of liver disease with hypoglycemia, cardiomyopathy, and skeletal myopathy. The membranebound VLCAD—as opposed to the soluble LCAD, whose specific metabolic role is unknown—is the main enzyme for initiating oxidation of free fatty acids that are derived from adipose stores, such as palmitic, stearic, and oleic acids. The fasting state may include coma. In its absence, however, the patients may exhibit hypotonia, hepatomegaly, and cardiomegaly. With an acute metabolic decompensation, urine organic acid analysis may demonstrate dicarboxylic aciduria. There may be a secondary carnitine deficiency with 232 PART VI Metabolic and Endocrine Disorders of the Newborn increased concentrations of long-chain fatty acids (LCFAs) bound to carnitine. VLCAD deficiency can be fatal; sudden death in early infancy has been reported. Therapy is directed toward replenishment of glucose, calorie administration, and treatment of any potential brain edema. The myopathy and cardiomyopathy, however, may proceed even in the absence of fasting. Most investigators suggest a reduction in dietary fat intake, supplementation in medium-chain triglycerides, and maintenance of an adequate intake of essential fatty acids. Diagnosis can be made through enzyme assay in cultured skin fibroblasts, but commonly by mutation analysis. Expanded newborn screening has expanded our horizons in this disorder as it has for MCAD deficiency. In addition to patients who may be homozygous for two mild mutations or compound heterozygous for a mild mutation, and who are at much reduced clinical risk, heterozygotes for a pathologic mutation and a wild type gene have been found and are presumably at no greater risk than others in the population. Short-Chain Acyl-Coenzyme A Dehydrogenase Deficiency The short-chain acyl-CoA dehydrogenase deficiency is a disorder mired in controversy (Roe and Ding, 2001; Stanley et al, 2006). Patients with a severe enzyme deficiency are infrequent, and the majority of them have no symptoms attributable to the disorder, as judged by the variability of symptoms in those ascertained symptomatically and the absence of obvious disease in similarly affected siblings. The situation is further complicated by extensive mutation analysis, especially in those ascertained through positive newborn screens. Because of the prevalence of common polymorphisms that impair but do not eliminate enzymatic activity, the majority of ascertained individuals have no symptoms and appear unlikely to develop them. This problem has led some people to declare the disorder harmless and to pay no attention to it. Other more sober voices accept it as a benign disease and disorder in the majority of patients, but recognize the possibility that a small minority may become symptomatic and treat them appropriately by warning the parents to react to unusual and unexpected symptoms. There tends to be some acrimony at meetings from members of the most extreme positions on either end of the spectrum. Acylcarnitine analysis on either newborn screen or later shows elevated C4 carnitine. Organic acid analysis shows elevated levels of ethylmalonic and methylsuccinic acids, with higher levels tending to occur with greater enzyme deficiencies. Diagnosis can be made by assaying the shortchain acyl-CoA dehydrogenase enzyme in cultured skin fibroblasts and by a fatty acid oxidation profile in these same cells. Mutation analysis is now the norm for this condition. Long-Chain 3-Hydroxy Acyl-Coenzyme A Dehydrogenase Deficiency The long-chain 3-hydroxy acyl-CoA dehydrogenase deficiency is associated with acute illness, fasting-induced hypoglycemia, hypoketosis, cardiomegaly, and muscle weakness (Roe and Ding, 2001; Stanley et al, 2006). As with VLCFA oxidative abnormalities, some older patients may have episodes of illness associated with elevated serum creatine phosphokinase levels and myoglobinuria. A few patients have had sensory motor neuropathy and pigmentary retinopathy. Half the patients do not survive. Some patients may have severe liver disease with fibrosis in addition to necrosis and steatosis. Women who are carriers for this disease may also manifest the HELLP syndrome when carrying an affected child. The diagnosis was made in symptomatic individuals and was suggested by symptoms and demonstration of longer chain 3-hydroxydicarboxylic acids on urine organic acid analysis. Enzymatic diagnosis can be made in lymphocytes or in skin fibroblasts. More commonly the diagnosis is confirmed by a combination of clinical biochemical abnormalities and DNA mutation analysis. The majority of moderate to severe cases are diagnosed by expanded newborn screening and follow-up. Treatment of this disorder has involved frequent highcarbohydrate feedings, dietary fat restriction, and supplementation with uncooked cornstarch. Administration of medium-chain triglycerides may be helpful. Carnitine and riboflavin have also been tried without benefit. A high level of parental vigilance is required to begin therapy quickly at signs of metabolic decompensation. Liver and neurologic disease can progress despite any intervention. The outcome of severely affected patients is guarded. Carnitine Transporter Defect and Deficiencies of Carnitine Palmitoyltransferase I and II and Acylcarnitine Translocase Carnitine is essential for fatty acid oxidation because transport of LCFAs into mitochondria depends on an adequate amount of carnitine and the presence of two enzymes that covalently link carnitine to fatty acid or remove the linkage and one transporter that carries it across the inner mitochondrial membrane (Roe and Ding, 2001; Stanley et al, 2006). These enzymes are carnitine palmitoyltransferase I and II and a carnitine translocase. Cellular levels of carnitine in turn depend on a sodium-dependent carnitine transporter. Primary carnitine deficiency appears to be associated with a carnitine transporter defect (Roe and Ding, 2001; Stanley et al, 2006). Patients with the transporter defect may present with symptoms in infancy or in childhood, but rarely in the newborn period. Earliest reports concern the extended newborn period or early infancy. The disease is characterized by hypoketotic hypoglycemia, hyperammonemia, elevations of transaminases, cardiomyopathy, and skeletal muscle weakness. In some of the older patients, cardiomyopathy may be the presenting sign. The characteristic laboratory finding in this disease is extremely low plasma carnitine levels. The total carnitine levels are usually less than 10 μmol/L in plasma. A dicarboxylic aciduria is not usually evident on urine organic acid analysis. Newborn screening has altered our view of this disorder. Most cases are ascertained with low carnitine on screening and are usually diagnosed with mutation analysis and the failure of carnitine levels to rise in the postnatal period. Most physicians treat the disorder with carnitine to raise CHAPTER 22 Inborn Errors of Carbohydrate, Ammonia, Amino Acid, and Organic Acid Metabolism plasma carnitine levels while the evaluation is in progress. Newborn screening has also ascertained asymptomatic and affected mothers and has altered our perception of this condition and its severity. In addition, largely asymptomatic mothers affected with an organic acidemia and low carnitine levels in the mother and fetus have been detected. This disorder is the only one in which pharmacologic administration of carnitine has dramatic effects on the clinical and laboratory abnormalities. The treatment is 100 to 200 mg/kg/day of l-carnitine. Repletion of plasma carnitine levels is the benchmark by which the efficacy of treatment is judged, but the degree of repletion of tissue levels (skeletal muscle and heart) is inferred but rarely demonstrated. Treatment is successful in the short, intermediate, and longer terms, but no information on the lifespan of these patients who appear healthy is available. Carnitine Palmitoyltransferase Type I Deficiency Carnitine palmitoyltransferase type I (CPT-I) is responsible for covalently linking LCFAs such as palmitate to carnitine. Although one patient with deficiency of CPT-I came to attention in the newborn period, most come to attention in early to late infancy (Roe and Ding, 2001; Stanley et al, 2006). The clinical findings are hypoketotic hypoglycemia, encephalopathy, and hepatomegaly; there is usually no evidence of cardiomyopathy or skeletal myopathy in CPT-I deficiency. Renal tubular acidosis, which is caused by impaired distal hydrogen ion secretion, has rarely been reported. Characteristic laboratory findings are the absence of dicarboxylic aciduria and high plasma levels of total carnitine and free carnitine; however, the plasma acylcarnitine profile is not abnormal. The definitive diagnosis rests on measuring CPT-I enzyme activity in cultured skin fibroblasts or mutation analysis. Frequent feeding, reduction of dietary fat, supplementation with medium-chain triglycerides, and avoidance of fasting all have been beneficial in the long-term management of patients with CPT-I deficiency. These patients are now often found by expanded newborn screening because of an abnormal ratio of carnitine to the carnitylated long-chain fatty acids. Acylcarnitine Translocase Deficiency Acylcarnitine translocase deficiency is an exceedingly rare defect. It was initially reported in a male infant who suffered a cardiac arrest at 36 hours of age in association with fasting stress and ventricular dysrhythmias (Roe and Ding, 2001; Stanley et al, 2006). Because of the failure to transport long-chain acylcarnitines across the intermitochondrial membrane after the synthesis by CPT-I, the patient had very low total plasma levels of carnitine, most of which was long-chain esterified carnitine; he also had recurrent episodes of hypoglycemia, vomiting, gastroesophageal reflux, and mild chronic hyperammonemia as well as severe skeletal myopathy and mild hypertrophic cardiomyopathy. The continuous nasogastric feeding of a low-fat, high-carbohydrate formula failed to normalize clinical abnormalities, and the patient died at 3 years of age. At that time, liver failure had also developed. Pathophysiologic findings in this patient suggested that accumulation of long-chain acylcarnitine species may be 233 toxic for several organs, including heart, liver, and skeletal muscle. In addition, the acute development of ventricular dysrhythmias may be related to accumulation of such species in cardiac tissue. It is often considered to be almost invariably fatal or very serious, but the probability is that more mild cases with reduced but not absent transporter activity have not been found. Carnitine Palmitoyltransferase Type II Deficiency There are two phenotypes of carnitine palmitoyltransferase type II (CPT-II) deficiency (Roe and Ding, 2001; Stanley et al, 2006). The enzyme is responsible for the hydrolysis of LCFA bound to carnitine after transport across the intermitochondrial membrane. Most of the patients reported with CPT-II deficiency have a mild deficiency of the enzyme and in adulthood exhibit episodes of muscle weakness and myoglobinuria brought on by prolonged exercise. The other phenotype is caused by a more serious enzyme defect and manifests in early infancy. The first detailed report was of a 3-month-old boy with hypoketotic hypoglycemia, coma, seizures, hepatomegaly, cardiomegaly, cardiac arrhythmias associated with an increase in long-chain acylcarnitine levels in tissues, and the absence of urinary dicarboxylic aciduria. The enzyme is also expressed in renal tissue and skeletal muscle. Renal dysgenesis has been noted in three patients (Zinn et al, 1991). Reports of onset in the newborn period associated with an abnormal physical appearance and MRI (Hug et al, 1991) and in later infancy have been recorded. Several mutations have been described. Decreased activity of the CPT-II enzyme may be demonstrated in cultured skin fibroblasts. DEFECTS IN KETONE METABOLISM After LCFAs are broken down first to medium-chain and finally to short-chain fatty acids such as acetoacetyl CoA, they must be converted in the liver to 3-hydroxy3-methylglutaryl CoA (HMG-CoA) before hydrolysis to acetoacetate, the ketone body used by the body for energy. Depending on the mitochondrial redox potential—that is, the ratio of the reduced form of nicotinamide adenine dinucleotide (NAD) to its oxidized form (NADH/NAD+)—some of the acetoacetate is converted to 3-hydroxybutyrate, and both ketone bodies are transported out of liver mitochondria and hepatocytes into blood, where they may be used by other tissues, especially brain. Acetoacetyl CoA derived from the last turn of the beta oxidation spiral together with acetyl CoA forms HMG-CoA in a reaction catalyzed by HMG-CoA synthetase. Normally, acetoacetyl CoA can also be hydrolyzed to acetyl CoA by the mitochondrial acetoacetyl-CoA thiolase. Patients with a deficiency of this thiolase do not have a defect in ketone body synthesis but, rather, metabolic acidosis associated with excess ketosis (Mitchell and Fukao, 2008; Stanley et al, 2006). The clinical features of the thiolase deficiency are variable. Severe acute metabolic decompensation has been reported in infants, but there are also asymptomatic adults with the disorder. The episodes are heralded by fasting or increased protein intake, because isoleucine is a precursor 234 PART VI Metabolic and Endocrine Disorders of the Newborn of 2-methylacetoacetyl CoA, which is also a substrate for the mitochondrial acetoacetyl-CoA thiolase enzyme. Therefore this block leads to a defect in distal catabolism of isoleucine and in processing of the precursor of ketone body formation—namely, acetoacetyl-CoA. Cardiomyopathy has been identified in rare patients. The characteristic urinary metabolite pattern detected on urine organic acid analysis is the presence of isoleucine metabolites, such as 2-methylacetoacetate, 2-methyl-3-hydroxybutyrate, and triglylglycine. During acute decompensation, lactate and the traditional ketone bodies are detected in excess amounts in the urine. Some older children have been mistakenly identified as having ketotic hypoglycemia because glycine may be elevated in plasma. Deficiency in the mitochondrial acetoacetyl-CoA thiolase can be demonstrated in cultured skin fibroblasts and by mutation analysis. Treatment of acute episodes consists of intravenous glucose and administration of alkali to correct metabolic acidosis, which may be severe. Long-term therapy involves mild protein restriction, avoidance of fasting, and prompt attention to any intercurrent illness or development of ketonuria. The synthesis of acetoacetate from HMG-CoA depends on the HMG-CoA lyase enzyme. A deficiency of this enzyme represents the most profound defect in ketone body synthesis (Roe and Ding, 2001; Stanley et al, 2006). Approximately one third of patients with this disease are in the first week of life. In this subgroup of patients, the onset is dramatic and the disease is catastrophic, being characterized by vomiting, lethargy, coma, seizures, hepatomegaly, hypoglycemia, little or no ketones in urine, and hyperammonemia. Most of the complications are related to severe effects of hypoglycemia on the CNS in addition to acidemia, which may be profound. The characteristic urine metabolites detected by GC-MS analysis are 3-hydroxy-3-methylglutaric acid, 3-methylglutaconic acid, and 3-hydroxyisovaleric acid. Only small, inappropriate amounts of acetoacetic acid and 3-hydroxybutyric acid may be detected. Lactate values may be elevated during the acute metabolic decompensation. Inherited as an autosomal recessive trait, the HMGCoA lyase deficiency can be demonstrated in cultured skin fibroblasts or by mutation analysis. Treatment of the acute episode consists of administration of intravenous glucose and alkali to correct metabolic acidosis. Most long-term therapy consists of a high-carbohydrate diet. Some patients treated with protein-restricted diets, but the most important element in long-term care is the avoidance of fasting. The last defect to be discussed in the area of disturbances in ketone body metabolism is the succinyl-CoA 3-ketoacidCoA transferase deficiency (Mitchell and Fukao, 2008; Stanley et al, 2006). In this disease, the ketone bodies, acetoacetic acid, and 3-hydroxybutyric acid are synthesized adequately in the liver, but they cannot be metabolized in the extrahepatic tissues because of the failure of activation of acetoacetate to acetoacetyl CoA by the transferase enzyme. Conversion to a CoA derivative is required for hydrolysis to acetyl CoA for final metabolism in the Krebs citric acid cycle. This disorder is rare, and most of the affected patients in the newborn period do not survive. Such patients usually exhibit severe ketosis and lactic acidosis. The hallmark of the disease is persistent ketosis, even after correction of overt metabolic acidosis and the institution of frequent feedings with the avoidance of fasting. Plasma values of acetoacetate and ­3-hydroxybutyrate are always mildly to moderately elevated, resulting in intermittent ketonuria. The most important aspect of therapy is the avoidance of fasting, during which acidosis and ketosis can be overwhelming. The gene that encodes this enzyme has been cloned, and several mutations have been detected. PRIMARY LACTIC ACIDOSIS The term congenital lactic acidosis (CLA) refers to a group of diseases in which impaired lactate metabolism is caused by a defect in the mitochondrial respiratory or electron transport chain (ETC), or the tricarboxylic acid (TCA; Krebs) cycle, in which there is a primary defect in pyruvate metabolism that secondarily leads to impaired lactate handling (see Figure 22-1) (De Meirleir et al, 2006; Munnich, 2006; Munnich et al, 2001; Robinson, 2001). However, there are a limited number of reports on inborn errors of the TCA cycle. Most of the information on CLA focuses on mitochondrial ETC, pyruvate dehydrogenase (PDH) complex, and pyruvate carboxylase (PC) defects, all of which have been reviewed (De Meirleir et al, 2006; Munnich, 2006; Munnich et al, 2001; Robinson, 2001) with more detailed information on conditions that must be treated more superficially in this chapter. Some patients with CLA present have overwhelming lactic acidosis in the neonatal period. In others, lactate may be elevated only in CSF, a “cerebral” lactic acidosis syndrome and present more indolently. Depending on the nature of the enzyme deficiency, lactate, pyruvate, and alanine levels can be elevated in the blood. The ratio of blood lactate to pyruvate (L:P ratio) can be helpful in distinguishing the different types of inborn errors. For example, the L:P ratio is often normal (10 to 20) in PDH complex and modestly elevated in PC deficiencies, but can be greatly elevated in an ETC defect. Unlike most of the disorders previously considered in this chapter, these conditions are not detectable by newborn screening and become apparent only when a patient’s symptoms are recognized. Pyruvate Dehydrogenase Deficiency Pyruvate Dehydrogenase Complex Deficiency PDH is a complex of three primary enzymes— plus a ­phosphatase, a kinase, and at least one other element of less known function—that combine as a multimer with a very large molecular weight and a number of copies of each enzyme. The first enzymatic step is a decarboxylation reaction catalyzed by a heterodimeric system consisting of the E-1α subunit, encoded by a gene on the ­X-chromosome, and E-1β, which is autosomally encoded as are all the other subunits in this complex. Defects in all the known genes have been reported, but mutations in the X-linked E-1α outnumber all others by far and may represent as much as 25% of known causes in patients with CLA. Severe PDH deficiency sometimes manifests in the neonatal period with profound lactic acidosis, elevated blood lactate and pyruvate, elevated plasma alanine, and congenital anomalies of the brain noted on MRI, including absent or underdeveloped corpus callosum, heterotopic migration deficits, and a somewhat typical dysmorphic appearance. Typically, the L:P ratio is normal and distinguishes it from disorders of the mitochondrial respiratory chain. CHAPTER 22 Inborn Errors of Carbohydrate, Ammonia, Amino Acid, and Organic Acid Metabolism The patients are hypotonic and respirator dependent and have a poor prognosis. The diagnosis is rarely known immediately, and the standard intravenous support of high glucose exacerbates the metabolic acidosis and worsens the outlook for the patient. The diagnosis is inferred when all the clinical biochemical data are collated and can be confirmed by an enzymatic deficiency in lymphocytes or cultured skin fibroblasts or by mutation analysis of the E-1α gene in particular. The enzyme assay is difficult in the most experienced hands, and it is difficult to understand the high residual activity that is often recorded. The majority of patients are more indolent on clinical presentation, with developmental delay that may resemble Leigh syndrome, and they may have a modest elevation of lactate, with pyruvate being the most telling biochemical marker of the disease. This group of patients often respond well biochemically to a high-fat and low-­carbohydrate diet. Fat, as acetyl CoA, enters the energy pathway after the block, whereas glucose must traverse the PDH reaction to provide all but minimal energy generation. Other defects of the PDH complex including two other subunits, the activating and deactivating enzymes, and a subunit X of unknown function are rare and usually result in chronic psychomotor retardation syndrome in late infancy and childhood. The E3 subunit defect causes a unique syndrome, because the subunit is important in the PDH complex, the BCKAD complex, and the α-ketoglutarate dehydrogenase complex. Therefore these patients have multiple deficiencies involving the BCAA as in MSUD, as well as Krebs cycle metabolites that are indicative of a block in the TCA cycle. Most of the patients are later than the newborn period and have severe progressive neurodegenerative disease. The key laboratory findings are elevations of lactic acid in blood, BCAAs in plasma, and detection of α-ketoglutarate in urine by urine organic acid analysis. PDH phosphatase deficiency is a rare cause of congenital lactic acidosis. Other than the E3 deficiency, other defects in PDH are responsive, at least biochemically to the high-fat, low-carbohydrate diet. Pyruvate Carboxylase Deficiency Pyruvate carboxylase is an enzyme that is involved in gluconeogenesis and that adds bicarbonate to pyruvate to form oxaloacetate, a compound also involved in replenishing intermediates of the tricarboxylic acid cycle. There are two main types of PC deficiency (De Meirleir et al, 2006; Robinson, 2001). Type A is characterized by lactic acidosis in the newborn period and delayed development. However, the disease is of a chronic nature. In type B, the catastrophic form of the disorder, the infant is acutely ill, usually in the first week of life, with encephalopathy, severe metabolic acidosis with lactic acidosis, and hyperammonemia (De Meirleir et al, 2006; Munnich et al, 2001; Robinson, 2001). The mortality rate in this form is high. As discussed earlier, PC is a biotin-containing enzyme. Most of the patients with a type B form of PC deficiency have been of French or English origin. Unlike patients with a type A defect, in whom the blood L:P ratio is normal because both lactate and pyruvate are comparably elevated, patients with the type B defect often have an elevated L:P ratio. Because of the importance of the PC product 235 oxaloacetate in providing adequate cellular levels of aspartate, citrulline metabolism in the urea cycle is defective, leading to elevations of plasma citrulline and plasma ammonium concentrations. Although PC is also an important enzyme in gluconeogenesis, hypoglycemia has not been commonly reported. The liver may be enlarged. There is no effective treatment for PC deficiency when it is associated with progressive neurodegeneration. The gene that encodes the PC subunits, of which four combine to make an active enzyme, has been cloned and sequenced. An expanding number of mutations have been identified and provide a basis for prenatal diagnosis. PC deficiency may be detected in cultured skin fibroblasts or in liver biopsy samples. Phosphoenolpyruvate Carboxykinase Deficiency Phosphoenolpyruvate carboxykinase enzyme also functions in gluconeogenesis, and there are two forms in liver, one in the cytosol and the other in the mitochondrial compartment. It is an exceedingly rare disorder. Patients do not usually come to attention until childhood with hypotonia, failure to thrive, hepatomegaly, lactic acidosis, and hypoglycemia. Although there is no adequate experience in identifying the optimal therapy for these patients, it is reasonable to assume that frequent feedings and avoidance of fasting are important in avoiding severe metabolic imbalance. Mitochondrial Respiratory Chain or Electron Transport Chain Defects Oxidative phosphorylation is the key process performed by the mitochondria of cells. Any inborn error of metabolism that involves the tightly coupled and regulated process of mitochondrial energy metabolism may have profound effects on health and disease, because oxidative phosphorylation is the process by which we convert nutrients into energy. The various derivatives of the nutrients, such as pyruvate and fatty acids, are converted to CO2 in mitochondria. The energy derived from such controlled chemical combustion is harnessed by allowing the reducing equivalents (in the form of NADH or the reduced form of flavin adenine dinucleotide [FADH2], which are derived from such metabolism) to combine with oxygen to form water, and in the process the synthesis of ATP is coupled to the orderly flow of electrons down the respiratory chain components. The important components in the mitochondrial respiratory chain are complex 1 (NADH dehydrogenase), complex 2 (ETF dehydrogenase), complex 3 (cytochromes b, c1) and the terminal complex in this chain, and complex 4, which is cytochrome c oxidase (COX) (Shoffner, 1995). In addition, there is a complex 5, or ATP synthetase, and an adenine nucleotide translocase, which permits transport of adenosine diphosphate into and ATP out of the mitochondria. Complex 2 is involved primarily in fatty acid oxidation and oxidation of succinate derived from the Krebs cycle, because the reducing equivalents extracted from fatty acids, glutaric acid, and succinate flow from ETF into complex 2. The polypeptides that compose these various complexes are derived from both the nuclear genes and the genes on mitochondrial DNA (mtDNA). Except for complex 2, mtDNA is important in production of the subunits of all the respiratory chain 236 PART VI Metabolic and Endocrine Disorders of the Newborn complexes. CLA involving the ETC components has been associated with both nuclear and mtDNA mutations. In addition to the actual complex components, there are many more genes responsible for the assembly of various subunits into the functional complexes, all of which are encoded in nuclear DNA (nDNA). There are estimates that defects in any one of more than 500 genes can result in an ETC deficiency. On the basis of molecular diagnostic testing, the oxidative phosphorylation diseases can be divided into the following four genetic groups (Shoffner, 1995): ll ll ll ll Group 1: nDNA mutations Group 2: mtDNA point mutations Group 3: mtDNA deletions and duplications Group 4: unidentified genetic defects The relationship between phenotype and mtDNA mutations is not straightforward, probably because of the phenomenon of heteroplasmy. Mitochondria with their unique mtDNA are inherited solely from the mother. Random segregation of mitochondria having mtDNA mutations leads to heteroplasmy and, ultimately, a variable concentration of defective mitochondria within cells and among tissues (Wallace et al, 1988). Much of our understanding of the detailed molecular mechanisms that contribute to or produce the ETC gene disturbances concern the mtDNA mutations, although the application of molecular methodology and increasingly inexpensive mass sequencing is allowing more definition of the nDNA disorders that compose the majority of the ETC disease seen in newborn and infancy periods. With the exceptions of the neurogenic muscle weakness, ataxia, and retinitis pigmentosa (NARP caused predominantly by mutations at position 8993 of the mitochondrial genome), only a minority of the mtDNA defects actually are known to manifest in the newborn period (Wong, 2007). Other examples of syndromes caused by mtDNA mutations are the MELAS (mitochondrial encephalopathy, lactic acidosis, and strokelike—episodes position 3243), MERRF (myoclonic epilepsy with ragged-red—fiber position A8344G), Leber hereditary optic neuropathy (Wallace et al, 1988), and sporadic deletion–duplication syndromes such as Pearson syndrome (Di Donato, 2009). The diseases that affect young infants are benign infantile mitochondrial myopathy, cardiomyopathy, or both; lethal infantile mitochondrial disease; lethal infantile cardiomyopathy; subacute necrotizing encephalomyopathy (SNE) or Leigh disease; Pearson syndrome; Alpers’ disease; and the most dramatic form, with presentation often in the first few days of life, during which an acid-base disturbance dominates the clinical picture (Carrozzo et al, 2007; Gibson et al, 2008). The hallmarks of mitochondrial disease are almost always multisystem involvement and unambiguous lactate acidemia or acidosis. Although some debate exists as to whether lactate elevation is present in all mitochondrial disorders that present in life, it is a virtual requirement for such a diagnosis in newborns and early infants. The nuclear encoded mitochondrial DNA depletion syndromes, which may include fatal hepatopathy, are caused by mutations in the SUCLG1, SUCLA2, MPV17, RRM2B, PEO1, TP, POLG1, DGUOK, and TK2 genes. The ribosomal translational defects (EFG1 and EFTμ genes) are the relatively new groups of mitochondrial diseases (Di Donato, 2009; DiMauro and Schon, 2008). Both are part of the group I genetic disease types. The diagnostic tools to unravel these disorders have proliferated recently. We assess lactate in the CSF after lumbar puncture, lactate (among other metabolites) in the brain by magnetic resonance spectroscopy (MRS) during an MRI, more readily obtained in a sick newborn than previously, and DNA technology that has become faster and less expensive and includes full sequencing of the mitochondrial genome. Older and still useful technologies such as muscle biopsy with histologic analysis by light and electron microscopy, mitochondrial complex assay on either fresh (preferred, but infrequently available) or flash-frozen tissue and skin fibroblast studies are still available. In fatal cases, a rapid autopsy and proper preservation of tissue specimens are essential. Finally, a newer and increasingly frequently diagnosed family of disorders includes mitochondrial depletion syndromes caused by a number of genetic defects in the mtDNA synthesis apparatus or the availability of nucleotides. A less frequent, newly recognized family of disorders are the coenzyme Q biosynthetic disorders, diagnosed reliably only by coenzyme Q10 levels in muscle, and not in plasma. Although the subject of much debate, there is no credible evidence that mitochondrial disease is treatable, with the obvious exception of a small minority caused by coenzyme Q deficiency. Given the inability to identify these patients quickly, it is reasonable to treat acutely ill patients with presumed mitochondrial disorders with high doses of coenzyme Q. However, indiscriminant use of this cofactor is to be discouraged. Benign Infantile Mitochondrial Myopathy, Cardiomyopathy, or Both Benign infantile mitochondrial myopathy is associated with congenital hypotonia and weakness at birth, feeding difficulties, respiratory difficulties, and lactic acidosis. In this poorly understood, developmental-like disorder, only skeletal muscle appears to be affected, and histochemical analyses show a COX deficiency that returns to normal levels after 1 to 3 years of age. An nDNA mutation in a gene important in a fetal isoform of an ETC polypeptide specific for muscle oxidative phosphorylation was hypothesized to be the cause of this problem. A developmental switch from the defective fetal gene to the adult form may be responsible for the gradual improvement. It was thought to be the only example of a developmental defect in oxidative phosphorylation that is probably nuclear encoded and in which the treatment is only supportive during the early newborn period to prevent death from respiratory disease. However, a recent study suggests that the etiology may be a maternally inherited, homoplasmic m.14674T>C mt_tRNAGlu: mutation (Horvath et al, 2009). The form also associated with cardiomyopathy may be a variant of the benign isolated myopathy and involves striated muscle in both skeletal and cardiac muscle. It manifests in the newborn period with lactic acidosis and a cardiomyopathy that improves during the first year of life. The exact gene defect is unknown. More attention must be paid to these two disease entities, because with early optimal medical care, affected infants may have an excellent prognosis. Lethal Infantile Mitochondrial Disease Infants with lethal infantile mitochondrial disease are severely ill in the first few days or weeks of life or in the extended newborn period. They exhibit hypotonia, CHAPTER 22 Inborn Errors of Carbohydrate, Ammonia, Amino Acid, and Organic Acid Metabolism 237 CASE STUDY A boy was born to nonconsanguinous, healthy parents after a full-term gestation. Biventricular hypertrophy with a predominant right-sided component had been observed on fetal ultrasonography at 24 weeks’ gestation. He was delivered by cesarean section because of variable decelerations, and he emerged without meconium staining or passage. Respiratory distress developed shortly after birth. An echocardiogram showed concentric right ventricular hypertrophy with elevated right ventricular pressure (70 mm Hg). Initial laboratory studies showed acute metabolic acidosis (blood lactate >18 mmol/L; arterial pH, 6.91). His weight was 2.36 kg (10th percentile), his length was 46 cm (5th percentile), and his head circumference was 31 cm (5th percentile). He had a right ventricular heave and an intermittent fourth heart sound. There was no liver enlargement. Muscle bulk was normal, and deep tendon reflexes were intact. Metabolic investigations showed persistent arterial lactic acidemia (range, 3 to 11 µmol/L) with increased pyruvate (range, 0.02 to 0.44 µmol/L). The L:P ratio ranged from 25 to 290. The activities of the respiratory chain complexes II, III, and IV were normal in cultured skin fibroblasts. A left quadriceps muscle biopsy was used for histochemical analysis and to prepare a 10% extract for respiratory chain studies. Histochemical analysis showed no ragged-red fibers with the modified Gomori trichrome stain, but a diffusely weak response to staining for cytochrome c oxidase (COX, ETC complex 4). Biochemical analysis showed markedly reduced COX activity, contrasting with normal activities of other complexes. Direct sequencing of the three COXs and all 22 transfer RNA genes of mtDNA, all the COX-assembly muscle weakness, failure to thrive, and severe lactic acidosis. Death often occurs by 6 months of age and almost always is associated with overwhelming lactic acidosis. Skeletal muscle shows lipid and glycogen accumulation and abnormally shaped mitochondria on electron microscopic examination. Hepatic dysfunction may be a prominent finding in these patients. Generalized proximal renal tubular dysfunction may occur, leading to the renal Fanconi syndrome. The ETC defects reported in these patients include defects in complexes 1, 3, and 4. Original reports concerned infants with a phenotype resembling severe Werdnig-Hoffman disease with COX deficiency and renal Fanconi syndrome. Lethal Infantile Cardiomyopathy Defects of the mitochondrial respiratory chain have varied neonatal presentations and are commonly caused by isolated COX deficiency (Shoffner, 1995). Clinical features of COX deficiency reflect involvement of one or more tissues and include encephalopathy, myopathy, cardiomyopathy, liver disease, and nephropathy (Munnich et al, 2001). The heterogeneous clinical features are due, in part, to the dual genetic control of COX. The three catalytic subunits— COX I, COX II, COX III—are encoded by mtDNA, whereas the remaining subunits (COX IV to COX VIII) are encoded by nDNA. In addition, the proper assembly of COX requires several nDNA-encoded proteins. The various syndromes caused by COX deficiency are only partially understood at the genetic level. Most infants with isolated complex IV deficiency are likely to have nDNA gene defects in COX assembly genes, including SURF1, SCO2, COX15, COX10, SC01, and LRPPRC (Di Mauro and Schon, 2008). Subacute Necrotizing Encephalomyelopathy or Leigh Disease Probably because of a failure to recognize the clinical signs, infants with SNE or Leigh disease usually come to clinical attention after the newborn period. This disease is nuclear genes known to harbor pathogenic mutations (SURF1, SC01, SC02, COX10, and COX15), and the two other nuclear ancillary genes (COX11 and COX17 ) showed no mutations. Southern blot analysis showed no deletions, but results were inconclusive regarding mtDNA depletion. However, sequencing of the two genes known to be associated with the myopathic (Saada et al, 2001) and the hepatocerebral (Mandel et al, 2001) mtDNA depletion syndromes (TK2 and dGK ) did not show any mutations. A series of echocardiograms documented biventricular, hypertrophic, nonobstructive cardiomyopathy. The patient died on the seventh hospital day after a sudden episode of hemoglobin desaturation. An autopsy was performed 4 hours after death. The right ventricle was found to be thickened and enlarged, with relative sparing of the left side. The pulmonary vasculature of the lungs showed hypertrophy that extended to the most distal vessels. Routine microscopic findings in skeletal muscle were normal, and the CNS was without microscopic or macroscopic abnormalities. This baby had a fatal syndrome defined clinically by prenatal cardiomyopathy and severe pulmonary hypertension in the newborn period. The syndrome was caused by isolated ETC complex 4 deficiency. The pathophysiology centered on the heart and lungs. How much disease was caused by the involvement of other organs remains unclear. This case is an example of a mitochondrial metabolic disorder, but without a specific molecular genetic cause. A more extensive evaluation by current methods might or might not have revealed a diagnosis. characterized as a progressive neurodegenerative disorder with severe hypotonia, seizures, extrapyramidal movement disorders, optic atrophy, and defects in automatic ventilation or respiratory control (Finsterer, 2008; Leigh, 1951). It is clear that there are many causes of SNE. As discussed earlier, PDH complex deficiency (PDHE1 alpha-subunit and PDHX1) can lead to Leigh disease. Patients with defects in the ETC have also been reported to have findings compatible with SNE. The mtDNA mutation causing NARP is an example. Reported nuclear gene defects include ETC complex I (NDUFS1), complex II (SDHA), complex IV (SURF1, COX15), coenzyme Q10 biosynthetic defects, mitochondrial ribosomal translational defects, and SUCLA2. Many neuropathologists believe that the diagnosis of SNE depends on an analysis of CNS tissue at autopsy. However, MRI characteristically shows bilateral symmetrical lesions of the basal ganglia as occurs in other mitochondrial disorders. There is no effective treatment for this disease, unless the cause is a specific inability to synthesize coenzyme Q10. It is possible that most of the patients with Leigh disease have disturbances in nuclear-encoded genes. Although, as discussed later, the NARP lesion caused by a group 2 mtDNA mutation is one important cause in early infancy. Clearly, this is not one disease entity, because the specific neuropathologic findings for SNE have also been reported in a patient with Menkes’ disease, in which there is a secondary ETC complex 4 deficiency, because copper is an important metal cofactor of COX. The following clinical findings have been noted in infants with SNE: optic atrophy, ophthalmoplegia, nystagmus, respiratory abnormalities, ataxia, hypotonia, spasticity, seizures, developmental delay, psychomotor retardation, myopathy, and renal tubular dysfunction. Some patients may manifest hypertrophic cardiomyopathy, liver dysfunction, and microcephaly. The neuropathologic lesions include demyelination, gliosis, necrosis, relative neuronal sparing, and capillary proliferation in specific brain lesions. 238 PART VI Metabolic and Endocrine Disorders of the Newborn There are lesions of the basal ganglia, which are bilaterally symmetrical, as well as of the brainstem, cerebellum, and the cerebral cortex to a lesser degree. Commonly, elevation in blood lactate is only slight to moderate, as well as intermittent, in this diverse group of patients. In some instances, lactate values may be elevated only in the CSF. The most commonly reported biochemical abnormalities are deficiencies in COX, complex 4 NADH dehydrogenase, or complex 1 and PDH. In a few rare patients, the abnormality in oxidative phosphorylation has been reported to be secondary to NARP mutation. This involves a T:C transition at base pair 8993 of the adenosine triphosphatase (ATPase) 6 gene, changing a leucine to a proline at position 156 in the ATPase 6 polypeptide. Investigators have speculated that defects such as those in COX and the NADH dehydrogenase, when associated with neuropathology of Leigh disease, are caused by nuclear gene mutations and not mtDNA gene defects such as the NARP point mutation in the ATPase 6 gene. The diseases in group 3 (see earlier), exemplified by the Kearns-Sayre and chronic progressive external ophthalmoplegia syndromes, are genetic but not familial and are caused by mtDNA deletions or duplications that are spontaneous mutations. These disorders do not usually come to clinical attention in infancy. The only example of such a mutation manifesting in early infancy is the Pearson syndrome. This disorder is systemic and primarily affects the hematopoietic system and pancreas function. The characteristics are severe macrocytic anemia with varying degrees of neutropenia and thrombocytopenia. Bone marrow examination shows normal cellularity, but extensive vacuolization of erythroid and myeloid precursors, hemosiderosis, and ringed sideroblasts. This disease of the bone marrow can lead to death in infancy. However, patients who are able to recover or who benefit from aggressive therapy may demonstrate other signs of this systemic disorder in late infancy or childhood, such as poor growth, pancreas dysfunction, mitochondrial myopathy, lactic acidosis, and progressive neurologic damage. Barth Syndrome Barth syndrome is an X-linked disorder associated with cardiomyopathy, skeletal muscle disease, and neutropenia (Yen et al, 2008). Skeletal muscle shows abnormal mitochondrial morphology. Important laboratory findings include decreased plasma-free carnitine, increased urinary excretion of 3-methylglutaconate on GC-MS analysis of urine organic acids, and decreased levels of serum cholesterol in early infancy. Positional cloning identified a gene for this disorder on Xq28 that encodes for a phospholipid remodeling enzyme, cardiolipin acyl transferase. Several mutations have been identified. It has been hypothesized that the organic acid 3-methylglutaconate accumulates because of defective mitochondrial transport. Patients must be supported from birth to early infancy. It is possible that if severe cholesterol deficiency can be avoided, affected infants may survive and may be relatively free of cardiomyopathy during childhood. Of diagnostic importance, not all patients with ­3-methylglutaconic aciduria have Barth syndrome. A few have isolated ­leucine-dependent 3-methylglutaconyl-CoA hydratase deficiency or Costeff syndrome, but most have ill-defined mitochondropathies. Unidentified Genetic Defects A number of diseases are believed to be caused by mitochondrial respiratory chain problems, but the specific mutations remain unknown. These disorders constitute the group 4 mutations, or the disorders of unknown inheritance (Shoffner, 1995). Alpers disease was once one such example. It has also been called progressive infantile poliodystrophy. Infants and children with this progressive disease experience progressive cerebral cortical damage, sometimes also involving the cerebellum, basal ganglia, and brainstem; in some, liver disease may progress to cirrhosis. The neuropathologic lesions consist of spongiform or microcystic cerebral degeneration, gliosis, necrosis, and capillary proliferation Seizures are prominent, including myoclonus. Laboratory abnormalities include abnormal NADH oxidation or complex 1 defects, impaired pyruvate handling, PDH complex deficiency, TCA cycle malfunction, and decreased mitochondrial cytochrome a + a3 content. This disorder is caused by mutations in one or more gene defects associated with mtDNA depletion, such as the autosomal recessive disease caused by mutations in the DNA polymerase gamma 1 (POLG1) gene and can usually be defined by mutation analysis. Early Lethal Lactic Acidosis In an unknown fraction of patients with primary disturbances in mitochondrial oxidative phosphorylation or ETC defects, massive lactic acidosis develops within 24 to 72 hours after birth. Commonly the condition is untreatable, because it is relentless and unresponsive to alkali therapy. Dialysis is a remedy but not a cure. Often, affected infants have no obvious organ damage early in the course or evidence of malformations; this is also true for infants with the PDH complex deficiency, which is probably a more common cause of overwhelming acidosis in the first week of life. In addition, acidemia per se can easily cause the coma or impaired cardiac contractility that may be encountered. Some infants have survived with aggressive therapy. The care of babies with these different forms of severe lactic acidosis almost always brings an ethical dilemma to the forefront for physicians and nurses of the neonatal intensive care unit as well as for the babies’ families. To further complicate the issues, enzymatic and molecular analyses usually are not immediately available. The disease in most patients probably remains idiopathic, and no DNA mutation, nuclear or mitochondrial, will be identified without extensive, expensive, and extremely inconvenient and Herculean efforts. A rigid approach to care is impractical and unwise. Decisions regarding management must be individualized, because the mitochondrial dysfunction and resultant pathophysiology can vary among infants. SUGGESTED READINGS Blau N, Hoffmann GF, Leonard J, et al, editors: Physicians guide to the treatment and follow-up of metabolic diseases, Heidelberg, 2006, Springer. Fernandes J, Saudubray JM, Van den Berghe G, et al: Inborn Metabolic diseases: diagnosis and treatment, Heidelberg, 2006, Springer. Scriver CR, Beaudet AL, Sly WS, et al: The metabolic and molecular bases of inherited disease. ed 8, New York, 2001, McGraw-Hill. Updated material is available online at www.ommbid.com. Complete references used in this text can be found online at www.expertconsult.com C H A P T E R 23 Lysosomal Storage, Peroxisomal, and Glycosylation Disorders and SmithLemli-Opitz Syndrome in the Neonate Janet A. Thomas, Carol L. Greene, and Gerard T. Berry Lysosomal storage diseases (LSDs), peroxisomal disorders, congenital disorders of glycosylation (CDGs), and Smith-Lemli-Opitz (SLO) syndrome are single-gene disorders, most of which demonstrate autosomal recessive inheritance. The combined incidence of LSDs has been reported to be 1 in 1500 to 8000 live births in the United States, Europe, and Australia (Fletcher, 2006; Meikle et al, 2006; Staretz-Chacham et al, 2009; Stone and ­Sidransky, 1999; Wenger et al, 2003; Winchester et al, 2000). The incidence of peroxisomal disorders is estimated to be more than 1 in 20,000. The most current estimate for SLO syndrome is 1 in 20,000, and a similar frequency of 1 in 20,000 is estimated for the congenital disorders of glycosylation. These four categories of metabolic diseases involve molecules important in cell membranes and share overlapping clinical presentations. Clinical presentations are heterogeneous, with a broad range of age at presentation and severity of symptoms. All are chronic and progressive. Age of onset varies from prenatal to adulthood, and severity can range from severe disability and early death to nearly normal lifestyle and life span. For each condition, interfamilial variability is greater than intrafamilial variability. The genetic and clinical characteristics of conditions in these categories that can manifest in the neonatal period (except Pompe disease, which is addressed in Chapter 22) are also summarized in Tables 23-1 to 23-3. Important presentations that should lead the neonatologist to consider these disorders in the differential diagnosis are as follows: 1. In utero infection—hepatosplenomegaly and hepatopathy, possibly with extramedullary hema­to­poiesis 2. Nonimmune hydrops fetalis, ichthyotic or collodion skin, or both 3. Neurologic only—early and often difficult to control seizures, hypertonia or hypotonia, with or without altered head size and with or without eye findings 4. Coarse facial features with bone changes, dysostosis multiplex, or osteoporosis 5. Dysmorphic facial features with or without major malformations 6. Rarely, known family history or positive prenatal diagnosis Only for the last three presentations are these conditions likely to be considered early in the differential diagnosis. Most babies with these conditions are born to healthy, nonconsanguineous couples with normal family histories, and these disorders are usually considered late, if at all, as in Case Study 1. LYSOSOMAL STORAGE DISORDERS Lysosomes are single-membrane–bound intracellular organelles that contain enzymes called hydrolases. These lysosomal enzymes are responsible for splitting large molecules into simple, low-molecular-weight compounds, which can be recycled. The materials digested by lysosomes and derived from endocytosis and phagocytosis, are separated from other intracellular materials by the process of autophagy, which is the main mechanism whereby endogenous molecules are delivered to lysosomes. The common element of all compounds digested by lysosomal enzymes is that they contain a carbohydrate portion attached to a protein or lipid. These glycoconjugates include glycoproteins, glycosaminoglycans, and glycolipids. Glycolipids are large molecules with carbohydrates attached to a lipid moiety. Sphingolipids, globosides, gangliosides, cerebrosides, and lipid sulfates all are glycolipids. The different classes of glycolipids are distinguished from one another primarily by different polar groups at C1. Sphingolipids are complex membrane lipids composed of one molecule each of the amino alcohol sphingosine, a long-chain fatty acid, and various polar head groups attached by a β-glycosidic linkage. Sphingolipids occur in the blood and nearly all tissues of the body, the highest concentration being found in white matter of the central nervous system (CNS). In addition, various sphingolipids are components of the plasma membrane of practically all cells. The core structure of natural sphingolipids is ceramide, a long-chain fatty acid amide derivative of sphingosine. Free ceramide, an intermediate in the biosynthesis and catabolism of glycosphingolipids and sphingomyelin, composes 16% to 20% of normal lipid content of stratum corneum of the skin. Sphingomyelin, a ceramide phosphocholine, is one of the principal structural lipids of membranes of nervous tissue. Cerebrosides are a group of ceramide monohexosides with a single sugar, either glucose or galactose, and an additional sulfate group on galactose. The two most common cerebrosides are galactocerebroside and glucocerebroside. The largest concentration of galactocerebroside is found in the brain. Glucocerebroside is an intermediate in the synthesis and degradation of more complex glycosphingolipids. Gangliosides, the most complex class of glycolipids, contain several sugar units and one or more sialic acid residues. Gangliosides are normal components of cell membranes and are found in high concentrations in ganglion cells of the CNS, particularly in nerve endings and dendrites. GMI is the major ganglioside in brain of vertebrates. 239 240 PART VI Metabolic and Endocrine Disorders of the Newborn CASE STUDY 1 C.J. was a 2200-g girl, born to a 24-year-old mother (third pregnancy, second viable child) after a 32-week gestation, by cesarean section performed for fetal distress. Pregnancy was complicated by the finding on ultrasonography of fetal hydrops and ascites and possible hepatosplenomegaly at 24 weeks’ gestation. Fetal blood sampling showed a hematocrit of 31% and elevations of γ-glutamyltransferase and aspartate transaminase values. Results of viral studies were negative, and chromosomes were normal. At delivery, the infant was limp and blue with a heart rate of 60 beats/min. Physical examination and chest radiograph showed marked abdominal distention, hepatosplenomegaly, multiple petechiae and bruises, a bell-shaped thorax, generalized hypotonia, talipes equinovarus, contractures at the knees, a large heart, and hazy lung fields with low volumes. Disseminated intravascular coagulopathy and evidence of liver disease developed rapidly, with elevated aspartate transaminase, γ-glutamyl-transferase, and increasing hyperbilirubinemia. The patient’s condition was maintained with a ventilator and treatment with antibiotics for possible sepsis. Results of evaluations for bacterial and viral agents were negative. Metabolic studies, including ammonia, lactate, very long chain fatty acids (VLCFAs), and urine amino and organic acids, yielded unremarkable measurements. The white blood cells were noted to have marked toxic granularity consistent with overwhelming bacterial sepsis or metabolic storage disease. The patient experienced continued cardiorespiratory deterioration, had bilateral pneumothoraces and pneumopericardium, and died on the third day of life. Consent for autopsy was obtained from the family. A standard autopsy was performed and showed the presence of large, membrane-bound vacuoles within hepatocytes, endothelial cells, pericytes, and bone marrow stromal cells, which are typical of a metabolic storage disorder. Similar cells were also found within the placenta. There was no evidence of an infectious cause. Unfortunately, because a lysosomal storage disorder was not considered as a possible cause at the time of death, no frozen tissue or cultured fibroblasts were available to pursue the diagnosis. As a result of efforts by a research laboratory and the recurrence of disease in the couple’s subsequent pregnancy, a diagnosis of β-glucuronidase deficiency, or mucopolysaccharidosis type VII, was confirmed. CASE STUDY 2 M.E. was born by normal spontaneous vaginal delivery, at full term according to dates based on early ultrasonography, with weight of 2.2 kg, length of 45 cm, and head circumference of 31.5 cm. On the basis of physical examination, gestational age was assessed as 36 weeks. A heart murmur was noted, and investigation showed the presence of a small ventricular septal defect with no hemodynamic significance. Submucous cleft palate was noted. Examination for dysmorphic features showed simple, posteriorly rotated ears, mild epicanthic folds, micrognathia, and unilateral simian crease. Tone was moderately decreased. Irritability and severe feeding problems were noted, and gavage feeding was required; growth was poor despite adequate calories. The results of a karyotype analysis were normal, and the results of studies for velocardiofacial syndrome were negative. Vomiting developed, and further evaluation showed no acidosis, hypoglycemia, or hyperammonemia. Liver-associated values and cholesterol level were normal, as were results of studies of amino acids, organic acids, and acylcarnitine profile. Vomiting became more severe and did not respond to elemental formula, and pyloric stenosis was detected. Feeding problems persisted after successful s­ urgical correction. Delivery of more than 140 kcal/kg by gavage was poorly tolerated, but resulted in weight gain; however, length and head growth remained poor. SLO syndrome was suggested despite the normal cholesterol value obtained on analysis in the hospital laboratory. Studies performed in a specialized laboratory showed the 7- and 8-dehydrocholesterol values to be elevated and the cholesterol value decreased. Cholesterol supplementation led to some improvement in behavior and feeding. A decrease to 110 kcal/kg/day was tolerated without worsening of growth, and weight for height gradually returned to normal. A review of records confirmed that the pregnancy had been accurately dated by ultrasonography at 10 weeks’ gestation, confirming that M.E. was small for gestational age and microcephalic at birth, with subsequent growth typical for SLO syndrome. The incorrect assessment of gestational age as 36 weeks on examination was found to result from a failure to appreciate the effect of hypotonia on the findings for gestational age. The family was counseled about autosomal recessive inheritance, including the availability of prenatal diagnosis. CASE STUDY 3 H.K. was born at term to healthy parents by cesarean section performed for breech presentation after an otherwise uncomplicated pregnancy. Hypotonia and dysmorphic features were noted in the delivery room, including inner epicanthic folds, flat occiput, large fontanels, shallow orbital ridges, low nasal bridge, micrognathia, redundant skin folds at the neck, and unilateral simian crease. Brushfield spots were present. Investigation of a heart murmur revealed patent ductus arteriosus and a small atrial septal defect. There was mild hepatomegaly but normal liver function, no acidosis, and no hypoglycemia. Suck was poor, and gavage feeding was required. Gangliosides function as receptors for toxic agents, hormones, and certain viruses, are involved in cell differentiation, and they can also have a role in cell-cell interaction by providing specific recognition determinants on the surface of cells. Ceramide oligosaccharides (i.e., globosides) are a family of cerebrosides that contain two or more sugar residues, usually galactose, glucose, or N-acetylgalactosamine. Glycosaminoglycans and oligosaccharides are essential constituents of connective tissue, parenchymal organs, cartilage, and the nervous system. Karyotype was normal and there was no evidence of trisomy 21 in blood in 50 interphase cells examined. The option of skin biopsy to search further for evidence of mosaicism for trisomy 21 was considered. Thyroid function values were normal. Urine amino and organic acid values were normal, as was the acylcarnitine profile. Plasma VLCFA analysis showed elevation consistent with a diagnosis of Zellweger syndrome, along with a typical increase in pipecolic acid value and impaired capacity for fibroblast synthesis of plasmalogens. The baby died at 3 months of age, and autopsy showed polymicrogyria and small hepatic and renal cysts. The family was counseled about autosomal recessive inheritance, including the availability of prenatal diagnosis. Glycosaminoglycans, also called mucopolysaccharides, are complex heterosaccharides consisting of long sugar chains rich in sulfate groups. The polymeric chains are bound to specific proteins (core proteins). Glycoproteins contain oligosaccharide chains (long sugar molecules) attached covalently to a peptide core. Glycosylation occurs in the endoplasmic reticulum and Golgi apparatus. Most glycoproteins are secreted from cells and include transport proteins, glycoprotein hormones, complement factors, enzymes, and enzyme inhibitors. There is extensive diversity in the composition and structure of oligosaccharides. CHAPTER 23 Lysosomal Storage, Peroxisomal, and Glycosylation Disorders and Smith-Lemli-Opitz Syndrome in the Neonate 241 CASE STUDY 4 M.J. had hypotonia at birth after an uncomplicated pregnancy. Minor dysmorphic features were noted, including high nasal bridge, large ears, and inverted nipples. Feeding difficulties were significant, and growth was poor. Findings on head ultrasonography were unremarkable, as were those of head magnetic resonance imaging, although the radiologist questioned whether the cerebellum might be slightly small. Results of a karyotype analysis were normal. Hypothyroidism, discovered on newborn screening, was promptly treated and closely monitored. There was no acidosis or hypoglycemia, and liver enzyme values were normal; results of amino and organic acid analyses and acylcarnitine profile were all normal. The baby was discharged on a diet providing 130 kcal/kg/day. On follow-up, growth remained poor, and development was severely delayed. At 6 months of age, she was admitted to the hospital for an episode of acutely altered mental status and low blood pressure. Mild acidosis, borderline elevations of lactate and ammonia, and significant elevation of liver enzymes all resolved over the course The degradation of glycolipids, glycosaminoglycans, and glycoproteins takes place especially within lysosomes of phagocytic cells, related to histiocytes and macrophages, in any tissue or organ. A series of hydrolytic enzymes cleaves specific bonds, resulting in sequential, stepwise removal of constituents such as sugars and sulfate, and degrading complex glycoconjugates to the level of their basic building blocks. Lysosomal storage diseases most commonly result when an inherited defect causes significantly decreased activity in one of these hydrolases. Other causes are failure of transport of an enzyme, substrate, or product. Whatever the specific cause, incompletely metabolized molecules accumulate, especially within the tissue responsible for catabolism of the glycoconjugate. Additional excess storage material may be excreted in urine. The mechanisms of cellular dysfunction and damage in the majority of LSDs remain unknown. Various hypotheses have been offered, such as a pivotal disturbance in the normal process of autophagy (Ballabio and Gieselmann, 2009; Kiselyov et al, 2007). In this pathophysiologic construct, endoplasmic reticulum membrane engulfment of cellular components, such as mitochondrial derivatives targeted for destruction, is perturbed. As a consequence, deleterious pathways become activated, leading to unwanted ubiquitination of targeted molecules and apoptosis. Lysosomal storage diseases are classified according to the stored compound. Clinical phenotype depends partially on the type and amount of storage substance. There are more than 50 different LSDs and a significant fraction, approximately 20 LSDs, may have manifestations in the newborn infant (Staretz-Chocham et al, 2009). The disorders selected for discussion in this chapter are all known to manifest in the neonatal period. CLINICAL PRESENTATIONS Table 23-1 summarizes the clinical characteristics of the neonatal presentations of lysosomal storage disorders. Niemann-Pick A Disease (Acute, Sphingomyelinase Deficient) Etiology Niemann-Pick A disease is caused by a deficiency of sphingomyelinase. Sphingomyelinase catalyzes the breakdown of sphingomyelin to ceramide and phosphocholine, and of the hospital stay. ­Cardiac ultrasonography showed mild ventricular dysfunction, which also resolved. Amino and organic acid values were normal, as was the acylcarnitine profile. Urine oligosaccharide levels showed an unusual pattern, and urine mucopolysaccharide values were normal. At 2 years of age, developmental delay remained marked, and hypotonia persisted with reflexes absent. The creatinine phosphokinase level was normal, but liver function values were again abnormal. Because mitochondrial disease was suspected, the patient was scheduled for liver biopsy, but clotting values were abnormal. A congenital disorder of glycosylation was suspected, and a transferrin assay confirmed the diagnosis. A review of neonatal records revealed a comment from a neurology consultant about the unusual distribution of fat on the buttocks and thighs of M.J. as a neonate. The family was counseled about autosomal recessive inheritance, including availability of prenatal diagnosis. its deficiency results in sphingomyelin storage within lysosomes. Cholesterol is also stored, suggesting that its metabolism is tied to that of sphingomyelin. Sphingomyelin normally composes 5% to 20% of phospholipid in liver, spleen, and brain, but in these disorders it can compose up to 70% of phospholipids. Patients with NiemannPick A disease usually have enzyme activity less than 5% of normal. Clinical Features Clinical features of this disorder may appear in utero or up to 1 year of age. Affected infants usually have massive hepatosplenomegaly (hepatomegaly greater than splenomegaly), constipation, feeding difficulties, and vomiting with consequent failure to thrive. Patients eventually appear strikingly emaciated with a protuberant abdomen and thin extremities. Neurologic disease is evident by 6 months of age, with hypotonia, decrease or absence of deep tendon reflexes, and weakness. Loss of motor skills, spasticity, rigidity, and loss of vision and hearing occur later. Seizures are rare. A retinal cherry-red spot is present in about half of cases, and the electroretinographic findings are abnormal. Respiratory infections are common. The skin may have an ochre or brownish yellow color, and xanthomas have been observed. Radiographic findings consist of widening of medullary cavities, cortical thinning of long bones, and osteoporosis. In the brain and spinal cord, neuronal storage is widespread, leading to cytoplasmic swelling together with atrophy of cerebellum. Bone marrow and tissue biopsy samples may show foam cells or sea-blue histiocytes, which represent lipid-laden cells of the monocyte-macrophage system. Similarly, vacuolated lymphocytes or monocytes may be present in peripheral blood. Tissue cholesterol levels may be threefold to tenfold of normal, and patients may have a microcytic anemia and thrombocytopenia. Death occurs by 2 to 3 years of age. Niemann-Pick C Disease Etiology Niemann-Pick C disease is caused by an error in the intracellular transport of exogenous low-density lipoprotein (LDL)–derived cholesterol, which leads to impaired esterification of cholesterol and trapping of unesterified cholesterol in lysosomes. The incidence may be higher than 242 TABLE 23-1 Lysosomal Storage Disorders in the Newborn Period: Genetic and Clinical Characteristics of Neonatal Presentation Neurologic Facies Findings Distinctive Features Eye Findings Cardiovascular Dysostosis Hepatomegaly/ Findings Multiplex Splenomegaly Defect Gene Location Molecular Findings Ethnic Predilection Onset NiemannPick A disease Early Frontal infancy bossing Difficulty feeding, apathy, deafness, blindness, hypotonia Brownish-yellow Cherry-red skin, xanthomas spot (50%) – – ++/+ Sphingomyelinase deficiency NiemannPick C disease Birth to 3 mo Normal Developmental delay, vertical gaze paralysis, hypotonia, later spasticity – – – – +/++ Abnormal ­cholesterol NPC1 gene at 18q11 Increased in ­esterification accounts French for >95% of cases; ­Canadians of HE1 gene mutaNova Scotia tions may account and for remaining Spanish cases Americans in the southwest United States Gaucher disease type 2 In utero to 6 mo Normal Congenital Poor suck and ichthyosis, swallow, weak ­collodion skin cry, squint, trismus, ­strabismus, opsoclonus, ­hypertonic, later ­flaccidity – – – +/++ Glucocerebrosidase deficiency 1q21; large number Panethnic of mutations known; five mutations account for approximately 97% of mutant alleles in the Ashkenazi population, but approximately 75% in the nonJewish population Krabbe disease 3-6 mo Normal Irritability, tonic spasms with light or noise stimulation, seizures, hypertonia, later flaccidity Optic atrophy – – –/– Galactocerbrosidase deficiency 14q 24.3-q32.1; >60 Increased in mutations with Scandinavian countries some common and in a large mutations in speDruze kindred cific populations in Israel Coarse Poor suck, weak Gingival cry, lethargy, ­hypertrophy, exaggerated startle, edema, rashes ­blindness, hypotonia, later spasticity Cherry-red spot (50%) – + +/+ β-Galactosidase deficiency 3 pter-3p21; hetero- Panethnic geneous mutations; common mutations in specific populations – Hepatomegaly in 50%, splenomegaly less common Lysosomal acid ceramidase 8p21.3-22; 9 disease- Panethnic causing mutations identified GM1 ganglio- Birth sidosis Farber disease type I 2 wk to 4 Normal mo Increased CSF protein Grayish opaci- Occasional Progressive psychoJoint swelling fication motor impairment, with nodules, surroundseizures, decreased hoarseness, ing retina reflexes, hypotonia lung disease, in some contractures, patients, fever, granulosubtle mas, dysphagia, ­vomiting, cherry-red increased CSF spot protein ASM gene at 1:40,000 in 11p15.1-p15.4 Ashkenazi Jews 3 of 18 mutawith carrier tions account for frequency of approximately 1:60 92% of mutant alleles in the Ashkenazi population PART VI Metabolic and Endocrine Disorders of the Newborn Disorder Normal Farber disease type IV (neonatal) Birth Normal Congenital sialidosis In utero Coarse, Mental retardation, to birth edema hypotonia Galactosialidosis In utero Coarse to birth Mental retardation, Ascites, edema, Cherry-red occasional deafness, inguinal hernias, spot, hypotonia renal disease, corneal telangiectasias clouding Wolman disease First weeks of life Mental deterioration Vomiting, diarrhea, steatorrhea, abdominal distention, failure to thrive, anemia, adrenal calcifications Infantile sialic acid storage disease In utero Coarse, to birth dysmorphic Mental retardation, hypotonia Ascites, anemia, diarrhea, failure to thrive I-cell disease In utero Coarse to birth Mental retardation, deafness Gingival hyperCorneal plasia, restricted clouding joint mobility, hernias Mental retardation, hypotonia — Severe corneal ­clouding, retinal degeneration, blindness Hernias Variable ­corneal clouding Normal Nodules not consistent findings Mucolipidosis Birth to type IV 3 mo Normal Mucopolysaccharidosis type VII Variable Mild to severe coarse- mental retardation ness In utero to childhood Joint swelling with Normal nodules, hoarsemacula, ness corneal opacities – Corneal opacities (1/3) – – Neonatal ascites, Corneal inguinal hernias, clouding renal disease – HSM less common than in type I 8p21.3-p22 Panethnic ++/++ Unknown Panethnic NEU 1 gene (sialidase) at 6p21 Panethnic – + +/+ Neuraminidase deficiency Cardiomegaly progressing to failure + +/+ Absence of a pro20q13.1 tective protein that safeguards neuraminidase and beta-galactosidase from premature degradation – – – +/+ Lysosomal acid lipase 10q23.2-q23.3; vari- Increased in Iradeficiency ety of mutations nian Jews and identified in non-Jewish and Arab populations of Galilee – Congestive heart failure + +/+ Defective transport SLC17A5 of sialic acid out of gene at 6q the lysosome Valvular disease, congestive heart failure, cor pulmonale ++ +++/+++ – – –/– Variable ++ Variable Panethnic Panethnic Lysosomal enzymes Enzyme encoded by Panethnic two genes; α and β lack mannosesubunits encoded 6-PO4 recognition marker and fail by gene at 12p; γ to enter the lysosubunit encoded some (phospho-­ by gene at 16p transferase deficiency, 3-subunit complex [α2 β2 γ2]) Unknown; some patients with partial deficiency of ganglioside sialidase MCOLN1 gene Increased in at 19p13.2­Ashkenazi 13.3 encoding Jews mucolipin; two founder mutations accounting for 95% of mutant alleles in Ashkenazi population β-Glucuronidase deficiency GUSB gene at 7q21.2-q22; heterogeneous mutations –, Not seen; +, typically present, usually not severe; ++, usually present, and moderately severe; +++, always present, usually severe; CSF, cerebrospinal fluid; HSM, hepatosplenomegaly. Panethnic 243 Birth to 9 mo (≤20 mo) CHAPTER 23 Lysosomal Storage, Peroxisomal, and Glycosylation Disorders and Smith-Lemli-Opitz Syndrome in the Neonate Farber disease types II and III 244 PART VI Metabolic and Endocrine Disorders of the Newborn 1 in 150,000 births (Wraith et al, 2009). Cell lines from patients can be divided into two complementation groups, NPC1 and NPC2, corresponding to different genes (­Millat et al, 2001). In each group, the primary defect is abnormal cholesterol esterification, but the enzyme responsible for cholesterol esterification—acetyl coenzyme A (CoA) acetyltransferase (ACAT)—is not deficient. The storage of sphingomyelin is secondary. It has been suggested that the defect is in transport of cholesterol out of the lysosome, making cholesterol unavailable to ACAT (Natowicz et al, 1995). Sphingomyelinase activity appears normal or elevated in most tissues, but is partially deficient (60% to 70%) in fibroblasts from most patients with this disorder. Storage of sphingomyelin in tissues is much less than in Niemann-Pick A or B disease and is accompanied by additional storage of unesterified cholesterol, phospholipids, and glycolipids in the liver and spleen. Only glycolipids are increased in the brain. Clinical Features The age of onset, clinical features, and natural history of Niemann-Pick C disease are highly variable. Onset can occur from birth to 18 years of age. Fifty percent of children with onset in the neonatal period have conjugated hyperbilirubinemia, which usually resolves spontaneously but is followed by neurologic symptoms later in childhood. In the severe infantile form, hepatosplenomegaly is common, accompanied by hypotonia and delayed motor development. Further mental regression is usually evident by the age of 1 to 1.5 years, in association with behavior problems, vertical supranuclear ophthalmoplegia, progressive ataxia, dystonia, spasticity, dementia, drooling, dysphagia, and dysarthria. Seizures are rare. Foam cells and sea-blue histiocytes may be found in many tissues. Neuronal storage with cytoplasmic ballooning, inclusions, meganeurites, and axonal spheroids are also seen. Death may occur in infancy or as late as the third decade of life. Niemann-Pick C disease can also manifest as fatal neonatal liver disease, often misdiagnosed as fetal hepatitis. Patients with mutations in the NPC2 gene (HE1) may have remarkable features consisting of pronounced pulmonary involvement leading to early death caused by respiratory failure (Millat et al, 2001). Gaucher Disease Type 2 (Acute Neuropathic) Etiology Three types of Gaucher disease have been defined. Type 1, the nonneuropathic form, is the most common and is distinguished from types 2 and 3 by the lack of CNS involvement. Type 1 disease most commonly manifests in early childhood, but may do so in adulthood. Type 2 disease, the acute neuropathic form, is characterized by infantile onset of severe CNS involvement. Type 3 disease, the subacute neuropathic form, is also late in onset with slow neurologic progression. Almost all types of Gaucher disease are caused by a deficiency of lysosomal glucocerebrosidase and result in storage of glucocerebroside in visceral organs; the brain is affected in types 2 and 3. Although there is significant variability in clinical presentation among individuals with the same mutations, there is a clear correlation between certain mutations and clinical symptoms involving the CNS (Beutler and Grabowski, 2001). The enzyme splits glucose from cerebroside, yielding ceramide and glucose. A few patients with Gaucher disease type 2 have a deficiency of saposin C, a cohydrolase required by glucocerebrosidase. Clinical Features Typically, the age of onset of Gaucher disease type 2 is approximately 3 months, consisting of hepatosplenomegaly (splenomegaly predominates) with subsequent neurologic deterioration. Hydrops fetalis, congenital ichthyosis, and collodion skin, however, are well-described presentations (Fujimoto et al, 1995; Ince et al, 1995; Lipson et al, 1991; Liu et al, 1988; Sherer et al, 1993; Sidransky et al, 1992). In a review of 18 cases of Gaucher disease manifesting in the newborn period, Sidransky et al (1992) found that eight of the patients had associated dermatologic findings and six patients had hydrops. The etiology of the association of such findings and Gaucher disease is unclear, although the enzyme deficiency appears to be directly responsible (Sidransky et al, 1992). Ceramides have been shown to be major components of intracellular bilayers in epidermal stratum corneum, and they have an important role in skin homeostasis (Fujimoto et al, 1995). Therefore Gaucher disease should be considered in the differential diagnosis for infants with hydrops fetalis and congenital ichthyosis. For the subset of patients in the prenatal period or at birth, death frequently occurs within hours to days, or at least within 2 to 3 months. Krabbe Disease (Globoid Cell Leukodystrophy) Etiology The synonym for Krabbe disease, globoid cell leukodystrophy, is derived from the finding of large numbers of multinuclear macrophages in cerebral white matter that contain undigested galactocerebroside. Disease is caused by a deficiency of lysosomal galactocerebroside β-galactosidase, which normally degrades galactocerebroside to ceramide and galactose. Deficiency of the enzyme results in storage of galactocerebroside. Galactocerebroside is present almost exclusively in myelin sheaths. Accumulation of the toxic metabolite psychosine, also a substrate for the enzyme, has been postulated to lead to early destruction of oligodendroglia. Impaired catabolism of galactosylceramide is also important in pathogenesis of the disease. Clinical Features Age of onset ranges from the first weeks of life to adulthood. The typical age of onset of infantile Krabbe disease is between 3 and 6 months, but there are cases of early onset in which neurologic symptoms are evident within weeks after birth. Symptoms and signs are confined to the nervous system; no visceral involvement is present. The clinical course has been divided into three stages. In stage I, patients who appeared relatively normal after birth exhibit hyperirritability, vomiting, episodic fevers, hyperesthesia, tonic spasms with light or noise stimulation, stiffness, and seizures. CHAPTER 23 Lysosomal Storage, Peroxisomal, and Glycosylation Disorders and Smith-Lemli-Opitz Syndrome in the Neonate Peripheral neuropathy is present, but reflexes are increased. Stage II is marked by CNS deterioration and hypertonia that progresses to hypotonia and flaccidity. Deep tendon reflexes are eventually lost. Patients with stage III disease are decerebrate, deaf, and blind with hyperpyrexia, hypersalivation, and frequent seizures. Routine laboratory findings are unremarkable except for an elevation of cerebrospinal fluid protein. Cerebral atrophy and demyelination become evident in the CNS, and segmental demyelination, axonal degeneration, fibrosis, and macrophage infiltration are common in the peripheral nervous system. The segmental demyelination of peripheral nerves is demonstrated by the finding of decreased motor nerve conduction. The white matter is severely depleted of all lipids, especially glycolipids, and nerve and brain biopsies show globoid cells. Death from hyperpyrexia, respiratory complications, or aspiration occurs at a median age of 13 months. GM1 Gangliosidosis Etiology Infantile GM1 gangliosidosis is caused by a deficiency in lysosomal β-galactosidase. The enzyme cleaves the terminal galactose in a β linkage from oligosaccharides, keratan sulfate, and GM1 ganglioside. Deficiency of the enzyme results in storage of GM1 ganglioside and oligosaccharides. Clinical severity correlates with the extent of substrate storage and residual enzyme activity. The same enzyme is deficient in Morquio disease type B. Clinical Features Age of onset ranges from prenatal to adult. Infantile or type 1 GM1 gangliosidosis may be evident at birth as coarse and thick skin, hirsutism on the forehead and neck, and coarse facial features consisting of a puffy face, frontal bossing, depressed nasal bridge, maxillary hyperplasia, large and low-set ears, wide upper lip, moderate macroglossia, and gingival hypertrophy. These dysmorphic features, however, are not always obvious in the neonate. A retinal cherry-red spot is seen in 50% of patients, and corneal clouding is often observed. Shortly after birth, or by 3 to 6 months of age, failure to thrive and hepatosplenomegaly become evident, as does neurologic involvement with poor development, hyperreflexia, hypotonia, and seizures. Cranial imaging shows diffuse atrophy of the brain, enlargement of the ventricular system, and evidence of myelin loss in white matter. The neurologic deterioration is progressive, resulting in generalized rigidity and spasticity and sensorimotor and psychointellectual dysfunction. By 6 months of age, skeletal features are present, including kyphoscoliosis and stiff joints with generalized contractures, and striking bone changes are seen—vertebral beaking in the thoracolumbar region, broadening of shafts of the long bones with distal and proximal tapering, and widening of the metacarpal shafts with proximal pinching of four lateral metacarpals. Tissue biopsy samples demonstrate neurons filled with membranous cytoplasmic bodies and various types of inclusions as well as foam cells in the bone marrow. Death generally occurs before 2 years of age. A severe neonatalonset type of GM1 gangliosidosis with cardiomyopathy has also been described (Kohlschütter et al, 1982). 245 Farber Lipogranulomatosis Etiology Farber lipogranulomatosis results from a deficiency of lysosomal acid ceramidase. Ceramidase catalyzes the degradation of ceramide to its long-chain base, sphingosine, and a fatty acid. Clinical disease is a consequence of storage of ceramide in various organs and body fluids. Clinical Features Four types of Farber lipogranulomatosis can manifest in the neonatal period. Type I, classic disease, is a unique disorder with onset from approximately 2 weeks to 4 months of age. Patients exhibit hoarseness progressing to aphonia, feeding and respiratory difficulties, poor weight gain, and intermittent fever caused by granuloma formation, and swelling of the epiglottis and larynx. Palpable nodules appear over joints and pressure points, and joints become painful and swollen. Later, joint contractures and pulmonary disease appear. Liver and cardiac involvement can occur, and patients can have a subtle retinal cherryred spot. Severe and progressive psychomotor impairment can occur, as can seizures, decreased deep tendon reflexes, hypotonia, and muscle atrophy. Affected patients die in early infancy, usually from pulmonary disease. Type 2, or intermediate, Farber lipogranulomatosis manifests from birth to 9 months of age as joint and laryngeal involvement and nodules. Death occurs in early childhood. Type 3 disease (mild) manifests slightly later, from approximately 2 months to 20 months of age, with survival into the third decade. Clinically types 2 and 3 are both dominated by subcutaneous nodules, joint deformity, and laryngeal involvement. Liver and pulmonary involvement may be absent. Two thirds of patients have a normal intelligence quotient. Type 4, or neonatal visceral, Farber lipogranulomatosis manifests at birth as hepatosplenomegaly caused by massive histiocyte infiltration of the liver and spleen, with infiltration also in the lungs, thymus, and lymphocytes. Subcutaneous nodules and laryngeal involvement may be subtle. Death occurs by 6 months of age. In all types of Farber lipogranulomatosis, tissue biopsy samples show granulomatous infiltration, foam cells, and lysosomes with comma-shaped, curvilinear tubular structures called Farber bodies. Cerebrospinal fluid protein may be elevated in patients with type 1 disease. Sialidosis Etiology Sialidosis is caused by a deficiency of neuraminidase, which is responsible for the cleavage of terminal sialyl linkages of several oligosaccharides and glycopeptides. The defect results in multisystem lysosomal accumulation of sugars rich in sialic acid. Clinical Features Type I sialidosis is characterized by retinal cherry-red spots and generalized myoclonus with onset generally in the second decade of life. Type II is distinguished from type I by the early onset of a progressive, severe phenotype with somatic features. Type II is often subdivided 246 PART VI Metabolic and Endocrine Disorders of the Newborn into juvenile, infantile, and congenital forms. Congenital sialidosis begins in utero and manifests at birth as coarse features, facial edema, hepatosplenomegaly, ascites, hernias, and hypotonia, and occasionally frank hydrops fetalis. Radiographs demonstrate dysostosis multiplex and epiphyseal stippling. Delayed mental development is quickly apparent. The patient may have recurrent infections. Severely dilated coronary arteries, excessive retinal vascular tortuosity, and an erythematous macular rash may also be features of this disease (Buchholz et al, 2001). Most patients are stillborn or die before 1 year of age. Age of onset for the infantile form of sialidosis ranges from birth to 12 months. Clinical features are coarse facial features, organomegaly, dysostosis multiplex, retinal cherry-red spot, and mental retardation. Death occurs by the second or third decade. In both types of sialidosis, vacuolated cells can be seen in almost all tissues, and bone marrow foam cells are present. Wolman Disease Etiology Galactosialidosis Wolman disease is caused by lysosomal acid lipase deficiency, which is an enzyme involved in cellular cholesterol homeostasis and responsible for hydrolysis of cholesterol esters and triglycerides. The result of enzyme deficiency is defective release of free cholesterol from lysosomes, which leads to upregulation of LDL receptors and 3-hydroxy-3methylglutaryl-CoA reductase activity. De novo synthesis of cholesterol and activation of receptor-mediated endocytosis of LDL then occur, leading to further deposition of lipid in lysosomes. The result is the accumulation of cholesterol esters and triglycerides in most tissues of the body, including the liver, spleen, lymph nodes, heart, blood vessels, and brain. An extreme level of lipid storage occurs in cells of the small intestine, particularly in the mucosa. In addition, neurons of the myenteric plexus demonstrate a high level of storage, with evidence of neuronal cell death, which may account for prominence of gastrointestinal symptoms (Wolman, 1995). Etiology Clinical Features Galactosialidosis results from a deficiency of two lysosomal enzymes, neuraminidase and β-galactosidase. The primary defect in galactosialidosis has been found to be a defect in protective protein–cathepsin A, an intralysosomal protein that protects the two enzymes from premature proteolytic processing. The protective protein has catalytic and protective functions, and the two functions appear to be distinct. Deficiency of enzymes results in the accumulation of sialyloligosaccharides in tissue lysosomes and in excreted body fluids. Clinical presentation of Wolman disease is within weeks of birth, with evidence of malnutrition and malabsorption, including symptoms of vomiting, diarrhea, steatorrhea, failure to thrive, abdominal distention, and hepatosplenomegaly. Adrenal calcifications may be seen on radiographs, and adrenal insufficiency appears. The presence of adrenal calcifications in association with hepatosplenomegaly and gastrointestinal symptoms is strongly suggestive of Wolman disease. Later, mental deterioration becomes apparent. Laboratory findings include anemia secondary to foam cell infiltration of the bone marrow and evidence of adrenal insufficiency. The serum cholesterol level is normal. Death usually occurs before 1 year of age. Clinical Features Galactosialidosis has been divided into three phenotypic subtypes based on age at onset and severity of clinical manifestations. Most cases occur in adolescence and adulthood, but early infantile and late infantile presentations occur. Patients develop early infantile galactosialidosis between birth and 3 months of age with ascites, edema, coarse facial features, inguinal hernias, proteinuria, hypotonia, and telangiectasias, and, occasionally, frank hydrops fetalis. Patients subsequently demonstrate organomegaly, including cardiomegaly progressing to cardiac failure, psychomotor delay, and skeletal changes, particularly in the spine. Ocular abnormalities can occur, including corneal clouding and retinal cherry-red spots. Death occurs at an average age of 8 months, usually from cardiac and renal failure. Galactosialidosis can be a cause of recurrent fetal loss or recurrent hydrops fetalis. Late infantile galactosialidosis manifests in the first months of life as coarse facial features, hepatosplenomegaly, and skeletal changes consistent with dysostosis multiplex. Cherry-red spots and corneal clouding may also be present. Neurologic involvement may be absent or mild. Valvular heart disease is a common feature, as is growth retardation, partially because of spinal involvement and often in association with muscular atrophy. Early death is not a feature of the late infantile form. Vacuolated cells in blood smears and foam cells in bone marrow are present in all forms of galactosialidosis. Infantile Sialic Acid Storage Disease Etiology Infantile sialic acid storage disease is caused by a defective lysosomal sialic acid transporter that is responsible for efflux of sialic acid and other acidic monosaccharides from the lysosomal compartment. The defective transporter results in greater storage of free sialic acid and glucuronic acid within lysosomes and increased sialic acid excretion. Clinical Features Infantile sialic acid storage disease often manifests at birth as mildly coarse features, hepatosplenomegaly, ascites, hypopigmentation, and generalized hypotonia. Mild dysostosis multiplex may be seen on radiographs. Failure to thrive and severe mental and motor retardation soon appear. Cardiomegaly may be present. Corneas are clear, but albinoid fundi have been reported (Lemyre et al, 1999). Vacuolated cells are seen on a tissue biopsy sample, and electron microscopy demonstrates swollen lysosomes filled with finely granular material. CNS changes include myelin loss, axonal spheroids, gliosis, and neuronal storage. Death occurs in early childhood. Infantile sialic acid storage disease can also manifest as fetal ascites, nonimmune CHAPTER 23 Lysosomal Storage, Peroxisomal, and Glycosylation Disorders and Smith-Lemli-Opitz Syndrome in the Neonate fetal hydrops, or infantile nephrotic syndrome (Lemyre et al, 1999). I-Cell Disease (Mucolipidosis Type II) Etiology In normal cells, targeting of enzymes to lysosomes is mediated by receptors that bind a mannose-6-­phosphate recognition marker on the enzyme. The recognition marker is synthesized in a two-step reaction in the Golgi complex. It is the enzyme that catalyzes the first step of this process, uridine diphosphate–N-­acetylglucosamine: lysosom­al enzyme N-acetylglucosaminyl-1-­phosphotransferase, that is defective in I-cell disease. As a result, the enzymes lack the mannose-6-phosphate recognition signal, and the newly synthesized lysosomal enzymes are secreted into the extracellular matrix instead of being targeted to the lysosome. Consequently, multiple lysosomal enzymes are found in plasma in 10- to 20-fold their normal concentrations. Affected cells, especially fibroblasts, show dense inclusions of storage material that probably consists of oligosaccharides, glycosaminoglycans, and lipids; these are the inclusion bodies from which the disease name is derived. This disorder is found more frequently in Ashkenazi Jews, because of a putative founder effect. Clinical Features I-cell disease can manifest at birth as coarse features, corneal clouding, organomegaly, hypotonia, and gingival hyperplasia. Birthweight and length are often below normal. Kyphoscoliosis, lumbar gibbus, and restricted joint movement are often present, and there may be hip dislocation, fractures, hernias, or bilateral talipes equinovarus. Dysostosis multiplex may be seen on radiographs. Severe psychomotor retardation, evident by 6 months of age, and progressive failure to thrive occur. The facial features become progressively more coarse, with a high forehead, puffy eyelids, epicanthal folds, flat nasal bridge, anteverted nares, and macroglossia. Linear growth slows during the first year of life and halts completely thereafter. The skeletal involvement is also progressive, with development of increasing joint immobility and claw-hand deformities. Respiratory infections, otitis media, and cardiac involvement are common complications. Death usually occurs in the first decade of life because of cardiorespiratory complications. Mucolipidosis Type IV Etiology Although mucolipidosis type IV is associated with a partial deficiency of the lysosomal enzyme ganglioside sialidase, a deficiency of mucolipin 1 (TRPML1), a member of the transient receptor potential TRMPL subfamily of channel proteins, is the cause of the disorder (Bargal et al, 2000; Sun et al, 2000). Mutations in the MCOLN1 gene result in lysosomal storage of lipids such as gangliosides, plus water-soluble materials such as glycosaminoglycans and glycoproteins in cells from almost all tissues. 247 Clinical Features The age of onset for mucolipidosis type IV ranges from infancy to 5 years. Presenting features are corneal clouding (may be congenital), retinal degeneration, blindness, hypotonia, and mental retardation. Survival of affected patients into the fourth decade of life has been reported (Chitayat et al, 1991). Cytoplasmic inclusions are noted in many cells, including those in conjunctiva, liver, and spleen, as well as fibroblasts. Mucopolysaccharidosis Type VII (Sly Disease) Etiology Sly disease is a member of a group of lysosomal storage disorders that are caused by a deficiency of enzymes catalyzing the stepwise degradation of glycosaminoglycans. Skeletal and neurologic involvement is variable. There is a wide spectrum of clinical severity among the mucopolysaccharidoses and even within a single enzyme deficiency. Most of these disorders manifest in childhood, but type VII is included in this chapter because of its well-­recognized neonatal and infantile presentations. Sly disease is caused by β-glucuronidase deficiency and results in lysosomal accumulation of glycosaminoglycans, including dermatan sulfate, heparan sulfate, and chondroitin sulfate, causing cell, tissue, and organ dysfunction. Clinical Features Sly disease can manifest as a wide spectrum of severity. Patients with the early-onset or neonatal form may have coarse features, hepatosplenomegaly, moderate dysostosis multiplex, hernias, and nonprogressive mental retardation. Corneal clouding is variably present. Frequent episodes of pneumonia during the first year of life are common. Short stature becomes evident. Granulocytes have coarse metachromic granules. A severe neonatal form associated with hydrops fetalis, and early death has been recognized frequently. Milder forms of the disease with later onset are also known. DIAGNOSIS, MANAGEMENT, AND PROGNOSIS Growing recognition of lysosomal storage disorders in the neonate has led to expansion of the spectrum of possible clinical presentation in the newborn period. Diagnostic tools and options for treatment also continue to advance. For example, efforts are currently underway to develop newborn screening for mucopolysaccharidoses (Whitley et al, 2002), with the goal to offer treatment with enzyme infusion or bone marrow transplantation (BMT) to affected babies (Vogler et al, 1999). The state of New York has implemented newborn screening for Krabbe disease using dried blood spots. The test uses a tandem massspectrometry–based enzyme analysis (Li et al, 2004a). This test has resulted in a fairly large number of positive newborn screens for Krabbe disease, most of which appear to be false positives, including enzyme perturbations that are not linked with clinical disease (Duffner et al, 2009). 248 PART VI Metabolic and Endocrine Disorders of the Newborn As a consequence, an expert advising panel, the Krabbe Consortium of New York State, has been generated to establish standardized clinical evaluation guidelines. The goal is to help physicians determine which infant with a positive newborn screen may express disease and require treatment, such as hematopoietic stem cell transplantation in early infancy. The neonatologist is urged to work closely with appropriate experts to explore diagnostic and treatment protocols on an individual basis. Larger panels of multiplex testing for various other LSDs are in the testing stages (Li et al, 2004b) and some states in the United States are poised to begin implementing LSD newborn screening. Currently a federal advisory committee actively reviews and makes recommendations to the U.S. Secretary of Health and Human Services about the introduction of new newborn screening tests in the United States, with the aim of vetting proposed tests for need, cost effectiveness, and availability of effective and timely therapy. Recognizing lysosomal storage disorders in the newborn period can be difficult, because they often mimic more common causes of illness in newborns, such as respiratory distress, nonimmune hydrops fetalis, liver disease, and sepsis. The initial step in the diagnosis of these disorders is to consider them in the differential diagnosis of a sick or unusual-appearing newborn. At times the phenotype may suggest a specific diagnosis, such as respiratory distress and painful, swollen joints in Farber lipogranulomatosis or gastrointestinal symptoms, hepatosplenomegaly, and adrenal calcifications in Wolman disease. Subtle dysmorphic features, coarsening of features, and radiographic evidence of dysostosis multiplex are also strong indications that lysosomal storage disorders should be considered. Routine laboratory findings are often normal or nonspecific. Affected infants do not have episodes of acute metabolic decompensation. Anemia and thrombocytopenia may be seen because of bone marrow involvement. Vacuolated cells may be found in peripheral blood, but the absence of this finding does not exclude lysosomal storage disease. Elevated cerebrospinal fluid protein is seen in Krabbe disease and Farber lipogranulomatosis type I. Nonimmune hydrops fetalis deserves special mention. The physician must consider LSDs as the cause of nonimmune hydrops fetalis or unexplained ascites in the affected newborn infant. The following LSDs are potential causes: sialidosis type II, MPS types VII and IV, ISSD, Salla disease, galactosialidosis, Gaucher disease type II, GM1 gangliosidosis, I-cell disease, Niemann-Pick disease types A and C, Wolman disease, and Farber’s disease (StaretzChacham et al, 2009). The mechanisms of edema are unclear. Furthermore, not all of the 13 LSDs routinely appear in the neonatal period. Directed analysis of urine is helpful for conditions in which characteristic metabolites are excreted in urine. One- or two-dimensional electrophoresis or thin-layer chromatography can detect excess excretion of urine glycosaminoglycans, oligosaccharides, or free sialic acid, but all urinary tests for the diagnosis of lysosomal storage disorders can have false-negative results. Examination of bone marrow or other tissues may demonstrate storage macrophages in Gaucher disease and in Niemann-Pick disease types A and C. Small skin or conjunctival biopsy specimens may demonstrate storage within lysosomes in most of these disorders. Definitive diagnosis for all lysosomal storage disorders, except for Niemann-Pick C disease, is confirmed by enzymatic assays in serum, leukocytes, fibroblasts, or a combination of these. The diagnosis of Niemann-Pick C disease requires measurement of cellular cholesterol esterification and documentation of a characteristic pattern of filipincholesterol staining in cultured fibroblasts during LDL uptake. Analysis of DNA mutations may be helpful for the diagnosis of Niemann-Pick C disease, Gaucher disease, and some other conditions, and it will become increasingly available for other conditions. An imperfect genotype-­ phenotype correlation impedes the use of mutation analysis as a prognostic tool. In addition, prenatal diagnosis is available for most lysosomal storage disorders through the use of enzyme assays performed on amniocytes or chorionic villus cells or measurements of levels of stored substrate in cultured cells or amniotic fluid. As mutation analysis becomes more prevalent, it will increasingly substitute for biochemical and enzymatic methods. These conditions must also be considered in the dying infant, and the neonatologist must be prepared to request the appropriate samples for diagnosis at the time of death. In surviving patients, treatment and management must be considered. All the lysosomal storage disorders are chronic and progressive conditions for which there is no curative treatment. Gene transfer therapy holds promise, but is not currently available for lysosomal storage disorders. With few exceptions, current standard medical management is supportive and palliative. Patients must be continually reassessed for evidence of disease progression and associated complications. These complications manifest at variable ages and can include hydrocephalus, valvular heart disease, joint limitation, and obstructive airway disease. For several disorders, particularly neonatal Gaucher disease and Niemann-Pick C disease, splenectomy may be indicated to improve severe anemia and thrombocytopenia. This procedure enhances the risk of serious infections, and it can accelerate the progression of disease at other sites. Patients with Krabbe disease may have significant pain of radiculopathy and spasms, and alleviation of that pain is important for the patient’s comfort. The administration of glutamic acid transaminase inhibitor, vigabatrin, has been used in a small number of patients with Krabbe disease, because part of the pathology may involve a secondary deficiency of γ-aminobutyric acid (Barth, 1995). Low-dose morphine has also been reported to improve the irritability associated with this disorder (Stewart et al, 2001). Enzyme replacement therapy with imiglucerase (Cerezyme), a recombinant enzyme, is available for Gaucher disease. Although enzyme replacement therapy has successfully reversed many of the systemic manifestations of the disease, it has been suggested that enzyme replacement therapy should not be given to patients with Gaucher disease type 2 who already have severe neurologic signs, because no substantial improvement has been demonstrated to occur in the neurologic symptoms of patients treated (Erikson et al, 1993; Gaucher disease, 1996). Bone marrow transplantation has been tried for a variety of lysosomal storage disorders. The rationale for the CHAPTER 23 Lysosomal Storage, Peroxisomal, and Glycosylation Disorders and Smith-Lemli-Opitz Syndrome in the Neonate procedure is that circulating blood cells derived from the transplanted marrow become a source of the missing enzyme. Results of bone marrow transplantation in disorders of glycosaminoglycans show that after successful engraftment, leukocyte and liver tissue enzyme activity normalizes, organomegaly decreases, and joint mobility increases. Skeletal abnormalities stabilize but do not improve. Whether brain function can be improved in patients with CNS disease remains questionable. Some patients maintained their learning capability or intelligence quotient, but others continued to deteriorate. Clinical experience and studies in animal models indicate that BMT before the onset of neurologic symptoms can prevent or delay the occurrence of symptoms, whereas there is no clear benefit if transplantation is performed when symptoms are already present (Hoogerbrugge et al, 1995). BMT in patients with nonneuropathic Gaucher disease can result in complete disappearance of all symptoms; however, the procedure is associated with significant risks (Hoogerbrugge et al, 1995) that must be balanced against lifelong enzyme replacement therapy. Currently it is unclear to what extent patients with the neuropathic types of Gaucher disease (types 2 and 3) would benefit from transplantation; therefore it is generally not recommended. BMT has also been attempted in a small number of patients with infantile Krabbe disease, Farber lipogranulomatosis, and Niemann-Pick A disease. The outcome after transplantation for these few patients has been poor, with continued disease progression and death. Krivit et al (2000) reported successful long-term bone marrow engraftment in a patient with Wolman disease that resulted in normalization of peripheral leukocyte lysosomal acid lipase enzyme activity. The patient’s diarrhea resolved; cholesterol, triglyceride; and liver function values normalized; and the patient attained developmental milestones. Lysosomal storage diseases are not all equally amenable to BMT, and the use of BMT as a treatment modality for most lysosomal storage disorders remains uncertain. In a small number of cases, BMT has been performed in utero after prenatal diagnosis showing an affected infant, and experimental protocols are available for families who wish to pursue this option. The goal of therapy for a dietary protocol proposed for the treatment of Wolman disease is reduced accumulation of storage material in intestine and phagocytes. The diet, which should be started as soon as the diagnosis of Wolman disease is suggested, consists of (1) discontinuing breastfeeding or feeding with a formula containing triglycerides and cholesterol esters and (2) keeping the infant on a fatty ester–free diet (Wolman, 1995). The diet should include all necessary vitamins, including fat-­ soluble vitamins. In addition, daily smearing of the skin of a different extremity with a small amount (10 to 50 μL) of sunflower or safflower oil or preferably soy, canola, flax, cod liver, or algal oil is required for preventing essential fatty acid deficiency, which complicates the restricted diet (Wolman, 1995). The absorption of fatty acids through skin spares the gastrointestinal tract from accumulation and is associated with the formation of phospholipids and triglycerides (Wolman, 1995). Preliminary results of this approach suggest that treatment appears to halt disease progression. 249 CONGENITAL DISORDERS OF GLYCOSYLATION ETIOLOGY Previously called carbohydrate-deficient glycoprotein syndromes, CDGs are a large and increasing family of genetic diseases resulting from deficient glycosylation of glycoconjugates, mainly glycoproteins and glycolipids. Most extracellular proteins, such as serum proteins (e.g., transferrin and clotting factors), most membrane proteins, and several intracellular proteins (e.g., lysosomal proteins), are glycosylated proteins. Glycosylation, the addition of sugar chains (glycans) to proteins, occurs in every human cell and serves a number of functions, including aiding in correct folding of the nascent protein, participating in cell adhesion phenomena, protecting against premature proteolytic destruction, and modifying biologic function (Grünewald, 2007). The glycans are defined by their linkage to the protein N-glycan or O-glycan. N-Glycosylation consists of the assembly of a glycan on and in the endoplasmic reticulum and its attachment to a particular asparagine of target proteins, followed by remodeling of this glycan mainly in Golgi (Jaeken and Matthijs, 2007); therefore it is a twopart process of assembly and processing. O-Glycosylation consists of assembly of a glycan and its attachment to a serine or threonine of a target protein, or the attachment of a monosaccharide (mannose, fructose, or xylose) to one of these amino acids. No processing pathway is present in O-glycosylation (Jaeken and Matthijs, 2007). Combined N- and O-glycosylation defects and lipid glycosylation defects have also been described (Jaeken and Matthijs, 2007). Given the ubiquitous occurrence of glycoproteins and the number of apparent genes involved in glycosylation (more than 200; approximately 1% of the human genome), it is not surprising that the number of described CDG defects is increasing rapidly and clinical manifestations are diverse (Jaeken, 2006; Jaeken et al, 2008; Morava et al, 2008b). Currently, 21 disorders in protein N-glycosylation, 12 in protein O-glycosylation, five in both protein N- and O-glycosylation, and two in lipid glycosylation have been described (Jaeken, 2006; Jaeken and Matthijs, 2007; ­Marklová and Albahri, 2007). The number of described disorders is expected to continue to grow. The disorders of N-glycosylation have been divided into two primary categories. CDG-I disorders result from defects in N-glycan assembly (designated CDG-Ia to -Im). On isoelectrofocusing of serum transferrin, the most widely used screening test for N-glycosylation disorders, a type I pattern is observed. This pattern is characterized by a decrease of anodal fractions and an increase of disialotransferrin and asialotransferrin. A type II pattern, showing an increase of trisialofractions, monosialofractions, or both is seen in CDG-II defects, which represent defects in N-glycan processing (designated CDG-IIa-IIf). O-Glycosylation defects have been found to be causative in a number of muscular dystrophies with reduced glycosylation of α-dystroglycan. These disorders, collectively referred to as α-dystroglycanopathies, encompass previously described disorders, such as Walker-­Warburg syndrome, muscle-eye-brain disease, Fukuyama congenital muscular 250 PART VI Metabolic and Endocrine Disorders of the Newborn dystrophy, and limb-girdle muscular dystrophy (Mercuri et al, 2009; Topaloglu, 2009). Six known or putative glycosyltransferase genes have been identified in these disorders (Godfrey et al, 2007). In general, this is a heterogeneous group of autosomal recessive disorders with a wide spectrum of clinical severity, and it shares the common pathologic feature of hypoglycosylated α-dystroglycan (Godfrey et al, 2007). α-Dystroglycan is a major component of the dystrophin-associated glycoprotein complex that forms a link between the actin-­ associated cytoskeleton and extracellular matrix. It is a highly glycosylated peripheral membrane protein that binds many of its extracellular matrix partners through its carbohydrate modifications (Godfrey et al, 2007). In the dystroglycanopathies, these modifications are either absent or reduced, resulting in decreased binding of ligands (Barresi and Campbell, 2006). O-Glycosylation defects have also been described in hereditary multiple exostoses syndrome, familial tumoral calcinosis, Schneckenbecken dysplasia, spondylocostal dysostosis type 3, Peters plus syndrome, and the progeria variant of EhlersDanlos syndrome (Grünewald, 2007; Jaeken et al, 2008; Jaeken and Matthijs, 2007). Combined N- and O-glycosylation defects are important because they appear to affect trafficking in the glycosylation machinery (Grünewald, 2007). The disruption of multiple glycosylation pathways is caused by mutations of the conserved oligomeric Golgi (COG) complex. This large complex spanning eight subunits plays a key role in protein transport between the endoplasmic reticulum and Golgi and within the Golgi complex (Grünewald, 2007; Jaeken, 2006). COG7 deficiency, first described in two siblings with poor intrauterine growth, dysmorphic features, encephalopathy, cholestatic liver disease, and perinatal asphyxia, shows a partial combined N- and O-glycosylation defect caused by decreased transport of CMP-sialic acid and UDP-galactose into Golgi and reduced activity of two glycosyltransferases involved in the galactosylation and sialylation of O-glycans (Jaeken, 2006). An autosomal recessive cutis laxa syndrome has recently been found to be also associated with a combined glycosylation defect (Morava et al, 2008a). Two disorders of lipid glycosylation have been described. Amish infantile epilepsy was the first identified and is caused by a defect of lactosylceramide α-2,3 sialyltransferase (GM3 synthase; Jaeken, 2006). This enzyme catalyzes the initial step in the biosynthesis of most complex gangliosides from lactosylceramide (Jaeken and Matthijs, 2007). The defect causes accumulation of lactosylceramide associated with decreased gangliosides of the GM3 and GD3 series (Jaeken, 2006). Glycosylphosphatidylinositol deficiency is the second disorder in glycolipid glycosylation described, but the first reported genetic defect in glycosylation of the glycosylphosphatidylinositol (GPI) anchor (Jaeken and Matthijs, 2007). Glycosylphosphatidylinositol-anchored proteins have heterogeneous functions as enzymes or adhesion molecules. Finally, the term CDG-x indicates syndromes with strong evidence for a glycosylation defect, but in which the defective gene has yet to be identified (Jaeken and Matthijs, 2007). This clinically heterogeneous group is also growing rapidly. CLINICAL FEATURES The phenotypic spectrum of CDG defects is extremely broad and ranges from mild to severe disease and from a single-organ system to multisystem disease. Clinical features alone are insufficient to define the CDG subtype. CDG should be considered a possible diagnosis in any unexplained clinical condition, but especially in multiorgan disease with neurologic involvement. Discussion of the clinical presentation of all forms of CDG is beyond the scope of this chapter (Table 23-2). CDG-Ia is the classic and most common presentation. The basic defect in CDG-Ia is a deficiency of the enzyme phosphomannomutase, which is required for the early steps of protein glycosylation (Jaeken, 2006; Jaeken et al, 2001; van Schaftingen and Jaeken, 1995). Patients with CDG-Ia at birth exhibit dysmorphic features consisting of a high nasal bridge, prominent jaw, large ears, and inverted nipples, feeding difficulties, and subsequent growth failure, hypotonia, lipocutaneous abnormalities (including prominent fat pads on the buttocks), and mild to moderate hepatomegaly. The clinical progression of this disorder is divided into four stages. In stage I—the infantile, multisystem stage—patients show evidence of multisystem involvement, including variable strokelike episodes, thrombotic disease, liver dysfunction, pericardial effusions and cardiomyopathy, proteinuria, and retinal degeneration. The coagulopathy likely stems from the number of clotting and anticlotting proteins that are N-linked glycoproteins. Mental retardation, peripheral neuropathy, and decreased nerve conduction velocities are observed. Strabismus and alternating esotropia are present in almost all patients, and retinitis pigmentosa and abnormalities of the electroretinogram are present in most. Cranial imaging shows varying degrees of cerebral, cerebellar, and brainstem hypoplasia. Electroencephalogram results are usually normal. Liver biopsy samples typically show steatosis and fibrosis, and multicystic changes in kidneys have been noted. Stage II, the childhood stage, is characterized by ataxia and mental retardation. Skeletal abnormalities may become more prominent, consisting of contractures, kyphoscoliosis, pectus carinatum, and short stature. Stage III, generally occurring in the teenage years, is characterized primarily by lower extremity atrophy. Adulthood, or stage IV, is characterized by hypogonadism. In general, patients have an extroverted disposition and happy appearance. Approximately 20% of patients die during the first year of life because of severe infection, liver failure, or cardiac insufficiency. Patients with CDG-Ib are unique among patients with these disorders. They may have vomiting, diarrhea, hypoglycemia, and liver disease (coagulopathy, hepatomegaly, hepatic fibrosis). In addition, they often have a proteinlosing enteropathy (also seen in CDG-Ih; Jaeken and Matthijs, 2007). Development is normal. The remaining forms of CDG-I and CDG-II are similar in presentation to type Ia. Additional features include seizures, normal cerebellar development, delayed myelination, optic atrophy, blindness, frequent infections, hypoventilation and apnea, and further dysmorphic features such as adducted thumbs, high-arched palate, coarse facies, widely spaced nipples, and low-set ears. 251 CHAPTER 23 Lysosomal Storage, Peroxisomal, and Glycosylation Disorders and Smith-Lemli-Opitz Syndrome in the Neonate TABLE 23-2 Common Congenital Disorders of Glycoprotein Types of CDG Findings Ia Ib Ic Id Ie IIa IIb x Enzyme defect Phosphomannomutase Phosphomannose isomerase α1,3Glucosyltransferase α1,3Mannosyl transferase Dol-P-Man synthase GlcNAc transferase 2 Glucosidase I Unknown Dysmorphic features + +/– +/– +/– + + + + Psychomotor retardation + – + + + + + + Hypotonia + +/– + + + + + + Cerebellar hypoplasia + – +/– – +/– – – – +/– +/– +/– + + + + – – Strabismus Optic atrophy Cortical blindness – – – Seizures Eye findings Strabismus, esotropia Liver disease + + – – + + + – Coagulopathy + + + – + + + – Multiorgan involvement, peripheral neuropathy, subcutaneous fat distribution, inverted nipples, strokelike episodes, cardiomyopathy, ataxia, microcephaly, hypothyroidism Proteinlosing enteropathy, cyclic vomiting, diarrhea, hypoglycemia Microcephaly, feeding difficulties, ataxia Stereotype behavior, frequent infections, ventricular septal defect, widely spaced nipples, delayed myelination Early death, generalized edema, hypoventilation, apnea, demyelinating polyneuropathy Leukocyte adhesion deficiency syndrome type II (guanosine diphosphate– fucose transporter); phenotype: elevated peripheral leukocytes, absence of CD 15, Bombay blood group phenotype, failure to thrive, recurrent infections, short arms and legs, simian crease Other Microcephaly, reduced responsiveness, adducted thumbs Microcephaly, delayed myelination Adapted from Westphal V, Srikrishna G, Freeze H: Congenital disorders of glycosylation: Have you encountered them? Genet Med 2:329-337, 2000. +, Present; –, absent; +/–, occasionally present. The phenotypic presentation of dystroglycanopathy is extremely variable. At the severe end of the spectrum are individuals with Walker-Warburg syndrome, muscleeye-brain disease, and Fukuyama congenital muscular dystrophy. These conditions are characterized by congenital muscular dystrophy with severe structural brain and eye abnormalities and death typically before the age of 1 year (Godfrey et al, 2007; Jaeken, 2006; Jaeken and Matthijs, 2007). Toward the more mild end of the spectrum are individuals in adulthood with limb-girdle muscular dystrophy with no brain or eye involvement (Godfrey et al, 2007; Jaeken and Matthijs, 2007). Intermediate phenotypes lie between the two extremes. The presence of carbohydrate-deficient transferrin in serum and cerebrospinal fluid is a distinctive biochemical feature of CDG. Laboratory findings are often nonspecific (e.g. elevated liver function studies or hypoalbuminemia), but concentrations of plasma glycoproteins such as α1antitrypsin, thyroxin-binding globulin, and transferrin are frequently abnormal (Marklová and Albahri, 2007). Clotting factors such as factors V, XI, II, X, antithrombin III, proteins C and S, and thyroid hormones (triiodothyronine, thyroxine, and reverse T3) are also frequently decreased (Marklová and Albahri, 2007; Morava et al, 2008b). Abnormal clotting is an important indicator of a glycosylation disorder (Morava et al, 2008b). 252 PART VI Metabolic and Endocrine Disorders of the Newborn DIAGNOSIS CDG should be considered in newborns with several of the following features: ll Neurologic signs, including hypotonia, hyporeflexia, or seizures ll Ophthalmic signs, including abnormal eye movements, cataracts, glaucoma, optic nerve atrophy, or retinitis pigmentosa ll Hepatic and gastrointestinal signs such as ascites or hydrops, hepatomegaly, diarrhea, and proteinlosing enteropathy ll Endocrinologic signs, including hyperinsulinemic hypoglycemia and hypothyroidism ll Signs of renal or cardiac disease ll Congential muscular dystrophy ll Congenital joint contractures ll Dysmorphic features, microcephaly, structural brain anomalies, or abnormal skin findings Isoelectrofocusing of serum transferrin is the screening method of choice, but only to detect defects of N-­glycosylation. Confirmatory enzyme assays or molecular studies are required to pinpoint the specific defect. Urine oligosaccharide analysis, Bombay blood phenotype, serum apoC-III screening, or membrane bound sialyl-LewisX antigen may be helpful in distinguishing subclasses of CDG II disease (Marklová and Albahri, 2007). If suspicion of CDG remains, further structural analysis of the lipid-linked oligosaccharide (LLO) and the peptide-protein N-linked oligosaccharide should follow, if available (Marklová and Albahri, 2007). Patients with CDG can often be identified through neonatal screening for congenital hypothyroidism, because of an associated thyroid-binding globulin deficiency and an increased thyroid-stimulating hormone level. Prenatal diagnosis by transferrin isoelectric focusing is not reliable. Prenatal diagnosis is possible in all types of CDG for which the molecular defect is known (Grünewald, 2007). The vast majority of CDG disorders are autosomal recessive disorders or presumed autosomal recessive disorders; hereditary multiple exostoses syndrome is autosomal dominant. TREATMENT AND MANAGEMENT The treatment and management for most types of CDGs are primarily supportive and palliative. There is no curative or corrective treatment. In infancy, evidence of multisystem involvement and the resulting complications must be treated promptly. There is substantial mortality in the first years of life because of severe infection or vital organ failure (Grünewald, 2007; Jaeken, 2006). The exception is treatment of CDG-Ib with oral mannose therapy. In this disorder, oral mannose effectively bypasses the impaired pathway and allows glycosylation to continue (Freeze, 1998; Jaeken, 2006; Jaeken and Matthijs, 2007). Therapy improves the protein-losing enteropathy and liver disease (Grünewald, 2007). Oral fucose therapy has also been used in patients with CDG-IIc, a GDP-fucose transporter defect (Grünewald, 2007). Therapy appears to improve the fucosylation of glycoproteins and to improve control of recurrent infections, but it has no effect on neurologic complications of the disorder (Grünewald, 2007). PEROXISOMAL DISORDERS DISORDERS OF PEROXISOME BIOGENESIS Peroxisomes are single-membrane–bound cellular organelles that contain no internal structure or DNA and are characterized by an electron-dense core and a homogeneous matrix. Peroxisomes are found in all cells and tissues except mature erythrocytes, and they are in highest concentration in the liver and kidneys. They are formed by growth and division of preexisting peroxisomes and are randomly destroyed by autophagy. Their half-life is 1.5 to 2 days. Peroxisomal proteins are encoded by nuclear genes, synthesized in cytosol, and imported posttranslationally into the peroxisome. The import of proteins into the peroxisome is mediated by receptors and requires adenosine triphosphate hydrolysis. Peroxisomes contain enzymes that use oxygen to oxidize a variety of substrates, thereby forming peroxide. The peroxide is decomposed within the organelle by the enzyme catalase to water and oxygen. This process protects the cell against peroxide damage through compartmentalization of peroxide metabolism within the organelle. Peroxisomes can also function to dispose of excess reducing equivalents and may contribute to thermogenesis, producing heat from cellular respiration (Gould et al, 2007). More than 50 enzymes have been found within peroxisomes (Gould et al, 2001). The proteins have multiple functions, both synthetic and degradative (Wanders et al, 2001). The primary synthetic functions are plasmalogen synthesis and bile acid formation. Plasmalogens constitute 5% to 20% of phospholipids in cell membranes and 80% to 90% of phospholipids in myelin. They are involved in platelet activation and may also protect cells against oxidative stress. Degradative functions include (1) β-oxidation of VLCFA (≥C23), fatty acids (down to C8 to C6), long-chain dicarboxylic acids, prostaglandins, and polyunsaturated fatty acids; (2) oxidation of bile acid intermediates, pipecolic acid and glutaric acid (intermediates in lysine catabolism), and phytanic acid; (3) deamination of d- and l-amino acids, (4) metabolism of glycolate to glyoxylate; (5) polyamine degradation (spermine and spermidine); and (6) ethanol clearance. At least 16 conditions caused by single peroxisomal enzyme deficiencies have been confirmed (Wanders et al, 2007). Peroxisomal disorders constitute a clinically and biochemically heterogeneous group of inherited diseases that result from the absence or dysfunction of one or more peroxisomal enzymes. Conditions in which multiple peroxisomal enzymes are affected can result from a disturbance of biogenesis or the organelle. Pathophysiology apparently involves either deficiency of necessary products of peroxisomal metabolism or excess of unmetabolized substrates. Disorders with similar biochemical defects may have markedly different clinical features, and disorders with similar clinical features may be associated with different biochemical findings. General features of peroxisomal disorders, each of which can manifest or be evident in the newborn period, are as follows: ll Dysmorphic craniofacial features ll Neurologic dysfunction, primarily consisting of severe hypotonia, possibly associated with hypertonia of the extremities and seizures ll Hepatodigestive dysfunction, including hepatomegaly, cholestasis, and prolonged hyperbilirubinemia CHAPTER 23 Lysosomal Storage, Peroxisomal, and Glycosylation Disorders and Smith-Lemli-Opitz Syndrome in the Neonate Rhizomelic shortening of the limbs, stippled calcifications of epiphyses, renal cysts, and abnormalities in neuronal migration may also be seen. Peroxisomal biogenesis disorders are composed of at least 12 complementation groups (Matsumoto et al, 2001). All involve defects in proteins targeted to the organelle. Genes and proteins required for peroxisomal biosynthesis are referred to as peroxins and are encoded by PEX genes. The PEX genes responsible for disease in most human patients are known, and more than 50% of patients with peroxisomal biogenesis disorders have mutations in PEX1 (Moser, 2000; Steinberg et al, 2006). This section We discusses the peroxisomal disorders that can manifest in the newborn period. Zellweger syndrome is the prototype of neonatal peroxisomal disease. It is a disorder of peroxisome biogenesis caused by failure to import newly synthesized peroxisomal proteins into the peroxisome. The proteins remain in the cytosol, where they are rapidly degraded. In this condition, peroxisomes are absent from liver hepatocytes or exist as “ghosts.” Neonatal adrenoleukodystrophy and infantile Refsum disease are also disorders of peroxisome biogenesis in which, as in Zellweger syndrome, disruption of function of more than one peroxisomal enzyme is demonstrable. A few residual peroxisomes, however, may be seen in the liver. These disorders represent a continuum of clinical severity. Rhizomelic chondrodysplasia punctata is caused by a defect in a subset of peroxisomal enzymes. In this disorder, liver peroxisomes are demonstrable and normal in number, but their distribution and structure are abnormal. 253 There is circumstantial evidence that in utero elevations of VLCFAs may be key to congenital CNS abnormalities. Powers and Moser (1998) proposed that VLCFAs and phytanic acid that accumulate in these peroxisomal disorders are incorporated into myelin and cell membranes, and that alteration of normal constituents of the membrane adversely affects membrane function. Specifically, these investigators suggest that abnormal constituents accelerate cell death and impede neuronal migration, accounting for the conspicuous CNS abnormalities in disorders of peroxisomal biogenesis. To date, four disorders of peroxisomal fatty acid β-oxidation have been defined: acetyl-CoA oxidase deficiency, d-bifunctional protein deficiency, peroxisomal thiolase deficiency, and 2-methylacyl-CoA racemase deficiency (Wanders et al, 2007). The clinical presentation of the first three disorders resembles that of biogenesis disorders. Individuals with the fourth disorder have a late-onset neuropathy. Clinical Presentations Table 23-3 summarizes the clinical features of disorders of peroxisome biogenesis that can present in the neonate. Zellweger Syndrome Zellweger syndrome is most often evident at birth, with affected babies having dysmorphic facial features including large fontanels, high forehead, flat occiput, epicanthus, hypertelorism, upward-slanting palpebral fissures, TABLE 23-3 Disorders of Peroxisomal Biogenesis in the Newborn Period Feature Zellweger Syndrome Neonatal ­Adrenoleukodystrophy Infantile Refsum ­Disease Rhizomelic Chondrodysplasia Punctata Onset Birth Birth to 3 mo Birth to 6 mo Birth Facies High forehead, large ­fontanels, upward-­ slanting palpebral fissures, hypoplastic supraorbital ridges, epicanthic folds, micrognathia, abnormal ears Milder features of ­Zellweger syndrome Epicanthic folds, midface ­hypoplasia, low-set ears Depressed nasal bridge, hypertelorism, ­microcephaly Neurologic ­findings Weakness, hypotonia, seizures, psychomotor retardation, sensorineural hearing loss Hypotonia, seizures, slow psychomotor ­development and ­neurodegeneration Mild hypotonia, normal early development followed by ­degeneration, ataxia, ­sensorineural hearing loss Severe psychomotor ­retardation Ophthalmologic findings Cataracts, glaucoma, corneal clouding, retinitis pigmentosa, optic nerve dysplasia, Brushfield spots Retinopathy Retinitis pigmentosa Cataracts Other findings Hepatomegaly, multicystic kidneys, congenital heart disease, growth failure, chondrodysplasia punctata Impaired adrenal function Hepatomegaly, anosmia, ­diarrhea Severe shortening of proximal limbs, joint contractures, ichthyosis Diagnosis ↑ plasma VLCFA, phytanic acid, pipecolic acid, and bile acid intermediates, ↓ plasmalogens Same as for Zellweger syndrome Same as for Zellweger ­syndrome ↑ phytanic and pipecolic acids, ↓ plasmalogens, normal VLCFA and bile acid intermediates VLCFA, Very-long-chain fatty acid; ↑, elevated; ↓, reduced. 254 PART VI Metabolic and Endocrine Disorders of the Newborn hypoplastic supraorbital ridges, abnormal ears, severe weakness and hypotonia, hepatomegaly, multicystic kidneys, and congenital heart disease. Seizures, feeding difficulties, and postnatal growth failure soon manifest. Ophthalmologic examination may detect cataracts, corneal clouding, glaucoma, optic atrophy, retinitis pigmentosa, and Brushfield spots. Somatic sensory evoked responses and electroretinograms are abnormal. Hearing assessment often shows an abnormal brainstem auditory evoked response consistent with sensorineural hearing loss. Skeletal radiographs demonstrate epiphyseal stippling, and cranial imaging shows leukodystrophy and neuronal migration abnormalities. Hepatic cirrhosis and severe psychomotor retardation occur later. Laboratory analysis may demonstrate abnormal liver function values, hyperbilirubinemia, or hypoprothrombinemia. Death usually occurs within the first year of life, the average life span being 12.5 weeks. Neonatal Adrenoleukodystrophy Clinically, neonatal adrenoleukodystrophy is similar to, but less severe than, Zellweger syndrome. Differences include less dysmorphology, absence of chondrodysplasia punctata and renal cysts, and fewer neuronal and gray matter changes. Patients with neonatal adrenoleukodystrophy may have striking white matter disease, however, and often show degenerative changes in adrenal glands. They also have slow psychomotor development followed by neurodegeneration that usually begins before the end of the first year of life. Disease progression is slower than that observed in Zellweger syndrome, and longer survival is usual, to an average of approximately 15 months of age. Infantile Refsum Disease Patients with infantile Refsum disease also have relatively mild dysmorphic features, such as epicanthic folds, midface hypoplasia with low-set ears, and mild hypotonia. Early neurodevelopment is normal, possibly up to 6 months of age, but then slow deterioration begins. Later, sensorineural hearing loss (100%), anosmia, retinitis pigmentosa, hepatomegaly with impaired function, and severe mental retardation are evident. Patients learn to walk, although their gait may be ataxic and broad based. Diarrhea and failure to thrive may also be seen. Chondrodysplasia punctata and renal cysts are absent. Neuronal migration defects are minor, and adrenal hypoplasia occurs. The life span of patients with infantile Refsum disease ranges from 3 to 11 years. Rhizomelic Chondrodysplasia Punctata Patients with rhizomelic chondrodysplasia punctata at birth have facial dysmorphia, microcephaly, cataracts, rhizomelic shortening of extremities with prominent stippling, and coronal clefting of vertebral bodies. The chondrodysplasia punctata is more widespread than in Zellweger syndrome and may involve extraskeletal tissues. Infants with this disorder have severe psychomotor retardation from birth onward and severe failure to thrive. In addition, patients may have joint contractures, and 25% experience ichthyosis. Neuronal migration is normal. Life span is usually less than 1 year. DISORDERS OF PEROXISOMAL β-OXIDATION d-Bifunctional protein deficiency is more rare than peroxisomal biogenesis disorders and results in a phenotype similar to Zellweger syndrome. In general, children have severe CNS involvement consisting of profound hypotonia, uncontrolled seizures, and failure to acquire any significant developmental milestones. Children are usually born full term without evidence of intrauterine growth restriction. Dysmorphic features, similar to those seen in Zellweger syndrome, are notable in most children. In most cases, neuronal migration is disturbed with areas of polymicrogyria and heterotopic neurons in the cerebrum and cerebellum. Death generally occurs before 1 year of age, but survival to at least 3 years of age is possible. Acetyl-CoA oxidase deficiency is less common. Patients exhibit global hypotonia, deafness, and delayed milestones with or without facial dysmorphic features. Patients may demonstrate early developmental gains, but then show regression of skills. Retinopathy with extinguished electroretinograms, failure to thrive, hepatomegaly, areflexia, and seizures have also been reported. One patient with peroxisomal thiolase deficiency has been described (Goldfischer et al, 1986). The child had marked facial dysmorphia, muscle weakness, and hypotonia. She demonstrated no psychomotor development during her 11 months of life. Autopsy showed renal cysts, atrophic adrenal glands, minimal liver fibrosis, hypomyelination in cerebral white matter, foci of neuronal heterotopia, and a sudanophilic leukodystrophy (Goldfischer et al, 1986). X-linked adrenoleukodystrophy is the most common peroxisomal disorder, because of altered function of a membrane transport protein ABCD1 that affects metabolism of VLCFAs, and does not usually present in the neonatal period. However, three patients with contiguous deletions involving the ABCD1 gene have exhibited a phenotype similar to peroxisomal biogenesis disorders (Corzo et al, 2002). DIAGNOSIS, MANAGEMENT, AND PROGNOSIS OF PEROXISOMAL DISORDERS The key to diagnosing peroxisomal disease is a high index of suspicion. Peroxisomal disorders should be considered in newborns with dysmorphic facial features, skeletal abnormalities, shortened proximal limbs, neurologic abnormalities (including hypotonia or hypertonia), ocular abnormalities, and hepatic abnormalities. Babies with abnormal visual, hearing, or somatosensory evoked potentials should also be considered for these diagnoses. Peroxisomal disorders are not associated with acute metabolic derangements or abnormal routine laboratory tests. Measurements of VLCFAs, phytanic acid, pipecolic acid, bile acid intermediates, and plasmalogens are required for diagnosis. Zellweger syndrome is associated with elevations of VLCFAs, phytanic acid, pipecolic acid, and bile acid intermediates, and a decrease in plasmalogen synthesis. Neonatal adrenoleukodystrophy and infantile Refsum disease have similar biochemical findings; however, the defect in plasmalogen synthesis and the degree of VLCFA accumulation are less severe. Laboratory findings in rhizomelic CHAPTER 23 Lysosomal Storage, Peroxisomal, and Glycosylation Disorders and Smith-Lemli-Opitz Syndrome in the Neonate chondrodysplasia punctata are elevations of phytanic and pipecolic acids, a decrease in plasmalogen, and normal levels of VLCFAs and bile acid intermediates. Therefore screening that uses only levels of VLCFAs fails to detect rhizomelic chondrodysplasia punctata. d-Bifunctional protein deficiency is associated with deficient oxidation of C23:0 and pristanic acid, leading to elevations of pristanic acid and, to a lesser extent, phytanic acid. This deficiency results in an elevated pristanic acid-to-phytanic acid ratio, which is generally not elevated in peroxisomal biogenesis disorders. Abnormal VLCFA and elevation of varanic acid (an intermediate metabolite in β-oxidation) are also seen. Accumulation of bile acid intermediates is a variable finding. Abnormalities in phytanic acid and plasmalogens are age dependent. The elevation of phytanic acid might not be demonstrable in young infants, and reduction in red blood cell plasmalogen levels may not be evident in children older than 20 weeks (Gould et al, 2007). A liver biopsy may be a useful adjunct diagnostic tool to assess for the presence or absence and structure of peroxisomes. Definitive diagnoses for all types of peroxisomal disease require cultured skin fibroblasts for measurement of VLCFA levels and their β-oxidation and, as needed, assay of the peroxisomal steps of plasmalogen synthesis, phytanic acid oxidation, subcellular localization of catalase, enzyme assays, and immunocytochemistry studies. Prenatal diagnosis with a variety of methods is available (Steinberg et al, 2005). DNA diagnosis is possible for most patients, but is “challenging for the Zellweger syndromes spectrum since 12 PEX genes are known to be associated with this spectrum of peroxisomal biogenesis disorders” (Steinberg et al, 2006). DNA study for deletions also has a role in diagnostic evaluations in some cases before counseling for recurrence risk, as demonstrated by the neonatal presentation of cases with deletion of the ABCD1 gene on the X chromosome (Corzo et al, 2002). The complexities of prenatal diagnosis emphasize the need for studies of tissue on the proband to determine the potential for and appropriate strategies to offer in prenatal diagnosis for any family. One of the more interesting recent developments in peroxisomal disease is the preliminary work on combined liquid chromatography-tandem mass spectroscopy application for blood spot–based newborn screening. It is targeted at conditions with abnormal VLCFAs, especially X-linked adrenoleukodystrophy, that typically manifest in childhood and for which presymptomatic therapy may alter the course of the usual progressive disease. The method is promising, but still under investigation ­(Hubbard et al, 2006). Treatment for all peroxisomal disorders in the newborn period remains supportive. These disorders are chronic, progressive diseases with no currently available curative therapy. Setchell et al (1992) described the effects of administration of primary bile acids on liver function in a 6-month-old infant with Zellweger syndrome. The effects included normalization of serum bilirubin and liver enzyme levels and a decrease in hepatic inflammation, bile duct proliferation, and canalicular plugs. The patient also showed an improvement in growth and neurologic function (Setchell et al, 1992). Martinez (1992) reported the use of docosahexaenoic acid ethyl ester in two patients with neonatal 255 adrenoleukodystrophy. Both patients had an increase in erythrocyte omega fatty acid levels and plasmalogens accompanied by significant improvement in clinical parameters, including alertness, motor performance, vocabulary, and visual evoked responses. Martinez et al (2000) also reported on the effects of docosahexaenoic acid supplementation in 13 patients with generalized peroxisomal disorders. The effects were normalization of blood docosahexaenoic acid levels, increased plasmalogen concentrations, decreased plasma VLCFAs, and improvement to near normal of liver enzymes. Although patients with severe neonatal Zellweger syndrome presentation did not benefit, patients with milder peroxisomal biogenesis disorders—in which clinical course is more variable—­ experienced improvement in vision, liver function, muscle tone, and social contact. Three patients showed normalization of brain myelin, and myelination improved in three others (Martinez et al, 2000). In addition, a “triple” dietary approach consisting of oral administration of ether lipids, decreased phytanic acid intake, and the oral administration of glyceryl trioleate and glyceryl trierucate (Lorenzo’s oil) leads to biochemical improvement and may have value in patients with mild forms of peroxisomal biogenesis defects (Gould et al, 2007). Although treatment protocols are available for infants affected with disorders of peroxisomal biogenesis, improvement in long-term outcome remains limited in all but the mildly affected patients. However, treatment involving diet and bone marrow transplant, when appropriate, has been demonstrated to significantly improve the outcome of some patients affected with adrenoleukodystrophy (Moser et al, 2001). SMITH-LEMLI-OPITZ SYNDROME ETIOLOGY Smith-Lemli-Opitz syndrome is a well-recognized autosomal recessive malformation syndrome, with an estimated incidence ranging from 1 in 10,000 to 70,000 in various populations (Porter, 2008; Yu and Patel, 2005). Because of the identification of an underlying biochemical defect, SLO syndrome has been reclassified as an inborn error of metabolism. In 1993, it was discovered that SLO syndrome is caused by a defect in cholesterol biosynthesis that results in low levels of cholesterol and elevated levels of 7-­dehydrocholesterol (7DHC) and its isomer, 8-dehydrocholesterol (8DHC). Patients have markedly reduced activity of the enzyme 7DHC reductase, the enzyme responsible for conversion of 7DHC to cholesterol (Porter, 2008; Salen et al, 1995; Waterham and Clayton, 2006), which is located on chromosome 11 (Kelley and Hennekam, 2000; Waterham and Clayton, 2006). Cholesterol is a major lipid component of cellular membranes such as myelin, and it is an important structural component of lipid rafts, which play a major role in signal transduction (Porter, 2008; Yu and Patel, 2005). In addition, bile acids, steroid hormones, neuroactive steroids, and oxysterols are all synthesized from cholesterol (Merkens et al, 2009; Porter, 2008). The possible role of these various pathways in the etiology of SLO syndrome is still being defined (Merkens et al, 2009; Porter, 2008). 256 PART VI Metabolic and Endocrine Disorders of the Newborn The cause of the clinical phenotype of SLO syndrome may be related to deficient cholesterol, deficient total sterols, the toxic effects of either 7DHC or compounds derived from 7DHC, or a combination of these factors (Porter, 2008; Yu and Patel, 2005). A single underlying pathologic mechanism is unlikely (Porter, 2008). The pivotal connection between cholesterol and SLO syndrome involves development and differentiation of the vertebrate body plan, although additional metabolic effects have not been ruled out. Among those proteins exerting decisive influence on patterning during embryogenesis are the Hedgehog proteins. One variant, Sonic hedgehog (Shh), becomes covalently linked to cholesterol at the protein’s amino-terminal signaling domain. This linkage is needed to restrict the locus of action of Shh to the region of plasma membrane. Although it was originally suggested that failure to modify Shh could account for multiplicity of structural abnormalities in patients with SLO syndrome (Farese and Herz, 1998), the process is thought to be more complex and to involve other signaling proteins such as Patched (PTCH) and Smoothened (SMO) (Farese and Herz, 1998; Kelley and Hennekam, 2000; Porter, 2008). CLINICAL FEATURES Recognition of the biochemical defect in SLO syndrome provided the diagnostic test required to recognize the most mild and most severe cases, substantially expanding the clinical spectrum of the condition. Classic SLO syndrome is often evident at birth; affected patients have microcephaly and facial dysmorphism, including bitemporal narrowing, ptosis, epicanthic folds, anteverted nares, broad nasal tip, prominent lateral palatine ridges, micrognathia, and low-set ears. Other features are 2- to 3-syndactyly of toes (found in 95% of patients), small proximally placed thumbs, and occasionally postaxial polydactyly and cataracts. Males usually have hypospadias, cryptorchidism, and a hypoplastic scrotum, but may have ambiguous or female genitalia. Pyloric stenosis, cleft palate, pancreatic anomalies, Hirschsprung disease, and lung segmentation defects have also been reported. Hypotonia progressing to hypertonia and moderate to severe mental deficiencies are also present. Feeding difficulties and vomiting are common problems in infancy. Irritable behavior and shrill screaming may also pose problems during infancy. Older children frequently have hyperactivity, self-injurious behavior, sleep difficulties, and autistic characteristics. Cranial imaging studies show (and autopsy confirms) defects in brain morphogenesis, including hypoplasia of frontal lobes, cerebellum, and brainstem, dilated ventricles, irregular gyral patterns, and irregular neuronal organization. Historically, approximately 20% of patients die within the first year of life, although others may survive for more than 30 years. Life expectancy appears to correlate inversely with the number and severity of organ defects and with the kinds and numbers of limb, facial, and genital abnormalities (Kelley and Hennekam, 2000; Tint et al, 1995). Developmental outcomes are also highly variable, ranging from severe mental retardation to normal intelligence. Development of treatment protocols for SLO syndrome may contribute to improvements in prognosis, but improved recognition of more mildly affected patients may explain the increasing reports of SLO syndrome in patients with mild mental retardation or normal intelligence (Jezela-Stanek et al, 2008). Testing for SLO syndrome has been suggested for all patients with idiopathic intellectual impairment, behavioral anomalies, or both, when associated with nonfamilial two- and three-toe syndactyly and failure to thrive (Jezela-Stanek et al, 2008). DIAGNOSIS The diagnosis of SLO syndrome is based on findings of elevated levels of 7DHC and 8DHC. Plasma cholesterol levels are usually but not always low; as many as 10% of patients at all ages have normal cholesterol levels (Kelley and Hennekam, 2000). In addition, the standard method for analysis of cholesterol in most hospital laboratories identifies 7DHC and 8DHC as cholesterol. Therefore most laboratories report normal cholesterol levels in patients who have low cholesterol, but elevations of 7DHC and 8DHC sufficient to bring the total level into the “normal” range (Kelley and Hennekam, 2000; ­Porter, 2008). The difference between mild and more severe disease appears to be one of degree; the enzyme defect is more severe and the block is more complete in patients with severe disease (Kelley and Hennekam, 2000; Tint et al, 1995). Clinical severity in SLO syndrome correlates best with either reduction in absolute cholesterol levels or the sum of 7DHC plus 8DHC expressed as a fraction of total sterol (Waterham and Clayton, 2006). Recently, maternal apolipoprotein E (ApoE) genotype was implicated in phenotype heterogeneity (Witsch-Baumgartner et al, 2004). Maternal ApoE2 genotypes were associated with a severe SLO syndrome phenotype, whereas ApoE genotypes without the E2 allele were associated with a milder phenotype (Witsch-Baumgartner et al, 2004; Yu and Patel, 2005). Confirmation of diagnosis via molecular analysis is available. Genotype-phenotype correlation, however, is relatively poor (Jezela-Stanek et al, 2008; Porter, 2008). Prenatal diagnosis is possible. A mother carrying an affected fetus may have an abnormally low unconjugated estriol value. Ultrasonography detects many but not all affected fetuses, and biochemical analysis of amniotic fluid and of chorionic villus samples is accurate and unambiguous in most cases (Kelley and Hennekam, 2000). A direct correlation between the level of 7DHC in amniotic fluid and clinical severity has been demonstrated; however, a similar correlation does not exist for the level of cholesterol (Yu and Patel, 2005). When molecular mutations are known, analysis may be useful. TREATMENT There are two goals of therapy for SLO syndrome: to increase the level of cholesterol in plasma and other body fluids and to lower the level of 7DHC. Treatment consists of providing exogenous cholesterol, in the form of either dietary cholesterol or cholesterol suspension, to replenish body stores of cholesterol and downregulate the patient’s endogenous cholesterol synthesis, thus decreasing the amount of 7DHC produced. A goal of cholesterol supplementation of 20 to 60 mg/kg per day was CHAPTER 23 Lysosomal Storage, Peroxisomal, and Glycosylation Disorders and Smith-Lemli-Opitz Syndrome in the Neonate initially advocated, but doses higher than 300 mg/kg per day have been used without adverse outcome (Irons et al, 1995; ­Kelley and Hennekam, 2000). In infants with SLO syndrome, the use of breast milk should be encouraged because it supplies approximately 133 mg/L of cholesterol (Irons et al, 1995). Providing bile acids to facilitate adequate absorption of dietary cholesterol is controversial, especially because fat malabsorption is unusual and certain bile acids can decrease tissue uptake of cholesterol (Kelley and ­Hennekam, 2000). Dietary cholesterol supplementation appears to restore both adrenal and bile salt deficiencies (Yu and Patel, 2005). Steroid replacement therapy may still be needed during times of stress or illness. Use of 3-hydroxy-3-methyl-CoA (HMG-CoA) lyase inhibitors such as lovastatin has also been suggested as a mechanism to decrease levels of 7DHC. However, because the results of animal studies suggest that downregulation of cholesterol synthesis might not decrease 7DHC, and because of concern about a decrease in the synthesis of other essential isoprenoid compounds, HMG-CoA reductase inhibitors are not routinely used in therapy (Kelley and Hennekam, 2000). Furthermore, this treatment could be theoretically detrimental in patients with SLO syndrome with little or no enzyme activity (Yu and Patel, 2005). Statins may also impair dietary cholesterol absorption (Merkens et al, 2009). In addition, a retrospective study of simvastatin use in patients with SLO syndrome failed to demonstrate a positive effect on anthropometric measures or behavior (Haas et al, 2007). Although many questions remain about optimal therapy and outcomes, therapeutic interventions appear to increase plasma cholesterol levels, decrease 7DHC levels, and improve irritability, behavior, and growth (Elias et al, 1997; Irons et al, 1995; Kelley and Hennekam, 2000; Waterham and Clayton, 2006). Parents reported children to be more alert, active, and happier during therapy. Therapy was well tolerated. Unfortunately, a study of 14 patients with SLO syndrome indicated that cholesterol supplementation had hardly any effect on developmental progress (Sikora et al, 2004). Treatment probably does not significantly change sterol levels in brain which are dependent on de novo cholesterol synthesis because of limited ability of cholesterol to cross the blood-brain barrier (­Porter, 2008; Waterham and Clayton, 2006). Direct delivery of cholesterol to the CNS by low-pressure catheter infusions has been proposed, but not tested (Yu and Patel, 2005). Gene therapy, the use of neuroactive steroids, and inhibition of glycosphingolipids are also being investigated as possible therapeutic options in SLO syndrome (Merkens et al, 2009). 257 For maximal benefit, it has been suggested that treatment should begin prenatally, because SLO syndrome has many features that are consistent with in utero involvement of the disease process. Antenatal supplementation by fetal intravenous and intraperitoneal transfusions of fresh frozen plasma were shown to increase fetal cholesterol in one patient (Irons et al, 1999). Treatment should otherwise begin as soon as possible after birth or as soon as the diagnosis is confirmed. Patients with severe SLO syndrome may need gavage or gastrostomy feeding for management of reflux and gastrointestinal dysmotility, and many have protein allergies and require elemental formulas. Growth is often a problem, but the temptation to overfeed must be avoided because overfeeding would contribute to feeding problems and could not rescue intrauterine growth restriction in severe SLO syndrome ­(Kelley and Hennekam, 2000). SUGGESTED READINGS Barth PG: Sphingolipids. In Fernandes J, Saudubray J-M, van den Berghe G, editors: Inborn Metabolic Diseases: Diagnosis and Treatment, ed 2, Berlin, 1995, Springer-Verlag, pp 375-382. Beutler E, Grabowski GA: Gaucher Disease. In Scriver CR, Beaudet AL, Sly WS, Valle D, et al, editors: The Metabolic and Molecular Bases of Inherited Disease, ed 8, New York, 2001, McGraw-Hill, pp 3635-3668. Fletcher JM: Screening for lysosomal storage disorders: a clinical perspective, J Inherit Metab Dis 29:405-408, 2006. Gould S, Raymond G, Valle D: The peroxisome biogenesis disorders. In Valle D, Beaudet AL, Vogelstein B, et al: The Metabolic and Molecular Bases of Inherited Disease, ed 8, New York, 2001, McGraw-Hill, pp 3181-3217. Grünewald S: Congenital disorders of glycosylation: rapidly enlarging group of (neuro)metabolic disorders, Early Hum Devel 83:825-830, 2007. Jaeken J: Congenital disorders of glycosylation. In Fenandes J, Saudubray JM, van den Berghe G, et al: Inborn Metabolic Diseases: Diagnosis and Treatment, ed 4, Heidelberg, 2006, Springer Medizin Verlag, pp 523-530. Jaeken J, Matthijs G, Carchon H, et al: Defects of N-glycan synthesis. In Scriver CR, Beaudet AL, Sly WS, et al: The Metabolic and Molecular Bases of Inherited Disease, ed 8, New York, 2001, McGraw-Hill, pp 1601-1622. Meikle PJ, Grasby DJ, Dean CJ, et al: Newborn screening for lysosomal storage disorders, Mol Genet Metab 88:307-314, 2006. Moser HW, Smith KD, Watkins PA, et al: X-linked adrenoleukodystrophy. In Valle D, Beaudet AL, Vogelstein B, et al: The Metabolic and Molecular Bases of Inherited Disease, ed 8, New York, 2001, McGraw Hill, pp 3257-3301. Porter FD: Smith-Lemli-Opitz syndrome: pathogenesis, diagnosis, and management, Eur J Hum Genet 16:535-541, 2008. Staretz-Chacham O, Lang TC, LaMarca ME, et al: Lysosomal storage disorders in the newborn, Pediatrics 123:1191-1207, 2009. Stone DL, Sidransky E: Hydrops fetalis: lysosomal storage disorders in extremis, Adv Pedatr 46:409-440, 1999. Wanders R, Barth P, Heymans H: Single peroxisomal enzyme deficiencies. In Valle D, Beaudet AL, Vogelstein B, et al: The Metabolic and Molecular Bases of Inherited Disease, ed 8, New York, 2001, McGraw Hill, pp 3219-3256. Waterham HR, Clayton PT: Disorders of cholesterol synthesis. In Fenandes J, Saudubray JM, van den Berghe G, Walters JH, editors: Inborn Metabolic Diseases: Diagnosis and Treatment, ed 4, Heidelberg, 2006, Springer Medizin Verlag, pp 414-415. Complete references used in this text can be found online at www.expertconsult.com C H A P T E R 24 Skeletal Dysplasias and Connective Tissue Disorders David L. Rimoin and George E. Tiller The skeletal dysplasias, or osteochondrodysplasias, are disorders of the development and growth of cartilage and bone. The connective tissue disorders involve abnormalities of the cells’ supporting and connecting structures in the matrix. In one series of 126,316 deliveries monitored over 15 years, the incidence of skeletal dysplasias was 2.14 in 10,000 (Rasmussen et al, 1996). With the growing use and accuracy of ultrasonography for prenatal care, a greater number of osteochondrodysplasias and connective tissue disorders are diagnosed prenatally. The skeletal dysplasias have been classified into 37 groups on the basis of radiologic criteria (Superti-Furga and Unger, 2007). Other classifications vary according to molecular, clinical, pathologic, and radiolographic criteria and may be confusing. For example, osteogenesis imperfecta (OI) can be classified as either a skeletal dysplasia or a connective tissue disorder. This chapter focuses on several of the more common skeletal dysplasias (see Table 24-1 for an expanded list) and connective tissue disorders that manifest prenatally or perinatally, but the discussion is not exhaustive. The osteochondrodysplasias have been reviewed extensively elsewhere (Cohen, 2006; SupertiFurga et al, 2001; Unger et al, 2007). There are a large number of different connective tissue molecules, including collagens (over two dozen types), elastin, fibrillin (two types), and microfibril-associated glycoproteins. These molecules are components of tissues such as bone, cartilage, skin, vascular media, tendon, ligaments, and basement membrane in many organs. The heritable disorders of connective tissue are varied, may be very dissimilar clinically, and may manifest in utero or at any age postnatally. Those that may manifest at birth include the infantile (neonatal) form of Marfan syndrome, congenital contractural arachnodactyly (Beals syndrome), cutis laxa, Ehlers-Danlos syndrome, and Menkes disease. CLINICAL SPECTRA OF DISORDERS WITH COMMON MOLECULAR BASES The number of clinically distinguishable skeletal dysplasias and connective tissue disorders is extensive. With advances in molecular knowledge, several different dysplasias have been recognized to have mutations in the same genes. In some of these disorders, clinical similarities noted previously suggested a common etiology. One such clinical spectrum includes achondroplasia, hypochondroplasia, severe achondroplasia with developmental delay and acanthosis nigricans (SADDAN), and thanatophoric dysplasia, all of which are caused by mutations in the fibroblast growth factor receptor 3 (FGFR3) gene (Bellus et al, 1995; Shiang et al, 1994; Vajo et al, 2000; Wilcox et al, 1998). Another spectrum of disorders includes Stickler syndrome, Kniest dysplasia, spondyloepimetaphyseal 258 dysplasia, spondyloepiphyseal dysplasias, hypochondrogenesis, achondrogenesis type II, and recessive multiple epiphyseal dysplasia, all of which are caused by mutations in the gene for collagen type II, COL2A1 (Spranger et al, 1994; Winterpacht et al, 1993). With other disorders, the common etiology is not as obvious clinically: diastrophic dysplasia, atelosteogenesis type II, and achondrogenesis type 1B are all caused by mutations in the diastrophic dysplasia sulfate transporter (DTDST) gene (Bonafe and Superti-Furga, 2007; Hastbacka et al, 1994, 1996; SupertiFurga et al, 1996). The obverse is also evident, wherein a specific clinical entity (e.g., multiple epiphyseal dysplasias) may be caused by a mutation in one of several genes—a concept known as genetic heterogeneity. APPROACH TO DIAGNOSIS An early and precise diagnosis is important for prognosis, optimal immediate- and long-term management, accurate genetic counseling about recurrence risk, and identification of other possibly affected family members or disease carriers. An example is the group of disorders with punctate calcifications (“stippling”) in epiphyses, called chondrodysplasia punctata. There are more than three types, each of which has a different cause and mode of inheritance: autosomal recessive, X-linked recessive, and X-linked dominant (see Table 24-1). As in any uncommon genetic condition, multiple factors may be required to arrive at the correct diagnosis: a complete physical examination, three-generation family history, radiologic studies, and biochemical or molecular tests. Most skeletal dysplasias cause short stature, which can be proportionate or disproportionate. The disproportion may be evident as a short-limbed or short-trunk form of dwarfism. If the limbs are affected, there may be segmental shortening of the upper arms and thighs (rhizomelia), forearms and legs (mesomelia), or hands and feet (acromelia). Most skeletal dysplasias that manifest at birth involve short limbs. Accurate measurements of length (on a firm surface), arm span, and head and chest circumferences must be plotted on standard growth curves, with calculation of upper and lower body segment ratios to objectively assess disproportion. Other skeletal characteristics can give important clues for specific disorders: ll Children with achondroplasia and thanatophoric dysplasia have large heads (macrocephaly). Cloverleaf skull deformity is present in some forms of thanatophoric dysplasia. ll A relatively long chest is seen in asphyxiating thoracic dystrophy. ll In achondroplasia the hand is short and the fingers form a trident configuration. In diastrophic dysplasia, there are distinctive “hitchhiker” thumbs. 259 CHAPTER 24 Skeletal Dysplasias and Connective Tissue Disorders TABLE 24-1 Skeletal Dysplasias Manifesting Prenatally or Perinatally Dysplasia Skeletal ­Features Nonskeletal Features Radiographic Features Inheritance; Gene Comments Lethal Achondrogenesis type IB Soft cranium; round face; short, round chest; very short limbs Polyhydramnios Poorly ossified calvarium; ribs short with fractures (beading); nonossified vertebrae; small pelvis; short broad femurs with metaphyseal spikes, short broad tibiae, and fibulae AR; DTDST (diastrophic dysplasia sulfate transporter) Same gene as diastrophic dysplasia and atelosteogenesis II Achondrogenesis type II, hypochondrogenesis Large head, flat face with cleft palate; short trunk; very short limbs (micromelia) Fetal hydrops; distended abdomen; Lack of vertebral mineralization; short limbs (all segments); enlarged cranium with normal ossification AD; COL2A1 (type II collagen) Same gene as for spondyloepiphyseal dysplasia congenita, spondyloepimetaphyseal dysplasia, Stickler syndrome, Kniest syndrome Asphyxiating thoracic dystrophy Normal face; narrow, long chest; variable limb shortening Lethal pulmonary insufficiency Normal calvarium and vertebrae; very short ribs with anterior cupping; short limbs with wide proximal femoral metaphyses; premature ossification of proximal femoral epiphysis AR Survivors may have renal disease; possibly a variant of short-rib polydactyly III Type I Flat face with cleft palate, micrognathia; very narrow chest; very short limbs (rhizomelic) with equinovalgus deformities; joint dislocations Prematurity; stillbirth Flat vertebrae with coronal and sagittal clefts, scoliosis, short ribs (11 pairs), small pelvis with enlarged sacrosciatic notch, short limbs, “drumstick” humeri and femurs, absent fibulas, short metacarpals triangular first metacarpals, dislocated knees AD; filamin B Filamin B mutations also seen in boomerang dysplasia, Larsen syndrome, and spondylocarpotarsal syndrome Type II Cleft palate, narrow chest, short limbs with dislocations, equinovarus deformities, gap between first and second digits Laryngeal stenosis; patent foramen ovale Occasional coronal and sagittal vertebral clefts; short ribs; normal sacrosciatic notch; short “dumbbell” humeri and femurs, small fibulas; large second and third metacarpals; small round midphalanges AR; DTDST (diastrophic dysplasia sulfate transporter) Same gene as for ­diastrophic dysplasia, achondrogenesis type IB, and some forms of multiple epiphyseal dysplasia (MED) Campomelic ­dysplasia Large cranium; small face with flat nose bridge, small chin (cleft soft palate); small, narrow chest; bowed thighs and legs, with dimple on leg Polyhydramnios, congenital cardiac abnormalities, female external genitalia in XY males Large dolichocephalic calvarium with shallow orbits; short and wavy ribs, often 11 pairs; hypoplastic scapula; small, flat vertebrae; tall, narrow pelvis; relatively long, thin limbs with bent femurs and short tibiae AD (most are new mutations); SOX9 Chondrodysplasia punctata, rhizomelic type 1 (RCDP1) Face flat; very flat nasal bridge and tip; proximal shortening of limbs Cataracts; joint contractures; ichthyosiform erythroderma Wide coronal vertebral clefts; short humeri and femurs; stippled epiphyses of long bones, pelvis and periarticular areas; trapezoid ilia AR; PEX7 (peroxisome biogenesis factor 7) Atelosteogenesis Short-rib ­polydactyly Survivors may live a few years, with severe growth and mental retardation; biochemical abnormalities are decreased RBC plasmalogens and increased phytanic acid Heterogeneous ­ciliopathies Continued 260 PART VI Metabolic and Endocrine Disorders of the Newborn TABLE 24-1 Skeletal Dysplasias Manifesting Prenatally or Perinatally—cont’d Nonskeletal Features Radiographic Features Inheritance; Gene Hydropic appearance, round flat face, micrognathia, extremely narrow chest, very short limbs, postaxial polydactyly Cardiac, renal, anal malformations Normal calvarium; very short, horizontal ribs; flat, wide intervertebral disc spaces; small pelvis; short limbs with lateral and medial metaphyseal spurs AR; DYNCH2H1 (dynein heavy chain 1B for type III only) Types II and IV Hydropic; short face, flat nose, CLP; low-set ears; narrow chest, protuberant abdomen; moderately short limbs Cardiac, renal, respiratory malformations Very short, horizontal ribs; normal pelvis and vertebrae; short limbs with round metaphyses; premature epiphyseal ossification; polydactyly AR; NEK1 (never in mitosis gene A-related kinase 1 for type II only) Thanatophoric dysplasia Large cranium, proptosis, flat nasal bridge, narrow chest, very short limbs (all segments) Polyhydramnios, hydrocephalus, brain anomalies, congenital cardiac abnormalities Large calvarium, short base, small foramen magnum, cloverleaf skull (type 2); short, splayed, cupped ribs; small, very flat, U-shaped vertebrae; short, small, flat pelvis; short, bowed limbs; metaphyseal flare with spike AD (most are new mutations); FGFR3 (fibroblast growth factor receptor 3) Same gene as for achondroplasia, hypochondroplasia, SADDAN Achondroplasia Large cranium; frontal bossing, flat nose bridge, short neck; slightly narrow chest; proximal limb shortening, short trident hands; short proximal and middle phalanges; joint laxity; thoracolumbar kyphosis Hypotonia: delayed motor milestones; spinal stenosis causes spinal compression; small foramen magnum can cause hydrocephalus and apnea Large calvarium, small foramen magnum, short base; diminished lumbosacral interpedicular space, short pedicles; short ribs with anterior cupping; short humeri and femurs; relatively long fibulas; metaphyseal flare; small iliac wings AD (most are new mutations); FGFR3 Same gene as for hypochondroplasia, SADDAN, and thanatophoric dysplasia Chondrodysplasia punctata, X-linked ­recessive Hypoplasia of the distal phalanges; severe hypoplasia of nose; short stature Cataracts; hearing loss; congenital ichthyosis, anosmia, and hypogonadism (in contiguous gene deletion patients) Distal phalangeal hypoplasia; stippled epiphyses of long bones; paravertebral stippling XLR; ARSE (­arylsulfatase E) Usually milder than X-linked dominant form; variable clinical severity, with neonatal death to longevity and diagnosis in adulthood Chondrodysplasia punctata, X-linked dominant (­ConradiHunermann ­syndrome) Asymmetric rhizomesomelia Congenital cataracts; ichthyosis; patchy alopecia Stippled epiphyses of long bones; paravertebral stippling; tracheal ­calcifications XLD; ESP (3β-hydroxy Δ8-Δ7 sterol isomerase) Severe form of disease; usually lethal in males; females vary from stillborn to mild (diagnosis in adulthood); elevated 8(9)-­cholestenol Diastrophic ­dysplasia Normal cranium; cleft palate; micrognathia; normal chest at birth; very short limbs; thumbs proximally placed and adducted (hitchhiker thumb); severe equinovarus of feet; limited movement of many joints Cystic masses in auricles (cauliflower ears) during infancy; deafness caused by lack or fusion of ossicles; narrow external auditory canal Premature ossification of rib cartilage; narrow L1-L5 interpedicular spaces; scoliosis; short limbs; disproportionately short ulna and fibula (mesomelia); broad flared metaphyses; ovoid first metacarpals; variable symphalangism of proximal interphalangeal joints AR; DTDST (diastrophic dysplasia sulfate transporter) Same gene as atelosteogenesis II, achondrogenesis type IB, and some forms of multiple epiphyseal dysplasia (MED); intrafamilial variability; normal life span if tracheomalacia or scoliosis do not impair respiratory function; normal intelligence Dysplasia Skeletal ­Features Types I and III Comments Nonlethal CHAPTER 24 Skeletal Dysplasias and Connective Tissue Disorders 261 TABLE 24-1 Skeletal Dysplasias Manifesting Prenatally or Perinatally—cont’d Nonskeletal Features Radiographic Features Inheritance; Gene Large cranium; flat face with large eyes, flat nasal bridge, cleft palate; short limbs with proximal shortening (more severe in lower limbs), enlarged joints, flexion contractures Infancy: tracheomalacia; childhood: myopia and retinal detachment, hearing loss, delayed motor development, normal intelligence Frontal and maxillary hypoplasia with shallow orbits; slightly short ribs; flat vertebrae with coronal clefts; small pelvis with irregular acetabular roof; short limbs with broad, flared metaphyses (dumbbell), lateral bowing of femurs and tibiae; slightly short and broad tubular bones of hands and feet; epiphyses at knees not ossified AD; COL2A1 (type II ­collagen) Same gene as for spondyloepiphyseal dysplasia congenita, spondyloepimetaphyseal dysplasia, Stickler syndrome, hypochondrogenesis, achondrogenesis type II Flat face, cleft palate, short limbs Infancy: tracheomalacia; childhood: myopia and retinal detachment, hearing loss, normal intelligence Frontal and maxillary hypoplasia, flat vertebrae, small pelvis with irregular acetabular roof, short limbs, normal hands and feet AD; COL2A1 (type II ­collagen) Same gene as for spondyloepimetaphyseal dysplasia, Stickler syndrome, hypochondrogenesis, achondrogenesis type II, Kniest syndrome Dysplasia Skeletal ­Features Kniest syndrome Spondyloepiphyseal dysplasia congenita Comments AR, Autosomal recessive; AD, autosomal dominant; RBC, red blood cell; CLP, cleft lip with or without cleft palate; XLR, X-linked recessive; XLD, X-linked dominant. Clubfeet may occur in diastrophic dysplasia, Kniest dysplasia, spondyloepiphyseal dysplasias, and OI type II. ll Postaxial polydactyly occurs in short-rib polydactyly and asphyxiating thoracic and chondroectodermal dysplasias. Occasionally, preaxial polydactyly can also occur in the short-rib dysplasias. ll Multiple joint dislocations can manifest at birth in Larsen syndrome, Ehlers-Danlos syndrome type VII, atelosteogenesis, and Desbuquois syndrome. The presence of extraskeletal abnormalities may provide additional clues to diagnosis, as follows: ll Cleft palate may occur in campomelic, Kniest, spondyloepiphyseal, short-rib polydactyly (Majewski), atelosteogenesis types I and II, hypochondrogenesis, and diastrophic dysplasias. ll Congenital cataracts are frequent in some forms of chondrodysplasia punctata. ll Congenital cardiac defects occur in short-rib polydactyly dysplasias and Ellis van Creveld syndrome. If the infant or fetus dies, specimens of cartilage and skin fibroblasts should be obtained for histochemical tests, biochemical assays, and molecular analysis; these can be used to make or confirm diagnoses and permit accurate future prenatal diagnosis. Even if the molecular or enzymatic basis of the condition is not understood at the time, the tissue may be useful in the future. If photographs and skeletal radiographs were not obtained premortem, they should be obtained postmortem. CLINICAL TESTING Presentation Radiographs of the entire skeleton, including the skull, hands, feet, and lateral spine, are essential for accurate diagnosis. Atlases dedicated to skeletal dysplasias are essential for this purpose (Lachman, 2006; Spranger et al, 2002), even to the experienced radiologist or neonatologist. Ultrasound images of the brain, heart, and kidneys may be helpful if anomalies in those organs are suspected. Detailed family history and measurements of family members may be helpful; more mildly affected members might have gone without a diagnosis. Molecular investigations may be necessary to arrive at the proper diagnosis; given their complexity, such analyses should be considered after consultation with a clinical geneticist. OI type II (perinatal lethal type) is estimated to affect 1 in 20,000 to 60,000 infants. Affected infants may be born prematurely, with low birthweight and disproportionately short stature. The limbs are short and bowed with extra, circular skin creases; the hips are abducted and flexed. The head is soft and boggy, and minimal calvarial bone can be felt. The sclerae are dark blue and the chest is narrow. The infant cries with handling because there are many fractures at different stages of healing. Sixty percent of affected babies are stillborn or die during the first day of life, and 80% die by 1 month. With the growing use of ultrasonography, affected fetuses may be detected in the early second trimester because of ll DISORDERS OF BONE FRAGILITY OSTEOGENESIS IMPERFECTA TYPES II AND III OI is characterized by increased bone fragility. There are classically four major clinical types: types II and III are the most severe, manifesting prenatally and perinatally (Byers, 2002; Steiner et al, 2005). However, fractures at birth can occur in OI type I. Further heterogeneity in OI has recently been described. 262 PART VI Metabolic and Endocrine Disorders of the Newborn A B FIGURE 24-1 (See also Color Plate 5.) Osteogenesis imperfecta type II. A, A 20-week fetus. The limbs are angulated and deformed from multiple fractures. B, Radiograph of fetus (20 weeks’ gestation) showing an absence of ossification in the calvarium, short telescoped or crumpled humeri and femurs, and short and wavy ribs with fractures. short and bowed or angulated limbs and narrow thoraces (Figure 24-1). OI type III (progressive deforming type) can manifest prenatally, perinatally, and in the first 2 years of life. Prenatal and perinatal clinical features resemble those in OI type II, but are less severe (Figure 24-2), and perinatal death is not uncommon. If not present at birth, fractures and deformations of the limbs develop in the first and second years. The highest prevalence of fractures in OI, up to 200, occur in type III. Extremely short stature, with adult heights of 92 to 108 cm, can result from microfractures in growth plates. The head may be large because the calvarium is soft with a large anterior fontanel. The sclerae may be blue initially, but are white by puberty. The head assumes a triangular shape, with a bossed, broad forehead and a tapered, pointed chin. Later in childhood, dentinogenesis imperfecta and hearing loss may develop. Severe kyphoscoliosis may occur, leading to cardiopulmonary compromise, which is the major cause of early death. Radiolographic Features Radiographs show the femurs in OI type II to be short, broad, and “telescoped” or “crumpled.” The tibiae are short and bowed or angulated, and the fibulae may be thin (see Figure 24-1, B). There is minimal to no calvarial mineralization. The acetabulae and iliac wings may be somewhat flattened. The ribs are short, wavy, and thin or broad, with “beading” from callus formation at fetal fracture sites. In OI type III, the femurs are short and deformed, but not crumpled as in OI type II (see Figure 24-2, B and C). The other long bones are thinner than usual, with healing fractures incurred in utero, bowing, and deformations. The calvarium is undermineralized with a large anterior fontanel, and there are many Wormian bones (small islands of bone in the suture spaces; see Figure 24-2, D). The ribs are thin and gracile. Etiology OI is most commonly caused by mutations in one of the two genes for type I collagen (COL1A1 and COL1A2), the predominant protein building block of bone. More clinically severe forms of OI are the result of qualitatively abnormal collagen synthesis, rather than decreased production (Byers, 2002). Inheritance A fetus or infant with OI type II or III is usually the result of a spontaneous dominant-acting gene mutation, but there is a small risk of recurrence (approximately 6%) in subsequent siblings because of parental somatic or gonadal mosaicism. The parent is usually asymptomatic but may have minimal manifestations, such as short stature. Most cases of OI are inherited as autosomal dominant traits, although rare recessive forms have been shown to be caused by mutations in the cartilage-associated protein CRTAP (Barnes et al, 2006), prolyl 3-hydroxylase (LEPRE1), and cyclophylin B (van Dijk et al, 2009). CHAPTER 24 Skeletal Dysplasias and Connective Tissue Disorders A 263 B C D FIGURE 24-2 Osteogenesis imperfecta type III. A, Neonate with normal face, short neck, slightly short limbs. B, Radiograph shows that the calvarium is undermineralized with Wormian bones. C, Radiograph shows the upper limbs, which have bowed humeri and callus in the ulnae. D, ­Radiograph shows lower limbs with moderately short, thick femora, and angulated tibiae and fibulae. (Courtesy Paige Kaplan, Children’s Hospital of ­Philadelphia, Philadelphia, Penn.) Differential Diagnosis Other lethal skeletal dysplasias may have similar abnormalities to those in OI type II and may be difficult to distinguish by prenatal ultrasonography; however, in experienced hands they can be differentiated based on several ultrasound findings. Krakow et al (2008) did a retrospective analysis of 1500 prenatally diagnosed cases of skeletal dysplasias. The three most common prenatal-onset skeletal dysplasias were osteogenesis imperfecta type 2, thanatophoric dysplasia, and achondrogenesis 2, accounting for almost 40% of cases. Postnatal radiographs clearly reveal distinctive differences among thanatophoric dysplasia, campomelic dysplasia, achondrogenesis, and perinatal hypophosphatasia, among others. Management If the diagnosis of OI is made prenatally, cesarean delivery has not been shown to decrease fracture rate or improve survival rate of severely affected fetuses (Cubert et al, 2001). Those severely affected with OI II are not expected to survive the neonatal period. In OI type III, the neonate needs careful handling to minimize pain and prevent further fractures. Analgesia alleviates pain. Consideration can 264 PART VI Metabolic and Endocrine Disorders of the Newborn be given to treatment with bisphosphonates (using intravenous pamidronate), which increase bone density, reduce the frequency of fractures and pain, possibly prevent short stature and deformations, and permit ambulation (Aström et al, 2007). It is prudent to treat only severely affected children in whom the clinical benefits outweigh potential long-term effects. Handling an Infant With Osteogenesis Imperfecta When changing the diapers of an infant with OI, place a hand behind the buttocks with the forearm supporting the legs. Similarly, when the infant is lifted the buttocks, head, and neck must be supported. The infant can be laid on a pillow to be carried. To transport the infant, an infant seat that reclines as much as possible and allows easy placement or removal should be used. The seat can be padded with egg crating or 1-inch foam. A layer of foam can be placed between the seat’s harnesses and the child for extra protection. The car seat must always be placed in the back seat. Sling carriers and “umbrella” strollers should not be used for infants with OI because they do not give sufficient leg, head, and neck support. PERINATAL HYPOPHOSPHATASIA Presentation Perinatal hypophosphatasia is a lethal condition characterized by short, deformed limbs, a soft skull, blue sclerae, and undermineralization of the entire skeleton, so that many bones cannot be visualized and may seem absent on radiography. In the skull, only the base can be visualized radiologically. There may be rachitic changes and fractures. Seizures that are responsive to pyridoxine may occur. There is polyhydramnios during pregnancy, and death can occur in utero. The disorder affects approximately 1 in 100,000 live births; neonatal death is common (Whyte, 2000). Radiolographic Features The radiologic features of perinatal hypophosphatasia include polyhydramnios (prenatal); underossification, especially of the calvarium and long bones (with marked variability); small thoracic cavity; short, bowed limbs; spurs in the middle portion of the forearms and lower legs; and dense vertebral bodies. Etiology Mutations in the ALPL gene are responsible for deficiency of the tissue-nonspecific isoenzyme of alkaline phosphatase (TNSALP), thus causing perinatal hypophosphatasia. The serum alkaline phosphatase (ALP) value is low. Serum values of inorganic pyrophosphate and pyridoxal 5′-­phosphate (putative natural substrates for TNSALP) may be elevated, and urinary phosphoethanolamine is elevated (Mornet and Nunes, 2007). TNSALP acts on multiple substrates: the essential function of TNSALP is in osteoblastic bone matrix mineralization. TNSALP hydrolyzes inorganic pyrophosphate to phosphate, thought to be critical in promoting osteoblastic mineralization. If TNSALP is deficient, there is extracellular accumulation of inorganic pyrophosphate, which inhibits hydroxyapatite crystal formation and mineralization of the skeleton. TNSALP is also needed for delivery of pyridoxal-5-phosphate into cells where it is a cofactor (vitamin B6). Inheritance Perinatal hypophosphatasia is inherited as an autosomal recessive trait, with a 25% recurrence risk in future pregnancies. Prenatal diagnosis is optimized through the use of ultrasonography, assay of TNSALP activity in amniocytes, DNA mutation analysis if the previously affected infant’s mutation was known, or a combination of these methods. Differential Diagnosis Differential diagnoses are osteogenesis imperfecta type II and achondrogenesis. Management Treatment is primarily supportive and directed toward minimizing pain and discomfort. Clinical trials with bonetargeted human recombinant enzyme replacement therapy are underway. FGFR3 SPECTRUM ACHONDROPLASIA Presentation Achondroplasia is the most common of the nonlethal chondrodysplasias; it affects 1 in 25,000 live births. It is characterized by short stature with short limbs, particularly rhizomelic (proximal) and acromelic (hands) shortening with trident hand configuration, large head with frontal prominence (“bossing”), flat nasal bridge and midface, long narrow trunk, joint laxity, and development of thoracolumbar kyphosis (“gibbus”) in infancy (Figure 24-3). The foramen magnum and cervical spinal canal may be narrow and can cause compression of the spinal cord. Standards have been published for foramen magnum size in achondroplasia (Hecht et al, 1985). Compression of the lower brainstem and cervical spinal cord can lead to hypotonia, central apnea, retardation, quadriparesis, and (rarely) sudden death (Pauli et al, 1984). Perinatal or infantile death can occur, but is unusual. Infants often sleep with their neck hyperextended, and symptoms can be exaggerated by neck flexion (Danielpour et al, 2007). Radiolographic Features The calvarium is large with a relatively small foramen magnum and a short base. The lateral cerebral ventricles may be large, but hydrocephalus is not a common CHAPTER 24 Skeletal Dysplasias and Connective Tissue Disorders A C 265 B FIGURE 24-3 Achondroplasia. A, Infant with achondroplasia has macrocephaly and proximal limb shortening (rhizomelia). B, Infant with achondroplasia exhibits frontal bossing and flat nasal bridge. C, Neonatal film of achondroplasia illustrates a large skull, a somewhat narrow chest, short vertebral bodies with a lack of lumbar interpediculate flare, and rhizomelia. (B, Courtesy Charles I. Scott, AI DuPont Institute, Wilmington, Del.) complication. The proximal long bones (humeri and femurs) are short, including the femoral neck. Fibulae are longer than tibiae. There is metaphyseal flaring. The hand is short with a trident configuration of the fingers, with short proximal and middle phalanges. Vertebrae are small and cuboid with short pedicles, and there may be anterior beaking of the first or second lumbar vertebrae; there is lack of flare of the interpedicular distance in the lumbar vertebrae. The pelvis has squared iliac wings (“elephant ear” appearance), a narrow greater sciatic notch, and flat acetabular roofs. Compression of the cervical cord, if present, can be ascertained with magnetic resonance imaging cerebrospinal fluid flow studies in flexion and extension. 266 PART VI Metabolic and Endocrine Disorders of the Newborn Etiology The cause of achondroplasia is a mutation of the gene for fibroblast growth factor receptor 3 (FGFR3). The FGFR3 protein is a membrane-spanning tyrosine kinase receptor, which may form dimers with other gene family members FGFR1, FGFR2, and FGFR4. The heterodimers serve as receptors for several fibroblast growth factors (Cohen, 2006). More than 97% of persons with achondroplasia have a mutation in the transmembrane domain of the FGFR3 gene, in which glycine is substituted by arginine (Gly380Arg; Shiang et al, 1994). The same gene is mutated at different sites in hypochondroplasia, thanatophoric dysplasia, SADDAN, Muenke craniosynostosis, and Crouzon craniosynostosis syndrome with acanthosis nigricans (Vajo et al, 2000). Histopathologic examination demonstrates a defect in the organization and maturation of the cartilage growth plates of long bones because of differing degrees of constitutive activation of the receptor. and given a copy of the guidelines for health supervision of children with achondroplasia issued by the American Academy of Pediatrics (Trotter et al, 2005). THANATOPHORIC DYSPLASIA Presentation Thanatophoric dysplasia is one of the most common lethal dysplasias (Unger et al, 2007), occurring in 1 in 45,000 births. It is characterized by extremely short limbs, long narrow trunk, large head with bulging forehead, prominent eyes, flat nasal bridge, wide fontanel, and occasionally cloverleaf skull deformity (Figure 24-4). It is differentiated into types I and II on the basis of radiologic features. Death occurs in the neonatal period from respiratory insufficiency. Polyhydramnios is common during pregnancy. Inheritance Radiographic Features The inheritance pattern in achondroplasia is autosomal dominant. Approximately 80% of cases are sporadic occurrences in a family, representing new mutations. Cases may be associated with advanced paternal age, with molecular confirmation that the new mutations are of paternal origin. Affected individuals are fertile, and achondroplasia is transmitted as a fully penetrant autosomal dominant trait, meaning that each person who inherits the mutant gene will manifest the condition. Femurs are short, flared at the metaphyses with a medial spike, and are bowed (type I) or straight (type II); other long bones are also short and bowed (see Figure 24-4). The calvarium is large with a short base and small foramen magnum; cloverleaf skull is sometimes present in type I and is severe in type II. Vertebrae are strikingly flat (platyspondyly) with a U- or H-shape in anteroposterior projection and uniform interpediculate narrowing. Ribs are short, cupped, and splayed anteriorly (Lachman, 2006). Differential Diagnosis Differential diagnoses are SADDAN (Vajo et al, 2000) hypochondroplasia (Bellus et al, 1995). Management The infant with achondroplasia is often hypotonic; together with the large head, the hypotonia leads to delayed motor milestones. Development of thoracolumbar kyphosis may be exacerbated by unsupported sitting before truncal muscle strength is adequate; therefore infants should not be carried in flexed positions (including soft slingcarriers and umbrella strollers). Rear-facing car safety seats should always be used. Most infants lose their kyphosis and develop lumbar lordosis when they begin walking. Hydrocephalus may occasionally develop during the first 2 years, so the head circumference and body length should be carefully measured and plotted on standard achondroplasia growth charts (Trotter et al, 2005). Routine imaging of the skull and brain is not recommended; however, development of hyperreflexia, hypotonia, or apnea may herald the development of clinically significant cord compression. Surgical decompression at the foramen magnum or the upper cervical spine may prevent neurologic damage, although most patients usually gain motor milestones late but spontaneously, because the foramen grows faster than the cord. The upper airway in individuals with achondroplasia is small, often leading to obstructive apnea, snoring, and chronic serous otitis media beyond infancy. Treatment may consist of tonsillectomy, adenoidectomy, and placement of myringotomy tubes. Parents should be counseled about the clinical and hereditary aspects of the disorder Etiology Thanatophoric dysplasia represents the severe end of the FGFR3 spectrum. In thanatophoric dysplasia type I, the most common mutation in the extracellular domain is a substitution of arginine at position 248 by cysteine (Arg248Cys), but other mutations have been described throughout the gene. In all studied cases of thanatophoric dysplasia type II, there is a substitution of lysine at position 650 by glutamate (Lys650Glu; Wilcox et al, 1998). Inheritance All cases of thanatophoric dysplasia, as with most cases of achondroplasia and hypochondroplasia, occur sporadically and result from new autosomal dominant mutations. Nevertheless, there may be a small risk of recurrence to siblings of a sporadic case, possibly caused by gonadal mosaicism. Differential Diagnosis Differential diagnoses are OI types II and III, achondroplasia (severe), achondrogenesis, and hypochondrogenesis. Management If the condition is diagnosed prenatally, the couple should receive genetic counseling and anticipate neonatal death. If the diagnosis is suggested after delivery and radiographically confirmed, management is solely supportive, with death from pulmonary insufficiency usually occurring within hours to days. CHAPTER 24 Skeletal Dysplasias and Connective Tissue Disorders 267 B A C FIGURE 24-4 (See also Color Plate 6.) Thanatophoric dysplasia. A, Neonate with thanatophoric dysplasia has a large head, narrow chest, short limbs, extra creases on the limbs, short hands with trident fingers, and angulated abducted thighs. B, Neonate with thanatophoric dysplasia has a face with a bossed forehead, flat nose bridge, short neck, very short limbs with extra creases, and trident fingers. C, Radiograph of an infant with thanatophoric dysplasia demonstrates a large calvarium, short ribs with anterior splaying, flat vertebral bodies (platyspondyly), and short bowed femurs with medial metaphyseal spike. (B, Courtesy Montreal Children’s Hospital, Montreal, Quebec, Canada.) COL2A1 SPECTRUM SPONDYLOEPIPHYSEAL DYSPLASIA CONGENITA from the spinal (spondylo-) and growth plate (epiphyseal) involvement. Congenita indicates that the condition is present from birth. Presentation Radiolographic Features Spondyloepiphyseal dysplasia congenita (SEDC) manifests with shortened neck, trunk and limbs, normal-sized hands and feet, flat facial profile, and occasional cleft palate and clubfoot (Unger et al, 2007). The name is derived Radiolographic features include ovoid or pear-shaped vertebral bodies in infancy, with platyspondyly more evident at a later age; odontoid hypoplasia evident in early childhood; midface hypoplasia; retrognathia; mild rhizomelia 268 PART VI Metabolic and Endocrine Disorders of the Newborn A D C B FIGURE 24-5 (See also Color Plate 7.) Spondyloepiphyseal dysplasia congenita. A, A 2-month-old infant demonstrating short neck, trunk, and limbs. Note the flat facial profile and normal size of hands and feet. B, Anteroposterior radiograph reveals platyspondyly and short chest. C, Lateral radiograph reveals platyspondyly. D, Upper limb radiograph reveals rhizomelia, mesomelia, and a normal-sized hand. and mesomelia (Figure 24-5); absent ossification of the os pubis; apparent decreased bone age caused by epiphyseal involvement; and development of coxa vara, variable kyphosis, and scoliosis in childhood Etiology SEDC is caused by mutations in the gene for type II collagen (COL2A1), the predominant protein building block of cartilage. Mutations in COL2A1 are also responsible for Kniest dysplasia, some forms of spondyloepimetaphyseal dysplasia and Stickler syndrome, and the perinatal lethal disorders achondrogenesis type II and hypochondrogenesis (Spranger et al, 1994; Tiller et al, 1995; Winterpacht et al, 1993). Inheritance SEDC is inherited in an autosomal dominant pattern. Offspring of affected individuals are at 50% risk for inheriting the disorder. Recurrence risk for unaffected parents is approximately 6%, because of parental gonadal mosaicism. Differential Diagnosis Differential diagnoses are the milder form of hypochondrogenesis and Morquio syndrome. Management Neonates may require intubation because of upper airway compromise. Care must be given when manipulating the cervical spine (as in endotracheal intubation), because of odontoid hypoplasia. C1-C2 fusion may be required in early childhood to stabilize the cervical spine. Annual hearing screens are recommended during childhood. Regular ophthalmologic evaluation (semiannually before school age) is essential to detect early development of retinal detachment and to manage myopia. Osteoarthritis is a common feature in early adulthood, often requiring hip arthroplasty. ACHONDROGENESIS II–HYPOCHONDROGENESIS Presentation The severe end of the COL2A1 spectrum manifests with fetal hydrops and maternal polyhydramnios, severe short trunk and limbs, and fetal or neonatal death caused by pulmonary hypoplasia. Radiolographic Features Radiolographic features include prenatal polyhydramnios, a large calvarium with normal ossification, midface hypoplasia, retrognathia, platyspondyly with underossification of the vertebral bodies (achondrogenesis II), short chest with a protuberant abdomen, marked shortening of all tubular bones (Figure 24-6), and small iliac wings. Etiology Achondrogenesis II–hypochondrogenesis is caused by mutations in the gene for type II collagen (COL2A1), the predominant protein building block of cartilage (Mortier et al, 2000). CHAPTER 24 Skeletal Dysplasias and Connective Tissue Disorders A FIGURE 24-6 Achondrogenesis type II. A, This 20-week fetus demonstrates small chest and short limbs. B, This radiograph demonstrates poor ossification of vertebral bodies and short limbs. B Inheritance All cases of achondrogenesis II–hypochondrogenesis are caused by spontaneous dominant-acting mutations in COL2A1. Recurrence risk has been reported as high as 6%, because of parental gonadal mosaicism (Forzano et al, 2007). Differential Diagnosis Differential diagnoses are achondrogenesis type I and osteogenesis imperfecta types II and III. Management If the condition is diagnosed prenatally, the couple should receive genetic counseling and anticipate neonatal death. If the diagnosis is suggested after delivery and radiographic confirmation is obtained, management is solely supportive, with death from pulmonary insufficiency usually occurring within hours to days. DTDST SPECTRUM DIASTROPHIC DYSPLASIA Presentation Newborns with diastrophic dysplasia exhibit limb shortening, cystic ear swelling, hitchhiker thumbs (developing several days after birth), spinal deformities (especially ­cervical kyphosis), and contractures of the large joints (Figure 24-7). Clubfoot and ulnar deviation 269 of the fingers may also be present. On occasion the ­disease can be lethal at birth, but most individuals survive the neonatal period (Bonafe and Superti-Furga, 2007). Radiolographic Features The most characteristic clinical and radiologic feature is the proximally-placed hitchhiker thumb, with ulnar deviation of the fingers. Cervical kyphosis is a frequent finding. Long bones are moderately shortened and thick, with mild metaphyseal flaring, rounding of the distal femur, and bowing of the radius and tibia. Severe talipes equinovarus may be present. Iliac wings are hypoplastic, with flat acetabular roofs. The chest can be narrow, bell shaped, or both. Narrowing (lack of flare) of the interpedicular distance in the lumbar spine is reminiscent of achondroplasia. Etiology Diastrophic dysplasia is caused by mutations in the diastrophic dysplasia sulfate transporter gene (DTDST), which also cause the lethal disorders achondrogenesis type 1B and atelosteogenesis type 2, as well as a rare recessive form of multiple epiphyseal dysplasia. The gene product is a sulfate-chloride exchanger of the cell membrane (Superti-Furga et al, 1996); this affects incorporation of sulfate into proteoglycans (mucopolysaccharides), especially chondroitin sulfate B–containing proteoglycans, which are prevalent in cartilage. 270 PART VI Metabolic and Endocrine Disorders of the Newborn A C B FIGURE 24-7 Diastrophic dysplasia. A, This infant has prominent eyes, small chin, slightly narrow chest, proximally placed angulated thumbs, and short limbs. B, Neonate profile showing small chin, swollen ears, and short neck. Note the proximally placed, angulated thumb. C, View of the ­neonate’s hand shows the proximally placed angulated thumb and mild syndactyly. (Courtesy Paige Kaplan, Children’s Hospital of Philadelphia, Philadelphia, Penn.) Inheritance Radiolographic Features Diastrophic dysplasia is inherited in an autosomal recessive pattern. Siblings of affected individuals are at 25% risk for inheriting an abnormal allele from both carrier parents. The long bones are extremely short, with square, globular, or triangular shapes and medial spikes in the metaphyses of the femurs (Figure 24-8). The calvarium and vertebrae are poorly ossified (type IB), and the ribs are short. Differential Diagnosis Differential diagnoses are atelosteogenesis type II (part of the DTDST spectrum), spondyloepiphyseal dysplasia, and arthrogryposis. Etiology Management Inheritance Mechanical ventilation may be required because of small chest circumference and a floppy airway. Maintenance of joint mobility and proper positioning through physical therapy is essential. Serial casting and or surgical correction of clubfeet may be required. Cervical kyphosis can impede endotracheal intubation and can result in cord compression, but may resolve spontaneously during infancy. Achondrogenesis 1B is inherited in an autosomal recessive pattern. Siblings of affected individuals are at 25% risk for inheriting an abnormal allele from both carrier parents. ACHONDROGENESIS 1B Presentation Achondrogenesis type IB is characterized by short stature, extremely short limbs, a relatively large head with a round face, short nose, small mouth, soft skull, and very short neck (Borochowitz et al, 1988). Polyhydramnios during pregnancy, premature delivery, and hydrops are common. The affected infant is stillborn or dies within hours of birth. Achondrogenesis type 1B is caused by mutations in the DTDST gene (see Diastrophic dysplasia, earlier). Differential Diagnosis Differential diagnoses are atelosteogenesis type II (part of DTDST spectrum), achondrogenesis type II, and hypochondrogenesis. Management If the condition is diagnosed prenatally, the couple should receive genetic counseling and anticipate neonatal death. If the diagnosis is suggested after delivery and radiographic confirmation is obtained, management is solely supportive, with death from pulmonary insufficiency usually occurring within hours to days. CHAPTER 24 Skeletal Dysplasias and Connective Tissue Disorders A B OTHER SKELETAL DYSPLASIAS CAMPOMELIC DYSPLASIA 271 FIGURE 24-8 A and B, Achondrogenesis type IB. Cervical, thoracic, and lumbar vertebral bodies are not ossified, the sacrum is not ossified, the ribs are short, and the limbs are extremely short with medial femoral metaphyseal spikes. (Courtesy Elaine Zackai, Children’s Hospital of Philadelphia, Philadelphia, Penn.) chest with thin, wavy ribs (with only 11 pairs); and scapular hypoplasia. Presentation Etiology Campomelic dysplasia is characterized by short stature (birth length of 35 to 49 cm), large dolichocephalic skull, large anterior fontanel, high forehead, flat face, widely spaced eyes with short palpebral fissures, low-set ears, cleft soft palate, micrognathia, relatively long and slender thighs and upper arms, short bowed legs with dimples in the midshaft (in most cases), narrow chest, and kyphoscoliosis (Figure 24-9). Sex reversal or ambiguous genitalia affects 75% of the chromosomal males; there may be internal and external genital abnormalities (from mild anomalies to complete sex reversal) in XY males (Meyer et al, 1997; see Figure 24-9, A and B). Absence of the olfactory bulbs and tracts as well as heart and renal malformations may occur. Death, usually in infancy, results from pulmonary hypoplasia, tracheomalacia, or cervical spinal instability. Survivors are usually globally developmentally delayed. A few more mildly affected people without bowed limbs have been reported (Unger et al, 2008). Campomelic dysplasia is caused by mutations in or near the SRY-related HMG-Box gene 9 (SOX9; Tommerup, 1993). SOX9, with homology to the SRY gene, is a transcription factor involved in both bone formation and testis development. Radiolographic Features The most characteristic finding is midshaft angulation (campomelia) of the femurs, although it is not a constant finding. Other features include hypoplastic, undermineralized cervical vertebrae and thoracic pedicles; narrow iliac wings with dislocated hips; brachydactyly; clubfeet; anterior bowing of tibia; bell-shaped Inheritance Campomelic dysplasia is an autosomal dominant trait. Most cases are new sporadic occurrences in a family; recurrence caused by gonadal mosaicism has been reported (Smyk et al, 2007). Differential Diagnosis Differential diagnoses are osteogenesis imperfecta types II and III, diastrophic dysplasia, kyphomelic dysplasia, thanatophoric dysplasia, and spondyloepiphyseal dysplasia congenita (severe). Management Survival beyond the newborn period is rare; therefore support is primarily directed toward comfort measures. In survivors, care must be given to the cervical spine, which may be unstable. Chromosomal studies to determine gender and pelvic ultrasonography to examine internal genitalia 272 PART VI Metabolic and Endocrine Disorders of the Newborn A B C FIGURE 24-9 Campomelic dysplasia. A, 46,XY female 22-week-old fetus with normal head, long philtrum, micrognathia, low-set ears, mild narrowing of chest, proximally placed thumbs, and bowed or angulated lower limbs resembling those of osteogenesis imperfecta type II but less shortened. The external genitalia are female. B, Neonate with the long-limb form of the disorder has a relatively large head, micrognathia, narrow chest, and bowing of lower limbs with characteristic dimpling of lower leg. C, Radiograph shows the narrow chest, the relatively long, thin limb bones with bowing of the femurs and tibiae, and a long, narrow pelvis. (Courtesy Paige Kaplan, Children’s Hospital of Philadelphia, Philadelphia, Penn.) may be performed. Cleft palate may be repaired in those able to feed orally, and clubfeet may require casting or surgical correction. CONNECTIVE TISSUE DISORDERS CONGENITAL (NEONATAL, INFANTILE) MARFAN SYNDROME Presentation Infants with neonatal congenital Marfan syndrome (cMS) have a long, thin body and can have an aged appearance because of a lack of subcutaneous tissue and wrinkled, sagging skin (Morse et al, 1990; Figure 24-10). The craniofacial features include dolichocephaly, deep-set eyes with large or small corneas (and occasionally cataracts), high nasal bridge, high palate, small pointed chin with a horizontal skin crease, and large simple or crumpled ears. The fingers and toes are long and thin (arachnodactyly). Some joints are hyperextensible, and others have flexion contractures causing clubfoot, dislocated hips, or adducted thumbs. Infants tend to exhibit hypotonia with low muscle mass. Lenses are usually not subluxated at birth. The most important cause of morbidity and mortality is severe cardiovascular disease, which affects almost every neonate with cMS—namely, mitral and tricuspid valve prolapse and insufficiency and aortic root dilatation. The ascending aorta may be dilated and tortuous. Many infants die in the first year of life from congestive heart failure. Survivors have chronic hypotonia and contractures, are unable to walk, and require many surgical procedures. Radiolographic Features Radiolographic features include pectus deformity, spontaneous pneumothorax, dural ectasia, aortic root dilatation, and mitral valve prolapse. Many of these features may not be present in the newborn period. Etiology Congenital MS is caused by mutations in the gene encoding fibrillin 1 (FBN1; Dietz et al, 1991). Fibrillin is a glycoprotein associated with microfibrils, which form linear bundles in the matrices of many tissues, such as aorta, periosteum, perichondrium, cartilage, tendons, muscle, pleura, and meninges. There are two regions in FBN1 in which many mutations causing cMS occur; these lie among exons 24 to 27 and exons 31 and 32 (Dietz, 2009). Molecular analysis does not yield mutations in all cases. Inheritance Marfan syndrome is an autosomal dominant disorder. Most neonates with cMS are sporadic occurrences within a family (Dietz et al, 1991; Morse et al, 1990). However, there is one well-documented neonate with cMS whose father had classic Marfan syndrome except for average height (Lopes et al, 1995). Differential Diagnosis Differential diagnoses are congenital contractural arachnodactyly (CCA), autosomal recessive cutis laxa, and Loeys-Dietz syndrome. CHAPTER 24 Skeletal Dysplasias and Connective Tissue Disorders 273 B A C FIGURE 24-10 Congenital Marfan syndrome. A, Neonate with long, thin trunk and limbs (particularly the feet), lack of adipose tissue, and multiple skin creases giving an aged appearance. The ears are large and simple, and the chin is small with a horizontal crease. There are flexion contractures at the joints. B, The neonate’s face shows laxity of skin, typical horizontal chin crease, and a pointed chin. The fingers are long with adduction contractures of the thumbs, which extend past the edge of the palm, and floppy wrists. C, Lateral view of the neonate’s head showing simple, large ears and redundant skin on the neck. (Courtesy Paige Kaplan, Children’s Hospital of Philadelphia, Philadelphia, Penn.) Management Patients require annual ophthalmologic and cardiac evaluation throughout childhood. Cardioselective beta-­blockers, such as atenolol, are often implemented at the first signs of aortic root dilatation. The angiotensin II antagonist losartan has also shown promise in this regard (Brooke et al, 2008). Children should be screened for the development of scoliosis. CONGENITAL CONTRACTURAL ARACHNODACTYLY (BEALS SYNDROME) Presentation CCA (Beals syndrome) is characterized by a thin, wasted appearance with minimal muscle and fat mass (similar to neonatal Marfan syndrome). Distinctive features include arachnodactyly with contractures of the large and small joints (Figure 24-11, A), as well as crumpled, overfolded helices of the external ear. Cardiovascular involvement is usually limited to mitral valve prolapse, but aortic root dilatation may occasionally develop (Godfrey, 2007). Radiolographic Features Features are nonspecific and include elongated proximal phalanges; contractures of digits, ankles, knees, and hips; thin, gracile tubular bones; and gradual development of kyphoscoliosis. Etiology CCA is caused by mutations in the fibrillin 2 gene. 274 PART VI Metabolic and Endocrine Disorders of the Newborn Inheritance CCA is inherited in an autosomal dominant manner, with many patients representing the result of spontaneous mutations. Offspring of affected individuals are at 50% risk for inheriting the condition. Gonadal mosaicism has been described in CCA (Putnam et al, 1997). Differential Diagnosis caused by mutations in the gene for type III collagen, COL3A1. The arthrochalasia type is caused by mutations in either gene for type I collagen (COL1A1 or COL1A2), which result in loss of the N-proteinase cleavage site of the protein. Inheritance Proper nutrition is essential to ensure adequate weight gain. Joint contractures respond to physical therapy, but occasionally surgical release may be required. Surveillance for development of spinal curvature and aortic root dilatation, although rare, are essential throughout childhood. Most types of EDS are inherited as autosomal dominant traits. Each child of an affected person has a 50% chance of inheriting and manifesting the disorder, although there can be marked intrafamilial variability (Malfait and de Paepe, 2005). One form of the arthrochalasia type, also referred to as dermatosparaxis, is inherited in an autosomal recessive pattern, and is caused by deficiency of procollagen N-peptidase. The kyphoscoliotic form (formerly type VI) is also inherited in an autosomal recessive pattern and is caused by a deficiency of lysyl hydroxylase, an enzyme that aids in crosslinking of collagen fibrils. EHLERS-DANLOS SYNDROMES Differential Diagnosis Presentation Differential diagnoses are cMS, congenital contractural arachnodactyly, and Larsen syndrome. Differential diagnoses are cMS, cutis laxa, and distal arthrogryposis. Management The Ehlers-Danlos syndromes (EDSs) are a clinically and genetically heterogeneous group of connective tissue disorders, which are characterized by varying degrees of joint and skin hypermobility, excessive bruising, abnormal wound healing, and fragility of tissues (Steinmann et al, 2002; Wenstrup and de Paepe, 2008). This group of disorders was reclassified in 1997 (Beighton et al, 1998). The classic type (formerly type I) and the arthrochalasia type (formerly type VII) are the most likely to manifest in the newborn period. Type I is often characterized by premature delivery of an affected fetus as a result of a rupture of the fragile amniotic membranes. The infant may be floppy and in the breech position. There may be joint laxity and joint instability. In type VII, the major involvement is in the ligaments and joint capsules. Large and small joints are hypermobile and dislocatable; severe congenital dislocation of hips occurs. In vascular (formerly type IV) EDS, the greatest danger is to the pregnant affected woman, for whom there is a high risk of uterine and arterial rupture. Although there is a 50% risk that the fetus will be affected, the problems of blood loss and prematurity are more important in the newborn period than the disorder itself. Radiolographic Features Radiolographic features are dependent on the particular type of EDS. Congenital hip dislocation may be evident on plain films. Hydronephrosis, bladder diverticula, and spontaneous pneumothorax may occur occasionally. Aortic dilatation and arterial aneurysms may be evident by echocardiography and other imaging modalities, but only occur in patients of school age or older. Etiology Mutations in two of the genes for type V collagen (COL5A1 and COL5A2) are demonstrable in some cases of classic EDS (Malfait and de Paepe, 2005). The vascular type is Management Trauma should be avoided because of skin fragility. Effective closure of surgical wounds is challenging because of a tendency for dehiscence (Wenstrup and Hoechstetter, 2004). CUTIS LAXA Presentation Cutis laxa is a genetically heterogeneous disorder, meaning that mutations in several different genes may be responsible for the phenotype. As such, the presentation can be highly varied. Infantile forms may exhibit loose, furrowed skin, a large anterior fontanelle, hypotonia, hernias, and congenital hip dislocation (see Figure 24-11, B) (Kaler, 2005; van Maldergem et al, 2011). Radiolographic Features Radiolographic features are in part dependent on the genetic form of the disorder. Nonspecific features include a large anterior fontanel, congenital hip dislocation, and hernias. The X-linked form may exhibit occipital horns. Arterial tortuosity, aortic root dilatation, and cortical and cerebellar anomalies may be seen in some forms, as well as gastrointestinal and urinary tract diverticula. Etiology The relatively mild, autosomal dominant form of cutis laxa is caused by mutations in the elastin gene, ELN. The X-linked recessive form (occipital horn syndrome) is caused by mutations in the ATP7A gene (allelic with Menkes syndrome). Autosomal recessive forms may be caused by mutations in the fibulin 4 (FBLN4) and fibulin CHAPTER 24 Skeletal Dysplasias and Connective Tissue Disorders A 275 B FIGURE 24-11 A, Congenital contractural arachnodactyly (Beals syndrome). This infant has a long, thin trunk and limbs, contractures of joints, and crumpled ears. B, Infant with cutis laxa. (Courtesy Montreal Children’s Hospital, Montreal, Quebec, Canada.) 5 (FBLN5) genes or the A2 subunit of the V-ATPase gene (ATP6V0A2). Biochemical clues as to the etiology in a particular patient may include decreased serum copper and ceruloplasmin (X-linked form), and abnormal serum sialotransferrin isoelectric focusing in cases caused by ATP6V0A2 mutations. Inheritance Because cutis laxa is genetically heterogeneous, modes of inheritance include autosomal dominant, autosomal recessive, and X-linked recessive. The latter two modes are usually responsible for forms with neonatal and infantile presentation. Differential Diagnosis Differential diagnoses are EDS, Menkes syndrome, gerodermia osteodysplastica, and de Barsy syndrome. Management Serious childhood complications include developmental delay, pulmonary emphysema, aortic root dilatation, and arterial tortuosity. Annual ophthalmologic and cardiac examinations are essential, and referral to special education programs may be indicated. MENKES SYNDROME Presentation Menkes syndrome often appears in the newborn period with nonspecific neurologic manifestations. Typically, developmental delay is evident in the first 2 to 3 months of life, with failure to thrive, seizures, and severe ocular manifestations. Changes in the appearance of the hair include hypopigmentation, brittleness, patchy alopecia, and twisted shafts seen by light microscopy (i.e., pili torti; Figure 24-12). Early death is common and may occur in infancy (Kaler, 2010). Serum copper and ceruloplasmin concentrations are low, and the plasma dopamine-to-­norepinephrine ratio may be elevated (Goldstein et al, 2009). Radiolographic Features Features may evolve during infancy and may include bladder diverticula (seen on bladder ultrasound and voiding cystourethrogram [VCUG]), tortuous vessels (on echocardiogram, magnetic resonance imaging with contrast), gastric polyps (on upper GI), metaphyseal spurring, osteopenia, and Wormian bones on plain radiographs (Lachman, 2006; see Figure 24-12). Etiology Menkes syndrome is caused by mutations in a coppertransporting adenosine triphosphatase gene, ATP7A (Kaler et al, 1994). This enzyme takes part in the final processing of a number of copper-dependent enzymes, including dopamine beta-hydroxylase, tyrosinase, lysyl oxidase, superoxide dismutase, and cytochrome c oxidase. As a result, several physiologic processes and cellular functions are impaired, including collagen cross-linking, pigment production, and neurotransmission (Goldstein et al, 2009). Inheritance Menkes syndrome is an X-linked recessive disorder; therefore only males are affected. Female carriers may exhibit pili torti in some hair shafts because of lyonization (Moore 276 PART VI Metabolic and Endocrine Disorders of the Newborn FIGURE 24-12 (See also Color Plate 8.) Menkes syndrome. A, Note blonde hair, fair complexion, and epicanthal folds in this 11-month-old Hispanic boy. B, Note multiple Wormian bones near the occiput. A and Howell, 1985). Sons born to carrier females have a 50% risk for manifesting the disease. Differential Diagnosis Differential diagnoses are cutis laxa (the occipital horn form is allelic), EDS, neonatal cMS, biotinidase deficiency, mitochondrial myopathies, nutritional copper deficiency, and organic aciduria. Management Early diagnosis allows for parenteral copper supplementation therapy (Kaler et al, 2008), but this is not effective in all patients. Patients should be monitored for the development of seizures, as well as a propensity for bone fragility, poor wound healing, and vascular fragility leading to excessive bleeding, hemorrhagic strokes, and B subdural hematomas. Bladder diverticula may result in urinary retention and urinary tract infections and should be surgically corrected. Patients are at risk for moderate to severe developmental delay, and they should be referred to infant stimulation and early intervention programs. SUGGESTED READINGS Aström E, Jorulf H, Söderhäll S: Intravenous pamidronate treatment of infants with severe osteogenesis imperfecta, Arch Dis Child 92:332-338, 2007. Brooke BS, Habashi JP, Judge DP, et al: Angiotensin II blockade and aortic-root dilation in Marfan’s syndrome, N Engl J Med 358:2787-2795, 2008. Kaler SG, Holmes CS, Goldstein DS, et al: Neonatal diagnosis and treatment of Menkes disease, N Engl J Med 358:605-614, 2008. Trotter TL, Hall JG: The Committee on Genetics: Health supervision for children with achondroplasia, Pediatrics 116:771, 2005. Wenstrup RJ, Hoechstetter LB: Ehlers-Danlos syndromes. In Cassidy SB, Allanson JE, editors: Management of Genetic Syndromes, ed 2, New York, 2004, Wiley-Liss, pp 211-224. Complete references and supplemental color images used in this text can be found online at www.expertconsult.com P A R T V  II Care of the Healthy Newborn C H A P T E R 25 Initial Evaluation: History and Physical Examination of the Newborn Jeffrey B. Smith The central focus of this chapter is the medical evaluation of the apparently well newborn in the first few days after birth in a hospital setting. Obtaining the history, performing the physical examination, and judging the significance of risk factors and findings all require skill and experience. For brevity, we mostly refer to the person evaluating the newborn as a pediatrician. In some institutions, the routine evaluation is performed by a general pediatrician, family practitioner, or pediatric nurse practitioner, and the neonatologist enters the well baby nursery only as a consultant. In other institutions, the neonatologist has direct responsibility for patients in the normal nursery in addition to the neonatal intensive care unit (NICU). In either setting, the neonatologist should develop and maintain proficiency in the evaluation of the normal newborn, a task that requires an approach and set of skills somewhat different than the evaluation of the NICU patient, who typically has a known symptom or high-risk condition. In the normal nursery, the vast majority of newborns are healthy and will not require medical intervention; therefore a primary goal is to identify the small minority with problems that have the potential to cause serious morbidity if not detected in a timely fashion. This goal includes identifying psychosocial and medical problems. For the healthy majority, the evaluation provides the basis for appropriate parental reassurance and education. For this reason, the emphasis of the discussion in this chapter is on common problems, variations of normal, and subtle abnormalities. Some symptoms and findings that suggest significant illness are also described, but the discussion in this chapter is not intended to provide a complete differential diagnosis or guide to management. Physical findings pertinent to specific diseases and organ systems are described with greater detail in other sections throughout the book. Evaluation of the dysmorphic infant is described in Chapter 19. Laboratory studies helpful in the routine care of the healthy newborn are discussed in Chapter 26. HISTORY The medical history of the newborn begins with pertinent information about the mother’s past medical and pregnancy history, the current pregnancy, and the family. Information about the current pregnancy, labor, and delivery is central at the time of delivery, but the infant’s postnatal history becomes progressively more important in the subsequent hours and days. An outline of basic components of the prenatal and newborn history is presented in Table 25-1. The history is potentially gathered from multiple sources, including records of prenatal outpatient visits and laboratory studies; the records of the mother’s current and prior hospitalizations; the delivery record; newborn records created by nurses and other personnel; direct communications from the obstetrician, midwife, and nurses; and interviews with the mother and other family members. Most of the prenatal history is collected and recorded by members of the obstetric team. Systems should be in place to ensure that the pediatrician responsible for the newborn is directly informed about high-risk conditions in a timely manner, but the pediatrician has an independent responsibility to review the information available in the maternal record. From a potentially large amount of information, a key task for the pediatrician is to efficiently identify and highlight the portions that are relevant to the newborn’s current situation and the task at hand. To ensure that important information is not overlooked, a systematic approach to the collection and recording of the history is essential. Structured data systems, preferably electronic, can help to ensure that essential information is not missed. However, information gathering must be prioritized appropriately, because the relative importance of specific parts of the history depends on the clinical situation. For example, the physician paged to attend the emergency delivery of a fetus in distress should focus information gathering in the few minutes available on what is directly relevant to preparations for resuscitation. At that moment, it is not necessary to know the results of prenatal testing of the mother for hepatitis B. For the stable infant sent to room-in with the mother and for the critically ill infant admitted to the NICU, it becomes important to obtain this information in the next few hours, so that hepatitis B vaccine can be administered within the recommended 12 hours after birth if the mother does not have a documented negative test result for hepatitis B infection. For the healthy newborn, history gathering at the time of the initial encounter after birth will emphasize the prenatal history (including maternal and family history), the delivery and neonatal transition, the initiation of feeding, and any symptoms or parental concerns that have manifested since birth. A major goal of the initial evaluation is 277 278 PART VII Care of the Healthy Newborn TABLE 25-1 Components of the Prenatal and Newborn History Category Typical Components of the History Maternal identification Name, medical record number, age, gravidity, parity, estimated gestational age Maternal medical history Allergies Significant past illnesses, hospitalizations, and surgeries Chronic illnesses, especially diabetes Chronic medications Psychiatric history Previous pregnancies Dates, routes of delivery, complications, and outcomes Breastfeeding history Family history General health status and history of family members Congenital anomalies, metabolic disorders, hearing impairment Food and other allergies of parents and siblings Psychosocial history Family structure, care of previous children Availability of support for parents (extended family, others) Living accommodations, access to transportation Language, racial or ethnic group, parental education level Smoking, alcohol, drugs of abuse Family strife, domestic violence Current pregnancy Estimated gestational age (and how determined) Singleton or multiple fetuses Fertilization history (e.g., assisted reproduction) Prenatal care (when started, number of visits, provider) Maternal blood type and screening for unusual isoimmune antibodies Screening for group B streptococcal colonization Other prenatal screening for infectious diseases* Glucose tolerance test results Results of maternal drug screening, if applicable Occupational or other exposure to teratogens Sexual contact with high-risk group Genetic screening tests Ultrasound studies and amniocentesis Complications, hospitalizations Prescription and nonprescription medications; herbal remedies Breastfeeding plans; plans for well-child care after discharge Labor Spontaneous, induced, augmented Duration of first and second stages Rupture of membranes: spontaneous versus artificial, duration Amniotic fluid: clear, meconium stained, bloody, foul smelling Signs of infection: fever, increased white blood cell count, uterine tenderness, tachycardia Antibiotics during labor and the indication Other medications during labor Fetal monitoring abnormalities Delivery and delivery room stabilization Route and presentation Indication for cesarean section, vacuum, or forceps, if applicable Stabilization and resuscitation measures; time to spontaneous breathing Apgar scores Placental abnormalities, if noted Cord blood gases, if obtained Ease versus difficulty of transition Medications administered (erythromycin eye drops, vitamin K) Abnormalities noted on initial examination Postnatal history Nursing or parental concerns Growth parameters (weight, length, head circumference, and percentiles) Vital signs, daily weights Glucose checks Activity and alertness Breastfeeding (frequency, duration, quality of latch) Bottle feeding (amount, frequency, reason) Voiding, wet diapers Meconium and transitional stools Procedures (lumbar puncture, circumcision) Results of laboratory tests and imaging studies *Prenatal screening performed in accordance with medical recommendations and local regulations commonly include tests for group B Streptococcus, hepatitis B, syphilis, gonorrhea, and for immunity to rubella; it may also include tests for chlamydia, tuberculosis, HIV, and hepatitis C. CHAPTER 25 Initial Evaluation: History and Physical Examination of the Newborn to identify risk factors for problems that may develop in the next few days, such as early-onset neonatal sepsis or exacerbated hyperbilirubinemia. Identification of psychosocial risk factors is also an important goal, and it remains important throughout the hospitalization. At the time of the predischarge evaluation, the key goal is to determine whether the infant can be discharged home safely. The prenatal and perinatal history will have already been documented and reviewed, so information-gathering for the predischarge evaluation will focus on the interval history, mostly collected from the nursing records and the parents. Because of the frequency and potential morbidity of early onset neonatal sepsis, the presence or absence of risk factors for sepsis should be assessed as part of the initial evaluation of every newborn. Historical risk factors for early onset sepsis include prolonged rupture of the fetal membranes (18 hours or more), a maternal body temperature of 38° C or higher, uterine tenderness, foul-smelling or purulent amniotic fluid, an elevated maternal white blood cell count or left shift, and fetal or maternal tachycardia. Maternal colonization with group B streptococci is also considered a risk factor unless adequate intrapartum prophylaxis was administered or the fetal membranes were intact until delivery by cesarean section (American Academy of Pediatrics, 1997). Poor tolerance of labor, manifested by an unexplained need for resuscitation or a slow transition, can be a nonspecific symptom of sepsis. Whereas gathering the history and performing the physical examination are distinct activities, described by convention in separate parts of a note, they are not performed in isolation. Knowledge of specific concerns, events, and risk factors in the history should prompt a more focused or detailed examination of the relevant body region or organ system than otherwise might be done. Conversely, specific questioning prompted by physical examination findings will often elicit information that was not volunteered in the earlier routine history gathering, such as a family history of a specific anomaly. When a note is written, the history, the physical examination, and laboratory studies are the data upon which the current assessment and plan of care are based. Once documented, the assessment and plan also become part of the patient’s ongoing history. PHYSICAL EXAMINATION Physical examination requires skills whose refinement can continue to reward and challenge the pediatrician and neonatologist throughout a professional lifetime. No matter how experienced the practitioner, every examination represents an opportunity to add to or refine one’s knowledge of significant abnormalities and of the wide range of variation of common, benign conditions. Nevertheless, the examiner knows that most newborns are in fact healthy; therefore one of the practical challenges of performing the routine newborn examination is to maintain a high level of vigilance and thoroughness throughout every examination. The examination of the healthy newborn entails a complete physical examination in the sense that all parts of the body are examined. However, no actual examination or its write-up can be complete in the sense of exhaustively exploring and explicitly documenting all possible findings. 279 The degree of detail that is appropriate is a matter of judgment that will vary with the presenting situation and the findings discovered during the examination. Distinct subtypes of the newborn examination include examinations in the delivery room, at admission, and at discharge from the normal nursery, in addition to examinations initiated in response to specific concerns. EVALUATION AT BIRTH At the time of birth, attention focuses on the initiation of air breathing and cardiorespiratory stability. The assessment of successful adaptation or a need for resuscitation is described in Chapter 28. The infant whose condition remains unstable or who has major anomalies that are apparent on initial inspection will be transferred to an NICU for further evaluation and management. DELIVERY ROOM DISPOSITION The infant whose condition stabilizes after delivery, with or without intervention, is evaluated to determine whether the infant can remain with the mother. This examination centers on assessment of the adequacy of the cardiorespiratory transition, and it also includes a basic inspection for congenital anomalies such as imperforate anus or other problems that may indicate a need for admission to an NICU or observation nursery. Passage of a thin catheter through both nostrils and into the stomach can be done in the delivery room to rule out choanal and esophageal atresia. For routine deliveries, the delivery room discharge examination is often done by the obstetrician or nursemidwife. If the pediatrician is present, the examination at this time can be expanded to serve as the nursery admission exam. NURSERY ADMISSION EXAMINATION For babies not requiring admission to an NICU or observation unit, this examination is done after the infant completes transition, and usually by 24 hours after birth. The purpose of a complete physical examination is to efficiently detect problems that, if present, are inapparent or may soon develop. This examination is the main focus of this chapter. The initial physical examination of the patient admitted to the NICU is similar, except that examination in the NICU should be initiated immediately after admission, and parts of the NICU examination may need to be modified or delayed because of physiologic instability or limited accessibility. Whether in the nursery or the NICU, the identification of historical risk factors or the detection of symptoms or abnormalities requires that the basic examination be expanded to focus additional attention on the areas relevant to the differential diagnosis. TARGETED OR PROBLEM-DIRECTED EXAMINATION When the physician is called to evaluate an infant because of specific symptoms or concerns, the examination will naturally focus on aspects relevant to those issues. Because 280 PART VII Care of the Healthy Newborn the nonspecific nature of many symptoms in the infant often implies a large differential diagnosis, even the targeted examination will often need a wide focus. The daily follow-up examination can be regarded as a variety of targeted examination, guided by the infant’s overall condition. For healthy infants remaining in the hospital for several days after birth, areas of active attention always include the infant’s neurologic and cardiorespiratory stability, hydration status, feeding and elimination behavior, and jaundice. NURSERY DISCHARGE EXAMINATION The discharge examination is similar in scope to the wellbaby admission examination, but with a slightly altered emphasis, based on the additional information provided by the period of observation in the hospital, and by the goal of determining whether the infant is ready for routine care at home. The discharge examination is still a complete physical examination that encompasses the entire body. However, it is not necessary for the discharge examination to duplicate all portions of the admission examination, provided that it was performed by the same person or specific aspects of the admission examination were sufficiently well documented by a trusted colleague. For example, it is unnecessary to repeat a search for physical anomalies that do not change with time, such as examination of the oral cavity for cleft palate, or to check again for red retinal reflexes if they were already found to be normal. Nevertheless, it is often more efficient to repeat the entire examination than to verify the completeness of an earlier examination by another. When the hospital stay is short, a combined admission-discharge examination is appropriate. EVALUATION OF GESTATIONAL AGE If the obstetric estimate of the gestational age is uncertain or appears unreliable, the gestational age of the newborn infant can be estimated based on physical examination criteria. No individual feature is a reliable guide to the gestational age, but scoring systems that use multiple features of physical and neuromuscular maturity have been evaluated extensively (Amiel-Tison, 1968; Dubowitz et al, 1970). The New Ballard Score is probably the most widely used in contemporary practice (Ballard et al, 1991). Detailed descriptions and a video demonstration of this examination are available at www.ballardscore.com. ENVIRONMENT OF EXAMINATION The environment in which the examination is performed can significantly affect the reliability of the examination via effects on the examiner or on the baby. The examiner must be aware of the limitations produced by a suboptimal environment and adjust the approach to the examination to compensate, arrange for the infant to be moved, or defer selected parts of the examination when appropriate. Important environmental considerations include lighting, room temperature, and the levels of background noise and other distractions. An often overlooked source of distraction is a situation that forces the examiner into a physically awkward or uncomfortable position, such as tall examiner bending over a low bed or examining table. If possible, the infant’s clinical state and tolerance for handling can also have a major effect on the conduct of the examination. The wise examiner will postpone the examination of a hungry, crying infant if possible, and try to return when the infant is fed and calm. The healthy infant is typically examined in a warmer bed in the delivery room or in a bassinet in the mother’s room or in the nursery. An open warmer bed provides the best access to the patient, allowing the infant to be kept warm while completely undressed during the examination. The presence of a warmer bed in the delivery room adds to the advantages of performing the nursery admission examination soon after delivery, but often this is not practical. Most examinations of the healthy newborn are done with the infant in a bassinet, starting with the infant dressed in at least a shirt and hat, and wrapped in blankets. An adequate well-baby examination can be performed under these conditions, but the sequence of the examination should be modified to minimize the time the infant is fully undressed. The examiner must take extra care to ensure that the examination is complete and the entire skin surface is visualized at some point during the examination. Performing the examination in the presence of the parents allows the examiner to show how the infant responds to handling, and to demonstrate immediately any findings that require explanation or reassurance. Watching the examination may stimulate the parents to ask questions that might otherwise not occur to them until later, and it provides the physician an immediate opportunity for further education. On the other hand, interruptions from parents and other family members can interfere with the examiner’s train of thought, risking inadvertent distraction and omissions. In large, busy nurseries, it may be more practical for the pediatrician to examine a series of infants in the nursery and then report the results of the examination to each set of parents afterward. Specific findings should be demonstrated to the parents at that time, if appropriate. Regardless of whether the parents are present for the examination, the results of the examination should be communicated promptly. Parents may be anxious about findings that the pediatrician believes are of little consequence, and vice versa. Prompt and sensitive communication of the examination findings helps to build parents’ trust, whereas delayed or poor communication can undermine the parents’ relationship with the physician and the hospital. CLINICAL APPROACH TO THE NEWBORN EXAMINATION An important challenge of the newborn examination is the need to maintain a high level of vigilance and thoroughness, while projecting an attitude of comfortable reassurance appropriate to the reality that the overwhelming majority of newborns seen in the well-baby nursery are in fact healthy. Although the eyes and fingers of the experienced examiner will detect many abnormalities by pattern recognition alone, even the most expert examiner can miss important findings unless a disciplined, systematic approach is used. However, the desire to ensure that no CHAPTER 25 Initial Evaluation: History and Physical Examination of the Newborn essential aspect of the examination is missed or slighted needs to be balanced with the infant’s limited tolerance for handling. It may seem easier to avoid omissions if the examiner always performs the examination in the same sequence, but a rigid approach that fails to adjust to the state and activity of the individual infant may yield a suboptimal examination that is inefficient and unnecessarily stressful for infant, parent, and examiner. The physical examination of the newborn includes measurements of the weight, length, and head circumference, which must be compared with standardized growth data (Figure 25-1) to determine whether the infant is small (<10th percentile), appropriate, or large (>90th percentile) for gestational age (Battaglia and Lubchenco, 1967). The body temperature, heart rate, and respiratory rate are typically recorded at regular intervals by nurses, but the pediatrician should also consciously evaluate the heart and respiratory rate at the time of examination. The blood pressure is not routinely measured in healthy newborns, but the blood pressure should be checked in all four extremities if the history or examination suggest a problem with the circulation. Of the standard physical examination techniques, the most important in the examination of the newborn is observation. Palpation and auscultation are also important, whereas percussion is of relatively limited use. The routine examination also includes specific physical maneuvers for examining the hips and for eliciting a variety of reflex responses. A stethoscope, an ophthalmoscope, and a tape measure are the only pieces of equipment generally needed. A source of light for transillumination is helpful for specific purposes. Pulse oximetry can enhance early detection of critical congenital heart disease, but further study is needed to determine whether this procedure should become a standard of care in the routine assessment of the neonate (Mahle et al, 2009). Sensitivity to the infant’s state and responses can greatly facilitate the examination and make it less stressful for the infant and less time-consuming for the examiner. The newborn infant generally responds more slowly to a stimulus than does an older child or adult, and the response 5.5 50 4 97 90 55 50 10 3 50 45 10 3.5 3 3 97 90 50 10 3 40 2.5 Head circ (cm) Birthweight (kg) 4.5 Length (cm) 90 5 tends not to remain localized to the area of the stimulus. For example, a gentle touch to the face or extremity of a sleeping infant, or merely loosening a blanket, typically results in some limited initial movement of the area touched, followed with a slight delay by a wave of movement that spreads to involve the whole body, and which may be accompanied by partial arousal. If the examiner waits a few seconds for this wave of response to fully subside before proceeding to move or touch the infant again, the infant will often settle and remain asleep. With a series of such gentle, well-spaced interactions, the examiner can often complete important portions of the examination without waking the infant. If, however, the examiner provides a new stimulus soon after the first, before the secondary movements have finished spreading and while the infant is partially aroused, the new stimulus is likely to reinforce the initial stimulation and provoke full awakening and a crying. The examination should begin with a deliberate pause to observe the infant in an undisturbed state. Before the infant is moved or undressed, the examiner should actively observe the infant and mentally note as much information as possible, including the state of alertness or sleep, color of visible portions of the skin and mucous membranes, respiratory rate and any audible sounds or signs of increased work of breathing, posture, spontaneous activity, and quality of cry. Throughout the examination, the examiner should alternate between focusing attention on specific elements of the examination (e.g., listen for a heart murmur, closely scrutinize a scalp lesion) and maintaining a global awareness of the infant’s overall responses and activity. The experienced examiner’s general impression of whether the patient is sick or healthy is an important part of the evaluation. If the infant is sleeping or in a quiet alert state, the examiner will usually begin by gently uncovering the chest. Depending on the type of clothing the infant is wearing, it may be best at first to lift the shirt or gown just enough to slide the stethoscope underneath. Because the heart sounds and a soft murmur, if present, can be obscured easily by 60 97 2 1.5 0 32 34 36 38 Weeks 40 42 44 281 35 30 25 32 34 36 38 Weeks 40 42 44 FIGURE 25-1 Fetal-infant growth chart for 32 to 44 weeks’ gestation. ( Redrawn from Fenton TR: A new growth chart for preterm babies: Babson and Benda’s chart updated with recent data and a new format. BMC Pediatr 3:13, 2003. The complete chart for 22 to 50 weeks’ gestation can be downloaded from http://members.shaw.ca/ growthchart.) 282 PART VII Care of the Healthy Newborn crying, the precordial area is usually auscultated first. If the infant remains quiet, the clothing can be removed further, and the stethoscope can be moved to other locations on the chest to expand the area of auscultation for breath and heart sounds, and then to the abdomen to listen for bowel sounds. If the infant is already crying or awakens and starts crying vigorously, auscultation can be deferred in favor of other parts of the examination. Auscultation of the chest can be performed later, after the infant has become quiet. If the infant is lightly dressed and is being examined on an open warming bed, all clothing can be removed after the initial auscultation of the chest and abdomen; this allows the entire body to be easily observed throughout the examination with a systematic head-to-toe approach. If the infant is in an open crib, he or she may become cold and upset if undressed fully at the beginning of the examination. In this environment, it is preferable to allow the infant to remain at least partially clothed for as long as possible. For example, start by examining the head, and after replacing the infant’s hat, open the shirt to finish examining the anterior chest and abdomen, closing but not buttoning or snapping the shirt while moving to examine the hips, genitalia, and lower extremities, and finally removing the shirt to examine the upper extremities and back. However, proceeding in this way requires extra attention on the part of the examiner to ensure that no parts of the examination are missed. An experienced examiner can accomplish a thorough examination of the newborn more quickly than a novice, largely because of the acquired skill in handling the infant and minimizing wasted or redundant motions. Although there is no substitute for experience, the trainee’s skill will develop more rapidly if close attention is paid to developing a fluid approach to the examination as a whole. The following paragraphs provide one example; each experienced examiner will develop his or her own preferred sequence. Further details of specific parts of the examination are described in subsequent sections. After the initial moment of observation and auscultation of the anterior chest and abdomen, begin examining the head. Using both hands, gently encircle and palpate the entire scalp while inspecting parts of the scalp that are visible in this position (the back of the head and neck will be inspected later). Gently turn the head to one side (noting any limitation of range of motion), and inspect that side of the head, including the ear, and then the side of the neck. Roll the infant’s head gently while working the fingers around the neck to the other side, palpating the neck and clavicular areas and applying gentle traction on the skin as needed to open the neck creases to allow full visualization. After reaching the other side of the neck, inspect the ear and side of the head. Next, proceed to the face, attending first to overall features of shape and symmetry and then focusing sequentially on the skin, eyes, nose, mouth, and oral cavity. With the shirt open in front, inspect and palpate the anterior chest. Observe the respiratory pattern, noting the presence or absence of retractions. Additional auscultation of the chest may be done at this time, if not completed earlier. Open the diaper, and inspect and palpate the abdomen and umbilicus. Palpate the femoral pulses. One may examine the genitalia and perineum at this time, or wait until after the examination of the hips and lower extremities, described next. Inspect the lower extremities. Check the range of motion of the hips and perform the Barlow and Ortolani maneuvers. Palpate the legs, sliding the examiner’s hands from the hips down to the ankles. The examiner should place the thumbs on the soles of the feet, with the fingers around the back of the ankles. From this position, the examiner can smoothly check the alignment of the feet, elicit the plantar grasp, and then slide the thumbs up along the outer edge of the soles to elicit Babinski’s reflex. Next, lift and abduct the legs into a frog-leg position to provide a full view of the perineum and anus. Inspect the genitalia by gently retracting the labia majora in females, or depressing the skin at the base of the penis in males. Inspect and palpate the scrotum and testes. After this, the diaper can be refastened. Attention next turns to the upper extremities. The shirt should be removed completely to allow unobstructed observation of the overall shape, symmetry, and movements of the arms and hands. With the infant supine, turn the head to elicit the asymmetric tonic neck reflex on each side. Palpate the whole arm gently, starting with one of the examiner’s hands on each of the baby’s shoulder’s, and then slide down to the baby’s hands, noting any swelling or discontinuities. Inspect the hands, fingers, nails, and palms. If the infant’s hand is tightly fisted, do not attempt to pry the fingers open. Instead, gently flex the wrist to 90 degrees, which will cause the fingers to relax naturally. Inspect the palms, and then elicit the palmar grasp reflex. Without releasing the baby’s hands, one can then perform the pull-to-sit maneuver. The examiner places his or her hand behind the infant’s head and neck to provide support as the infant is gently lowered back toward the bed. When the infant’s head and shoulders are a few inches from the bed, the examiner drops his or her hand rapidly to elicit the Moro reflex. Next, place the hands on either side of the chest, under the arms at the shoulders, and raise the baby to an upright position, noting the strength and tone of the shoulder muscles. Lower the infant, still in an upright position, and try to elicit the supporting and stepping reflexes. Turn the infant to a prone position, suspended on the examiner’s hand. Observe the infant’s posture and tone, and elicit the incurvation response. Inspect the infant’s back, from the vertex of the head down to the sacrum (pulling the diaper down, if needed). Gently place the infant back in the crib, and dress the infant. The red reflex examination can be performed at this time, or at any suitable moment when the eyes are open spontaneously. The preceding description assumes routine findings throughout. The examiner should always be prepared to deviate from his or her preferred sequence in response to changes in the infant’s state or if abnormalities are found that require a more extended examination of particular features. However, it remains important for the examiner to complete the entire examination and not become distracted by a prominent finding. On occasion, portions of an examination need to be deferred; this is more common in the NICU, because of patient instability, than in the well-baby nursery. In either environment, any limitations of the examination should be clearly documented, with the need for reexamination listed explicitly in the plan of care. CHAPTER 25 Initial Evaluation: History and Physical Examination of the Newborn SKIN The entire skin surface should be inspected during the course of the examination, with attention to the color, moisture, temperature, texture, and elasticity of the skin, the pattern and depth of skin creases, and the presence and character of any local alterations or lesions. The skin appendages—nails and hair—are inspected along with the skin, with attention to the color, size, shape, and any lesions of the nails, and to the growth pattern, color, texture, and distribution of scalp and body hair. Lighting must be adequate and consistent. Natural light is best, although not always available. Phototherapy lights must be turned off during the examination. Typically the skin in different regions of the body is examined sequentially during the course of the routine examination, but if abnormalities are seen, it can be helpful to reexamine the infant in a fully undressed state and in a warm environment, so that all parts of the body can be directly compared. The size, shape, location, distribution, and time course of lesions are all important for differential diagnosis. Skin color is a composite of the infant’s basic skin pigmentation, the adequacy of perfusion, the amounts of oxygenated and deoxygenated hemoglobin in the local circulation, and, at times, internal staining of skin by extravasated blood or pigmented molecules such as bilirubin. The skin can also be stained externally by in utero exposure to meconium or postnatally by foreign substances applied to the skin. The newborn is typically much more pink or red than the mother, because of the healthy infant’s high hemoglobin level. With excessive hemoglobin levels, as in polycythemia or twin-twin transfusion, a deeply red or purple-red color (plethora) is evident. Pallor may be caused by anemia or poor perfusion. Central oxygenation is usually best evaluated by observing the color of the tongue and oral mucous membranes, because the color of lips can be misleading. Visual detection of central cyanosis requires approximately 5 g of desaturated Hb per 100 mL of blood, and may not be apparent in the presence of significant anemia (Snider, 1990). Cyanosis of the perioral area, hands, and feet (acrocyanosis), which is common in the first 24 hours after birth or if the infant is cold, is due to relatively poor perfusion of those areas, resulting in increased O2 extraction by the tissues and an increase in the concentration of deoxyhemoglobin. Because skin color is dependent on perfusion, it can be a sensitive indicator of systemic perfusion. However, the infant’s peripheral color can vary markedly and rapidly with activity and the local environmental temperature. Transient mottling of skin of the extremities and trunk (cutis marmorata) is common in newborn infants in response to cold. If mottling does not resolve with warming, other conditions should be considered, including hypertension and hypothyroidism. Harlequin color change is a striking but infrequently observed transient asymmetry of color and perfusion, in which one side of the body is vasodilated, with a clear line of demarcation at the midline. The yellow-orange color of jaundice caused by unconjugated bilirubin becomes obvious if serum bilirubin levels are sufficiently high, but it can be difficult to detect at more moderate levels. Typically, jaundice first becomes apparent in the face and then progresses distally as bilirubin levels continue to rise. The detection of mild-to-moderate jaundice is easier if the infant’s foot is lifted so that the face and 283 foot can be seen side by side in the same visual field, with the foot providing a built-in control for the infant’s basic skin pigmentation. This is particularly helpful in the black or brown infant, in whom mild jaundice is more difficult to detect than in the white infant, and in the Asian infant whose skin tones can exaggerate the appearance of mild jaundice. A transient period of generalized erythema may be noted a few hours after birth, usually by the parents, often during a bath or with vigorous crying, resolving within minutes to an hour. Sometimes called erythema neonatorum, it appears to be associated with successful completion of neonatal transition of the circulation, and it rarely recurs with the same intensity after the first episode (Fletcher, 1998). Skin creasing is affected by both development and movement. Creasing increases if movement is constrained, as in the fetus near term. An absence of creasing on the palms or soles of a term or near-term neonate can be caused by a prolonged lack of movement secondary to a neuromuscular disorder. Edema stretches the skin and obscures normal skin folds and creases. Edematous skin pits with gentle pressure. The distribution of edema is affected by gravity and thus changes with position. In contrast, lymphedema (as in Turner syndrome) tends to accentuate the creases, and is less affected by gravity (Fletcher, 1998). Injuries caused by iatrogenic trauma in utero or at delivery are common; they include scars from injury by an amniocentesis needle, minor lacerations and abscesses from fetal scalp electrodes or fetal blood gas sampling, scalpel lacerations during cesarean section, and ecchymoses and abrasions from forceps or vacuum. Trauma can also occur from pressure on the fetus during labor, particularly over bony prominences. Such pressure may be involved in the development of subcutaneous fat necrosis, an uncommon condition hypothesized to be caused by hypoxic injury to fat and manifested by firm, subcutaneous nodules and plaques (Cohen, 2008). The lesions usually resolve spontaneously after several months, but can become inflamed and fluctuant and are occasionally associated with the late development of significant hypercalcemia. Aplasia cutis congenita is a focal lesion characterized by congenital absence of some or all layers of skin (Figure 25-2) and can FIGURE 25-2 Aplasia cutis congenita. Lesions are usually single, but in this case there are two adjacent lesions near the vertex—one bullous and one membranous. (Reprinted from Rudolph AJ: Atlas of the newborn, vol 4, Hamilton, Ont, Canada, 1997, BC Decker, p 30.) 284 PART VII Care of the Healthy Newborn BOX 25-1 C  ommon and Uncommon Causes of Skin Lesions in Neonates PUSTULAR, VESICULOPUSTULAR, AND VESICULOBULLOUS LESIONS ll Common or benign: erythema toxicum neonatorum, transient neonatal pustular melanosis, miliaria crystallina, miliaria rubra, sucking blisters, neonatal acne (benign cephalic pustulosis) ll Infectious: herpes simplex, varicella, staphylococcal pustulosis, bullous impetigo, congenital candidiasis, syphilis, scabies ll Chronic or recurrent: epidermolysis bullosa, mastocytosis, epidermolytic hyperkeratosis, acropustulosis of infancy ll Positive Nikolsky’s sign: epidermolysis bullosa, staphylococcal scalded skin syndrome ll Other: incontinentia pigmenti NODULES AND PLAQUES Common or benign: milia, Ebstein’s pearls, Bohn’s nodules, sebaceous hyperplasia ll Yellow: sebaceous nevus, juvenile xanthogranuloma ll Brown or black: congenital pigmented nevus, epidermal nevus ll Other: subcutaneous fat necrosis, dermoid cyst, fibroma, infantile myofibromatosis, hamartomas, malignant tumors, leukemia ll PAPULOSQUAMOUS AND SCALING LESIONS Common or benign: physiologic desquamation ll Healthy infant: atopic dermatitis, contact dermatitis, seborrheic dermatitis, local candida dermatitis, psoriasis ll Ill infant: acrodermatitis enteropathica, Langerhans cell histiocytosis, syphilis ll Other: ichthyosis syndromes, collodion baby, harlequin baby ll EROSIONS AND ULCERATIONS Common or benign: sucking blisters, traumatic injury (e.g., scalp electrode, diaper erosions, reaction to adhesives) ll Other: aplasia cutis congenita, herpes simplex, epidermolysis bullosa, toxic epidermal necrolysis ll ALTERED PIGMENTATION Common or benign: Mongolian spots, transient neonatal pustular ­melanosis, isolated café au lait macules ll Increased pigmentation: Mongolian spots, transient neonatal pustular melanosis, café au lait macules, lentigines, incontinentia pigmenti ll Decreased pigmentation: ash leaf macule, nevus depigmentosus, ­piebaldism, albinism ll Purpuric or erythematous: petechiae, dermal hematopoiesis (“blueberry muffin” lesions), neonatal lupus erythematosus ll VASCULAR AND LYMPHATIC LESIONS Common or benign: Nevus simplex or salmon patch, petechiae on the presenting part, small hemangioma ll Other vascular: complicated hemangioma, vascular malformation, port-wine stain ll Lymphatic: cystic hygroma, lymphangioma, lymphedema ll Data from Cohen B: Pediatric dermatology, ed 3, Philadelphia, 2005, Mosby; Eichenfield LF, Frieden IJ, Esterly NB, editors: Neonatal dermatology, (See also Color Plate 9.) ed 2, Philadelphia, 2008, Elsevier Saunders; and Fletcher MA: Physical diagnosis in neonatology, Philadelphia, 1998, Lippincott-Raven. occur sporadically or in association with chromosomal defects or other malformations (Kos and Drolet, 2008). It occurs most often on the scalp near the vertex and can be mistaken for a traumatic injury. Some common and uncommon skin conditions that may be seen in the newborn are outlined in Box 25-1. Fortunately, most of the skin findings encountered during the routine examination are due to a relatively few common FIGURE 25-3 (See also Color Plate 9.) Erythema toxicum neonatorum with erythematous macules, wheals, and pustules. Pustules predominate in this example. At times, patchy or confluent areas of erythema occur without pustules. (Reprinted from Eichenfield LF, Frieden IJ, Esterly NB, editors: Neonatal dermatology, ed 2, Philadelphia, 2008, Elsevier Saunders, p 88.) conditions, mostly benign, that are described in the following paragraphs. If unusual or unfamiliar lesions that fall outside the examiner’s “comfort zone” are observed, consultation with an appropriate subspecialist should be arranged. Erythema toxicum neonatorum, or erythema toxicum, is the most common skin rash in the newborn, occurring in up to 70% of term infants (Howard and Frieden, 2008; Lucky, 2008). Lesions can be present at birth, but in most cases the lesions are first noted at 1 to 2 days. The lesions are seen predominately on the face and trunk, but can appear anywhere on the body except the palms and soles. New lesions may continue to appear for approximately 1 week while older lesions resolve. The typical lesion is an isolated elevated erythematous papule or pustule 1 to 2 mm in diameter, surrounded by an irregular area of erythema 1 to 3 cm in diameter (Figure 25-3). When extensive, the lesions can occur in clusters or become nearly confluent. Diagnosis can usually be made by appearance alone, but a scraping of the pustule will reveal an almost pure infiltrate of eosinophils if confirmation is needed in atypical cases. Transient neonatal pustular melanosis is a self-limited process with lesions that evolve through three distinct phases (Howard and Frieden, 2008; Lucky, 2008; Figure 25-4). Initially, a superficial vesicopustule appears, which after rupturing leaves a fine collarette of scale around the unroofed pustule, which is without erythema. The final stage is a hyperpigmented macule that gradually disappears. The condition is most common in African American infants. In some cases, only the second or third stages are seen at birth, the initial stage presumably having occurred in utero. Miliaria crystallina is the result of superficial obstruction of the sweat ducts, producing small, crystal-clear vesicles that resemble water droplets (Figure 25-5). It is mostly seen in warm climates or febrile infants. The vesicles are fragile and can be removed by wiping the skin with a soft damp cloth (Howard and Frieden, 2008; Lucky, 2008). Miliaria rubra, also called heat rash or prickly heat, results from sweat duct obstruction deeper in the epidermal layer, and is more CHAPTER 25 Initial Evaluation: History and Physical Examination of the Newborn A C 285 B D FIGURE 25-5 (See also Color Plate 10.) Miliaria crystallina. The tiny, clear vesicles resemble water droplets, with no signs of inflammation. (Reprinted from Rudolph AJ: Atlas of the newborn, vol 4, Hamilton, Ont, Canada, 1997, BC Decker, p13.) common after the first week. Occasionally miliaria rubra will look sufficiently pustular as to mimic lesions caused by staphylococcal, candidal, or herpes simplex. Sucking blisters and calluses result from vigorous sucking on a hand or forearm in utero, and they can cause a tense, fluid-filled blister, which when ruptured forms an erosion or a callus. The lesion is most often solitary, but may be bilateral, and is without inflammation. A sucking pad or callus may develop on the lips postnatally as a result of vigorous and frequent nursing. Neonatal cephalic pustulosis, or neonatal acne may occasionally be seen at birth, but more typically appears later, with a mean age of onset of 2 to 3 weeks (Howard and Frieden, 2008; Lucky, 2008). It is characterized by inflammatory, erythematous papules and pustules located primarily on the cheeks with extension over the face and FIGURE 25-4 Transient neonatal pustular melanosis. The first stage consists of small superficial pustules, without inflammation (A). Collarettes of scale, the second stage, may be seen at birth without evident pustules (B), or may develop postnatally after pustules have ruptured (C). Small hyperpigmented macules remain in the final stage (D), gradually fading over weeks or months. (Reprinted from Eichenfield LF, Frieden IJ, Esterly NB, editors: N ­ eonatal dermatology, ed 2, Philadelphia, 2008, WB Saunders, p 89.) FIGURE 25-6 (See also Color Plate 11.) Neonatal cephalic pustu­ losis (neonatal acne). Small, red papules and pustules are seen on the cheeks and forehead, with some extension into the scalp. Comedones are absent. (Reprinted from ­Eichenfield LF, Frieden IJ, Esterly NB, editors: Neonatal dermatology, ed 2, Philadelphia, 2008, WB Saunders, p 90.) into the scalp (Figure 25-6). It can be difficult to distinguish clinically from miliaria rubra, but both are benign conditions and biopsy is not warranted. Infectious causes of skin lesions are relatively uncommon in the immediate newborn period, but often need to be considered in the differential diagnosis. Superficial skin infections of Staphylococcus aureus are rarely present at birth, but can develop within the first few days. Thrush and candidal diaper dermatitis, common later but not immediately after birth, usually pose no diagnostic difficulty. Congenital candidiasis, which is uncommon, typically manifests within the first day with a papulovesicular eruption that progresses to pustules, followed by crusting and desquamation (Figure 25-7). Lesions can be widespread and may appear on any part of the body, including the palms and soles (Carder, 286 PART VII Care of the Healthy Newborn FIGURE 25-7 (See also Color Plate 12.) Congenital ­candidiasis. The rash may be a diffuse, erythematous pustular eruption (A) or have diffusely distributed but distinct pustules (B). In premature infants, a diffuse scaldlike erythematous dermatitis may be seen (not shown). (Reprinted from Eichenfield LF, Frieden IJ, Esterly NB, ­editors: ­Neonatal dermatology, ed 2, Philadelphia, 2008, Elsevier Saunders, p 214.) A B A A B FIGURE 25-9 (See also Color Plate 14.) Desquamation on the palms (A) and soles (B) of an infant with congenital syphilis. (Reprinted from Rudolph AJ: Atlas of the newborn, vol 4, Hamilton, Ont, Canada, 1997, BC Decker, p 108.) B FIGURE 25-8 (See also Color Plate 13.) Herpes simplex. A, The first signs of herpes infection in this neonate were eroded vesicles at the corner of the mouth. B, Herpetic vesicles on the face, scalp, and ear of an infant with ­respiratory distress and hepatitis. (Reprinted from Cohen B: Pediatric dermatology, ed 3, Philadelphia, 2005, Mosby, p 36.) 2008; Darmstadt et al, 2000). Neonatal herpes simplex is unlikely in the first few days after birth, unless the fetal membranes were ruptured for many days, but it must be considered whenever vesicles are seen in the newborn. The skin lesions of herpes simplex start as small, 2- to 4-mm vesicles on an erythematous base, become pustular after 1 to 3 days, and develop an eschar (Friedlander and Bradley, 2008; Figure 25-8). The mucosal lesions are usually shallow ulcerations. The “blueberry muffin” skin lesions of congenital rubella and cytomegalovirus, caused by dermal hematopoiesis, are unlikely to be seen as an isolated finding. Affected infants typically have multiple stigmata of congenital infection including growth retardation, microcephaly, and hepatosplenomegaly. Cutaneous findings of congenital syphilis are present in only a minority of infected infants, but they classically involve the palms, soles, and perioral and anogenital areas (Dinulos and Pace, 2008; Howard and Frieden, 2008). The rash is highly variable, taking papulosquamous, vesiculobullous, macular erythematous, annular, and polymorphous forms. Desquamation limited to the palms and soles, with no rash or peeling elsewhere, is suggestive of congenital syphilis (Figure 25-9). Milia, Ebstein’s pearls, and Bohn’s nodules are epidermal inclusion cysts (Lucky, 2008; Figure 25-10). Milia are tiny epidermal inclusion cysts seen primarily on the face and scalp in small numbers; they are smooth, firm, white papules with no associated erythema. They may be present at birth or appear somewhat later, and they generally resolve spontaneously within a few months. Larger inclusion cysts, which usually occur singly, are called pearls; the foreskin and the ventral surface of the penis and scrotum are common locations. Epidermal inclusion cysts located on the palate are called Ebstein’s pearls, and those on the alveolar ridge are called Bohn’s nodules. Sebaceous hyperplasia occurs mostly on the face, especially on the nose and upper lip. It is due to hypertrophy of the sebaceous glands caused by androgenic hormonal stimulation in utero, and it gradually resolves over several weeks. It is characterized by sheets of smooth, yellowwhite papules with the regular spacing of involved follicles and no surrounding erythema (Figure 25-11). Physiologic desquamation occurs in many full-term infants. These infants will have some dry skin with fine CHAPTER 25 Initial Evaluation: History and Physical Examination of the Newborn 287 A FIGURE 25-11 (Supplemental color version of this figure is available online at www.expertconsult.com.) Sebaceous hyperplasia. Sheets of tiny, white-yellow follicular papules, without inflammation, are seen on the nose. (Reprinted from Eichenfield LF, Frieden IJ, Esterly NB, editors: ­Neonatal dermatology, ed 2, Philadelphia, 2008, Elsevier Saunders, p 87.) B C FIGURE 25-10 Epidermal inclusion cysts. Milia are most commonly seen on the face (A), but can occur anywhere on the body. Ebstein’s pearls occur on the midline of the hard palate, most commonly near the junction with soft palate (B). Bohn’s nodules are found along the gum margins and on the lateral palate (C). Dental lamina cysts (not shown) are similar inclusions located on the crest of the alveolar ridge. (Reprinted from Fletcher MA: Physical diagnosis in neonatology, Philadelphia, 1998, Lippincott-Raven, p124 [A]; and Eichenfield LF, Frieden IJ, Esterly NB, editors: Neonatal dermatology, ed 2, Philadelphia, 2008, Elsevier Saunders, pp 503-504 [B and C].) desquamation at 1 to 2 days of age, particularly on the hands and feet. A more exaggerated variety of physiological desquamation is common in postmature infant (born at greater than 41 6⁄7 weeks post-menstrualage). These infants are often born with dry, thickened skin that cracks and peels extensively and then normalizes spontaneously over the course of approximately 1 week. The condition is usually not difficult to distinguish from the rare inherited disorders of cornification (ichthyosis) that manifest in the neonatal period (Irvine and Paller, 2008). These disorders include Harlequin ichthyosis (caused by mutations in the ABCA12 gene) and “collodion babies” who appear encased in a thickened, shiny skin that resembles collodion, a phenotype associated with a number of different ichthyotic conditions. Mongolian spots (dermal melanocytosis) are macular areas of a slate grey, blue-grey, blue-black, or deep brown color. The distinctive appearance is due to the presence of melanocytes located in the dermis, instead of their typical site at the dermal-epidermal junction. The spots are most commonly located on the lower back and buttocks, but can occur elsewhere. Mongolian spots in this location—which are common in East Asian, East African, Native American, and Polynesian infants—tend to fade over several years, whereas similarly appearing lesions that occur elsewhere on the body may never resolve (Gibbs and Makkar, 2008; Lucky, 2008). Dermal melanocytosis in the area of the first and second divisions of the trigeminal nerve is called the nevus of Ota. The nevus of Ito is a similar lesion occurring on the neck, upper back, and shoulders in the area of the posterior supraclavicular and lateral brachial cutaneous nerves. Unlike Mongolian spots, these nevi do not become less pigmented with time, and malignant melanoma or malignant nevi can develop within them, though rarely. Café au lait macules are tan-brown macules that can occur anywhere on the body. They are common as an isolated finding in the newborn and are generally benign, but can be markers of other conditions (Gibbs and Makkar, 2008). The presence of six or more café au lait macules greater than 5 mm in diameter is considered presumptive evidence of neurofibromatosis type 1. Multiple large café au lait macules with irregular borders may be a manifestation of the McCune-Albright syndrome. The salmon patch (or nevus simplex) is a benign vascular lesion consisting of erythematous macules or patches frequently seen at the nape of the neck (“stork bite”), on the eyelids and glabella (“angel’s kisses”), and somewhat less frequently on the nose and upper lip (Lucky, 2008; Figure 25-12). The lesions are caused by dilated capillaries in the upper dermis, with normal overlying skin. Those on the face usually fade or resolve completely within 1 to 2 years, but 25% to 50% of those on the neck persist throughout life (Lucky, 2008). After vaginal delivery of a healthy infant, it is not unusual or concerning to see petechiae on the presenting part (i.e. the part of the fetus closest to the pelvic inlet of the birth canal at the onset of labor) or on other areas of the body subjected to localized pressure during delivery. A tight nuchal cord can cause extensive ecchymoses and petechiae on the entire head. Widespread petechiae are abnormal, however, and are suggestive of thrombocytopenia or platelet dysfunction. Hemangiomas are soft, pink-red, compressible vascular tumors composed of proliferating endothelial cells. Small 288 PART VII Care of the Healthy Newborn A FIGURE 25-13 This infant with the Sturge-Weber syndrome has a port-wine stain in the distribution of the ophthalmic division of the facial nerve. (Reprinted from Cohen B: Pediatric dermatology, ed 3, ­Philadelphia, 2005, Mosby, p 49.) B FIGURE 25-12 (See also Color Plate 15.) Salmon patches are commonly seen on the glabella, eyelids, nose, or upper lip, either singly or in all these locations (A), and on the nape of the neck (B). (Reprinted from Eichenfield LF, Frieden IJ, Esterly NB, editors: Neonatal dermatology, ed 2, Philadelphia, 2008, Elsevier Saunders, p. 95.) hemangiomas are recognized at or shortly after birth in 1% to 3% of healthy term infants and become apparent in 10% of all infants by 1 month old (Cohen, 2005). Hemangiomas are more common in female infants by 2:1 to 9:1 as compared with males. The incidence is much higher (22% to 30%) in preterm infants weighing less than 1000 g, but only slightly higher (15%) in those weighing 1000 to 1500 g (Amir et al, 1986; Enjolras and Garzon, 2008). Infantile hemangiomas typically undergo a period of growth beyond that of surrounding tissues for 6 to 12 months, typically followed by spontaneous involution. Approximately 25% regress by 2 years of age, 40% to 50% regress by 4 years of age, 60% to 75% regress by 6 years of age, and 95% regress by adolescence (Cohen, 2005). Before the hemangioma becomes obvious, careful examination may reveal a precursor lesion manifesting as telangiectasias surrounded by an area of pallor or as pale, erythematous, or bruiselike macules and patches. In contrast, vascular malformations are nonproliferative lesions, usually present at birth. The Kasabach-Merritt phenomenon occurs in association with specific types of large vascular tumors that cause platelet trapping and severe thrombocytopenia, and is not associated with the true hemangioma of infancy (Enjolras and Garzon, 2008). In the rare Klippel-Trenaunay syndrome, capillary malformation and varicose veins are associated with overgrowth of an affected limb. Port-wine stains are capillary malformations evident at birth as pink or red patches that grow proportionately with the child and persist throughout life (Enjolras and Garzon, 2008). The initial pink-red color typically changes to a deeper red or purple hue with age. Approximately 10% of port-wine stains that involve the area supplied by the ophthalmic (V1) branch of the trigeminal nerve (Figure 25-13) are associated with seizures, arterial brain malformations, and ocular abnormalities that constitute the Sturge-Weber syndrome (Enjolras et al, 1985). HEAD The head is mainly examined by inspection and palpation. The overall size, shape, and features of the skull should be noted. The head circumference (maximal occipital-frontal circumference) should be measured in every neonate and plotted on an appropriate growth chart (see Figure 25-1). The measurement of the head circumference is subject to error and can be significantly affected by molding of the skull during labor, so the head circumference should be measured again if the result appears discordant with the visual examination or with the infant’s weight and length, and repeated after molding has resolved. The soft tissue of the scalp should be examined for swelling, ecchymoses, and other evidence of injury because of the forces of labor, and for iatrogenic injuries including those from application of a vacuum device, 289 CHAPTER 25 Initial Evaluation: History and Physical Examination of the Newborn placement of scalp electrodes, fetal blood sampling, and lacerations with a scalpel during cesarean section. The quantity of scalp hair present in the newborn is highly variable, but abnormalities in the distribution, texture, and patterning of the hair are potentially informative. The hair usually forms a single whorl near the vertex, but double whorls occur in approximately 5% of newborns. Abnormal placement or the presence of more than two whorls may be a marker of abnormal brain development (Smith and Gong, 1974). The infant skull is composed of several bony plates, separated by sutures and fontanels. This structure allows the skull to deform during labor (Figure 25-14). The entire surface of the skull should be palpated to identify the location and size of the major fontanels and assess for discontinuities. The soft tissue over a fontanel should normally be flat; a raised or bulging fontanel suggests that intracranial pressure is increased. It is common to find a palpable discontinuity at a suture, because of vertical displacement of a skull bone relative to its neighbor as a result of molding. In some cases, one bone can truly override the other. Such discontinuities at a suture must be distinguished from the step-off of a displaced skull fracture. Premature fusion of one or more of the cranial sutures, or craniosynostosis, results in a variety of abnormal skull shapes, depending on which sutures are involved (Fletcher, 1998; Volpe, 2008). Fusion of the saggital suture produces a narrow skull elongated in the anterior-posterior dimension (scaphocephaly or dolichocephaly). Fusion of the coronal sutures causes a widened skull that is shortened in the anterior-posterior dimension (brachycephaly). Unilateral closure of either a coronal or lambdoid suture causes an oblique deformity (frontal or occipital plagiocephaly, respectively). Closure of a metopic suture produces a triangular skull with a prominent, narrow forehead (trigonocephaly). Depending on the timing of the fusion, the skull shape may be abnormal at birth or become visibly deformed later. The fused suture typically has a palpable elevation or ridge, which must be distinguished from displacement of a normal suture caused by molding. In the absence of craniosynostosis or overriding, normal mobility at a suture can be verified by gently applying alternating pressure to the bones on either side of the suture line. Caput succedaneum is a diffuse edematous swelling of the scalp caused by pressure during delivery that results in fluid accumulation external to the periosteum. The caput is boggy, has diffuse edges, is not limited by suture lines, A B and most commonly is located over the vertex. It is usually present at birth and resolves over several days. In contrast, a cephalohematoma is caused by hemorrhage under the periosteum; it forms a distinct, firm bump laterally that does not cross suture lines ­(Figure 25-15). The cephalohematoma may not appear until several hours after birth, and it often increases in size over the first 12 to 24 hours. It typically remains palpable for 2 to 3 weeks, during which time it may develop a calcified rim. Unlikely a cephalohematoma, a subgaleal hematoma is not confined by the coronal suture posterior fontanel (lambda) lambdoid suture mastoid fontanel sphenoidal fontanel A posterior fontanel parietal tuberosity sagittal suture anterior fontanel (bregma) coronal suture B metopic suture FIGURE 25-14 The major sutures and fontanels of the newborn skull. (Reprinted from Fletcher MA: Physical diagnosis in neonatology, Philadelphia, 1998, Lippincott-Raven, p 175.) FIGURE 25-15 (See also Color Plate 16.) Anterior (A) and ­posterior (B) views of a large cephalohematoma under the periosteum of the right parietal bone. (Reprinted from Fletcher MA: Physical diagnosis in neonatology, Philadelphia, 1998, Lippincott-Raven, p 185.) 290 PART VII Care of the Healthy Newborn periosteum and can involve massive blood loss. Depending on the volume of blood that has accumulated, it can be palpated as a firm or fluctuant mass with poorly defined edges that may extend onto the neck or forehead. Large subgaleal hematomas are uncommon, but potentially life threatening. Craniotabes (or craniomalacia) is a softening of the skull, most commonly involving the parietal bones near the vertex. Gentle pressure on the involved bone produces a sudden collapse, with recoil when the pressure is released, similar to the way a ping-pong ball collapses when squeezed. It has usually been attributed to localized bone resorption or interference with ossification caused by prolonged pressure on the fetal skull from the maternal pelvis, but it may be associated with maternal vitamin D deficiency in some populations (Yorifuji et al, 2008). A localized defect in the skull can occur in association with aplasia cutis congenita. A small soft-tissue mass or bulge at the occiput or in the midline of the forehead near the bridge of the nose may be caused by an encephalocele protruding through a small defect in the skull. Large encephaloceles and major neural tube defects will be obvious if present. Hydranencephaly, if suspected, can be detected by transillumination of the skull. FACE The face and facial expressions should be observed throughout the course of the entire examination. In addition, there should be a moment when attention is directed to the overall arrangement and proportion of the features of the face as a whole. It is helpful to view the face from the front, in profile, and looking down from the top of the head. Important characteristics to note include the symmetry and expressiveness of the face, both at rest and during crying, and the relative sizes, spacing, and orientation of the features. After this overview, specific attention should be paid to the ears, eyes, eyebrows, eyelids, nose and nasal passages, lips, palate, and mouth. Suggested abnormalities should be described as precisely as possible. Measurements of the facial features are not part of the routine newborn examination, but comparison with reference nomograms can be helpful if dysmorphism is suspected (see Chapter 19). EARS The external ear is a relatively common site of minor anomalies. The external ear develops from the first and second branchial arches, from which the six hillocks of His grow, migrate, and fuse to form the pinna. Most of the pinna derives from the second branchial arch. The third arch contributes the tragus, antitragus, and anterior wall of the canal (Figure 25-16). The position and orientation of each ear should be noted, as well as the size, shape, and structure of the helix, and any signs of trauma. The ear is palpated to assess the firmness and recoil of the cartilage and to detect any masses or abnormalities of texture. The skin near the pinna is examined for skin tags, pits, and sinus tracts. Minor variations in the shape of the ear and ear lobe are common, and they usually have only cosmetic significance. A lop ear is characterized by downward folding of the superior helix caused by underdevelopment of Helix Triangular fossa Antihelix Tragus External auditory meatus Concha Lobule Antitragus FIGURE 25-16 Anatomy of the external ear. (Reprinted from Fletcher MA: Physical diagnosis in neonatology, Philadelphia, 1998, Lippincott-Raven, p 285.) the upper third of the pinna; this needs to be distinguished from folding caused by positioning in utero. A cup ear is a protruding ear with an excessively concave concha. Microtia implies a severely dysplastic and malformed ear, and it is often associated with abnormalities of the middle ear and other malformations. Assessment of hearing by observation of behavioral responses has been essentially replaced in the routine examination by automated testing of hearing before hospital discharge (American Academy of Pediatrics, 2007; US Preventive Services Task Force, 2008). Otoscopy to inspect the eardrum is not part of the routine examination in the immediate newborn period. The eardrum is more horizontal in neonates than in adults, lying almost parallel to the external canal. Movement of the eardrum is more difficult to detect than in older infants, and visualization of the eardrum is often impeded by the presence of amniotic debris, vernix, or blood in the ear canal. Moreover, findings typical of otitis media in older infants (dullness of the tympanic membrane, decreased light reflex and translucence, and diminished mobility) may be present in healthy newborns. EYES As part of the inspection of the face, the size, spacing, and orientation of the eyes, lids, and eyebrows should be assessed. In microphthalmos the entire eye is small, whereas microcornea can be an isolated finding in an otherwise normal eye. Hypertelorism implies that the spacing between the bony orbits is excessively wide. In telecanthus, the inner canthi are displaced laterally, giving a false impression of hypertelorism. Hypotelorism is often associated with holoprosencephaly, Trisomy 13, and other genetic abnormalities. The palpebral fissures are downslanting if the medial canthi are higher than the lateral canthi, and upslanting if the lateral canthi are higher. Ptosis can be difficult to detect reliably in the neonate CHAPTER 25 Initial Evaluation: History and Physical Examination of the Newborn unless it is severe or unilateral. It may be congenital or caused by trauma or inflammation. Transient episodes of gaze disconjugation are common in otherwise healthy newborns, but persistent eye deviation requires follow-up. Obstruction at both ends of the lacrimal sac produces a mucocele (or dacryocele), which appears as a bluish subcutaneous mass that can be mistaken for a hemangioma or an encephalocele. Mucoceles can sometimes extend into the nasal passages and cause respiratory distress. The routine examination of sclera, cornea, and internal portions of the newborn eye is often limited by the presence of significant eyelid edema in the first few days after birth. If the corneas and conjunctivae cannot be adequately visualized during spontaneous eye opening or with gentle lid retraction, it is normally preferable to defer further examination of the eyes until the swelling recedes, rather than to attempt to retract the lids forcibly. A coloboma is a congenital defect in the formation of the eye, which can affect any or all of the external or internal structures of the eye. Colobomas of the iris, most frequently located inferomedially and producing a keyhole-shaped pupil, are the most common type visible on external examination. As part of the routine examination, the lower lid of each eye should be retracted so that the entire pupil is visible, if not the entire cornea. If a coloboma of the lid or iris is seen, ophthalmologic evaluation of internal structures of the eye is advisable. Cloudy corneas represent glaucoma until proven otherwise and require prompt ophthalmologic evaluation even if obvious enlargement of the cornea and globe (buphthalmos) is not present. Small subconjunctival hemorrhages are common in newborns, particularly after vaginal delivery, and do not indicate trauma unless other findings are present. An examination of the red reflex of the eyes should be performed in all neonates before discharge from the nursery and at subsequent health supervision visits (American Academy of Pediatrics, 2008). Anything that interferes with the transmission of light through the normally transparent parts of the eye on its path to the fundus and back to the examiner will result in an abnormality of the red reflex. Problems potentially detected by the red reflex examination include cataracts, aqueous and vitreous opacities, and retinal abnormalities including tumors and chorioretinal colobomas. High-refractive errors and strabismus may also produce abnormalities or asymmetry of the red reflex. Dark spots in the red reflex, a markedly diminished reflex, the presence of a white reflex, or asymmetry of the reflexes are indications for immediate referral to an ophthalmologist experienced in neonatal evaluation (American Academy of Pediatrics, 2008). The red reflex examination is to be rated as normal when the reflections of the two eyes viewed both individually and simultaneously are equivalent in color, intensity, and clarity, and there are no opacities or white spots within the area of either or both red reflexes (American Academy of Pediatrics, 2008). The red reflex examination is vital for early detection of visionthreatening and potentially life-threatening abnormalities, including retinoblastoma. Unfortunately, the sensitivity of the routine red reflex examination for retinoblastoma is low, and most retinoblastomas are first detected by family members rather than pediatricians, despite routine screening (Abramson et al, 2003). 291 NOSE The nose should be inspected for deformities, with attention to the size, shape, and symmetry of the various components of the nose, including the nostrils, columella, alae nasi, and the bridge, root, and tip. Asymmetries of the nose caused by in utero compression are common and need to be distinguished from malformations. Passage of a catheter beyond the nasopharynx via each nare, usually done in the delivery room, rules out choanal atresia but does not ensure that the size of the nasal passages is adequate for normal breathing. The quality and volume of air flow through each nostril can be evaluated by listening with the bell of a stethoscope or by observing the deflection of a wisp of cotton placed under the nostril. With the mouth closed, the infant should be able to breath comfortably via each nostril separately; this can be assessed by briefly occluding each nostril with the examiner’s fingertip. If unilateral atresia or stenosis is present, the infant will exhibit signs of distress when the patent nostril is occluded. Flaring of the alae nasi may occur as the sole or initial symptom of mild respiratory distress, or it may accompany grunting and retractions. MOUTH AND ORAL CAVITY The lips, perioral area, and oral cavity should be inspected both at rest and during crying. The shape of the philtrum should be evaluated when the mouth is relaxed, because stretching of the upper lip during crying can give a false impression of a flat philtrum, which is suggestive of fetal alcohol syndrome. Perioral cyanosis is common and benign in normal infants in the immediate newborn period, whereas cyanosis of the tongue and mucous membranes is always abnormal and requires immediate investigation. The mucous membranes will be moist and shiny with saliva if the infant is well hydrated. Excessive oral secretions can be caused by esophageal atresia or impaired swallowing. An excessively short frenulum that restricts protrusion of the tongue can be a cause of feeding difficulty. In the routine examination, the oral cavity is usually inspected without the use of a tongue depressor or other instruments. Epstein’s pearls, Bohn’s nodules, and dental lamina cysts (Figure 25-10) are common findings that should be indicated to the parents as benign. With patience and some adjustment of the head position, the entire palate and much of the pharynx can often be visualized. The tonsils are normally inconspicuous in neonates. The uvula is short but highly mobile. The soft palate elevates during crying if cranial nerves (CNs) IX and X are intact. Inspection of the oral cavity is supplemented by palpation with a gloved finger to assess the shape and integrity of the palate, and to feel for natal teeth and masses. Although clefting of the lip and anterior palate will be obvious at a glance, an isolated cleft of the posterior palate may be missed unless deliberately sought by palpation. Eliciting the sucking reflex allows the strength and coordination of sucking to be assessed. In the routine examination of a vigorous, alert infant, elicitation of a gag reflex is unnecessarily upsetting to the infant and is normally avoided. The gag reflex should be tested if the infant is neurologically depressed or has difficulty swallowing. 292 PART VII Care of the Healthy Newborn NECK It is usually convenient to combine palpation of the neck for lymphadenopathy and masses with palpation of the clavicles. The entire skin surface of the neck should be visualized and palpated, while turning the head and retracting the skin to open the neck creases and folds. The range of motion of the head and neck is evaluated at the same time. Congenital muscular torticollis at birth is commonly but not invariably accompanied by a palpable fibrous tumor (fibromatosis colli) in the shortened sternocleidomastoid muscle. Nonmuscular causes of torticollis include tumors of the posterior fossa or cervical spine and malformations of the cervical spine. A short neck, low hairline at the back of the head, and restricted mobility of the upper spine are characteristic features of the Klippel-Feil syndrome. Redundant skin or a webbed neck may be seen in Trisomy 21 and in Turner and Noonan syndromes. Cystic hygromas are soft, fluctuant masses that transilluminate and are usually unilateral. Branchial cleft cysts or sinuses are also found laterally, from the level of the mastoid to the center of the sternocleidomastoid muscle. Thyroglossal duct cysts are located in the midline high in the neck or under the chin. Further investigation is needed if the larynx or trachea are displaced from the midline, or if enlargement of the thyroid gland is suspected. CHEST WALL Although evaluation of the heart and lungs is a central focus of the examination of the chest, the skin, soft tissue, and bony structures of the thorax should not be neglected. The position of the nipples and the presence of any accessory nipples should be noted. The definition and stippling of the areola and the size of the breast bud are developmental features helpful as part of scoring for gestational age estimation. Transient galactorrhea occurs in approximately 5% of term neonates (Madlon-Kay, 1986). Variations in the shape of the xiphoid process are common, and parents can be reassured that a prominent or bifid xiphoid is benign and will usually become much less apparent as the infant grows. A mildly depressed sternum (pectus excavatum) or protuberant one (pectus carinatum) is usually of no clinical consequence. A small, bell-shaped chest in an infant with respiratory distress may reflect lung hypoplasia or a disorder of skeletal growth. An increase in the anterior-posterior diameter of the chest (barrel chest) may reflect an increase in the intrathoracic volume caused by air trapping from meconium aspiration or pneumothorax. Palpation of the chest wall may reveal irregularities or tenderness, and crepitus may be felt at the site of a fractured clavicle or rib. Crepitus can also be caused by dissection of air into the subcutaneous tissue from a pneumothorax or pneumomediastinum, but this is an unlikely occurrence in an asymptomatic infant. LUNGS AND RESPIRATION The respiratory examination begins with observing the color of the tongue and mucous membranes, respiratory rate, breathing pattern, and work of breathing. Respiratory problems are unlikely to be found in an infant who is centrally pink and breathing comfortably at a relaxed rate. In the normal newborn, the abdomen expands smoothly with each contraction of the diaphragm, while the chest moves inward slightly. The respiratory rate of the newborn infant is highly variable when the infant is awake, changing with activity such as feeding and crying. Tachypnea during sleep is more clearly associated with respiratory problems than is tachypnea during awake states. During a routine examination of the healthy infant, it is not necessary to measure the exact respiratory rate, as long as it is clearly within a normal range in an asymptomatic infant, but the respiratory rate should be determined if tachypnea or an unusually slow rate is seen. Because short pauses and brief periods of rapid breathing are common in normal newborns, accurate measurement of the respiratory rate requires counting for a full minute, preferably when the infant is asleep or at least not crying. During crying, the quality and vigor of vocalization are assessed, and the infant is observed for changes in color and perfusion. Central cyanosis that only appears during crying may be caused by cardiac or respiratory disease and requires further evaluation. Cyanosis that resolves during crying may be due to choanal atresia or stenosis, apnea, or hypoventilation. The classic symptoms of respiratory distress are nasal flaring, grunting, and retractions. Nasal flaring and mild grunting are common in the immediate postnatal period, but in the healthy newborn they should resolve within 15 to 20 minutes after birth. Increasing respiratory distress caused by decreasing lung compliance is typically reflected in a progression from nasal flaring, or mild tachypnea, or both; to nasal flaring plus mild or intermittent grunting; and then to flaring, grunting, and increasingly severe retractions. The respiratory rate generally decreases as the work or effort of breathing increases, as indicated by the development of grunting and increasing retractions. When respiratory distress is mild, intermittent grunting at a slower respiratory rate may alternate with periods of mild tachypnea. As grunting becomes more severe, the expiratory phase becomes increasingly prolonged. The length of the grunt, rather than its loudness, correlates with the severity of distress. Intermittent mild grunting can be misinterpreted by parents as crying. The rhythm of grunting and its occurrence at the end of expiration are key features that help to distinguish it from other vocalizations. Retractions require a forceful inspiratory effort and decreased lung compliance, and they may be absent or less prominent than expected in an infant with neuromuscular depression. Nasal congestion, airway obstruction, and airway secretions can produce sounds that are audible without a stethoscope. Noisy or congested nasal breathing and intermittent sneezing not associated with upper respiratory infection is common in the first few days after birth. A hoarse cry suggests an abnormality affecting the vocal cords. Because intubation of vigorous infants born through meconiumstained amniotic fluid is no longer routine (Halliday and Sweet, 2001), hoarseness or stridor caused by vocal cord trauma in healthy term infants is less common than previously. Inspiratory stridor is due to narrowing or partial obstruction of the upper airway. The presence and loudness of the stridor depends on respiratory effort as well as the extent of airway narrowing, so that stridor worsens with forceful inspiration during crying. Stridor during CHAPTER 25 Initial Evaluation: History and Physical Examination of the Newborn crying in an infant with no respiratory distress when quiet is often due to tracheolaryngomalacia, and it is usually benign. Stridor that is present during quiet breathing or present during both inspiration and expiration suggests the presence of a more significant airway obstruction that requires further evaluation. In the routine examination of the infant who is observed to be centrally pink and breathing comfortably in room air, brief auscultation of the chest of a sleeping or quiet infant is usually sufficient to ensure that the breath sounds are clear, and that air entry is adequate and equal bilaterally. Sounds are not well localized in the neonate, and the infant might not remain quiet for long, so attention to the quality of the breath sounds is usually more helpful than attempting to compare multiple sites. Detection of abnormal lung sounds including crackles, wheezes, and rhonchi requires further assessment. If more detailed examination is indicated, auscultate over the four major quadrants anteriorly, on the sides, and on the upper and lower back bilaterally. Diaphragmatic hernia manifesting in the neonatal period usually causes significant respiratory distress, but rarely a small diaphragmatic hernia is detected by the presence of bowel sounds in the chest in an asymptomatic infant. Spontaneous cough, which is abnormal in neonates, is most commonly caused by infection or aspiration. Percussion of the chest, rarely done as part of routine examination of newborn, can be useful for estimating the position of the upper margin of the liver. Percussion can also be used to detect a large effusion or lung consolidation, but infants with these conditions will have other symptoms of respiratory distress, so the diagnosis will rely on imaging studies and not the physical examination. Transillumination can be useful for supporting a rapid diagnosis of pneumothorax in a distressed infant, but it is not reliable for detecting a small pneumothorax that produces minimal symptoms. Respiratory symptoms are sensitive but nonspecific indicators of illness in the newborn, because alterations in respiration (including apnea) can accompany illness of many different etiologies. Common causes of subtle or mild respiratory distress detected in the routine evaluation include retained fetal lung fluid (transient tachypnea of the newborn), spontaneous pneumothorax, neonatal sepsis, pneumonia, meconium or amniotic fluid aspiration, and congenital heart disease. Any infant with respiratory distress should be transferred to an NICU or observation nursery for further evaluation, monitoring, and treatment. CARDIOVASCULAR The evaluation of the cardiovascular system during the routine examination has two major goals: to assess the current status of the circulation and to detect signs of congenital heart disease, particularly the critical, ductaldependent forms that can produce rapid clinical deterioration in the newborn period. Although detection of heart disease is important, abnormal circulatory findings in the newborn are more often secondary to other problems, including sepsis, hypovolemia, anemia, and hypoglycemia. The cardiovascular system undergoes marked changes after birth involving the transition to air breathing, the progressive decrease of pulmonary vascular resistance, and the closure of the ductus arteriosus. These changes affect 293 the physical examination of healthy infants and those with congenital heart disease; therefore the time after birth is always an important consideration in the interpretation of the examination. Physical examination of the cardiovascular system is not limited to examination of the heart and pulses. It also requires attention to the infant’s general behavior and activity, to respiratory symptoms, to the color of the tongue and oral mucosa, and to the temperature, color and perfusion of all regions of the body. These features will usually be inspected at different times during the course of the routine examination, but they should be reevaluated during the heart and chest examination if cardiovascular abnormalities are suspected. Important features of the pulses are the rate, rhythm, volume and character. The heart rate of the resting well newborn averages 120 to 130 beats/min, with high variability. Transient sinus tachycardia during vigorous crying is common, but persistent tachycardia with a rate greater than 160 beats/min suggests a need for further investigation. A low resting heart rate caused by sinus bradycardia (80 to 100 beats/min) during sleep is common in healthy full-term infants. Isolated premature beats can be noted occasionally in otherwise healthy infants and are almost always benign. The femoral pulses in a normal neonate are not easy to palpate, and they are easily obliterated with pressure. They usually receive a grade of 2 on the traditional scale of 0 to 4. A patent ductus arterious will not cause bounding pulses until the pulmonary vascular resistance has dropped enough to allow significant left-to-right shunting. Uniformly weak pulses suggest a low output state, usually accompanied by signs of poor perfusion. If a decrease or delay of the femoral relative to the brachial pulses is detected, measurement of the blood pressure in all four extremities can reveal a gradient in the blood pressure caused by coarctation of the aorta. However, normal pulses and four-limb blood pressures do not rule out a coarctation if the ductus arteriosus remains open. The precordial area and heart are examined by inspection, palpation, and auscultation. A precordial impulse can be visible in normal newborns, especially during activity, but visible prominence of the precordial area together with a palpably increased cardiac impulse suggest cardiomegaly or a hyperdynamic state. Displacement of the cardiac impulse to the right suggests dextrocardia or a shift in the mediastinum. Thrills are rarely palpable in the newborn. Auscultation of the heart should be performed with specific questions in mind, with attention attuned sequentially for heart sounds, clicks, murmurs, and other abnormal sounds. Most sounds and murmurs in the newborn are relatively high-pitched, so the diaphragm is usually used initially. The bell can be used for further evaluation of lowpitched sounds, if needed. The first heart sound, produced by closure of the mitral and tricuspid valves, is usually single in newborns and best heard in the precordial area. The second sound, produced by closure of the aortic and pulmonary valves, is usually best heard at the left upper sternal border. With focused attention, slight splitting of the second sound and its variation with respiration can be appreciated. Third and fourth heart sounds are abnormal in the newborn. Systolic ejection clicks may be heard in some normal neonates in the first several hours after birth, but ejection clicks are abnormal after this period (Johnson, 1990). 294 PART VII Care of the Healthy Newborn The timing, location, intensity, radiation, quality, and pitch are important characteristics of heart murmurs that can help to distinguish physiologic from pathologic murmurs. With repeated examination, physiologic heart murmurs may be detected in as many as 60% of infants during the first 48 hours (Johnson, 1990). These murmurs are most commonly systolic ejection murmurs that are transient and soft (grade I or II), and they are usually attributed to flow through a closing ductus arteriosus or to increasing flow across the pulmonic valve as pulmonary vascular resistance drops. Vibratory systolic murmurs resembling Still’s murmur can be heard in some newborns. A murmur detected on a routine newborn examination that is not clearly physiologic needs further evaluation. Detection of a suspicious murmur is the most common reason for further evaluation in an otherwise asymptomatic newborn, but absence of a murmur does not rule out congenital heart disease. Examples include transposition of the great arteries and atrioventricular (AV) canal defect. A decrease or loss of a murmur can be an ominous sign that accompanies clinical deterioration associated with ductal closure in an infant with ductal-dependent pulmonary or systemic blood flow. Pathologic murmurs detected in the first few hours of life are usually caused by obstruction of ventricular outflow such as aortic stenosis or pulmonic stenosis; they are crescendo-decrescendo murmurs, usually grade II or III. This type of murmur may also be caused by subaortic stenosis associated with hypertrophic cardiomyopathy in a macrosomic infant of a mother with diabetes. Murmurs associated with defects that produce left-to-right shunting usually appear after a few days, when the pulmonary vascular resistance has dropped sufficiently. Pansystolic murmurs (which include and obscure the first heart sound) that are heard soon after birth are most commonly caused by AV valve insufficiency, whereas left-to-right shunting through a ventricular septal defect produces a pansystolic murmur that typically appears only after 1 to 2 days. Diastolic murmurs are rare in newborns. A continuous murmur in a neonate usually represents an aortopulmonary communication or an arteriovenous fistula. The location of an extrathoracic arteriovenous fistula may be revealed by auscultation of a bruit, most commonly in the head or liver. ABDOMEN Examination of the abdomen begins with observation of the configuration, fullness, and movement with respiration of the abdominal wall. Major abdominal wall abnormalities such as omphalocele, gastroschisis, prune belly syndrome, and bladder extrophy will be obvious on initial inspection in the delivery room and may be diagnosed prenatally. Although a large omphalocele will never be missed, a small one may produce only a slight widening of the umbilicus and proximal umbilical cord (Figure 25-17). If such an omphalocele is not detected in the delivery room before the umbilical cord is clamped and cut, the intestine within it may be damaged. The abdomen of the neonate ranges from flat to moderately protuberant, with substantial variation depending on feeding and the passage of gas and meconium. A markedly distended abdomen suggests the possibility of significant ascites, a large mass, or FIGURE 25-17 A small omphalocele that could be injured if the cord were to be clamped too close to its insertion. (Reprinted from Rudolph AJ: Atlas of the newborn, vol 4, Hamilton, Ont, Canada, 1997, BC Decker, p 109.) an intestinal obstruction. A proximal obstruction such as esophageal or duodenal atresia does not cause abdominal distension, however. A sunken or scaphoid abdomen may be seen in the infant with respiratory distress caused by a diaphragmatic hernia. The umbilicus should be inspected for meconium staining, for signs of infection, and for the rare occurrence of pallor and edema or the visible discharge of urine caused by a patent urachus. At 1 or 2 days after birth, slight redness of the periumbilical skin is common, because of irritation from the cord clamp, and needs to be distinguished from an omphalitis or cellulitis. Counting of the umbilical vessels is best done in the delivery room, on the freshly cut cord. If the infant is asleep or resting quietly, it is prudent to auscultate for bowel sounds before proceeding to palpate the abdomen. Palpation should be initially gentle and superficial, to detect any signs of tenderness and the presence of an enlarged liver or spleen. Tenderness must be distinguished from the tendency of the infant to stiffen the abdominal muscles in reaction to the touch of the examiner’s fingers, which are usually colder than the infant’s skin. If the examiner’s fingers remain in gentle contact within the same spot, the temperature difference will disappear in a moment, and the baby will usually relax and allow the examiner to proceed without a struggle. In the healthy infant, the liver edge may be at or slightly above the right costal margin, or palpable 1 to 2 cm below. The spleen is rarely palpable unless it is enlarged. Gentle palpation of the lower abdomen can detect an enlarged bladder, which is the most common cause of a midline abdominal mass in neonates. Deep palpation to detect small masses or enlargement of the kidneys is most easily done soon after birth, before the infant has fed much, and when the infant is quiet. However, a satisfactory examination can be done even in a crying infant by keeping the fingers in position, and gradually increasing the depth of palpation each time the infant briefly relaxes the abdominal muscles while taking a breath between cries. It is helpful to support the flank with one hand while palpating for the kidney with the other, or to palpate with the thumb while supporting the flank with fingers of the same hand. Percussion of the CHAPTER 25 Initial Evaluation: History and Physical Examination of the Newborn abdomen is not particularly helpful in routine examination, but it can sometimes help to define the boundaries of an enlarged liver or bladder. GENITALIA AND PERINEUM Brief inspection after delivery is usually sufficient to identify the infant as male or female. The evaluation and management of the infant with ambiguous genitalia is discussed in Chapter 92. In both male and female infants, a soft swelling or bulge in the inguinal area may be due to an inguinal hernia. The bulge typically appears or increases in size during crying and is easily reduced with gentle pressure when the infant relaxes. The perineum is inspected to locate the anus and assess the tone of the anal sphincter. Absence of a normal anal opening should be detected as part of the initial evaluation in the delivery room. However, external observation of an apparently normal anus does not guarantee internal patency of the anus, which is best confirmed by the normal passage of meconium. The genitalia are mainly examined by inspection, supplemented by palpation, with the infant in a supine, frogleg position. In the newborn male, the foreskin normally covers the entire head of the penis, which is adherent to the glans. The urethral opening is usually hidden by the foreskin and need not be visualized if the foreskin is intact. The foreskin is typically incomplete if hypospadias is present, which allows the abnormal position of the urethral opening to be identified easily. Congenital chordee, a ventral angulation of the head of the penis, may accompany hypospadias or occur in isolation. Chordee can be missed unless the examiner straightens the penis by gently retracting the skin along the shaft towards the base of the penis. In infants who have a generous pad of subcutaneous fat at the base of the penis, this maneuver also helps to avoid a false impression that the penis is short. Dribbling of urine or a weak stream, if observed, is suspicious for bladder dysfunction or urethral obstruction. The scrotal sac and inguinal areas are palpated to locate the testes and assess their size. The scrotal rugae usually appear at approximately 36 weeks’ gestation and cover the entire scrotum at term. Enlargement of the scrotal sac is most commonly caused by a hydrocele. Transillumination can help to distinguish a hydrocele from swelling because of congenital testicular torsion or other masses. Bowel sounds may be audible in a scrotum enlarged by an inguinal hernia. During the inspection of the female genitalia, the examiner must gently retract the labia majora laterally to allow full visualization. The sizes and positions of the labia minora, clitoris, urethra, and vaginal opening should be noted. Relative prominence of the labia minora is normal in preterm infants. Partial labial fusion and an increase in the size of the clitoris may represent virilization caused by congenital adrenal hyperplasia or related endocrine abnormalities. The posterior fourchette should be at least 1 cm from the anal opening. Enlargement of the uterus because of hydrometrocolpos may produce a protruding perineal mass. Vaginal tags and mucoid vaginal discharge are common at birth, resulting from exposure to maternal estrogen. A slightly bloody vaginal discharge (pseudomenses) caused by hormonal withdrawal is common in healthy females during the first week after birth. 295 BACK The back is inspected for asymmetry or abnormal positioning of the shoulders, ribs, and hips. Major neural tube defects and large masses such as a large sacrococcygeal teratoma will be detected prenatally or on initial inspection in the delivery room. In the routine examination, the lumbosacral area should be inspected carefully for the presence of deep or unusual dimpling of the skin over the sacrum, for sinus tracts, for unusual tufts of hair, and for small masses such as a lipoma or hemangioma, any of which may be associated with spina bifida occulta or tethering of the spinal cord. The spine is inspected for straightness and palpated for the integrity and alignment of the posterior spinous processes. MUSCULOSKELETAL SYSTEM Routine assessment of the musculoskeletal system begins with plotting the height, weight, and head circumference on appropriate growth charts. Detailed measurements of the limbs and trunk can be made if growth appears disproportionate, but they are not part of the routine examination. The infant’s posture, muscle mass, and movements should be observed. As the different parts of the body are surveyed during the course of the examination, any limitation in the normal range of joint motion and any localized swelling or tenderness should be noted. Major limb or skeletal malformations will usually be appreciated on initial inspection in the delivery room, but minor anomalies such as supernumerary digits, syndactyly, or nail hypoplasia can be missed if a disciplined approach is not taken. Lack or restriction of normal movement of an extremity can be caused by trauma, most commonly a fractured humerus or clavicle, or by an intrinsic abnormality of the joint or limb. Although they generally require no treatment, fractures of the clavicle are sufficiently common that the clavicles should be specifically examined in every newborn. Crepitus and tenderness at the site of a clavicle fracture may be more easily detected if the examiner palpates a clavicle with one hand while using the other to elevate and rotate the ipsilateral shoulder. Deformations caused by in utero positioning are not always easy to distinguish from malformations on an initial examination. Mild inward bowing of the lower legs and feet is common in newborns. If an inward-turning foot can be brought easily to a neutral position, a clubfoot deformity is unlikely, and the angulation of the foot can be expected to normalize spontaneously. EXAMINATION OF THE HIPS Assessment for developmental dysplasia of the hip (DDH) is an important component of the examination of every newborn (American Academy of Pediatrics, 2000; Sewell et al, 2009). Frank dislocations, partial dislocations, and instability of the humoral head in the socket can be detected by examination in the newborn period. Nevertheless, DDH is not always detectable at birth. Statistical risk factors for DDH include a positive family history, female sex, and breech presentation. In addition to the Ortolani and Barlow maneuvers, the examination also involves a focused inspection for indirect clues related to the possible presence of DDH. 296 PART VII Care of the Healthy Newborn During general inspection of the infant, attention should be paid to the resting posture and spontaneous movements; any asymmetry or unusual positioning of the legs should be noted. Asymmetry of the gluteal or femoral skin folds may be a sign of unilateral hip dislocation. A unilateral dislocation may also produce an apparent inequality of leg length, seen either with the legs in extension (Thomas’ sign) or with the feet flat on the bed and the knees bent (Galeazzi’s sign). A restricted range of motion, particularly abduction, is a clue that may detect either unilateral or bilateral dislocations. The dislocation of the femoral head during a Barlow maneuver or its relocation during an Ortolani maneuver produces what is called a clunk that must be distinguished from high-pitched clicks of the hip that are common and benign. Clunks and clicks are perceived by palpation and are not actually audible. The key element defining the clunk is a distinct sensation of abrupt movement of the femoral head as it passes over the rim of acetabulum and drops into or out of the socket. A dislocated or dislocatable hip has the distinctive clunk, whereas a subluxable hip is characterized by a feeling of looseness or sliding without a distinct clunk (American Academy of Pediatrics, 2000). Both maneuvers are performed with the infant supine, starting with the legs held in neutral rotation and the hips flexed to 90 degrees but not more. Each hip is examined separately. For the Ortolani maneuver, the index and middle fingers of the examiner’s hand are placed along the greater trochanter with the thumb placed along the inner thigh near the knee. The other hand stabilizes the pelvis. The examiner gently abducts the hip by rotating the thumb outward while lifting anteriorly with the fingers. A distinct sensation of movement is felt when a posteriorly dislocated hip relocates during abduction. For the Barlow maneuver, the index and middle fingers are also positioned along the greater trochanter, but the examiner’s hand is rotated so that the base of the thumb is on top of the knee. The maneuver is performed by adducting the leg until the knee is in the midline, and then applying gentle pressure to the knee in a downward direction along the adducted femur. A clunk is felt if this maneuver induces the femoral head to exit the acetabulum posteriorly. Minimal force is used in either maneuver. The examination is considered positive for DDH if a clunk of dislocation or reduction is elicited during the Barlow or Ortolani maneuvers, in which case prompt referral to an orthopedist is strongly recommended. The examination is considered equivocal if the Barlow and Ortolani test results are negative, but warning signs such as asymmetric creases, apparent or true leg length discrepancy, or limited abduction are found. In this case, the recommended next step in evaluation is a follow-up examination of the hips by a pediatrician at 2 weeks ­(American Academy of Pediatrics, 2000). NEUROLOGIC EXAMINATION The neurologic system is assessed during the course of the general newborn examination and by performing some specific maneuvers to elicit neonatal reflexes. Because the healthy infant’s responses vary with the state of alertness, and because the infant’s tolerance for prolonged examination is limited, eliciting a perfect response to each maneuver should not be expected. Fortunately, there is sufficient redundancy in the routine examination that it is not difficult in most cases for the examiner to be satisfied that the infant is neurologically normal. When the examination is not fully reassuring, repeating selected parts of the examination at a later time may be more helpful in clarifying findings than attempting an extended examination at one time. The newborn screening examination includes assessment of the infant’s alertness, spontaneous activity, posture, muscle tone and strength, head control, and responses to manipulation and handling. The overall assessment of alertness, tone, and activity is one of the most important components of the newborn examination. Diminished alertness, tone, or spontaneous activity are sensitive but nonspecific indicators of illness that are much more likely the result of other causes, such as neonatal sepsis, than a specific neurologic abnormality. The typical healthy newborn is easily awakened from sleep and remains alert through the remainder of the routine examination, shifting among states of quiet alertness, active alertness, and crying. The examiner’s initial attempt to quietly auscultate the breath and heart sounds of a sleeping infant is often foiled by the infant’s prompt arousal and crying. The healthy newborn demonstrates a vigorous cry when upset, but is able to self-console or to be consoled with holding, sucking, or feeding. An infant who is stuporous or difficult to arouse is clearly abnormal and needs further evaluation, as does an infant who is unusually irritable or inconsolable. The typical newborn will be rather alert for several hours after delivery, but may then provoke parental concern by becoming relatively sleepy and uninterested in feeding for the remainder of the first 24 hours. As long as the infant continues to be easily arousable and the examination result remains otherwise normal, parents can be reassured that the infant will likely begin feeding much more vigorously on the second day. Decreases in tone or alertness occurring 1 day or more after birth in a previously vigorous infant are not normal and require prompt investigation for sepsis or other problems, including inborn errors of metabolism that can cause progressive symptoms starting a few days after birth because of accumulation of toxic metabolites. Because the newborn cannot respond to verbal questions or commands, assessment of the infant’s sensory functions in the routine examination depends on observation of the strength and quality of the infant’s movements in response to handling during the examination and to specific local stimulation, such as the elicitation of the palmar and plantar grasp reflexes and the rooting and sucking reflexes. Assessment of vision in the routine newborn examination is limited to observing pupillary constriction and a blink response to light, which are subcortical responses, and to observing the infant’s visual attentiveness. The healthy term newborn is expected to be able to visually fixate on and follow the examiner’s face, but failure to observe this response during the course of the routine examination is common, because it can take more time and patience to elicit than are available. Hearing can be evaluated behaviorally by observing the infant’s responses to the ringing of a bell and other sounds. However, behavioral testing can only detect profound, bilateral hearing loss, and it has been effectively replaced in the routine newborn examination by automated hearing tests (American Academy of Pediatrics, 2007; US Preventive Services Task Force, 2008). CHAPTER 25 Initial Evaluation: History and Physical Examination of the Newborn The assessment of motor function relies on the observation of the infant’s posture and spontaneous movements, observation of the infant’s general responses to stimulation and handling, and the elicitation of specific reflexes. The healthy term newborn normally maintains a resting posture with elbows, hips, and knees strongly flexed. The scarf sign, forearm recoil, square window of the wrist, the heal-to-ear maneuver, and popliteal angle are measures of tone and flexibility commonly scored in the gestational age assessment (Ballard et al, 1991; Dubowitz et al, 1970). The active tone and strength of upper extremity muscle groups are routinely assessed by eliciting the palmar grasp and by arm traction during the pull-to-sit maneuver. Lower extremity strength and tone are assessed during the general examination of hips and feet, and by testing for the supporting reaction, stepping reflex, Babinski’s reflex, and ankle clonus. The neck flexors may be evaluated during the standard pull-to-sit maneuver, or by lifting the shoulders to pull the baby to a sitting position. The neck extensors can be evaluated by tilting the infant forward from a sitting position, or evaluated along with truncal tone by suspending the infant in a prone position with the examiner’s hand under the chest (ventral suspension maneuver). When observing the infant’s movements, the examiner should note any asymmetry and pay attention to qualitative characteristics such as the relative smoothness versus jerkiness or tremulousness of movement. Unusual jitteriness may be a symptom of hypoglycemia or hypocalcemia, especially in infants of mothers with diabetes, infants that are small or large for gestational age, or infants who were exposed to opiates or other drugs in utero and are experiencing withdrawal. Passive restraint of an extremity should inhibit jittery movements, but will not stop the rhythmic contractions of clonic seizure activity. In addition to clonic movements, signs of neonatal seizures can include tonic posturing, repetitive stereotyped movements of the face or extremities, tonic horizontal eye deviation or nystagmoid jerking, staring or blinking, apnea, and unexplained changes in heart rate or blood pressure. Although it is uncommon to observe an actual seizure during the routine newborn examination, the pediatrician or neonatologist is frequently called to evaluate an infant because of concern about unusual movements seen by a parent or nurse. In the otherwise healthy newborn who is fully alert and responsive, jerky or abrupt movements that are evoked by stimulation (e.g., startles), that can be suppressed by passive restraint (e.g. jitteriness), and that are not accompanied by autonomic changes or changes in alertness are unlikely to be seizures. Transient episodes of disconjugate gaze are also not unusual in normal newborn infants, particularly when the infant is entering or awakening from sleep. Neonatal Reflexes The neonatal or primitive reflexes frequently tested during routine examination of the newborn include the Moro reflex, the asymmetrical tonic neck reflex, truncal incurvation (Galant reflex), the palmar and plantar grasp reflexes, the Babinski’s reflex, and the placing and stepping reflexes. The Moro reflex can be elicited following the pull-tosit maneuver, by lowering the infant until there is only a slight space between the neck and bed, and then allowing 297 the infant to fall back suddenly. Alternatively, the Moro response can be elicited by the “drop” method: the examiner lifts the baby completely off the bed, supporting the head and trunk with both hands and keeping the baby supine, and then rapidly lowers the baby by approximately 4 to 8 inches. The complete Moro reflex involves a quick bilateral abduction of the arms and extension of the forearms with full opening of the hands, followed by smoother and slower return of the hands toward the midline, with curling of the fingers. The startle reflex is similar to the Moro reflex, but without full extension or hand opening, and may occur spontaneously or be evoked by a sudden noise or movement. The position of the neck affects the tone of the extremities via the asymmetrical tonic neck reflex. This should be kept in mind during observation of the infant’s spontaneous movements, because it can cause a false impression of asymmetry if the position of the neck is not taken into account. To test the asymmetric tonic neck reflex, turn the infant’s head 90 degrees to one side for 15 seconds, keeping the infant lying on the back with the shoulders horizontal. In the complete response, the ipselateral arm and leg will extend and the contralateral arm and leg will flex, producing the “fencing” posture. The test is then repeated with the head turned to the other side. Observation of the complete response is reassuring, but its absence is not necessarily abnormal, because partial and unidirectional responses are common in normal newborns. However, an unusually sustained or exaggerated response is abnormal. The truncal incurvation reflex is elicited with the infant held in ventral suspension by stroking lightly down the back on one side, and then the other. The normal response is for the infant to curve the spine strongly, concave toward the stimulated side. The supporting, placing, and stepping reflexes are elicited with the infant held upright. The supporting reaction is elicited by lowering the infant vertically until both feet touch the surface of the bed or table. A positive response, usually seen after a slight delay, is partial extension at the hips and knees, as though the infant is attempting to stand and support his or her weight. The stepping reaction is tested by lowering the infant so that one foot touches the surface, with the infant tilted slightly forward. The infant should flex that leg and extend the other, as though taking a step. The response can sometimes be sustained for several alternating steps. The placing reaction is elicited by lifting the infant to bring the dorsum of one foot in contact with the underside of a table or bassinet edge. In a positive response, the infant lifts the foot up and places it on the top surface. Babinski’s reflex, which is normal in neonates, consists of dorsal flexion of the big toe and spreading of the other toes in response to stroking the foot laterally. Firm pressure on the sole of the foot, in contrast, will elicit a plantar grasp. The deep tendon reflexes are usually not elicited during routine examination of the well newborn, but they are helpful as part of a more complete examination if neurologic abnormalities are suspected. The pectoralis, biceps, brachioradialis, thigh adductor, crossed adductor, knee jerk, and ankle jerk reflexes are the most readily elicited (Volpe, 2008). Ankle clonus may be elicited by quickly dorsiflexing the foot, which in a healthy term infant should produce no more than approximately 298 PART VII Care of the Healthy Newborn five beats of alternating extension and flexion with rapidly decaying intensity. Brachial Plexus Injury Although fortunately not common, brachial plexus injury is one of the more frequent neurologic abnormalities found in the otherwise healthy newborn, occurring in about 0.5 to 2 per 1000 live births. Often there is a history of difficult delivery because of shoulder dystocia. Brachial plexus injury can occur in isolation or in conjunction with fractures of the clavicle or humerus. Typical findings in Erb’s palsy are an inability to abduct and externally rotate the shoulder, flex the elbow, and supinate the forearm because of injury to C5-C6 (Volpe, 2008). If C7 is involved, wrist and finger extension are also weak (Figure 25-18). A distal brachial plexus injury involving C8-T1 causes weakness of wrist and finger flexion. Cranial Nerves Observation and maneuvers during the routine examination provide at least partial assessment of all the cranial nerves (Volpe, 2008) except for the olfactory nerve (CN I), which is not routinely tested. Visual attentiveness, the ability to fix and follow, eyelid closing in response to light, and the pupillary light reflex require the optic nerve (CN II). The pupillary light reflex also tests the oculomotor nerve (CN III). Extraocular movements are controlled by CNs III, IV, and VI. Vertical gaze is difficult to assess in neonates, but horizontal gaze in both directions can usually be verified by observing spontaneous eye movements or by the rotation test. The examiner performs the rotation test by turning in place while holding the infant upright and supporting the back of the head. If vestibular functions (CN VIII) are intact, the infant will turn the head in the direction of rotation or, if the head is restrained, turn the eyes in that direction. Eliciting rooting or a facial grimace in response to touching the face tests the sensory portion of the trigeminal nerve (CN V). The corneal reflex can be tested by touching the cornea with a wisp of sterile cotton, but this is not part of the routine newborn examination. CNs V, VII, IX, X, and XII are all involved in normal sucking and swallowing. The motor portion of CN V controls the muscles of mastication, which are involved in the jaw-closing phase of the suck and are assessed during elicitation of the suck when the infant bites down on the examiner’s finger. Although not part of the routine examination, the masseter or jaw-jerk reflex can be elicited by placing the forefinger of one hand on the infant’s relaxed chin and tapping it with the forefinger of the other hand. FIGURE 25-18 Hand and arm position in an infant with Erb’s palsy involving C5, C6, and C7. (Reprinted from Fletcher MA: Physical diagnosis in neonatology, Philadelphia, 1998, Lippincott-Raven, p 450.) The facial nerve (CN VII) is required for pursing of the lips in sucking, as well as normal facial expression and tone. Compression of the facial nerve by in utero positioning or by forceps injury are common causes of unilateral facial weakness in the newborn period (Figure 25-19). CNs IX and X are needed for normal swallowing and the gag reflex. CN XII is involved in the milking action of the tongue during sucking and swallowing. Taste, mediated by CN VII (anterior two thirds of the tongue) and IX (posterior third) is not evaluated in the routine examination. Sternocleidomastoid function (CN XI) is assessed by evaluating head flexion and lateral rotation. Because head control is often poor in normal newborns, it can be difficult to detect an abnormality unless it is unilateral. However, an asymmetric abnormality of a sternocleidomastoid is caused more commonly by torticollis than dysfunction of CN XI. The vestibular portion of CN VIII is assessed by the rotation test, as noted previously, or by the doll’s eye maneuver. The auditory portion of CN VIII is best assessed using brainstem auditory evoked responses, rather than behavioral responses to auditory stimulation. CHAPTER 25 Initial Evaluation: History and Physical Examination of the Newborn A C 299 B FIGURE 25-19 (See also Color Plate 17.) Unilateral facial weak­ ness in two infants. There is mild facial asymmetry with flattening of the nasolabial folds at rest (A and C) and more obvious asymmetries of the grimace and eye closing during crying (B and D). The weakness is on the infant’s left side in A and B, and on the infant’s right side in C and D. (Reprinted from Fletcher MA: Physical diagnosis in neonatology, Philadelphia, 1998, Lippincott-Raven, p 457.) D SUGGESTED READINGS Books Cohen B: Pediatric Dermatology, ed 3, Philadelphia, 2005, Mosby. Eichenfield LF, Frieden IJ, Esterly NB, editors: Neonatal Dermatology, ed 2, Philadelphia, 2008, Saunders. Fletcher MA: Physical Diagnosis in Neonatology, Philadelphia, 1998, LippincottRaven. Rudolph AJ: Atlas of the Newborn, Hamilton, Ontario, 1997, B.C. Decker. Volpe JJ: Neurology of the Newborn, ed 5, Philadelphia, 2008, Saunders. Guidelines and Policy Statements Early detection of developmental dysplasia of the hip: clinical practice guideline. Committee on Quality Improvement, Subcommittee on Developmental Dysplasia of the Hip. American Academy of Pediatrics, Pediatrics 105:896-905, 2000. Hospital stay for healthy term newborns: policy statement. American Academy of Pediatrics, Pediatrics 125:405-409, 2010. Red reflex examination in neonates, infants, and children: Joint policy statement. American Academy of Pediatrics Section on Ophthalmology, American Association for Pediatric Ophthalmology and Strabismus, American Academy of Ophthalmology, American Association of Certified Orthoptists, Pediatrics 122:1401-1404, 2008. Universal screening for hearing loss in newborns: US Preventive Services Task Force recommendation statement, Pediatrics 122:143-148, 2008. Online Resources Lehmann CU, Cohen BA: Dermatlas: an online collaborative education tool, The Internet Journal of Dermatology 1, 2002. Available at http://dermatlas.med.jhmi. edu/derm/. Ballard JL, Khoury JC, Wedig K, Wang L, Eilers-Walsman BL, Lipp R: New Ballard Score, expanded to include extremely premature infants, J Pediatr 119:417-423, 1991. Descriptions and video demonstrations by JL Ballard. Available at www.ballardscore.com. Pediatric NeuroLogic Exam, Available at http://library.med.utah.edu/pedineurologic­ exam/html/home_exam.html. A tutorial with video demonstrations and descriptions by PD Larsen and SS Stensaas. Complete references and supplemental color images used in this text can be found online at www.expertconsult.com C H A P T E R 26 Routine Newborn Care James A. Taylor, Jeffrey A. Wright, and David Woodrum Two central paradoxes underlie the care of healthy newborn infants. First, although birth is perhaps the oldest and most natural of all human processes, the infant mortality rate has been extraordinarily high until recently. The second paradox in providing newborn care is that neonates are at once the healthiest and most vulnerable patients in medicine. Recent medical history is replete with examples of the pendulum swinging too far in each direction around these paradoxes. The promotion of scheduled feeding using infant formulas rather than breastfeeding is an example of the medicalization of neonatal care. Conversely, recent resistance to treatments that prevent uncommon but disastrous conditions represent a denial of the benefits provided by medical care. Therefore optimal care of a normal neonate is an attempt to balance these competing forces. Systems of care should be designed to support the concept that newborn infants are extraordinarily healthy and require little intervention beyond promotion of breastfeeding. Interventions for which there is evidence that the benefits far outweigh the risks should be provided as unobtrusively as possible. Simultaneously, while promoting natural care for these newborns, health care providers need to be vigilant for the early identification of neonates who are at risk for conditions such as dehydration, sepsis, and severe hyperbilirubinemia. The goal of this chapter is to provide an evidence base for the promotion of healthy newborn care by parents, the rationale for monitoring full-term neonates for various conditions, a risk-benefit analysis of common treatments, and the significance of common prenatal and postnatal findings. Rather than providing a comprehensive prescription for the care of these infants, it is hoped that the reader will integrate the information provided in this chapter with expert opinion and clinical experience to determine the proper care of healthy newborns. INITIAL ASSESSMENT Timing of the initial assessment of a full-term newborn is dependent on the condition of the infant and parental preference. In most instances a health care professional who is present at the birth will make a general appraisal of the infant and alert the child’s provider if there is an acute problem necessitating an immediate evaluation. Usually the neonate will be healthy and the assessment can be timed so as not to interfere with breastfeeding, bonding with the family, and routine care. Before examining a healthy newborn, the mother’s medical history should be reviewed to identify issues that would affect the care or prognosis of the infant. For example, a history of diabetes in the mother would lead to glucose testing in the neonate. Maternal drug use should be assessed for possible teratogenic effects, possibility of 300 symptoms of withdrawal in the infant, and compatibility with breastfeeding. It is important to review the pregnancy history and focus on estimated gestational age, the results of screening for genetic conditions, and the results of prenatal ultrasound examinations. Perinatal events such as type of delivery, length of time that membranes were ruptured, and Apgar scores should also be reviewed. Finally, it is critical to review the mother’s social history to insure that the newborn will be raised in a nurturing environment and to identify high-risk situations for which interventions are needed before or shortly after discharge from the newborn nursery. The results of several laboratory tests commonly performed on pregnant women will determine the need for treatment or monitoring during the newborn nursery stay. These tests include maternal HIV, hepatitis B surface antigen status, and syphilis. The mother’s blood type, Rh status, and antibody test results are useful in identifying newborns with an increased risk of hyperbilirubinemia. It is important to note the results of testing for maternal colonization with group B Streptococcus (GBS) and the type and timing of antenatal antibiotic prophylaxis in mothers with a positive test result for GBS. The infant’s weight, length, and head circumference should be measured shortly after birth and plotted on a standardized chart. Although the most common reason for a significant discrepancies among weight, height, and head circumference percentiles is an inaccurate measurement, a valid discrepancy warrants close clinical observation or testing. Glucose testing may be indicated for newborns found to be small or large for gestational age. If the estimated gestational age of the infant is inconsistent with the growth parameters a formal evaluation using a Dubowitz-Ballard assessment of gestational age may be helpful. When examining a newborn infant for the first time, the initial focus is directed toward an overall assessment of the child’s health. Observation and auscultation of the chest allows for detection of an irregular heart rate, murmur, or acute lung condition such as pneumothorax. The heart rate and respiratory rate can be measured. Normal values for heart and respiratory rate in a newborn infant are 100 to 160 beats/min and 35 to 60 breaths/min. Evaluation of skin color may be useful for identifying cyanotic congenital heart disease or pulmonary conditions in a neonate. If uncertainty exists about the presence of cyanosis, oxygen saturation can be measured quickly with a pulse oximeter. The newborn’s tone, general posture, and movement should be assessed; abnormalities may be indicative of an acute or chronic central nervous system problem or sepsis. There are several purposes for the rest of physical examination, including maternal reassurance and education about normal variations. Asymptomatic conditions can be CHAPTER 26 Routine Newborn Care 220 BOX 26-1 C  linical Signs Compatible With Hypoglycemia 200 ll 180 ll ll Plasma glucose (mg/dl) 160 ll ll 140 ll 120 100 ll * ll (35) 80 (55) (51) (69) (51) ll (26) (55) (40) (49) ll (52) 40 6 68 –1 6 –9 2 –7 8 –4 4 –2 97 4 73 3 49 2 25 1 12 0 Age in hours FIGURE 26-1 Predicted plasma glucose values during the first week of life in healthy term neonates appropriate for gestational age. (Adapted from Srinivasan G et al: Plasma glucose values in normal neonates: a new look. J Pediatr 109:114, 1986.) diagnosed, such as some types of congenital heart disease or developmental dysplasia of the hips. Finally, the examination is a useful screen for rare but serious conditions. ROUTINE TESTING GLUCOSE The fetal blood glucose level is approximately 80% of the maternal level. After birth and separation of the infant from its major energy supply, the infant’s glucose falls by an average factor of 0.5. Over the next several hours, it gradually increases to a level approaching that of older infants (Figure 26-1). Critical factors involved in this normal adaptive process include transient inhibition of the infant’s insulin secretion and an increase in the counter regulatory hormones, including growth hormone, cortisol, epinephrine, and glucagon (Polin et al, 2004). The end result is the promotion of liver glycogen breakdown, gluconeogenesis, and tissue lipolysis. Given this normal sequence of events, there is no reason to routinely screen the blood glucose in infants who are products of uncomplicated pregnancies, labor, and delivery. Clinical scenarios that might be expected to alter the normal sequence of events and indicate the need for early glucose screening include: ll ll ll Infants with midline facial anomalies, which might be markers for pituitary deficiency Infants demonstrating hepatomegaly, suggesting glycogen storage disease Any infant showing clinical signs of hypoglycemia (Box 26-1) The treatment approach to confirmed hypoglycemia depends on the glucose level, the presence of symptoms, or both (see Chapter 94). ( ) Number of samples * Mean & 95% confidence interval 20 ll Poor feeding Lethargy Hypotonia Irritability Tremor Seizure-like activity Apnea (52) 60 ll 301 A pregnancy history of maternal treatment with β-agonists or intrapartum intravenous glucose Infants of diabetic mothers Infants demonstrating intrauterine growth restriction Premature infants Infants delivered after in utero or intrapartum fetal distress NEWBORN METABOLIC SCREENING Newborn screening for metabolic disorders began in 1962 when 29 states participated in a trial of testing for phenylketonuria. With the implementation of screening programs, criteria were proposed for determining which conditions should be screened. It was recommended that only disorders that were important health problems be included in screening programs. The condition should be detectable before the onset of significant symptoms. Importantly, a specific treatment to prevent adverse clinical consequences from the disorder should be available, and the screening program for the condition should be cost effective (Tarini, 2007). Based on these criteria, conditions such as congenital hypothyroidism and congenital adrenal hyperplasia were slowly added to newborn screening tests in many states, and subsequently conditions such as sickle cell disease were added. Although there is no specific treatment for sickle cell disease, there was evidence that the use of a newborn screening program to identify infants with the disorder led to early initiation of penicillin treatment, which resulted in fewer deaths from sepsis than when disease was identified at the onset of symptoms (Vichinsky et al, 1988). Given the demonstrable effectiveness of early identification, sickle cell disease met criteria for newborn screening. The advent of tandem mass spectrometry in the 1990s revolutionized newborn metabolic screening. With this technology it is possible to test for a multitude of conditions using a small sample of blood. In 1995, the average number of conditions included in state-mandated screening programs was eight; by 2005 this had increased to 19, with some states testing for up to 46 conditions. Unfortunately, this increase in newborn screening has been controversial. Some of the conditions included do not meet the established criteria for screening, in that there are no known effective treatments, and in some cases it is not known whether the targeted condition always leads to disease. In addition, with increasing numbers of tests come increasing numbers of false-positive results, with the 302 PART VII Care of the Healthy Newborn resulting parental anxiety and potential for the overuse of medical services (Tarini et al, 2006). In an attempt to define a rational list of disorders for which newborn screening is appropriate, the American College of Medical Genetics used an iterative process to identify 29 “core conditions” that should be included in mandatory screening programs (Watson et al, 2006). As of 2009, approximately 25 of these conditions were included in the screening panels of at least 48 states (National Newborn Screening and Genetics Resource Center, 2009). The most common disorders included in newborn metabolic screens in the United States are congenital hypothyroidism (1 case per 3000 to 4000 infants) and sickle cell disease (American Academy of Pediatrics [AAP] Newborn Screening Task Force, 2000). The incidence of phenylketonuria is approximately 1 in 15,000. For many of the core conditions for which screening is recommended, incidence rates are in the range of 1 in 100,000 to 1 in 200,000 (Kaye et al, 2006). For some disorders, the incidence rate is unknown. The expansion in newborn screening programs presents a challenge to health care providers. It is difficult to remain knowledgeable about all of the conditions that are screened, the incidence of the various conditions, and their natural histories and treatments. Fortunately, many states have informative Web sites providing up-to-date information for parents and health care professionals. In addition, the American College of Medical Genetics maintains online information for health care practitioners, including synopses of conditions commonly included in screening programs and appropriate management of infants with a positive result (accessible at http://www.acmg.net/resources/ policies/ACT/condition-analyte-links.htm). HEARING SCREEN Newborn hearing screening has become nearly universal in the United States. Thirty states plus Guam, Puerto Rico, and the District of Columbia have established mandatory early hearing screening programs, and 17 states plus Guam, Puerto Rico, and the District of Columbia require all health insurers to cover the test (National Conference of State Legislatures, 2009). Newborn hearing screening and early intervention is endorsed by the AAP (AAP Joint Committee on Infant Hearing, 2007), and the U.S. Preventive Services Task Force (2008). Many nurseries now use a two-step screen, first using automated otoacoustic emissions screening followed by auditory brainstem response done in those who fail the otoacoustic emissions screening. Before universal screening, there were questions about the utility of newborn hearing screening, including whether the false-positive rate creates more harm than the benefit of detecting a small number of infants with hearing loss (approximately 1 in 1000 newborns) and whether early intervention is effective (Keren et al, 2002). The specificity of a two-step screening program for hearing loss approaches 0.99. However, even with this high specificity, there at least six false-positive screens for every true positive screening result, because of the low rate of hearing loss (Nelson et al, 2008). The effect of false-positive screens has been minimally investigated. In one study, at 6 months of age, parents of newborns with false-positive screens continued to express worries about their child’s hearing after subsequent testing confirmed normal hearing. Fortunately, using standardized measures, there was no evidence of an increase in general anxiety these parents compared with parents of babies with normal newborn hearing screens (van der Ploeg et al, 2008). False-negative screens are also a concern, because the equipment used is designed for screening and because most auditory brainstem response screening is designed to detect moderate or greater hearing loss. As a result, there is a chance that the screen will be falsely negative in approximately 2% of newborns (Johnson et al, 2005), and pediatricians should continue surveillance of hearing status during childhood. The results of recent studies have shown improved reading and communication skills in hearing-impaired children identified during the period of universal newborn hearing screening. Furthermore, hearing-impaired children who enroll in treatment programs during the first 3 months of life have better language outcomes at school age (McCann et al, 2009; Vohr et al, 2008). Investigators in a Belgian study found that a cause for hearing loss can be determined in approximately half of children identified by newborn screening, of which 60% have genetic origins and 19% have cytomegalovirus infection (Declau et al, 2008). It is challenging to ensure that all children with positive newborn screening test results have confirmatory audiologic testing and begin treatment by 3 to 6 months of age. The number of patients lost to follow-up remains problematic in many areas, and services such as amplification may take time to arrange. Efforts must be made to improve these problems for hearing screening programs to be maximally effective. PRENATAL ULTRASOUND SCREENING FOR BIRTH DEFECTS Second-trimester ultrasound screening for fetal anomalies has become increasingly routine. Major fetal organ system abnormalities can, for the most part, be identified and referred for appropriate fetal and neonatal management. There are, however, a number of ultrasound findings that have a variable natural history, may or may not be markers for serious conditions, and do not always result in a definitive prenatal workup. These findings often do not fit within the pediatric lexicon, and they can present a challenge to the pediatrician regarding parent counseling and determining management in the neonatal period. Central Nervous System Findings Choroid plexus cysts are found in 1% to 3% of second-trimester fetal ultrasound examinations. They are transient, functionally benign in nature, and resolve spontaneously before term. If one or more choroid plexus cysts are found in isolation on prenatal ultrasound examination, no adverse effect on fetal growth and development has been noted. Therefore, without other risk factors, no further evaluation is needed in an infant with this isolated finding and a benign prenatal course. Choroid plexus cysts are believed to be a “soft marker” for aneuploidy (particularly trisomy CHAPTER 26 Routine Newborn Care 18) when associated with other fetal anomalies or with maternal risk factors such as maternal age. In such situations, begin an appropriate prenatal evaluation, such as karyotyping (DiPietro et al, 2006; Lopez and Reich, 2006; Sohaey, 2008b). Agenesis of the corpus callosum is reported to occur in 0.01% to 0.7% of unselected postnatal populations. Aneuploidy has been reported in 10% to 20% of children with this prenatal ultrasound finding. Major organ system abnormalities are reported to occur in up to 60% of such fetuses. Notably, when absence of the corpus callosum is an isolated fetal ultrasound finding, the reported rate of a relatively normal developmental outcome ranges from 50% to 75%. Well-conducted, long-term studies are lacking. Postnatal follow-up for infants with a history of a finding of agenesis of the corpus callosum on prenatal ultrasound examination should include, at a minimum, close long-term clinical assessment (Chadie et al, 2008; Fratelli et al, 2007; Woodward, 2008). Mild ventriculomegaly is a relatively uncommon fetal ultrasound finding that may be a soft marker for aneuploidy, fetal infection, or other central nervous system abnormalities. As such, it is recommended that serial imaging studies be undertaken, in some cases including a more extensive workup. In the presence of a benign fetal assessment, most infants appear to do reasonably well after delivery. Close pediatric developmental followup and serial imaging studies are important to consider (Leitner et al, 2009; Melchiorre et al, 2009; Sohaey and Filipek, 2008). Cardiac Findings Echogenic cardiac focus is an incidental ultrasound finding in 3% to 4% of healthy fetuses. Notably, there is an increased incidence of twofold to threefold in Asian populations. It is said to be a soft marker for chromosomal abnormalities when associated with other screening abnormalities. If the results of a physical examination of a newborn are normal and there are no other ultrasound findings, no further evaluation is needed (Borgida et al, 2005; Koklanaris et al, 2005; Ouzounian et al, 2007; Sohaey, 2008b). Gastrointestinal Findings Echogenic bowel, when noted to be present during a second-trimester ultrasound examination and determined to be grade 0 or 1 (i.e., less echogenic than bone) is a normal variant. Density greater than that of bone (grade 2 to 3) is abnormal and is potentially a marker for cystic fibrosis, trisomy 21, gastrointestinal anomalies, or in utero infection. Isolated echogenic findings are considered benign with a good prognosis and require no special workup or postnatal management (Al-Kouatly et al, 2001; Patel et al, 2004; Sohaey, 2008a). Cholelithiasis is an uncommon third-trimester fetal ultrasound finding that needs to be differentiated from hepatic calcification. Cholelithiasis is a benign condition requiring no special evaluation or treatment. An imaging examination at 1 year of age for a child with this prenatal finding may be helpful in documenting expected resolution (Agnifili et al, 1999; Sohaey, 2008b). 303 Hepatic calcifications are uncommon fetal ultrasound findings; they are often isolated, single and, in a low-risk pregnant patient, of no significance. When numerous, hepatic calcifications may be markers for fetal aneuploidy, infection, meconium peritonitis, hepatic tumor, or vascular insult. Twenty percent are associated with some form of fetal pathology. Neonatal management depends on the prenatal workup and the clinical presentation in the newborn period (Oh, 2008; Simchen et al, 2002). Urinary Tract Findings Mild fetal pelviectasis is one of the more common abnormalities detected by second-trimester ultrasound examination, with a reported incidence of 0.5% to 5% in unselected pregnant populations. Diagnostic criteria vary, but generally include a second-trimester renal pelvis diameter of 4 to 10 mm and 7 to 10 mm during the third trimester. Some authors consider mild fetal pelviectasis to be a soft marker for trisomy 21. When mild fetal pelviectasis is an isolated finding, the prognosis is good and the condition usually resolves either in utero or during early childhood. In a meta-analysis, it was reported that 11% of children with a history of mild fetal pelviectasis demonstrated postnatal pathology. Authors of a prospective cohort followup study reported uropathy in 18% of the patients. There is a lack of consensus as to postnatal management, with some authorities recommending close clinical assessment and others recommending follow-up renal ultrasound examinations at approximately 1 week and 1 month of life (Coelho et al, 2007; Lee et al, 2006; Sohaey and Arnold, 2008). CAR SEAT TRIALS The observation that preterm infants had episodes of hypoxia when monitored in car seats led the AAP to recommend in 1991 that preterm infants should be observed and monitored for apnea, bradycardia, or oxygen desaturation in their car safety seat before hospital discharge—the so-called car seat challenge (AAP Committee on Injury and Poison Prevention and Committee on Fetus and Newborn, 1991, 1996; Bull et al, 1999). In the United States, the car seat challenge expanded to include late preterm infants, most of whom did not have respiratory problems during their newborn hospital stay. It has been reported that 25% of late preterm infants do not fit securely into standard car safety seats, and 12% of healthy late preterm infants have apneic or bradycardic events in their car seats (Merchant et al, 2001). The authors of a Cochrane Review questioned whether car seat trials actually prevent morbidity or mortality, and whether there are adverse effects of not passing this test, such as prolonging the hospital stay or creating parental anxiety. Their review did not disclose any randomized trials, and they concluded that “it is unclear whether undertaking a car seat challenge is beneficial or indeed whether it causes harm” (Pilley and McGuire, 2006). Since then, there has been one randomized trial in healthy term infants comparing car seats to car beds, and no differences were found in rates of oxygen desaturation or apneic 304 PART VII Care of the Healthy Newborn events (Kinane et al, 2006). Further study is needed to determine whether car seat trials in late preterm neonates are warranted. ROUTINE AND COMMON MEDICAL TREATMENTS PREVENTION OF OPHTHALMIA NEONATORUM AND CONJUNCTIVITIS Approximately15% to 20% of babies will develop conjunctivitis in the first few weeks of life. Conjunctivitis can be caused by a sexually transmitted bacteria, normal skin or nasopharyngeal flora, or chemical irritation (Krohn et al, 1993). In addition, eye discharge can be caused by obstruction of the nasolacrimal duct rather than from the conjunctivitis. The most worrisome infection is that from Neisseria gonorrhea, which can be invasive to the cornea in a matter of hours and lead to blindness. Despite effective preventive measures known since the 1880s, there are still thousands of children blinded by this infection worldwide each year. Most states in the United States have laws or regulations that require administration of topical antibiotic ointment to the conjunctivae of babies within a few hours of birth. This practice has been effective in reducing the cases of blindness caused by gonococcal conjunctivitis. It is moderately effective in preventing conjunctivitis caused by chlamydia. Parents may question the need to expose all babies to eye medication, especially if the mother has been tested and found to be without gonorrhea or chlamydia. Some countries have stopped routine administration of eye prophylaxis. In those countries, an increase in infection, primarily caused by chlamydia, has been noted. With informed consent, parents can opt out of eye prophylaxis for their newborn. There is lay literature recommending the instillation of colostrum or breast milk into the eyes of babies to prevent or treat conjunctivitis. Although colostrum has antimicrobial action, its efficacy has not been adequately studied. Povidone-iodine solution has been shown to be more effective and cause less irritation than erythromycin ointment. It is also less expensive, but is not yet approved for this use by the U.S. Food and Drug Administration (­Isenberg et al, 1995). Whether instillation of povidoneiodine in the eyes may affect newborn thyroid screens—as has been reported with use of this solution on umbilical cord stumps—is not clear (Lin et al, 1994). VITAMIN K Vitamin K is necessary for biologic activation of several human proteins, most notably coagulation factors II, VII, IX, and X. Because placental transfer is limited, cord blood levels of vitamin K1 (phylloquinone) are 30-fold lower than maternal levels. Intestinal bacteria synthesize menaquinone (vitamin K2), which has 60% of the activity of phylloquinone. However, neonates have a decreased number of bacteria in their gut that manufacture vitamin K2; levels of this form of vitamin K are not found in the livers of infants until they are 2 to 3 months old. Therefore newborn infants are deficient in vitamin K at birth and are at risk for significant bleeding. Fortunately intramuscular vitamin K rapidly activates clotting factors, greatly decreasing this risk. Three presentations of vitamin K–deficient bleeding (VKDB) have been described. Early VKDB manifests in the first 24 hours after birth. It is not prevented by postnatal administration of vitamin K, and it usually occurs in infants born to mothers who are taking medications that inhibit vitamin K. Common medications that inhibit vitamin K include many anticonvulsants, isoniazid, rifampin, warfarin, and some antibiotics such as cephalosporins. Early VKDB is frequently serious because of intracranial and intraabdominal hemorrhage. It is estimated that in neonates at risk for early VKDB, the incidence is as high as 12% (Van Winckel et al, 2009). Classic VKDB occurs in infants during the first week of life. Although the clinical presentation is often mild, blood loss can be significant, and intracranial hemorrhages have been reported. Although estimates vary, the incidence of classic VKDB in the absence of vitamin K supplementation is likely 0.25% to 1.7% (AAP Committee on Fetus and Newborn, 2003). Late VKDB occurs in infants between the ages of 2 and 12 weeks and is usually severe. The mortality rate from late VKDB is approximately 20%, and 50% of children with this disorder have intracranial hemorrhages. Late VKDB is associated with exclusive breastfeeding. Human milk contains only 1 to 4 μg/L of vitamin K, whereas commercially available formula contains 50 μg/L or greater. In neonates who do not receive supplemental vitamin K, the incidence of late VKDB is estimated at 4.4 to 7.2 in 100,000 (or 1 in 15,000 to 1 in 20,000; Van Winckel et al, 2009). Vitamin K administered shortly after birth is effective in preventing classic and late VKDB. Since 1961 the recommended dose of vitamin K for term infants born in the United States has been 1 mg given intramuscularly. However, the results of a study suggesting an association between intramuscular vitamin K given at birth and childhood cancer created controversy regarding this practice (Golding et al, 1990, 1992). The results of subsequent studies have indicated conclusively that there is no increased risk for solid tumors in children given intramuscular vitamin K; however, the possibility that there is a slightly increased risk for leukemia cannot be excluded (Puckett and Offringa, 2000). Because of concerns regarding an increased risk of childhood cancers, a switch to oral vitamin K occurred in some countries, but not in the United States. It is apparent that a single oral dose of vitamin K has an efficacy similar to an intramuscular dose in preventing classic VKDB, but offers less protection against late VKDB. Repeated doses of an oral vitamin K preparation until an infant is 8 to 12 weeks old increases the efficacy of this route of administration. However, it is not clear that even multiple doses of an oral formulation of vitamin K are as effective as a single intramuscular dose, given at birth. In a multinational review, rates of late VKDB in infants receiving various regimens of oral vitamin K were mostly in the range of 1.2 to 1.8 in 100,000 compared with 0 cases in 325,000 children receiving an intramuscular dose (Cornelissen et al, 1997). The oral regimens assessed included a dose at birth of 1 mg. Reported rates of VKDB in infants who received CHAPTER 26 Routine Newborn Care 2 mg orally at birth, with doses repeated subsequently, are lower but still somewhat higher than in neonates treated with intramuscular vitamin K (Busfield et al, 2007; Von Kries et al, 2003). Early data from the Netherlands, where infants received 1 mg orally at birth and 25 mcg daily for up to 12 weeks, suggested that this regimen was as efficacious as an intramuscular dose (Cornelissen et al, 1997). However, in a subsequent study from the Netherlands, the rate of late VKDB was 3.2 in 100,000 in a group of infants with undiagnosed biliary atresia who had been treated with this dosing schedule (van Hasselt et al, 2008). Unrecognized cholestatic liver disease is a significant risk factor for VKDB. Finally, no cases of late VKDB were found among 396,000 Danish infants who received an oral dose of 2 mg of vitamin K at birth and 1 mg weekly until the age of 3 months (Hansen et al, 2003). Risks from intramuscular vitamin K include pain at the injection site and the possibility of a serious medication error. The risks of a significant complication from the injection are probably negligible; in one study, zero significant complications were reported after 420,000 injections (Von Kries, 1992). In the United States, oral administration is complicated by the lack of an oral vitamin K preparation being licensed for newborns. In some settings, infants have received the intramuscular preparation orally; however, tolerance may be a problem, and the efficacy of this preparation when given orally might not be comparable to the oral formulations used in Europe. In addition, compliance with repeated doses of oral vitamin K in infants may be suboptimal. Finally, it is unknown whether the use of repeated administration of oral vitamin K in the dose range of 1 to 2 mg each week is associated with an increased risk of childhood cancers. For parents who have questions regarding the best method to prevent classic and late VKDB, the clinician is advised to discuss the pros and cons of intramuscular versus oral vitamin K. If the parents choose oral administration, a dose of 2 mg of vitamin K should be given shortly after birth, with subsequent doses until the infant is at least 4 weeks old if he or she is breastfed. In a policy statement, the AAP suggests that, if an oral vitamin K formulation becomes licensed for use in the United States, a dose of 2 mg can be given at birth and repeated at 1 to 2 weeks of age and at 4 weeks of age for neonates whose parents decline intramuscular vitamin K (AAP Vitamin K Ad Hoc Task Force, 1993). CIRCUMCISION Neonatal circumcision is a polarizing issue for health care professionals and parents. Those who favor routine circumcision highlight health benefits such as decreased urinary tract infections (UTIs), reduced risk of penile cancer, and possibly lower rates of sexually transmitted diseases, including HIV (Schoen, 2008). Those who oppose the procedure indicate that the number of circumcisions needed to be performed to prevent one of these outcomes (i.e., the number needed to treat [NNT]) is large, the risks of the procedure balance out these benefits, circumcision leads to loss of sexual sensation, and subjecting a neonate to a painful procedure without clear benefits may be unethical (Andres, 2008). 305 It is clear that circumcision reduces the risk of UTI by approximately 10-fold; however, given the low incidence of UTI in male infants, 100 boys need to be circumcised to prevent one UTI. Similarly, although circumcision has been shown to prevent penile cancer, it is an extremely rare condition and the NNT is about 900 (Christakis et al, 2000). There has been recent interest in circumcision as a method for preventing HIV. The results of studies in three African countries indicate that circumcision reduces the risk of HIV infection by 56%. Because the incidence of HIV in these countries is high, the calculated NNT is approximately 72 (Mills et al, 2008). Given the same efficacy, the NNT to prevent one case of HIV infection in Canada, where the incidence is 13 in 100,000, or in the United States, where the incidence is 23 in 100,000, is greater than 5000 (Andres, 2008; Hall et al, 2008). Circumcision is generally a safe procedure. Although some increased bleeding is reported after 1% of circumcisions, the rate of significant complications is approximately 0.2% (Christakis et al, 2000; Gee and Ansell, 1976; Wiswell and Geschke, 1989). Bleeding, sometimes requiring suturing of a vessel, is the most common significant complication, followed by penile injury and infection. Infection is more common after a circumcision using a Plastibell rather than a Gomco clamp; hemorrhage is reportedly similar after either technique (Gee and Ansell, 1976). Circumcision is an uncomfortable experience for the neonate. Small amounts of sucrose solutions can be offered to the baby for soothing. Pain from the actual surgery can be significantly decreased with the use of a dorsal penile nerve block or ring block. In one study, 65% of infants who had received a dorsal nerve block had no or minimal response to the initial clamping of the foreskin (Taeusch et al, 2002). A poor cosmetic outcome can be caused by removal of too little foreskin. It has been estimated that 1-9.5% of circumcisions are redone because of parental concern regarding the appearance. In a prospective study, among boys younger than 3 years who had been circumcised using either a Plastibell or with a Mogen clamp, the glans was fully exposed in only 35.6%. However, in older circumcised males the glans was fully exposed in more than 90% (Van Howe, 1997). This finding suggests that parents of a circumcised infant should be counseled that the vast majority of properly done circumcisions will lead to an acceptable cosmetic appearance over time. In the United States, the Gomco clamp is the most commonly used apparatus for performing circumcisions, followed by the Plastibell and Mogen clamp (Stang and Snellman, 1998). The use of the Mogen clamp leads to shorter procedures and, reportedly, less pain and bleeding than the other techniques (Kurtis et al, 1999; Taeusch et al, 2002). However, less foreskin is removed with the Mogen clamp than with the other two techniques (Alanis and Lucidi, 2004). HEPATITIS B VACCINE The implementation of routine hepatitis B immunization during infancy has been associated with a dramatic decrease in the incidence of this infection. Between 1990 306 PART VII Care of the Healthy Newborn (before routine vaccination of infants) and 2004 the overall incidence of acute hepatitis B in the United States declined by 75%, and by 94% among children and adolescents (Centers for Disease Control and Prevention, 2005). The Centers for Disease Control and Prevention and the AAP recommend that the initial dose of the three-dose hepatitis B immunization series be given in the newborn nursery; however, this recommendation is far from being followed universally. In 1999, it was estimated that 54% of newborns in the United States received a dose of hepatitis B vaccine at birth. However, at that time a recommendation was made to suspend birth dosing until vaccines were made that did not contain thimerosal (a preservative containing mercury). Even after a thimerosal-free hepatitis B vaccine was produced, many newborn nurseries in the United States did not immediately resume newborn immunization programs. This issue, combined with the reluctance of some parents to have their newborns immunized, has led to continued suboptimal rates of administration of the birth dose of hepatitis B vaccine (Clark et al, 2001). Among infants born in the United States between 2003 and 2005, it was estimated that 50.1% received a dose of the vaccine by the age of 3 days (Centers for Disease Control and Prevention, 2008). There are at least two advantages of providing the first dose of hepatitis B vaccine during the newborn nursery stay. First, newborns who receive a dose at birth are more likely to complete their hepatitis B immunization series on time than those who receive a first dose later (Yusuf et al, 2000). Second, because a dose of hepatitis B vaccine given within 12 hours of birth can prevent vertical transmission of hepatitis B infections in 75% to 90% of cases, early provision of immunization serves as a safety net in cases where there has been an error in identifying a mother who is positive for hepatitis B surface antigen (Centers for Disease Control and Prevention, 2005). The main disadvantage of providing a dose of hepatitis B vaccine during the nursery stay is that it can complicate documentation of hepatitis B immunization status in a child by increasing the number of vaccination providers. There is no evidence that administration of a dose of hepatitis B vaccine at birth leads to more evaluations for sepsis because of adverse events related to the immunization. ONGOING CARE UMBILICAL CORD CARE AND SINGLE UMBILICAL ARTERY Umbilical cord care recommendations vary from “dry cord care” to the use of dyes and cleansing with alcohol, soap and water, or antiseptics. Concern over the possible toxic effects of dye and antiseptics led many hospitals in the United States to adopt the dry cord care method of cord care. Unfortunately this method may be responsible for causing an increase in the risk for omphalitis (Janssen et al, 2003; Simon and Simon, 2004). In addition, the results of a randomized trial in Nepal indicate that cord care using topical chlorhexidine reduces the risk of developing omphalitis; however, that population may be at higher risk compared with those born in the United States (Mullany et al, 2006). Because omphalitis is rare, and its more severe complication necrotizing fasciitis is even rarer, large trials are needed to determine which cord care regimen is best for preventing these complications. Until such trials are conducted, there is no clear advantage of one regimen over another. Providers caring for newborns need to keep these diagnoses in mind and encourage parents to report redness around the umbilical cord stump. A single umbilical artery is detected in about 4 out of 1000 births. There is an association of the single umbilical artery with a number of congenital anomalies, including renal or genitourinary malformations, cardiac malformations, and chromosomal anomalies such as Down syndrome. In this era of near universal use of prenatal fetal ultrasonography, any associated anomalies are usually discovered (Deshpande et al, 2009; Johnson and Tennenbaum, 2003). Unless there are abnormalities noted on physical examination, there is no need to repeat diagnostic ultrasound examinations after birth in a child with a single umbilical artery. BREASTFEEDING There is voluminous evidence that the optimal feeding for normal neonates is human milk provided via the mother’s breast. From this most preferred feeding there is a continuum to least preferred feeding that includes: maternal milk given by an artificial method, donor human milk (although this is rarely used by most health care facilities), and commercially available infant formula. It is incumbent on health care professionals and health care systems to vigorously promote the optimal feeding for normal newborns. This promotion includes education to prospective mothers, knowledge of the characteristics of human milk and the normal course of lactation, organization of health care facilities to optimize the initiation of breastfeeding after delivery, early recognition of suboptimal breastfeeding and interventions to correct problems, and adoption of a parsimonious list of contraindications to breastfeeding. Nutritional Composition of Human Milk Although the composition varies among mothers, by age of the child, and even within a feeding, the nutritional components of mature human milk can be summarized (Picciano, 2001). The caloric content of human milk is approximately 0.67 Kcal/mL (20 Kcal/ounce). Approximately 50% of these calories are provided by the lipid components. Importantly, human milk contains omega-3 fatty acids. Lactose is the principal carbohydrate, providing approximately 40% of the calories. The protein in human milk is divided into two categories, casein and whey, that are provided in a 40:60 ratio. The sodium content of human milk is low, averaging approximately 7 mEq/L. With the exception of vitamins D and K, the amounts of minerals and micronutrients available are all adequate for optimal infant growth. The evolutionary advantage afforded by the relatively low concentrations of vitamin D and vitamin K in human milk is unclear. Healthy newborns receive surprisingly little breast milk in the first few days of life. Average intake of colostrum during the first day of life is approximately 10 mL/kg and CHAPTER 26 Routine Newborn Care 20 mL/kg during the second 24 hours of life (Casey et al, 1986; Evans et al, 2003). Maternal milk production dramatically increases during the period from 36 to 96 hours of life; this increase in quantity is accompanied by a change from colostrum to mature milk (Casey et al, 1986; Saint et al, 1984). The increase in milk production is perceived by the mother as breast fullness. In one study, mothers noted that their breasts were noticeably fuller when their infants were an average of 53 hours old. This finding was closely correlated with the mean age (58 hours) when infants transferred greater than 15 mL of milk from the breast (Dewey et al, 2003). Given the low volume of milk provided initially, neonates have a decrease in weight and an increase in serum sodium during the first few days of life (Marchini and Stock, 1997). Average maximal weight loss in breastfed infants is 5% to 7% of birthweight and occurs between 48 and 72 hours of life (Macdonald et al, 2003; Marchini and Stock, 1997; Rodriguezet al, 2000). With the onset of copious production of mature milk, neonates begin to gain weight and their serum sodium levels fall (Marchini and Stock, 1997). Infants fed human milk regain their birthweight, on average, by the age of 8.3 days; 97.5% have regained their birthweight by 21 days of age (Macdonald et al, 2003). Health Benefits of Human Milk There is copious research on numerous health benefits for infants who are breastfed. The list of benefits that have been found include: decreased incidence of conditions such as gastrointestinal infections, lower respiratory tract disease, otitis media, hypertension, obesity, diabetes, allergies, and asthma; improved cognitive development; and reduced risk of sudden infant death syndrome (Hoddinott et al, 2008). Unfortunately, because it is unethical to conduct individual randomized controlled trials on ­breastfeeding, most of these benefits have been documented in observational studies. Such studies are subject to numerous problems that might bias the results toward the null hypothesis or an overestimation of the benefits (Kramer et al, 2009). Perhaps the biggest problem with observational studies on breastfeeding is that there is significant confounding by socioeconomic status, for which it is difficult to adequately account in analyses. The best evidence of the health benefits of breastfeeding comes from a large trial conducted in Belarus. Health care professionals at participating hospitals were randomly assigned to receive training on promoting breastfeeding using the Baby-Friendly Hospital Initiative (BFHI) developed by the World Health Organization and the United Nations Children’s Fund (BFHI is further described in a following section) or to continue to provide standard care. A total of 17,046 mother-infant dyads were enrolled in the study (Kramer et al, 2001). All participating infants were born at 37 weeks’ gestation or later and weighed more than 2500 g; all the mothers intended to breastfeed their infants. The rate of any breastfeeding at 3 months old was significantly higher among infants born in BFHI hospitals than at control sites (43% and 6%, respectively). Rates of breastfeeding were also significantly higher at BFHI locations at 6 and 12 months of age. Most pertinent, rates of 307 BOX 26-2 B  aby-Friendly Hospital ­Initiative: Ten Steps to Successful ­Breastfeeding 1. Maintain a written breastfeeding policy that is routinely communicated to all health care staff. 2. Train all health care staff in the skills necessary to implement this policy. 3. Inform all pregnant women about the benefits and management of breastfeeding. 4. Help mothers initiate breastfeeding within 1 hour of birth. 5. Show mothers how to breastfeed and how to maintain lactation, even if they are separated from their newborns. 6. Give infants no food or drink other than breast milk unless medically indicated. 7. Practice “rooming-in”—allow mothers and infants to remain together 24 hours a day. 8. Encourage unrestricted breastfeeding (i.e., on demand). 9. Give no pacifiers or artificial nipples to breastfeeding infants. 10. Foster the establishment of breastfeeding support groups and refer mothers to them on discharge from the hospital or clinic. gastrointestinal infection and atopic dermatitis were significantly lower in babies who had been born at BFHI hospitals. These data provide strong evidence for a significant beneficial effect of breastfeeding. Among children enrolled in the original study who were assessed at 6.5 years of age, the verbal intelligence quotient of those born at BFHI hospitals was 7.5 points higher than those born at control sites (Kramer et al, 2008). Whether this finding was related to the components in the breast milk or from the physical act of breastfeeding is unclear; regardless of the mechanism, the effect is profound. In other follow-up studies, no difference between the two groups of children were found in the rates of asthma, allergy, or obesity (Kramer et al, 2007); however, the design of the study provides a conservative estimate of the effects of breastfeeding. Promotion of Breastfeeding Given the demonstrable improvements in outcomes, health care providers should promote breastfeeding as the preferred method of feeding newborn infants and facilitate its initiation during the newborn nursery stay. As a comprehensive program, implementation of the 10 steps of the BFHI (Box 26-2) has been shown to significantly increase the rates of breastfeeding (Kramer et al, 2001). There is evidence to support the efficacy of each of the 10 separate steps, although for some of the steps, such as excluding pacifiers, the evidence is contradictory (Cramton et al, 2009; O’Connor et al, 2009). From a practical standpoint there are several evidencebased interventions during the newborn nursery stay that increase the rate or prolongation of breastfeeding. These interventions include the use of frequent demand feedings as opposed to a rigid feeding schedule, early skin-to-skin contact between mother and infant, professional advice on breastfeeding techniques, and exclusion of commercial formula from discharge packs (Anderson et al, 2003; Britton et al, 2007; Donnelly et al, 2000; Renfrew et al, 2000). 308 PART VII Care of the Healthy Newborn Breastfeeding Problems The vast majority of difficulties with breastfeeding are related to a delay in transfer of adequate quantities of human milk to the infant. Delayed lactogenesis occurs in 20% to 30% of mothers; however, in most instances the primary problem is related to infant breastfeeding behaviors in the first few days of life. Although the clinical correlate is not well defined, weight loss of greater than 10% of birthweight is considered excessive in breastfed infants. Studies on breastfed neonates indicate that approximately 10% of infants lose more than 10% of their birthweight during the first few days of life (Dewey et al, 2003). It is important to identify mother-infant dyads who are at risk for breastfeeding problems, so that early interventions can prevent excessive weight loss. Mothers with previous breast surgery, particularly breast reduction, are at increased risk of primary insufficient lactation. Flat or inverted nipples can also make breastfeeding more difficult. Prolonged labor and cesarean delivery have been associated with delayed onset of milk production (Dewey et al, 2003). The adequacy of breastfeeding behaviors can be assessed during the newborn nursery stay using scoring systems such as the Infant Breastfeeding Assessment Tool. Low scores on this measure during the first day of life are moderately predictive of excessive weight loss in the neonate (Dewey et al, 2003). Decreased numbers of voids and stools in the newborn are also helpful in identifying children with breastfeeding problems, but this information is most useful after day 3 of life. Supplementation of Breastfeeding It is usually unnecessary to provide any nutrition or fluid to full-term breastfed infants beyond human milk. Dextrose water or commercial formula may be needed in neonates with hypoglycemia who are not responsive to breastfeeding. Supplementation may also be indicated in newborns who have lost more than 10% of birthweight or have decreased urine and stool output. Supplementation should be considered a temporary intervention, and its provision should not interfere with the onset of successful breastfeeding. Contraindications to Breastfeeding The few absolute contraindications to breastfeeding include maternal HIV infection, untreated tuberculosis in the mother, evidence of current cocaine use or antimetabolite drugs in the mother, and galactosemia in the neonate (Gartner et al, 2005). There are a myriad of other drugs for which there is concern regarding long-term neurodevelopmental outcomes in the infant. Selective serotonin reuptake inhibitors are commonly used to treat depression and anxiety in young women. Among drugs in this category, sertraline and paroxetine are thought to be the safest for use in breastfeeding mothers, whereas fluoxetine and citalopram are believed to have the most potential for toxicity in the neonate (Field, 2008). Overall there have been few adverse effects noted with use of any of these drugs, and generally the potential risks associated with these medications are thought to be outweighed by the benefits of breastfeeding (Field, 2008). Similarly, although methadone is detectable in the breast milk of women receiving this medication, serum levels in neonates are low and unlikely to cause a significant effect (Jansson et al, 2008). Hepatitis C virus RNA has been found in the milk of mothers infected with this virus. Despite this finding, transmission of infection via breastfeeding has not been documented. Maternal hepatitis C infection is not considered a contraindication to breastfeeding (Gartner et al, 2005). BOTTLE FEEDING Commercial formula that provides adequate nutrition, vitamins, and minerals is available for infants of mothers who do not wish to breastfeed their infants or when breastfeeding is contraindicated or impossible. There are three major categories of formula used in neonates: cow milk–based, soy, and hydrolyzed formula. Of these, cow milk–based formula is used most commonly. The main carbohydrate in cow milk–based formula is lactose. Soy formulas were developed for infants with a possible cow’s milk allergy. Because the main carbohydrate in soy formulas is sucrose or corn syrup, soy formula can be used in neonates with possible galactosemia. Protein hydrolysate formulas were initially developed for use in infants who are highly intolerant to cow’s milk protein (Kleinman, 2009). These formulas are purported to lead to fewer allergies in babies and children than does cow milk–based formula, but the evidence for this is limited (Osborn and Sinn, 2006). All extensively hydrolyzed formulas are lactose free (Kleinman, 2009). These formulas are indicated in infants with definitive evidence of cow’s milk protein allergy because 10% to 14% of them also have a soy allergy (Bhatia and Greer, 2008). Standard preparations of formulas available for use in healthy term neonates provide 0.67 Kcal/mL. Most formulas are fortified with 10 to 12 mg/L of iron; however, some low-iron cow milk–based formulas are available. It is recommended that all formula-fed newborns receive the iron-fortified products. Vitamin D at the concentration of approximately 400 IU/L is provided in all the commercially available formulas (Kleinman, 2009). Mothers who elect to bottle-feed report feeling unsupported for their decision by health care professionals, and up to 50% feel pressured to breastfeed (Lakshman et al, 2009). Although the benefits of breastfeeding should be provided to mothers who have not decided how to feed their babies, the role of health care providers is to support the decision of those who have elected to bottle-feed. It is also important to provide practical education about bottle-feeding to these parents; this is frequently neglected in many newborn nurseries (Lakshman et al, 2009). Newborns who are bottle-fed can feed ad libitum beginning shortly after birth. Average formula intake in term newborns during the first day of life is 15 to 20 mL/kg and 40 to 45 mL/kg during the second day. During the nursery stay, neonates who are formula fed typically lose less weight than breastfed infants (Dollberg et al, 2001). ANTICIPATORY GUIDANCE A primary duty of providers of newborn care is to ensure that parents of new infants have the knowledge and skills to provide for normal growth and development. Parents CHAPTER 26 Routine Newborn Care who are taught about normal newborn development and behavior have more realistic expectations about the work involved and look upon their child with more fondness. Conversely, before discharge from the newborn nursery, it is important to assess the parents’ ability to provide a safe and nurturing environment for the neonate. Parents showing concerning behaviors, possibly leading to abuse or neglect, should have supervision and interventions to help them, possibly leading to termination of parental rights (Davidson-Arad et al, 2003; Wattenberg et al, 2001). There are several major challenges for parents of normal newborns: sleep deprivation, learning to calm a crying infant, significant life changes, and the new worries that come with being responsible for a totally dependent being. Postpartum depression is more common and of longer duration than previously thought, and it occurs in at least 10% of mothers. This condition is related to sleep deprivation and has major and long-lasting effects on infant homeostasis and development (Chaudron, 2003). Anticipatory guidance should be given to help prepare new parents for the common tasks of newborn care and to educate them about the many normal variations in newborn behavior. Learning how to soothe a baby is one of the first needed parenting tasks. Providers can help by giving suggestions to reduce crying and to better cope with infants who are more sensitive and harder to soothe (Barr et al, 2009). Most parents have questions about feeding, elimination, bathing, cord care, genital care, jaundice, and common rashes. There are numerous checklists of educational topics that can be overwhelming to new parents. In addition, learning styles can vary, with some preferring written materials and while others preferring audio-visual materials or hands-on demonstration. Ideally, education should be targeted toward the topics of interest and with the appropriate materials for learning style (Dusing et al, 2008). Mothers are often not in a good learning state in the immediate postpartum period because of pain, postpartum hormonal changes, and the stress of being in a hospital. However, there is heightened receptivity to change during this period, so attempts to teach or make lifestyle changes (e.g., smoking cessation) may be more effective. There is some evidence that providing parental education using tools such as interactive video and computers may be superior to traditional teaching (Snowdon et al, 2009; Trepka et al, 2008). Given the obstacles to providing meaningful education during the nursery stay, it is probably better for practitioners to focus on a few key points of anticipatory guidance rather than reciting a litany of instructions. There is also a philosophical choice in deciding whether to emphasize the overall health of a newborn or to concentrate on prevention or identification of illness. There is little evidence for the efficacy of most anticipatory guidance provided to parents during the newborn nursery stay. A notable exception is the advice to put infants to sleep in the supine position (described in the following section). There is also emerging evidence that education about the normality of inconsolable crying in infants helps parents cope with this 309 stressful situation, and it could reduce the risk of shaken baby syndrome (Barr et al, 2009). Sleep Position With the exception of immunizations, no child health intervention in the past two decades has resulted in a larger decrease in postneonatal infant mortality than the “Back to Sleep” campaign. The remarkable change in the predominant sleep position of infants from prone to supine has led to a 30% to 50% reduction in the rate of sudden infant death syndrome (SIDS) in the United States (AAP Task Force on Sudden Infant Death Syndrome, 2005). A multipronged effort including brochures, public service announcements, and education provided by health care professionals was used to affect the change in sleep position (Willinger et al, 2000). Obviously education provided to parents during the newborn nursery stay is a crucial determinant of the sleep position of an infant. In addition to providing education, there is evidence that parents model sleep position for their babies after how they saw nurses and physicians place their neonate in the bassinet in the newborn nursery (AAP Task Force on Sudden Infant Death Syndrome, 2005; Colson and Joslin, 2002). Therefore it is crucial that neonates are placed on their backs to sleep in the newborn nursery. In addition, there is an additive effect of both physicians and nurses recommending the supine sleep position (Willinger et al, 2000). In addition to supine position, there are other factors related to the sleep environment that can effect the risk of SIDS in a newborn infant. It is recommended that infants sleep on firm surfaces and without excessive bedding such as pillows. Many experts also recommend against ­co-sleeping between parents and infants; however, this topic is controversial and the evidence is somewhat contradictory. Similarly, although use of a pacifier has been found to reduce the risk of SIDS, there is a reluctance to recommend these devices because of concerns about reducing breastfeeding (AAP Task Force on Sudden Infant Death Syndrome, 2005; Willinger et al, 2000). DISCHARGE AND FOLLOW-UP For infants born in the United States at 35 weeks’ gestation or later, the average length of the initial hospital stay is 48 to 52 hours (Datar and Sood, 2006; Kuzniewicz et al, 2009; Paul et al, 2009). Because approximately 50% of infants born by vaginal delivery are discharged before the age of 48 hours, and because up to 40% of those born by cesarean delivery are discharged before 72 hours of age, a large proportion of neonates are discharged before the age of 3 to 4 days, when bilirubin levels typically peak and breastfeeding is well established (Paul et al, 2006). It is recommended that infants discharged before 48 hours have a follow-up appointment with a provider within 48 hours (AAP Committee on Fetus and Newborn, 2004). This follow-up can be accomplished either by a visit to a health care provider or via a home nursing visit. Risk factors for readmission after an initial hospital stay of less than 48 to 72 hours include gestational age less than 39 weeks (and especially less than 37 weeks), 310 PART VII Care of the Healthy Newborn primiparous mother, and Asian race (presumably because of an increased risk of hyperbilirubinemia) (Burgos et al, 2008; Grupp-Phelan et al, 1999; Liu et al, 1997; Paul et al, 2006). Consideration of a longer nursery stay is suggested for infants with one or more of these risk factors. In addition, early discharge is not recommended for term newborns who have not voided, passed at least one stool, or demonstrated adequate breastfeeding (AAP Committee on Fetus and Newborn, 2004). However, there is little evidence to support these recommendations. COMMON PROBLEMS DURING THE NURSERY STAY HYPOTHERMIA AND HYPERTHERMIA Upon leaving the womb, a newborn is immediately challenged with maintaining a normal body temperature. If a neonate is not quickly dried at birth, he or she may lose up to 1° C body temperature per minute. Healthy term babies are able to increase heat production through glycogenolysis and nonshivering thermogenesis for minutes to a few hours, depending on environmental conditions (Aylott, 2006). Babies typically have a decline in body temperature during the first hour of life with a gradual increase during the following 12 hours (Li et al, 2004). By the second day of life, the infant’s body temperature becomes more stable, but heat loss can occur again with bathing or other stresses (Takayama et al, 2000). Being too cold or too hot causes metabolic stress to the newborn, so efforts to maintain a steady and neutral thermal environment should be provided. The best practice is to dry the baby immediately after delivery and place the infant skin-to-skin with the mother. Although the AAP and the American College of Obstetricians and Gynecologists jointly recommend keeping infants’ core temperatures within the narrow range of 36.5° to 37° C, in one study of healthy term newborns the average temperature was 36.5° C, with a normal range from 36.0° to 37.9° C (Takayama et al, 2000). Thin babies tend to have lower body temperatures, and heavier babies tend to have higher body temperatures. Hypothermia should be managed by placing the baby skin-to-skin with a parent or under a radiant warmer. Standard practice at most nurseries is to measure axillary temperatures, probably because of reports in the 1960s and 1970s of perforations caused by rectal thermometers; however, axillary temperatures may not always accurately reflect core temperature (Hutton et al, 2009). An elevated body temperature at birth generally reflects the intrauterine temperature and is not usually a sign of sepsis (Baumgart, 2008). Isolated hyperthermia during labor is associated with neonatal encephalopathy, occurring in approximately 1 in 2000 births (Blume et al, 2008). After the first 3 to 4 days of life, increased temperatures are most likely caused by dehydration from suboptimal breast milk supply (Maayan-Metzger et al, 2003). A single increased temperature in an otherwise normally behaving newborn is not a strong predictor of infection, but has been reported as a sign of intracranial hemorrhage (Fang et al, 2008). ELIMINATION Voiding Approximately 15% of healthy newborns void at the time of delivery, and 95% void by 24 hours of age. The cause of delayed voiding is likely a consequence of stress on the infant during labor and delivery (Vuohelainen et al, 2007; 2008), which is a protective mechanism for the baby. Normally no intervention is needed once homeostatic adaption to extrauterine life is stable. The differential diagnosis of delayed voiding (defined as no urine output by 24 to 48 hours of age) includes renal and postrenal causes. With the frequent use of prenatal ultrasound examination, it is unusual that a significant renal anomaly is discovered because of a delay in voiding. Most infants with bilateral renal agenesis have other findings, such as oligohydramnios or Potter sequence. Unilateral renal agenesis does not usually give symptoms of decreased urine output. Renal vascular thrombosis can also cause anuria, and babies with this condition are usually ill. Severe cystic kidney disease can involve urinary outflow obstruction. The diagnosis of cystic kidneys is usually made after the newborn period, or it is found incidental to evaluation of other anomalies and not because of delayed voiding. Postrenal causes of delayed voiding include neuropathic bladder dysfunction and anatomic obstruction of urinary flow by anomalies in ureters, the bladder, or the urethra. Persistent or recurrent bladder distention after catheterization is found with occult lower spinal cord anomalies. Presacral teratoma or other tumors can cause compression and urinary blockage as well. In male infants, there is the possibility of posterior urethral valves. Physical findings of loose abdominal skin or musculature and a distended bladder suggest this diagnosis. In a healthy-appearing newborn with a normal result on prenatal ultrasound examination, the absence of enlarged kidneys or palpable suprapubic mass, allowing up to 72 hours for a spontaneous first void, will avoid excessive testing. In fussy neonates, infants with other genitourinary abnormalities, enlarged kidneys, or distended bladder, testing should begin immediately. Ultrasound examination of bladder, kidneys, and posterior urethra is often diagnostic. Normal newborns have decreased renal concentrating ability and excessive extracellular free water at birth. As a result, neonates will continue to void despite low intake of fluids. This process is normal, with excess fluid volumes being intrinsically protective against dehydration. Conversely, delayed voiding is not indicative of dehydration in the first 72 hours of life. Defecation Similar to the first void, the first passage of meconium occurs by an average of 7 hours of age. One third of newborns pass meconium before their first feeding. Late preterm newborns tend to pass meconium later than term infants, and 32% of preterm infants do not pass meconium in the first 2 days of life. In the first few days of life, intake is not well correlated with meconium output. However, CHAPTER 26 Routine Newborn Care the number of wet and soiled diapers reflects adequacy of breast milk production on day 4 of life. Fewer than four soiled diapers on day 4 correlates with inadequate milk production (Nommsen-Rivers et al, 2008). By 2 weeks of age, breastfed infants pass feces more frequently than bottle-fed infants; they also have larger variability in time between bowel movements (Sievers et al, 1993). Because 99.7% of healthy newborns pass meconium by 34 hours of age, those who are delayed beyond that time deserve extra vigilance during examination to avoid missing obstructions, such as an imperforate anus (Metaj et al, 2003). A baby with abdominal distention or vomiting and delayed stooling deserves evaluation for a possible gastrointestinal tract obstruction. JAUNDICE There are few conditions in newborn infants that create as much controversy and clinician and parental angst as does hyperbilirubinemia. Since the discovery of phototherapy in 1956 and its integration into medical care in the 1960s, the standard management of neonatal jaundice in the United States has gone through three distinct phases. Until the early 1990s, clinicians visually monitored full-term neonates during their 2- to 5-day newborn nursery stay and obtained serum bilirubin levels on those with significant jaundice. Phototherapy was begun when the total bilirubin was 15 mg/dL, and an exchange transfusion was indicated if the level rose to 20 mg/dL (Watchko et al, 1983). The wisdom of this approach was challenged by several significant events. First, there was a growing awareness of the increase in costs associated with the prolonged hospitalization of a healthy infant for phototherapy to treat a relatively modest level of hyperbilirubinemia. Second, Kemper et al (1989, 1990) conducted a series of studies documenting that mothers of neonates who received phototherapy were at risk of overmedicalizing their children. Finally and most significantly, informal and formal reviews of data on jaundice in full-term newborns, without hemolytic disease, revealed that the risk of kernicterus in such infants was extraordinarily low (Newman and Maisels, 1992; Watchko et al, 1983). Based on this evidence, a “kinder and gentler” approach to the management of hyperbilirubinemia in term infants was advocated, leading to the AAP practice parameter in 1994 (AAP Provisional Committee for Quality Improvement and Subcommittee on Hyperbilirubinemia, 1994). Under this guideline, phototherapy for a healthy, 72-hour-old, full-term newborn was not definitively recommended unless the serum bilirubin was 20 mg/dL or greater. Unfortunately, publication of this guideline corresponded to a shortening of the nursery stay by term infants to as little as 24 hours. Therefore infants were discharged home before their levels of bilirubin peaked at 3 to 4 days of life, and there were numerous reports of infants with extremely high bilirubin levels and a general impression that the incidence of kernicterus was increasing. In retrospect, it does not appear that the incidence of kernicterus increased, but case reports and anecdotal evidence led to significant consternation by clinicians, parents, and quality assurance organizations (Burke et al, 2009). 311 Because newborns are usually discharged well before bilirubin levels reach their peak, it is clear that predictive models were needed to assess risk in newborns who are discharged early. Bhutani et al (1999) developed a nomogram based on data from neonates in whom serum bilirubin levels were measured multiple times. Infants who had initial bilirubin levels above the 95th percentile for any time period were significantly more likely to have “significant hyperbilirubinemia” detected in subsequent bilirubin measurements. These data were used in the development of the 2004 AAP practice guideline (AAP Subcommittee on Hyperbilirubinemia, 2004b). With this iteration of the guideline, clinicians are provided hourly guidance on levels of bilirubin for which phototherapy or exchange transfusions are indicated. Separate curves for low-risk, medium-risk, and high-risk neonates have been developed. Internet-based tools are available that provide the appropriate management for an infant with a specific bilirubin at a specific hour of life. Although the AAP bilirubin nomogram provides useful information, there are some caveats. First, in the study by Bhutani et al (1999), 39.5% of infants with an initial bilirubin level greater than the 95th percentile had significant hyperbilirubinemia at subsequent testing. Therefore more than 60% of such infants had less severe hyperbilirubinemia when retested. The nomograms for medium- and high-risk infants are based almost exclusively on expert opinion rather than actual data. Given these limitations, it is important for the clinician to determine management of a jaundiced neonate on an individual basis, remaining cognizant of the infant’s medical condition, social situation, and parental preferences. There are numerous neonatal conditions that increase the risk for hyperbilirubinemia (Dennery et al, 2001); chief among these is hemolysis secondary to maternal antibodies to red blood cell antigens. Thankfully, hemolysis secondary to antibodies to Rh factor is rare because of proper management of Rh-negative mothers. Because there is no way to prevent hemolysis from an ABO incompatibility, it may be useful to test cord blood from neonates born to mothers with blood type O for blood type and the presence of antibodies on their red cells (i.e., Coombs’ test) at birth. The increase in bilirubin secondary to ABO compatibility is highly variable, even with a positive direct Coombs’ test. Some infants will have an early and dramatic rise in serum bilirubin and evidence of hemolysis, whereas in others no effect can be detected clinically. In addition to ABO incompatibility, some women will have antibodies to minor red cell antigens that can usually be diagnosed prenatally. In most instances the increases in bilirubin associated with antibodies against minor antigens are mild. Other neonatal conditions that are risk factors for hyperbilirubinemia include bruising secondary to birth trauma and polycythemia. Decreased intake of breast milk can lead to decreased passage of stool. Because intestinal bacteria break down conjugated bilirubin to the unconjugated form, a decrease in stooling can lead to increased reabsorption of this unconjugated bilirubin (enterohepatic circulation). Breastfeeding is a significant risk factor for hyperbilirubinemia particularly when intake is limited. The propensity for developing significant jaundice is variable in 312 PART VII Care of the Healthy Newborn different racial groups. Asian and American Indian infants are at the highest risk for significant hyperbilirubinemia (Dennery et al, 2001). Finally, late preterm infants are at significantly increased risk for significant hyperbilirubinemia and kernicterus. Traditionally, visual assessment has been used to judge whether a newborn infant has significant jaundice. This method of assessment is moderately accurate, but may be sufficient to rule out the need for serum bilirubin testing in many full-term newborns with no risk factors for jaundice (Moyer et al, 2000; Riskin et al, 2008). Transcutaneous bilirubinometers offer a noninvasive method for screening for hyperbilirubinemia. Depending on the technology and brand used, these instruments generally provide estimates of transcutaneous bilirubin (TcB) that correlate well with serum values. However, in practice it is not clear whether categorical estimates of the severity of neonatal jaundice using the transcutaneous meter are actually better than a visual assessment (Kaplan et al, 2008). Because the TcB may be lower or higher than the serum level in actual clinical practice, it should be considered as a screening tool only. Serum bilirubin testing remains the standard on which management decisions are based. Transcutaneous testing may be warranted for infants with risk factors for significant hyperbilirubinemia, even in the absence of significant jaundice. Unless levels are high enough to require an exchange transfusion, phototherapy is effective for treating an infant with significant hyperbilirubinemia. Although repeated measurements of direct bilirubin are not cost effective, one assessment is helpful before or immediately after initiating phototherapy to rule out direct hyperbilirubinemia (Newman et al, 1991). An assessment of the potential for hemolysis as the etiology of the elevated bilirubin, possibly including a review of maternal and infant blood type, direct Coombs’ test, hematocrit, reticulocyte count, and red cell morphology may also be useful. There is no conclusive evidence as to whether continuous phototherapy leads to more rapid reduction in serum bilirubin levels than does intermittent treatment (AAP Subcommittee on Hyperbilirubinemia, 2004b; Lau and Fung, 1984). Unless bilirubin levels are approaching exchange transfusion levels, it is probably reasonable to discontinue treatment for several minutes to 1 hour at frequent intervals to allow parents to feed and hold their baby. Serial bilirubin measurements are needed to determine the adequacy of therapy and to determine when phototherapy can be discontinued. A rebound bilirubin level obtained 24 hours after discontinuation of phototherapy may be helpful in some clinical situations. RESPIRATORY COMPLICATIONS The term or late preterm fetus accomplishes the transition from dependency on the placenta to the newborn cardiorespiratory system, for the most part, without incident. After birth, pulmonary blood flow increases, fetal shunts reverse and begin to close, spontaneous breathing effort is initiated, and fetal lung fluid is cleared. Effective cardiorespiratory function, as represented by an absence of respiratory distress (nasal flaring, grunting, chest wall retractions, a respiratory rate of greater than 60 per minute) and an oxygen saturation in the middle 90s should be established by several hours of age (Levesque et al, 2000; O’Brien et al, 2000). This normal sequence of events fails to occur in 2% to 8% of infants born at 34 weeks; gestation or later (Farchi et al, 2009; Hansen et al, 2008; Yoder et al, 2008). It is important to keep in mind that initial presenting symptoms are relatively nonspecific. Agrawal et al (2003) studied a large number of consecutive births in an attempt to determine the frequency and nature of different early-onset respiratory disorders and found that more than half did not meet specific diagnostic criteria. When confronted with a neonate with early onset respiratory symptoms, the most important diagnostic considerations include: ll ll ll ll ll ll ll ll Complex structural cardiac system anomalies; incidence estimated to be between 0.2% and 0.3%, often but not always identified by in utero imaging studies Diaphragmatic hernia, incidence estimated to be between 0.04% and 0.08%, commonly identified by second-trimester ultrasound (de Buys Roessingh and Dinh-Xuan, 2009) Respiratory distress syndrome (RDS); incidence estimated to vary between 0.45% and 2.4%, depending on the population studied; risk increased in the late preterm and infants delivered by cesarean section, particularly if accomplished before labor (Jain et al, 2009; Tita et al, 2009; Yoder et al, 2008) Persistent pulmonary hypertension of the newborn; incidence of 0.1% to 0.3%; often occurs in association with other acute respiratory conditions; questionable increased risk with maternal selective serotonin reuptake inhibitor treatment (Andrade et al, 2009; Chambers et al, 2006; Konduri and Kim, 2009) Meconium aspiration syndrome; incidence reported to vary between 2% and 9% of infants delivered through meconium-stained amniotic fluid (7% to 20% of all deliveries); risk increased in infants delivered after 40 weeks’ gestation, intrapartum distress, or both (Bhutani, 2008; Liu and Harrington, 2002) Spontaneous pneumothorax; incidence between 0.1% and 0.8%; infants born by cesarean delivery may be at increased risk (Benterud et al, 2009; Zanardo et al, 2007) Transient tachypnea of the newborn (TTNB); incidence variable between 0.3% and 3.9%; risk factors include late prematurity and cesarean section; initial diagnosis sometimes difficult to differentiate from pneumonia and early RDS (Guglani et al, 2008; Jain et al, 2009; Tita et al, 2009; Yoder et al, 2008) Pneumonia; incidence difficult to determine, with one recent retrospective report estimated rate at 0.3%; risk factors include maternal chorioamnionitis and prolonged ruptured membranes; sometimes difficult to differentiate from RDS and/or TTNB (Yoder et al, 2008) A review of the maternal history—particularly pregnancy, labor, and delivery—can provide useful diagnostic information. For example, the results of a second-trimester ultrasound examination could reveal the possibility of a cardiac defect or diaphragmatic hernia. A positive CHAPTER 26 Routine Newborn Care maternal GBS test result without adequate treatment, prolonged rupture of amniotic membranes, or evidence of chorioamnionitis suggests the possibility of pneumonia. For infants with respiratory distress born by cesarean section before the onset of labor, a diagnosis of RDS should be considered and assessment of gestational age should be performed. Finally, TTNB is a diagnosis of exclusion; it is prudent to rule out other causes before it is considered as the cause of respiratory distress in a term neonate. In most cases minimal initial diagnostic efforts for a term newborn with unsuspected respiratory distress should include a chest radiograph and assessment of the arterial oxygen saturation. The results of these studies, in combination with maternal history, should provide information helpful to: (1) establish initial management, such as the need for supplemental oxygen, continuous monitoring, or both; (2) determine the need for further work-up or treatment, possibly including an echocardiogram, laboratory testing, and treatment for possible sepsis; or (3) give a referral for further specialty consultation, intensive care, or both (in severe cases). CARDIOVASCULAR ISSUES Congenital heart disease is a relatively common condition in newborns, with an estimated incidence of 81 cases per 10,000 live births (Reller et al, 2008). Ventricular septal defect (VSD) is by far the most common defect, accounting for more than 30% of all cases. The increasing accuracy of prenatal ultrasound examination has greatly improved the early diagnosis of complex congenital heart disease. The results of population-based reviews indicate that the sensitivity of routine prenatal ultrasound examination in identifying selected congenital defects is as high as 70% and ranges as high as 85% for hypoplastic left heart (Chew et al, 2007; Rasiah et al, 2006). For mothers at high risk of delivering a newborn with congenital heart disease, the use of fetal echocardiography is helpful for delineating the anatomy and significance of specific lesions. However, many of the most common defects, particularly VSD, are not typically detected prenatally. In the absence of a prenatal diagnosis, detection of congenital heart disease is by physical examination. Sequential examinations are most helpful. At birth, many babies have loud murmurs that are thought to be from either a closing ductus arteriosus or tricuspid regurgitation (Silberbach and Hannon, 2007). These murmurs are transient and not indicative of disease. Conversely, murmurs associated with VSDs may not be heard for several days when the pressures on the right side of the heart have dropped enough to permit a significant shunting of blood from left to right. Although the ratio of pathologic to benign murmurs is higher in newborns than in older children, most of the murmurs heard during the newborn nursery stay in a healthy neonate are not clinically significant. Characteristics that increase the likelihood that a murmur signifies the presence of congenital heart disease include an intensity of 3/6 or greater, a harsh quality, occurrence during all of systole or into diastole, and being heard best at the lower sternal border or right upper border (Mackie et al, 2009). In a healthy newborn, the 313 most common presentation of congenital heart disease is a somewhat harsh systolic murmur that is heard best at the lower left sternal border in an asymptomatic infant, indicative of a VSD. In addition to auscultation, it is helpful to assess a newborn with a murmur for dysmorphic features and other anomalies, because these findings increase the likelihood that the murmur is indicative of congenital heart disease. It is important to evaluate the adequacy of femoral pulses to rule out coarctation of the aorta. Femoral pulses may be difficult to palpate in a neonate; if there is uncertainty, upper and lower extremity blood pressures can be measured. Although it is not a good screening tool, measurement of oxygen saturation may be helpful in diagnosing a cyanotic lesion in a child with a murmur or other signs of heart disease, particularly because hypoxia is frequently difficult to detect in newborns (Mahle et al, 2009). Chest radiographs and electrocardiograms are usually of limited value in evaluating healthy newborns with murmurs (Mackie et al, 2009; Oeppen et al, 2002). Full-term neonates frequently have alterations in cardiac rhythm and rate. Heart rates in full-term infants can range as high as 200 beats/min or as low as 80 beats/min. These values are usually indicative of normal variation and are not clinically meaningful unless there are other signs of illness or if there is a lack of variability in rate with stimulation or attempts at calming the newborn. Arrhythmias are also not uncommon, occurring in approximately 1% of newborns (Oeppen et al, 2002). By far, the most common arrhythmia in a healthy-appearing, full-term newborn is from premature atrial contractions (Larmay and Strasburger, 2004; Southall et al, 1981). These contractions are almost always benign and are usually transient. If there is concern about an irregular rhythm in a newborn, an electrocardiogram can be obtained. With premature atrial contractions, the irregular beat is initiated by a P wave. Although the QRS complex may be widened, it is always be preceded by the P wave. In most cases no further workup is needed. Cardiology consultation may be warranted if premature atrial contractions are persistent or if widened QRS complexes are seen on an electrocardiogram. POSSIBLE NEONATAL SEPSIS Group B Streptococcus Screening and Intrapartum Antibiotics Since the 1970s, GBS has been a major cause of neonatal sepsis (Schuchat, 1998). The implementation of intrapartum antibiotic prophylaxis (IAP) to prevent early-onset GBS disease in neonates was associated with an 80% decrease in the rate of infection (Phares et al, 2008). The currently recommended strategy to prevent GBS disease is to obtain rectovaginal cultures on all pregnant women at 35 to 37 weeks’ gestation and to administer penicillin or ampicillin during labor to those colonized with the bacteria. In situations where the mother’s GBS status is unknown before the onset of labor, IAP is advised for those with certain risk factors for neonatal infection (Box 26-3; Schrag et al, 2002). A third strategy may become available with the development of rapid strep testing, which can be done intrapartum. Overall, it is estimated that IAP can reduce 314 PART VII Care of the Healthy Newborn BOX 26-3 R  isk Factors for Group B ­Streptococcus Sepsis in ­Newborns ll ll ll ll ll ll ll ll ll ll ll Delivery before 37 weeks’ gestation Maternal fever during labor (body temperature >38.5° C) Prolonged rupture of membranes (>18 hours) Chorioamnionitis (maternal fever >38.5° C, tender uterus, fetal tachycardia) Prior child with group B Streptococcus disease Group B Streptococcus infection during pregnancy (urinary tract infection or bacteremia) Young maternal age African American race Hispanic ethnicity Meconium-stained amniotic fluid Newborn low absolute neutrophil count (e.g., immature:total ratio >0.2) the risk of early-onset GBS disease by 83% (Ohlsson and Shah, 2009). The overall rate of early-onset disease in the United States is approximately 0.3 to 0.4 cases per 1000 live births (Centers for Disease Control and Prevention, 2009; Van Dyke et al, 2009). Unfortunately, despite an effective strategy to prevent the infection, the rate of GBS disease in the United States increased between 2003 and 2006 (Centers for Disease Control and Prevention, 2009). Much of this increase was seen in African American infants. Although the rate of infection was 2.8-fold higher in premature infants than term newborns, 72% of the cases of GBS disease during this period were in full-term neonates. In addition, although 93% of mothers of infants in this series with GBS had been screened, IAP was administered to only 20%. Finally, although GBS cultures are highly accurate, there is always the possibility of a false-negative screen. In one review, 61% of infants with GBS disease were born to mothers with a negative test result before delivery (Van Dyke et al, 2009). These statistics highlight the need for continued vigilance for signs of GBS infection in term newborns by health care providers, even in an era of surveillance and IAP. Guidelines for prevention of GBS sepsis indicate that neonates are adequately treated only if IAP is administered at least 4 hours before delivery (Schrag et al, 2002). There is no clinical evidence to support this recommendation; it is based on expert opinion and is used to provide a margin of safety. However, there is evidence that penicillin given 2 hours before delivery is 90% effective in preventing GBS sepsis, and IAP provided less than 2 hours prior to delivery may be less effective (Illuzzi and Bracken, 2006). Blood levels are higher in the neonate than in the mother, even 30 minutes after a dose of penicillin, but the clinical importance of this has not been studied (Colombo et al, 2006). Term infants whose mothers have received IAP for a positive GBS screen at least 4 hours before delivery can be safely discharged at 24 hours of age if there are no signs or symptoms of infection (Ohlsson and Shah, 2009). One area of consternation is how long to observe term newborns when antibiotics were not given more than 4 hours before delivery. In the current guidelines it is recommended that such infants be monitored in the hospital for at least 48 hours after birth. The pressure for early newborn discharge has led some to compromise this period of observation. Most infants with sepsis manifest it early in life—many at birth, and nearly all by 12 hours of age (Escobar et al, 2000). Among a group of 172 term infants with documented early-onset GBS infection, 95% had presenting symptoms within 24 hours after delivery. More importantly, among the 33 neonates in this study that had GBS infection despite IAP, 31 had signs of infection before 24 hours of age (93.9%; 95% confidence interval, 79.8% to 99.3%; Bromberger et al, 2000). These data suggest, but do not definitively indicate, that some term newborns born to GBS positive mothers who received IAP less than 4 hours before delivery can be discharged by 48 hours of age. This decision is best made on an individual basis considering all risk factors, examination of the baby, vital sign stability, the results of any available laboratory tests, and parental wishes. Assessment of Term Newborn for Possible Sepsis The decision to remove a baby from the care of his or her parents during the period of initial homeostasis, and bonding should not be done without significant concern for the well-being of the infant. These decisions can have permanent effects on the parent-child relationship (Paxton and Nyington, 2001; Pearson and Boyce, 2004). Conversely, neonatal sepsis is such a dangerous and potentially rapidly progressive condition that any suggestive signs and symptoms cannot be ignored. It is between these extremes that the provider dwells and must make decisions often based on uncertainty and vague information. The strongest predictor of sepsis is whether the baby is ill. This predictor can be somewhat challenging to determine given the limited repertoire of behaviors in newborns. Neonates typically have erratic temperature, poor feeding, periods of lethargy, and crying, all of which can also be symptoms of infection. Some of the earliest signs of infection include resting tachypnea or tachycardia, especially with reduced heart rate variability and transient decelerations (Griffin et al, 2005). This level of early identification requires continuous monitoring which, although routinely used on premature infants, is not routinely performed on term newborns. Because the infection entry portal is often the lungs, increased respiratory rate or increased work of breathing are the most common symptoms noted (Andersen et al, 2004). Fever is not a common symptom in an infected newborn, and usually babies with high temperatures are found to have other causes, especially after 24 hours of age (Maayan-Metzger et al, 2006). Decreased skin perfusion or mottling and decreased or increased tone are also worrisome signs. GBS is not the only cause of sepsis in newborns. Coliforms were the dominant cause in the 1950s, and current causes may involve a natural cycle in the dominant bacteria causing sepsis. There are also sporadic cases of sepsis caused by enterococci, Serratia spp., and gram-negative bacteria such as Escherichia coli, Proteus spp., and Klebsiella spp. Although the decision of whether to evaluate or treat a neonate for possible sepsis cannot be informed only by CHAPTER 26 Routine Newborn Care the GBS status of the mother, and because other causes of infections are rare, most guidelines focus on GBS sepsis risk. In general, the physical examination is as good as laboratory testing in identifying ill newborns (Escobar et al, 2000). Laboratory tests can aid in the prediction that a baby is ill, but testing is also viewed as invasive by parents, so it is often helpful to carefully explain the tests and how the results will be used to make decisions. There is no one single test that has shown superiority, and opinion on testing varies. Among the most useful tests are the absolute neutrophil count, immature-to-total neutrophil ratio, and the procalcitonin level (Carrol et al, 2002; van Rossum et al, 2004). The C-reactive protein or erythrocyte sedimentation rate may help, but have lower predictive values (Galetto-Lacour et al, 2003). The decision to start testing or treating a term newborn for suspected sepsis should be multivariate and include: (1) risk factors during pregnancy, labor, and delivery (see preceding section); (2) current age in hours of the newborn; (3) presence of concerning signs and symptoms; and (4) physical examination. One single finding, such as an elevated temperature in an otherwise normally behaving newborn, is not usually sufficient evidence to start intravenous antibiotics. If the status of a term newborn is not rapidly changing, the use of close observation and periodic reassessment of risk factors and findings may prevent unnecessary testing or treatment. 315 SUGGESTED READING Burke BL, Robbins JM, Bird TM, et al: Trends in hospitalizations for neonatal jaundice and kernicterus in the United States, 1988-2005, Pediatrics 123: 524-532, 2009. Bhutani VK, Johnson L, Sivieri EM: Predictive ability of a predischarge hourspecific serum bilirubin for subsequent significant hyperbilirubinemia in healthy term and near-term newborns, Pediatrics 103:6-14, 1999. Dewey KG, Nommsen-Rivers LA, Heinig MJ, et al: Risk factors for suboptimal infant breastfeeding behavior, delayed onset of lactation, and excess neonatal weight loss, Pediatrics 112:607-619, 2003. Nommsen-Rivers LA, Heinig MJ, Cohen RJ, et al: Newborn wet and soiled diaper counts and timing of onset of lactation as indicators of breastfeeding inadequacy, J Hum Lact 24:27-33, 2008. Escobar GJ, Li DK, Armstrong MA, et al: neonatal sepsis workups in infants >/=2000 grams at birth: a population-based study, Pediatrics 106:256-263, 2000. Kramer MS, Aboud F, Mironova E, et al: Breastfeeding and child cognitive development: new evidence from a large randomized trial, Arch Gen Psychiatry 65:578-584, 2008. Kramer MS, Chalmers B, Hodnett ED, et al: Promotion of Breastfeeding Intervention Trial (PROBIT): a randomized trial in the Republic of Belarus, JAMA 285:413-420, 2001. Kuzniewicz MW, Escobar GJ, Newman TB: Impact of universal bilirubin screening on severe hyperbilirubinemia and phototherapy use, Pediatrics 124: 1031-1039, 2009. Lee RS, Cendron M, Kinnamon DD, et al: Antenatal hydronephrosis as a predictor of postnatal outcome: a meta-analysis, Pediatrics 118:586-593, 2006. Phares CR, Lynfield R, Farley MM, et al: Epidemiology of invasive group B streptococcal disease in the United States, 1999-2005, JAMA 299:2056-2065, 2008. Van Dyke MK, Phares CR, Lynfield R, et al: Evaluation of universal antenatal screening for group B streptococcus, N Engl J Med 360:2626-2636, 2009. Tarini BA, Christakis DA, Weich HG: State newborn screening in the tandem mass spectrometry era: more tests, more false-positive results, Pediatrics 118: 448-456, 2006. Complete references and supplemental color images used in this text can be found online at www.expertconsult.com C H A P T E R 27 Newborn Screening Inderneel Sahai and Harvey L. Levy Newborn screening is directed primarily at disorders in which the clinical complications develop postnatally. In metabolic diseases, these complications result from biochemical abnormalities that appear after birth, when the infant is no longer protected by fetal-maternal exchange. For example, the infant with phenylketonuria (PKU) has a normal blood phenylalanine level at birth, but within a few hours demonstrates hyperphenylalaninemia. The infant with congenital hypothyroidism (CH) is also protected in utero, most likely from placental transfer of maternal thyroxine (T4). If the hyperphenylalanemia in PKU is not controlled by diet or the hypothyroidism in CH is not corrected by supplemental T4, the infant begins to show signs of developmental delay and subsequently becomes mentally retarded. If therapy begins during the first weeks of life, mental retardation in both disorders is prevented. PKU was the first metabolic disorder known to benefit from dietary therapy. This fact was established by the middle 1950s. By the late 1950s, it was evident that the diet could prevent mental retardation if initiated in the neonatal period. Detecting PKU in all affected infants at that early age, before irreversible brain damage occurred, then became the challenge. This challenge entailed neonatal screening for a biochemical marker of the disease. In 1962, Guthrie developed a simple bacterial assay for phenylalanine that required only a small amount of whole blood soaked into filter paper (Guthrie and Susi, 1963). Therefore infants in newborn nurseries could be routinely tested for PKU in blood specimens obtained by lancing the heel and blotting the drops of blood onto a filter paper card. This filter paper blood specimen (dried blood spot specimen) could be mailed to a central laboratory for PKU testing. An increased concentration of phenylalanine in the specimen indicated PKU in the infant. By the middle 1960s, many states had established routine newborn screening programs for PKU using the Guthrie method. Infants with PKU were identified in larger numbers than anticipated and were showing normal development while receiving treatment (O’Flynn, 1992). The success of PKU screening led to the addition of tests for other metabolic diseases, including galactosemia, maple syrup urine disease (MSUD), and homocystinuria. These additional tests could be performed on the same blood specimen obtained for PKU screening. Over time, a test was added for the endocrine disorder congenital hypothyroidism, followed by screening tests for sickle cell disease, congenital adrenal hyperplasia (CAH), biotinidase deficiency, cystic fibrosis, and others. In 1990, tandem mass spectrometry (MS/MS) was first applied to the Guthrie specimen, opening a new era in newborn screening (Levy, 1998; Millington et al, 1990). This technology allowed for the accurate detection of numerous biochemical markers with a single assay, thereby making redundant several assays traditionally used in screening 316 for metabolic disorders. Furthermore, it enabled analysis of biomarkers not detectable by previous methodologies, thus greatly expanding the spectrum of conditions identifiable in the neonate (Levy and Albers, 2000). By the early 1990s, Naylor (Chace and Naylor, 1999) and Rashed et al (1995) began using this technology to routinely screen neonates for more than 20 biochemical disorders with high specificity and an extremely low rate of false-positive results. Currently most programs in the United States and screening programs in Europe and elsewhere have integrated MS/MS into newborn screening, and many are screening for more than 50 individual conditions (http://genes-r-us.uthscsa.edu). In order to standardize screening panels nationwide, the American College of Medical Genetics has recommended a panel of 29 core conditions for screening (Table 27-1). An additional 25 secondary conditions were listed for which test results could be reported (American College of Medical Genetics, 2006). These secondary conditions, revealed in the course of screening for the 29 core conditions, are not well known or documented treatment is unavailable. Molecular diagnostic techniques (i.e., DNA analysis) are also commonly used in newborn screening, predominantly as secondary tests for diseases such as cystic fibrosis or medium-chain acyl-CoA dehydrogenase deficiency (MCADD). Molecular testing substantially improves the positive predictive value of a primary screening result that is based on metabolite testing only (Ranieri et al, 1994; Wilcken et al, 1995; Ziadeh et al, 1995). With the initiation of pilot screening for severe combined immunodeficiency syndrome wherein the markers measured are DNA molecules (i.e., T cell receptor excision circles), the role of molecular technology has extended into primary marker analysis (Baker et al, 2009; Vogt, 2008). The use of molecular testing in screening will most likely expand further with advances in DNA technology. Screening for neuroblastoma, the most common solid tumor of childhood, was formerly being performed in Japan and some European programs measuring vanillylmandelic acid and homovanillic acid in urine collected from infants (Sawada, 1993). However, this screening was abandoned because it failed to identify individuals with poor-prognosis neuroblastoma, instead identifying infants who had a form of neuroblastoma that either spontaneously regressed or could be effectively treated after clinical detection (Woods et al, 2003). SCREENING PROCEDURE SPECIMEN The blood specimen is generally obtained from the heel of the infant. This simple sampling method, conceived and introduced by Guthrie and Susi (1963), has had an CHAPTER 27 Newborn Screening 317 TABLE 27-1 Core Disorders Recommended for Screening by American College of Medical Genetics Disorder Acronym Primary Marker Beta-ketothiolase deficiency (mitochondrial acetoacetyl CoA thiolase deficiency) BKT C5:1/C5OH Cobalamin defects A, B CBL (A,B) C3 Isovaleric acidemia* IVA C5 Glutaric aciduria I GA-I C5DC 3-Hydroxy 3-methylglutaryl-CoA lyase deficiency* HMG C5OH/C5-3M-DC Multiple carboxylase deficiency* MCD C3/C5OH 3-Methylcrotonyl-CoA carboxylase deficiency 3MCC C5OH Methylmalonic aciduria (mutase)* MMA C3 Propionic acidemia* Fatty acid oxidation defects PA C3 Carnitine uptake defect (carnitine transporter defect) CUD C0 Long-chain hydroxyacyl-CoA dehydrogenase deficiency* LCHAD/D C16OH/C18:1OH Medium-chain acyl-CoA dehydrogenase deficiency MCAD/D C8 Trifunctional protein deficiency* TFP C16OH/C18:1OH Very-long-chain acyl-CoA dehydrogenase deficiency Amino acid disorders VLCAD/D C14:1/C14 Argininosuccinic aciduria (argininosuccinate lyase deficiency)* ASA ASA Citrullinemia I (argininosuccinate synthase deficiency)* CIT-I Citrulline Phenylketonuria PKU Phenylalanine Maple syrup urine disease* MSUD Leucine Homocystinuria HCY Methionine Tyrosinemia type I TYR-I Tyrosine Biotinidase deficiency BIOT Biotinidase activity Galactosemia* GALT Total galactose, GALT activity Congenital adrenal hyperplasia* CAH 17-Hydroxyprogesterone Congenital hypothyroidism CH T4, TSH Sickle cell anemia HbSS Hb variants Sickle cell disorder HbS/C Hb variants Hemoglobin S/β-thalassemia HbS/betaTh Hb variants Cystic fibrosis CF Immunoreactive trypsinogen Hearing HEAR Metabolic Disorders Detected Using Tandem Mass Spectrometry Organic acid disorders Other Metabolic Disorders Endocrine Disorders Hemoglobin Disorders Other Disorders GALT, Galactose-1-phosphate uridyl transferase; T4, thyroxine; TSH, thyroid-stimulating hormone. *Can manifest acutely in the first week of life. enormous effect on newborn screening. The specimen is easily obtained and easily and inexpensively sent by mail to a central testing facility. There are no complications in obtaining the specimen from the newborn, contrary to early fears that its collection would lead to infection or result in excessive bleeding. SPECIMEN COLLECTION PROCEDURE The blood specimen should be obtained from the lateral or the medial side of the heel (Figure 27-1). Blood should be applied to only one side of the filter paper card, but it should saturate each circle on the card. Contamination of the filter paper specimen with iodine, alcohol, petroleum jelly, stool, urine, milk, or a substance such as oil from the fingers can adversely affect the results of the screening tests. In addition, exposure to heat and humidity can inactivate enzymes and produce false results. The specimen should be dried in air at room temperature for at least 3 hours before being placed in an envelope. Specimens are sometimes collected in capillary tubes, by venipuncture of a dorsal vein or from a central line, and then spotted on filter paper. There is little or no substantial difference in analyte levels between blood collected 318 PART VII Care of the Healthy Newborn FIGURE 27-1 Hatched areas at medial and lateral sides of the heel in this drawing of the sole indicate the proper sites for a heel stick in the newborn. directly from the heel and that collected by any of these other methods (Lorey and Cunningham, 1994). However, there is the danger of introducing amino acids from total parental nutrition solutions given through a central line into blood collected from this line, resulting in a falsepositive increase in amino acids or interference in some molecular assays by the heparin from the line. In general, it is preferable that blood for screening be spotted on filter paper directly from the heel. TIMING OF COLLECTION Newborn screening encompasses a gamut of conditions, each with its own ideal screening period during which there is the greatest chance of diagnosing the disorder and before onset of symptoms. As a result, it is worth noting that recommendations on the timing of specimen collection, although appropriate for the majority, may not be ideal for all conditions on the screening panel. In congenital adrenal hyperplasia, in which the symptoms can manifest within the first week of life, the optimal time for collection of the specimen is within 24 to 48 hours after birth. Formerly there was concern that with the newborn screening specimen collected early, often during the first day of life, some infants with metabolic disorders or CH might not have a sufficient degree of abnormality for identification. However, MS/MS methodology with its improved sensitivity and specificity has considerably improved the reliability of screening for metabolic conditions in early specimens (Chace and Naylor, 1999) and using thyroid-stimulating hormone (TSH) as the primary marker for CH, or as a second-tier test when T4 is the primary marker, has similarly allowed early screening for CH to be reliable. Specimen collection timings vary around the world. In Europe and Australia, most screening specimens are collected within 48 to 72 hours, whereas specimens in the United Kingdom are not collected until the infant is 5 to 8 days old (U.K. Newborn Screening Progam Centre, 2008). In the United States, most screening specimens are collected within 24 to 72 hours after birth. The specimen should be obtained from every newborn infant before nursery discharge or by the third day of life, whichever is first. In infants whose initial specimen was obtained within the first 24 hours of life, as may happen with the practice of early nursery discharge, a second blood specimen should be obtained no later than 7 days of age to be certain that a diagnosis is not missed. Special circumstances require specific attention to newborn blood specimen collection. Premature infants or those with very low birthweight, as well as infants who are sick and those in neonatal intensive care units (NICUs), are at a risk of unreliable screening owing to factors such as the unique physiology of the infant, therapeutic interventions, and a focus on critical activities in caring for the very sick neonate. Consequently, a single specimen is inadequate for screening in this subpopulation, and additional specimens should be collected for retesting. Serial screening with collection of three specimens—upon admission to the NICU, between 24 and 48 hours of life, and at discharge or at 28 days of life, whichever is sooner—has been proposed as an adequate and efficient protocol for this population (Clinical and Laboratory Standards Institute, 2009). In addition, some programs recommend screening every month until discharge for babies who continue to remain in the NICU. A blood specimen should be collected from any infant who is being transferred to a different hospital or to a NICU, regardless of age. The first specimen should be collected before transfer, and a second specimen at the receiving hospital by 4 days of age. This dual collection policy covers the infant from whom a newborn specimen might not have been obtained in the turmoil that frequently accompanies the transfer of neonates. In a newborn who is to receive a blood transfusion, a screening specimen should be collected before transfusion, and a second specimen should be collected 2 days after the transfusion. In addition, a third screening specimen should be obtained 2 months after the transfusion, when most of the donor red blood cells (RBCs) have been replaced. This practice ensures reliable testing for analytes present in RBCs, if a pretransfusion specimen has not been obtained. Newborn screening tests are usually performed in a centralized state, provincial, or regional laboratory. In a regional program, the specimens may be received by the state program and then delivered to the regional state or private laboratory, or they may be sent directly to the regional laboratory. In either case, the individual state programs serve as the state data and follow-up centers. SCREENING TESTS The testing procedure begins with the punching of small discs (each usually 3 mm in diameter) from the screening specimen. These small discs are then analyzed by various methodologies for the individual markers being sought. Amino acids and acylcarnitines, the markers for the majority of screened metabolic conditions, are simultaneously measured by MS/MS (Rinaldo et al, 2004). MS/MS is CHAPTER 27 Newborn Screening superior in terms of accuracy of measurement of the individual analytes when compared with alternate methods originally used for screening amino acids, such as bacterial assays or fluorometric techniques. Immunoassays, including fluoroimmunoassay and enzyme-linked immunosorbent assays, are used to test for endocrinopathies such as congenital hypothyroidism and CAH, for infectious diseases such as congenital toxoplasmosis and human immunodeficiency virus (HIV) seropositivity, and for cystic fibrosis. Hemoglobin electrophoresis of blood eluted from the filter paper disc is used for sickle cell disease screening. An enzyme assay is often used to screen for galactosemia and is always used to screen for biotinidase deficiency. Molecular assays are applied to quantify T cell receptor excision circles in dried blood to screen for severe combined immunodeficiency syndrome. More commonly, molecular assays are used to identify known mutations in certain disorders as a second-tier test. Several platforms, such as DNA microarrays and microsphere-based assays, can multiplex several molecular and immunological assays for high-throughput screening and are being used by screening programs (Dobrowolski et al, 1999; Green and Pass, 2005; McCabe and McCabe, 1999). SECONDARY TESTS An abnormal finding on a newborn screening test is not diagnostic of a disorder. Abnormalities in the newborn specimen can be transient or artifactual. Accordingly, when an abnormality is identified, the original specimen is retested for the analyte that was abnormal. Additional tests can be performed by the screening laboratory to substantiate the finding and improve the specificity of screening (Matern et al, 2007; Rinaldo et al, 2006). In screening for congenital hypothyroidism, many programs initially measure T4 in the original newborn blood specimen. Specimens in which a low T4 level is found are further tested for an increased level of TSH, which would indicate congenital hypothyroidism. A normal TSH level would suggest transiently low T4, a common finding in premature infants, or T4-binding globulin deficiency. Some screening programs have adopted screening protocols in which the primary analysis is for TSH, and T4 is measured as a second-tier test in specimens with high concentrations of TSH. Similarly, in screening for galactosemia an elevated galactose measurement in a specimen can trigger the analysis of galactose-1-phosphate uridyltransferase (GALT) enzyme activity as a second-tier test. Second-tier molecular testing is also performed in some screening laboratories. For example, in screening for cystic fibrosis, an initial out-of-range primary marker prompts DNA analysis to identify several specific pathogenic mutations (Comeau et al, 2004; Rock et al, 2005; Wilcken et al, 1995). Screening programs following this two-tiered immunoreactive trypsinogen-DNA approach can identify up to 99% of patients with cystic fibrosis and report a positive predictive value ranging between 1/9.5 to 1/25 (Grosse et al, 2004). Molecular assays to detect disease-causing mutations are currently used as second-tier tests for several other disorders, and their use is likely to expand with advances in DNA technology. These uses include testing for the prevalent c.985A>G mutation in MCADD, the 319 predominant E474Q mutation in long-chain 3-hydroxyacyl-CoA dehydrogenase deficiency (LCHADD; Dobrowolski et al, 1999), and a panel of several GALT mutations in galactosemia screening. The final interpretation of the screening results is based on the primary analysis and, if available, the results of second-tier testing. However, it is important to realize that screening is not intended to be diagnostic; abnormal screening results must be supported by confirmatory investigations. These studies require additional specimens and are performed by clinical laboratories or sometimes by the screening laboratory. PHYSICIAN CONTACT FOR ABNORMAL RESULTS Table 27-2 indicates disorders or other reasons for abnormal screening results, sorted according to the primary analyte usually used to screen for the condition. For example, a low T4 level together with an elevated TSH concentration indicates congenital hypothyroidism, and a marked elevation of 17-hydroxyprogesterone (17-OHP) indicates the likelihood of CAH. An elevation of an acylcarnitine could indicate an organic acid or fatty acid oxidation disorder. Any infant for whom such a screening result is reported should be evaluated by the primary care provider as soon as possible to facilitate the next steps towards the confirmation and management of the disorder. However, several conditions screened for are extremely rare, and primary health care professionals might not have sufficient information available to be able to direct appropriate intervention in screen-positive infants. To overcome the challenge, readily accessible, one- to two-page explanations of the possible disorders represented by the abnormality and the recommended confirmatory tests, known as ACT sheets and confirmatory algorithms (Kaye et al, 2006), are available on the Web site of the American College of Medical ­Genetics (www.acmg.net/resources/policies/ACT/condi tion-analyte-links.htm). Although all specimens with a metabolite concentration that crosses its threshold are considered screen-positive, all screen-positive results are not associated with the same likelihood of being associated with a disorder. Most infants with a positive screening result that is only mildly abnormal are less likely to have a disorder (see later discussion of false-positive results) than are infants with analyte concentrations that are several fold above the cutoff. Applying a uniform approach for all positive results in terms of urgency of intervention or battery of tests suggested can result in unnecessary parental anxiety and medical costs. However, if recommendations for further action and workup are customized in accordance with the potential significance of the abnormality, both parental anxiety and costs associated with false-positive results can be reduced. To achieve this goal, some programs subcategorize positive screening results. The New England Newborn Screening Program uses primary marker concentrations, second-tier analyses, biomarker profiles for markers analyzed by MS/ MS, and acuity of the likely disorder to subcategorize outof-range screening results (Sahai et al, 2007). The primary care providers are supplied with category-based, customized fact sheets when a positive screening result is reported 320 PART VII Care of the Healthy Newborn TABLE 27-2 Disorders or Other Reasons for Abnormal Screening Results Sorted by Primary Marker Analyzed Marker Disorders Possible Causes of False Positives Markers Analyzed Using Tandem Mass Spectrometry Free carnitine (C0) Low CUD Poor feeding, maternal CUD Free carnitine (C0) High CPT-I Carnitine supplementation Propionylcarnitine (C3) High MMA, PA, CBL (A,B) Hemolysis,* maternal biotin deficiency, carrier of associated disorders Butyrylcarnitine (C4) High SCADD/EE/IBDD Hypoglycemia from other causes,* FIGLU elevation Tigylcarnitine (5:1) High BKT deficiency VLBW neonate Isovalerylcarnitine (C5) High IVA Antibiotics containing pivalic acid, IVA/MBCD carrier, VLBW neonate, neonate receiving total parental nutrition, FAS hemoglobin profile* Glutarylcarnitine (C5DC) High GA-I MCADD carrier, twin or multiple births* 3-Methylglutarylcarnitine (C5-3M-DC) High HMG Severe respiratory distress, neonates receiving ECMO* Hydroxyisovalerylcarnitine (C5OH) High 3MCC Maternal MCC, maternal biotin deficiency Octanoylcarnitine (C8) High MCADD MCADD carrier, MCT supplementation Tetradecenenoylcarnitine (C14:1) High VLCADD Carrier Hydroxyhexadecanoylcarnitine (C16OH) High LCHADD/TFP Carrier Hexadecanoylcarnitine (C16) High CPT-II Severe hemolysis* Arginine (Arg) High ARG Hyperalimentation Argininoscuccinic acid (ASA) High ASA — Citrulline (Cit) Low OTC/CPS/NAGS Poor feeding* Citrulline (Cit) High CIT-I Hyperalimentation Phenylalanine (Phe) High PKU Hyperalimentation, specimen contaminated with artificial sweetener* Leucine (Leu) High MSUD Hyperalimentation, hydroxyprolinemia Methionine (Met) High HCY Hyperalimentation Tyrosine (Tyr) High TYR-I Prematurity, transient immaturity of enzyme Biotinidase activity (Bio) Low BIOT Exposure of specimen to heat, improper drying Total galactose (T-Gal) High GALT GALE deficiency, GALK deficiency, contamination with milk/cream Activity GALT Low GALT Exposure of specimen to heat 17-Hydroxyprogesterone (17-OHP) High CAH Physiologic stress (seen commonly in NICU babies), VLBW, EDTA in specimen Markers analyzed by assays other than tandem mass spectrometry T4 / TSH Low/High CH Neonates in NICU, maternal thyroid medications Immunoreactive trypsinogen (IRT) High CF VLBW neonate, NICU T cell receptor excision circles (TREC) Low SCID Other immunodeficiencies, Di George syndrome, heparin in specimen* BIOT, biotinidase deficiency; BKT, beta-ketothiolase deficiency; CAH, congenital adrenal hyperplasia; CBL, cobalamin defect; CF, cystic fibrosis; CH, congenital hypothyroidism; CIT-I, citrullinemia I; CPT, carnitine palmitoyltransferase deficiency; CUD, carnitine uptake defect; ECMO, extracorporeal membrane oxygenation; EDTA, ethylenediamine ­tetraacetic acid; EE, ethylmalonic encephalopathy; FIGLU, formiminoglutamic; GA-I, glutaric aciduria I; GALE, uridine diphosphate galactose-4-epimerase; GALK, galactokinase; GALT, galactosemia; HCY, homocystinuria; HMG, 3-hydroxy 3-methylglutaryl-CoA lyase deficiency; IBDD, Isobut yryl-CoA dehydrogenase deficiency; IVA, isovaleric acidemia; LCHADD, long-chain 3-hydroxyacyl-CoA dehydrogenase deficiency; MBCD, 2-methylbutyryl-CoA dehydrogenase deficiency; MCC, methylcrotonyl-CoA carboxylase; MCT, medium-chain ­triglycerides; MSUD, maple syrup urine disease; MMA, methylmalonic ­aciduria; MCADD, medium-chain acyl-CoA dehydrogenase deficiency; NICU, neonatal ­intensive care unit; PA, propionic acidemia; PKU, phenylketonuria; SCADD, Short-chain acyl-CoA dehydrogenase; SCID, severe combined immune deficiency; TFP, trifunctional protein deficiency; TYR-I, tyrosinemia type I; VLBW, very low birthweight; VLCADD, very-long-chain acyl-CoA dehydrogenase deficiency. *Associations observed in the New England Newborn Screening Program. (Sahai and Eaton, 2008). These sheets include information on disorders associated with the marker, estimated likelihood of being affected, clinical presentations of likely disorders, factors contributing to false positives, and recommendations for further management. The follow-up recommendations can range from immediate admission to a hospital, where further evaluation and therapy for the illness can be initiated without delay, to simply repeating the filter paper analysis on a sample collected a few days later. Other programs approach this problem differently, but with the same goal in mind, providing the primary care providers with the information needed to put the result in the appropriate context for the family. Any infant for whom an abnormal screening result is reported should be seen as soon as possible and evaluated with a careful history and physical examination. When specific guidelines based on the individual results are not provided by the screening program, and the infant is ill or CHAPTER 27 Newborn Screening the likely disorder manifests acutely within the first few days of life (see Table 27-1), a specialist should be contacted. The infant may need to be admitted to the hospital where further evaluation and therapy for the illness can be initiated without delay. If the infant is active and alert with good feeding and shows no abnormal signs on initial evaluation, and the suspected disorder does not require immediate attention, a second filter paper blood specimen can be obtained and sent to the screening laboratory for repeated testing, or confirmatory testing can be performed on a less urgent basis. In many cases, confirmatory testing or referral to a specialist is required only if the second test indicates the presence of a disorder. However, the follow-up of an initial positive screening result can vary. In some programs, more specific confirmatory testing is the first response to a presumptive positive newborn screen, with a less intense time frame for individuals in whom the level of suspicion is lower. The physician should contact the screening laboratory when an infant whose screen has been reported as normal or whose screening results have not yet been reported has symptoms that suggest a metabolic disorder. The screening laboratory can check the results in the infant’s newborn specimen. If the testing has been completed and the newborn specimen is retained in storage, the laboratory may wish to recover the specimen and repeat the tests. The physician should also contact the screening laboratory for the results of repeated tests and inform the family of the results as soon as possible. If the second result is normal, the duration of the family’s anxiety may be shortened. SCREENED DISORDERS Following are brief summaries of the categories of the most common disorders detected by newborn screening. There is no attempt to describe any of the disorders in detail or their rare variants. METABOLIC DISORDERS Amino Acid Disorders The amino acid disorders are caused by an enzymatic defect in the catabolic pathway of amino acids, with consequent accumulation of specific amino acids above the block. Screening relies on the detection of these elevated amino acids in the newborn specimen. The clinical manifestations may be a result of the toxic effects of the accumulating amino acid and metabolites produced by alternate pathways, a deficiency of the products of the normal pathway, or both. PKU is the best-known example of an amino acid disorder and is the paradigm for screened disorders in the newborn. In addition to PKU, MSUD and homocystinuria are historically significant in the context of screening, because they are among the original metabolic conditions for which screening was performed before the introduction of MS/MS and therefore the expansion of screening. Screening for other amino acid disorders, such as the urea cycle defects and tyrosinemia type I, became possible only with the advent of MS/MS technology. 321 In PKU, the cardinal screening feature is an increased level of phenylalanine. PKU should always be identified by newborn screening. If untreated, patients with PKU experience severe mental retardation and other neurologic abnormalities. The average incidence of the disorder is approximately 1 in 12,000 live births. With screening by MS/MS, PKU can reliably be identified as early as the first day of life (Chace et al, 1998). Not all infants with an elevation of phenylalanine have PKU. Mild elevations can indicate benign mild hyperphenylalaninemia. Liver disease, such as that associated with galactosemia, tyrosinemia type I or citrin deficiency, can also produce increased phenylalanine. Therefore treatment for PKU should never be started on the basis of a positive screening test result alone. The dietary therapy is complicated and can be hazardous to an infant who does not have PKU. If the screening test reveals a marked elevation of phenylalanine (≥6 mg/dL) and the metabolic profile (including reduced tyrosine) is supportive of PKU, the infant should be referred directly to a metabolic center for confirmatory testing and prompt consideration of dietary treatment. If the screening level of phenylalanine is only slightly increased, retesting a second specimen before initiating a complete diagnostic work-up may suffice. The primary indicator for MSUD in the newborn blood specimen is an increase in leucine. To improve the specificity for MSUD of an abnormal leucine measurement, some screening programs also report the ratio between leucine and a reference amino acid or alloisoleucine analysis as a second-tier test (Matern et al, 2007). The average incidence of classic MSUD is 1 in 185,000. MSUD can be a fulminant disease associated with severe ketoacidosis, vomiting, and lethargy, and it can progress rapidly to coma and death. Consequently, the finding of a substantially increased leucine level in the newborn blood specimen should prompt an immediate telephone call from the screening program to the attending physician. If the infant is ill, he or she should be transported immediately to an NICU at a medical center with a metabolic specialist. Confirmatory plasma and urine specimens should be obtained, and emergency therapy should be initiated. Plasma amino acid analysis in an infant with MSUD will show marked increases in leucine, isoleucine, and valine as well as alloisoleucine (the branched-chain amino acids). The urine specimen will test strongly positive for ketones and will contain large quantities of the branched-chain ketoacids and amino acids. The characteristic odor reminiscent of maple syrup, which appears earliest in cerumen and only later in urine, will probably be detected on a cotton-tipped swab inserted in the infant’s ear. Milder variants of MSUD can be missed by newborn screening (Bhattacharya et al, 2006). A newborn with the intermediate variant might not have a blood leucine elevation, or the increase may be so mild as to be below the cutoff value. In the intermittent variant, the blood leucine concentration is normal in the newborn period, becoming elevated only in later infancy or childhood during acute metabolic episodes precipitated by febrile illness or surgery. Individuals with homocystinuria are clinically normal at birth but, if untreated, may develop ectopia lentis (dislocation of the lens), thromboembolism, osteoporosis, and mental retardation. The worldwide frequency of all forms 322 PART VII Care of the Healthy Newborn of homocystinuria has been estimated at 1 in 344,000, but may be considerably higher (Skovby et al, in press). The newborn blood screening marker for detecting homocystinuria is an increased level of methionine. Homocysteine can be measured as a second-tier analysis to improve specificity (Matern et al, 2007). The diagnosis of homocystinuria may be missed if the blood methionine concentration is not elevated at the time the newborn specimen is collected (Whiteman et al, 1979). Reducing the cutoff value for methionine can substantially increase the frequency of identified infants (Peterschmitt et al, 1999), but may also result in an increased number of false-positive results. A high methionine level alone is not diagnostic of homocystinuria. Liver disease of a nonmetabolic nature can produce a strikingly high methionine level, as can isolated hypermethioninemia (methionine-S-adenosyltransferase [MAT] I/III deficiency), a metabolic disorder that may be benign. Two additional rare disorders also produce hypermethioninemia: glycine-N-methyltransferase deficiency associated with liver disease (Luka et al, 2002; Mudd et al, 2001) and S-adenosylhomocysteine hydrolase deficiency, which may result in developmental delay and hypotonia (Baric et al, 2004). Furthermore, transient hypermethioninemia may occur in newborn infants. Confirmation of the disorder requires quantitative amino acid analyses of plasma and urine. In the infant with homocystinuria, homocystine is usually detectable in plasma and urine, plasma total homocysteine is increased as is methionine, and cystine is reduced. In isolated hypermethioninemia, methionine is markedly increased in plasma, but there is no detectable homocystine in plasma or urine and the plasma cystine concentration is normal. Hypermethioninemia secondary to liver disease owing to tyrosinemia type I, or to nonspecific liver disease, is usually accompanied by increased tyrosine. The three urea cycle disorders currently screened by MS/MS analysis are citrullinemia, argininosuccinic acidemia, and arginase deficiency. Citrullinemia and argininosuccinic acidemia produce hyperammonemia, often in the neonatal period, accompanied by poor feeding, tachypnea, lethargy, and vomiting. Respiratory alkalosis is characteristic. Severe hyperammonemia in the newborn is a medical emergency and should trigger prompt consultation with a metabolic specialist or referral to an NICU at a metabolic center. Discontinuation of protein and the provision of intravenous fluids with high caloric content are the first steps to take. L-Arginine or L-citrulline, as well as the “scavenger drugs” sodium phenylbutyrate and sodium benzoate, may be administered. Hemodialysis might be required to control the neurotoxic hyperammonemia, which can lead to irreversible brain damage, coma, and death. It is hoped that with early identification through newborn screening, patients with urea cycle disorders will be protected by presymptomatic therapy in the neonatal period. Arginase deficiency can also present acutely with hyperammonemia as described earlier, although more frequently it manifests as developmental delay and spastic diplegia in childhood with a milder degree of hyperammonemia (Crombez and Cederbaum, 2005). Tyrosinemia type I is an amino acid disorder that can be diagnosed by finding an elevation of succinylacetone in MS/MS analysis (Allard et al, 2004). However, the preanalytic processing required for succinylacetone is more involved than that required for the amino acids and acylcarnitines. As a result, succinylacetone is not currently measured by all screening programs. These programs may rely on elevations of tyrosine for identification of this disorder. Unfortunately, moderate elevations of tyrosine that are transient occur frequently in neonates, especially those who have low birthweights and are sick, necessitating frequent requests for repeated screening with virtually no detection of tyrosinemia type I. Unfortunately moderate transient elevations of tyrosine occur frequently in neonates, especially those who have low birthweights and are sick, necessitating frequent requests for repeated screening. Virtually no cases of tyrosinemia type I have been detected based on elevated tyrosine because almost all infants with tyrosinemia type I have had normal tyrosine levels when screened (Frazier et al, 2006). Consequently, the newborn detection of tyrosinemia type I by a tyrosine marker alone is ineffective. Tyrosinemia type I leads to liver and renal tubular disease and can later result in hepatocellular carcinoma. It is treated with administration of 2-(2-nitro-4-trifluoromethylbenzoyl)-1,3-cyclohexanedione (NTBC; Orfadin) and a diet low in phenylalanine and tyrosine. Tyrosinemia types II and III are identified by an increased level of tyrosine in newborn screening. Tyrosinemia type II can result in mental retardation, painful hyperkeratoses, and keratoconjunctivitis, and it is treated by a low tyrosine-phenylalanine diet. Tyrosinemia type III seems to be benign, although developmental delay has been reported occasionally (Ellaway et al, 2001). Medium-Chain Acyl-CoA Dehydrogenase Deficiency and Other Fatty Acid Oxidation Disorders The fatty acid oxidation disorders include those in which the long-chain fatty acids cannot traverse the mitochondrial membranes to be oxidized within the mitochondrial matrix (i.e., the carnitine palmitoyltransferase disorders) and those that constitute defects in fatty acid oxidation per se. In either category, the problem is the inability to fully oxidize fatty acids. Fatty acid oxidation is essential to supply energy as adenosine triphosphate via the Krebs cycle and as ketones in the presence of a low supply of glucose. The disorders involving defective transport concern carnitine, whereas those with defective oxidation are named according to the enzyme that is deficient (see Table 27-1). The clinical consequence of these disorders is fasting intolerance resulting in hypoketotic hypoglycemia, ­lethargy, hyperammonemia, metabolic acidosis, hepatomegaly, and sometimes sudden death. Cardiomyopathy is an additional feature of very-long-chain acyl-CoA dehydrogenase deficiency (VLCADD) and long-chain hydroxyacyl CoA dehydrogenase deficiency (LCHADD). Each fatty acid disorder is associated with a specific or almost specific acylcarnitine pattern on MS/MS analysis. The most common fatty acid oxidation disorder is MCADD. Tragically, before newborn screening was available, this disorder was often diagnosed only retrospectively after a sudden unexplained death, usually when postmortem examination revealed a fatty liver. This devastating outcome and a frequency of 1:15,000 to 1:20,000 (Zytkovicz CHAPTER 27 Newborn Screening et al, 2001), comparable with that of PKU, made MCADD the primary reason for the addition of MS/MS technology to newborn screening. To reduce the rate of false-positive results in screening for MCADD, some programs have added molecular testing of the c.985A>G MCAD mutation as second-tier screening. Because this mutation occurs in as many as 90% of persons with MCADD, this additional analysis of a newborn blood specimen with an elevation of the octanoylcarnitine marker for MCADD substantially improves the predictive value of the screening abnormality (Zytkovicz et al, 2001). The treatment for fatty acid oxidation disorders is avoidance of fasting with high-carbohydrate, low-fat feedings and, of critical importance, prompt attention to acute illnesses in which vomiting occurs. Carnitine supplementation may be beneficial. Medium-chain triglycerides (i.e., MCT oil) is given for the long-chain disorders VLCADD and LCHADD. Any infant with a fatty acid oxidation disorder should be evaluated at a metabolic center. Most of these disorders are treatable, but screening enables early diagnosis and genetic counseling for the family even when early treatment may not be effective, such as in neonatal carnitine palmitoyltransferase II deficiency (CPT-II) (Albers et al, 2001a). Organic Acid Disorders Organic acid disorders are a heterogenous group of disorders with a combined frequency of approximately 1 in 50,000 (Zytkovicz et al, 2001). Many of them can be identified through MS/MS screening (see Table 27-1). The marker for this disease group, as for the fatty acid oxidation disorders, is an abnormal acylcarnitine pattern. If a screening result suggests an organic acidemia, a metabolic specialist should be consulted immediately. The major organic acid disorders identified in newborn screening are propionic acidemia, the methylmalonic acidemias, and isovaleric acidemia. The organic acidemias can manifest in the neonatal period with a life-threatening, sepsis-like picture of feeding difficulties, lethargy, vomiting, and seizures. Metabolic acidosis virtually always accompanies this presentation, and hyperammonemia is common. In this situation, protein administration should be discontinued and replaced by administration of intravenous fluids with high caloric content and carnitine. The hyperammonemia rarely requires specific treatment. The effects of early diagnosis and treatment on the clinical and neurologic development of individuals affected by an organic acid disorder are currently under investigation (Albers et al, 2001b). Galactosemia Galactosemia typically manifests in the neonatal period as failure to thrive, vomiting, and liver disease (Hughes et al, 2009). Death from bacterial sepsis, usually caused by Escherichia coli, occurs in a high percentage of untreated neonates (Levy et al, 1977). The average incidence of the disorder is 1 in 62,000. Some screening programs use a metabolite assay for total galactose (galactose and galactose-1-phosphate) to detect galactosemia. Other programs screen the newborn 323 specimen with a specific enzyme assay for activity of GALT, undetectable in severe galactosemia. The enzyme assay identifies only galactosemia, whereas the metabolite assay also identifies other galactose metabolic disorders, such as deficiencies of galactokinase and epimerase. Severe neonatal liver disease and portosystemic shunting caused by anomalies in the portal system can also increase the galactose level. The most rapid confirmatory test for a positive result in galactosemia screening is urine testing for reducing substance. In almost all cases of severe galactosemia, this test produces a strongly positive reaction. Galactosemia with residual GALT activity, however, may be accompanied by a negative or only slightly positive result for urine reducing substance. If the urine contains reducing substance and the infant has clinical signs of galactosemia (e.g., poor feeding, jaundice, hepatomegaly), a blood specimen for confirmatory testing should be collected, and milk feeding (breast or formula) should be discontinued with substitution of a nonlactose formula such as soy or elemental. The infant should then be referred immediately to a pediatric metabolic center, where confirmatory testing should include the measurement of RBC galactose-1-phosphate and an enzyme assay for RBC GALT activity (Levy and Hammersen, 1978). Molecular testing for GALT mutations may also be performed. Approximately 70% of the patients with galactosemia carry the Q188R mutation (Elsas and Lai, 1998). Variant galactosemia is produced by mutations such as S135L. The N314D mutation produces the benign Duarte variant. If the urine test is negative for reducing substance, the newborn screening result is most likely to be false-positive or to indicate a benign GALT enzyme variant (e.g., Duarte variant). Nevertheless, urine-reducing substance may be absent in infants with clinically significant variants of galactosemia. Consequently, follow-up testing should be performed for all infants with an initial positive galactosemia screening result. Biotinidase Deficiency Biotin recycling is necessary for the maintenance of sufficient intracellular biotin to activate carboxylase enzymes. Biotinidase is a key enzyme in biotin recycling. Lack of biotinidase activity results in reduced carboxylase activities and an organic acid disorder known as multiple carboxylase deficiency (Wolf and Heard, 1991). The clinical features of the disorder are developmental delay, seizures, hearing loss, alopecia, and dermatitis. The developmental delay and seizures usually manifest at 3 to 4 months of age. Death during infancy has also been reported. Initiating biotin therapy in early infancy, when the disorder is presymptomatic, seems to prevent all the features of biotinidase deficiency. For this reason, a screening test has been developed and added to newborn screening in a number of newborn screening programs throughout the world (Hart et al, 1992). The frequency of identified newborns in these programs has a wide range, from 1:30,000 to 1:235,000. The average frequency is approximately 1 in 70,000. Almost all identified infants have been asymptomatic and have remained normal with biotin treatment. 324 PART VII Care of the Healthy Newborn ENDOCRINE DISORDERS Congenital Hypothyroidism Congenital hypothyroidism is the most common disorder identified by routine newborn screening. It is found in 1:3000 to 1:5000 screened infants (Dussault, 1993). The major clinical features of untreated congenital hypothyroidism are growth retardation and delayed cognitive development leading to mental deficiency. If treatment with pharmacologic doses of T4 is initiated early, growth and mental development are normal. Two screening approaches are used (Pass and Neto, 2009). One method is primary screening for low T4 with secondary screening for high TSH. The second method is primary screening for high TSH levels. Either procedure reliably identifies congenital hypothyroidism. Nevertheless, affected infants can be missed with either approach. This situation may be due to a lack of the identifying marker abnormality at the time of specimen collection. Specifically, the T4 level during the first 24 hours of life in an affected infant might not yet be sufficiently decreased for identification because of persistence of maternally transmitted T4. Moreover, in the premature infant with congenital hypothyroidism, it might take 2 weeks or more for a TSH elevation to develop (Larson et al, 2003). The reported false-positive rates of screening for congenital hypothyroidism range from approximately 0.05% to as high as 4% (Pass and Neto, 2009). Infants with falsepositive results have transiently low T4 or elevated TSH levels. Many of those with low T4 levels are premature infants with a normal TSH concentration or infants with perinatal stress and elevated TSH levels. To avoid missing congenital hypothyroidism, screening programs require a second blood specimen from each of these infants. In addition to false-positive results, a low T4 level with a normal TSH value can result from benign T4-binding globulin deficiency (Dussault, 1993; Mandel et al, 1993) or hypothyroidism secondary to pituitary deficiency. Infants with a positive screening test result should not be labeled as having congenital hypothyroidism until diagnostic testing confirms the disorder. This is especially true if the TSH concentration reported by the screening program is normal. If congenital hypothyroidism is confirmed, however, administration of T4 should be started without delay to prevent irreversible brain damage. Congenital Adrenal Hyperplasia CAH caused by steroid 21-hydroxylase deficiency occurs in 1:16,000 to 1:20,000 births (White, 2009). Infants with the salt-losing form of CAH can rapidly become hyperkalemic and die precipitously, often without a specific diagnosis. The clinical diagnosis may be suspected in the newborn girl because of ambiguous genitalia. However, the diagnosis is usually not suspected in boys and in girls with atypical forms of CAH in which ambiguous genitalia may not occur. Infant girls with ambiguous genitalia might not be recognized as having CAH if the ambiguity is not obvious, or they could be misassigned as boys if the ambiguity is advanced. Because accurate gender assignment and initiation of hormone therapy as soon as possible are critical to a favorable prognosis in CAH, newborn screening is important in leading to early diagnosis and prompt therapy with pharmacologic doses of hydrocortisone. Consequently, testing for CAH has been incorporated into routine newborn screening in most programs. Screening is based on identifying elevated levels of 17-OHP, the preferred substrate for 21-hydroxylase, and is usually measured using an immunoassay. Unfortunately, compared with other neonatal screening assays, the specificity of the CAH assay is low and false-positive results in newborn screening for CAH are relatively common, with the rate often as high as 0.5%. The finding may be due to a truly increased 17-OHP, as in perinatal stress and early specimen collection (within the first 24 hours of life), or to cross-reacting steroids, such as in prematurity and low birthweight (al Saedi et al, 1996). Cross-reacting steroids are produced by residual fetal adrenal cortex or result from decreased metabolic clearance by an immature liver. A decidedly increased level of 17-OHP suggests CAH, and the infant should be referred to a pediatric endocrinologist for management. A second blood specimen is usually requested from infants found to have slight to moderately increased 17-OHP. If the infant shows signs of illness or has ambiguous genitalia, serum electrolytes should be measured. If these results indicate hyponatremia and hyperkalemia, the infant should be hospitalized without delay, and the electrolyte imbalance should be corrected immediately. Pediatric endocrinology consultation should also be sought. It may be possible to improve the positive predictive value by second-tier screening using DNA-based methods or liquid chromatography followed by tandem mass spectrometry, but currently these methods are not widely used by newborn screening programs. SICKLE CELL DISEASE In most newborn screening programs in the United States, the blood specimen is routinely tested for hemoglobin abnormalities. The major goal of this testing is to identify infants with sickle cell disease so that they can be given penicillin prophylaxis to prevent pneumococcal septicemia. Additional benefits of early detection are early referral to a comprehensive sickle cell program and early education and genetic counseling for parents (Smith and Kinney, 1993). Unfortunately, the long-term complications are not yet preventable. Sickle cell screening is usually performed by means of hemoglobin electrophoresis of blood eluted from a disc of the Guthrie specimen. This procedure identifies sickle cell disease, sickle cell trait, and several other abnormal hemoglobins. Other than sickle cell disease, most of these abnormalities are benign. It is especially critical to differentiate the common and benign sickle cell trait from the much rarer sickle cell disease (homozygosity for S hemoglobin). For example, sickle cell disease affects approximately 1 in 600 African American persons, whereas sickle cell trait (carrier status for S hemoglobin) is present in 1 in 12. Infants with sickle cell trait do not have complications and should not be stigmatized as having sickle cell disease. When sickle cell disease is confirmed, penicillin prophylaxis should be initiated as soon as possible, and the infant should be referred to a sickle cell disease center or CHAPTER 27 Newborn Screening hematologist. The combination of screening and careful follow-up has been highly effective in preventing pneumococcal sepsis in infants with sickle cell disease. CYSTIC FIBROSIS The frequency (1:2000 to 1:3000) and severity of cystic fibrosis explain its inclusion in routine newborn screening. As with sickle cell disease, therapy that can prevent the ultimate complications of cystic fibrosis is not yet available. However, early and usually presymptomatic diagnosis through screening leads to early nutritional therapy, pancreatic enzyme replacement, and antibiotic prophylaxis for pulmonary infection. Data from newborn screening suggest better growth, prevention of vitamin deficiency in early infancy, and some advantage in terms of pulmonary status later in life in children identified by screening (Farrell et al, 2001; McKay and Wilcken, 2008; Southern et al, 2009). Other benefits of newborn screening are identifying the parents’ genetic potential for producing additional children with cystic fibrosis and, through presymptomatic identification, allowing the family to avoid months or years of delay in the correct diagnosis of a child with chronic respiratory problems or poor growth (Farrell and Mischler, 1992). The analyte marker in newborn screening for cystic fibrosis is increased immunoreactive trypsinogen (IRT). Transient increases in IRT are common in healthy newborns as a result of perinatal stress or for unknown reasons. Consequently, the rate of false-positive results in cystic fibrosis screening using only IRT is relatively high. To reduce this rate, screening programs have adopted a second-tier DNA analysis for a panel of cystic fibrosis mutations when the specimen has increased IRT (Ferec et al, 1995). Despite this two-tiered approach to screening, a substantial number of infants who do not have cystic fibrosis must undergo a sweat test or complete DNA sequencing of the cystic fibrosis gene before the diagnosis can be eliminated. SPECIFIC ISSUES IN NEWBORN SCREENING CRITERIA FOR NEWBORN SCREENING Under the auspices of the World Health Organization, Wilson and Jungner (1968) published a set of criteria for screening for disease that have generally been accepted as required for population screening. The ten criteria state that (1) the condition be an important health problem, (2) there be accepted treatment, (3) facilities for diagnosis and treatment be available, (4) there be a recognizable latent or early symptomatic stage, (5) there should be a suitable test, (6) the test should be acceptable to the population, (7) the natural history of the condition should be understood, (8) there should be a policy prescribing whom to treat, (9) the cost of case finding should be economically balanced in relation to medical care as whole, and (10) case-finding should be a continuing process. These criteria were developed at a time when newborn screening was in its beginning stages and with screening for adult disorders in mind. For example, regarding to the 325 first criterion, Wilson and Jungner recognized that the term important is relative; whereas diabetes was prevalent, treatment might not influence outcome while PKU was rare, but it warranted screening because of the serious consequences that would be prevented by early discovery and treatment. There is a question as to the current relevance of the Wilson-Jungner criteria for newborn screening. Ideally, all the criteria should be applied to newborn screening. However, advances in technology have caused this application to be questioned (Green and Pollitt, 1999; Levy, 1999). As an example, screening using MS/MS allows detection of serious disorders for which there may not be acceptable or agreed upon therapy, or in which the natural history is largely unknown, challenging the criterion that the disorder included in screening have acceptable treatment. How is this dilemma resolved? The answer is not yet available. It is hoped that the experience and findings from expanded newborn screening will be used to develop a new set of criteria that will apply to newborn screening. These criteria will likely retain the essence of the Wilson-Jungner compilation, but with important modifications that could be applied to any new screening venture. FALSE-POSITIVE RESULTS The majority of positive results in newborn screening, particularly when this result is only mildly or moderately abnormal, are not due to a disorder. Unfortunately, because screening is primarily based on a quantitative measure, false-positive results must be addressed. Metabolite concentrations vary among individuals, and the distribution curves of the markers from affected and unaffected populations are expected to be different (Figure 27-2). In screening, a value that separates the two distribution curves is established as a cutoff. The cutoff values of the quantitative biomarkers are established by the individual screening laboratories, and they can vary among the different laboratories because of variations in the testing technology. In general, the cutoff value should be set such that it is greater than the 99th percentile of the concentration in normal neonates and less than the 5th percentile of the concentration in affected neonates (Rinaldo et al, 2006). However, for disorders that are extremely rare, the population of affected neonates may be so small that establishing an appropriate cutoff becomes a challenge. In such cases, the laboratory may empirically set a cutoff at 3 to 4 SD from the population mean and adjust the values with experience to minimize the false positives without compromising the sensitivity. Specimens in which the concentration crosses the established cut-off are considered screen-positive. The metabolite concentration in a majority of unaffected individuals is below the cutoff value, but in a small proportion it crosses this threshold. The positive screening results that are not caused by a disorder are considered false positives. Because of some degree of overlap in the distribution curves of the affected and unaffected populations, these false positives cannot be entirely eliminated without compromising the sensitivity of screening. Furthermore other physiologic factors, such as immaturity of enzymes or stress and therapeutic interventions, can skew 326 PART VII Care of the Healthy Newborn Normal Population Affected Unaffected TP FN FP TN Test Positive Test Negative Number of Neonates True Negatives Sensitivity  [TP] / [TP  FN] Specificity  [TN] / [TN  FP] Positive Predictive Value  [TP] / [TP  FP] False Positives Affected False Negatives True Positives Mean (Normal) Cut-Off Mean (Affected) Analyte Concentration FIGURE 27-2 Distribution of quantitative markers measured in screening. For most conditions, the distribution in the unaffected or normal population overlaps that in the affected individuals. The number of false positives (FP), false negatives (FN), true positives (TP), and true negatives (TN) depends on the established cutoff. the concentrations of certain metabolites and lead to falsepositive screening results. Currently, indicators measured using immunoassays (e.g.,17-OHP in CAH and T4 in congenital hypothyroidism) or enzymatic activity (low GALT activity in galactosemia) are associated with the highest false-positive rates. The false-positive results are more common in preterm and low-birthweight infants than in full-term infants. For example, up to 85% of preterm infants have transiently low T4 levels (Paul et al, 1998). Transient increases in 17-OHP are another common abnormality in infants who are preterm or have low birthweights or have experienced perinatal stress (Pang and Shook, 1997). In addition, transient tyrosinemia is commonly observed in preterm and low-birthweight infants, although it can also occur in fullterm infants (Levy et al, 1969). Artifacts produced in the collection or transport of the Guthrie specimen account for some false-positive results. As mentioned in the discussion of the specimen collection procedure, collection of the specimen from a central line can result in mixing with amino acids in total parental nutrition solution and a false increase of amino acids in the specimen. Contamination with milk (or any drink containing milk) can result in a false elevation in galactose and the mistaken suspicion of galactosemia. Prolonged exposure to heat can reduce the activity of GALT in the specimen and produce a false impression of galactosemia when the enzyme assay is used to screen for this disorder. This error is common during the summer, especially when the specimens remain in a mailbox for a period of time. Some factors known to be associated with false positives are shown in Table 27-2 (Sahai et al, 2009). With the substitution of MS/MS for the traditional bacterial or specific assays, such as those for PKU and MSUD, the number of false-positive results is distinctly lower. For PKU screening, the false-positive rate is reported at 0.05% compared with 0.23% by earlier methods (Levy, 1998). In addition, with MS/MS the multiplexed approach allows for a profile of analytes rather than a single analyte level (e.g., phe/tyr ratio vs. only a phenylalanine level for PKU identification), further improving the specificity of screening (Chace et al, 1998; Schulze et al, 1999). Therefore the average positive predictive value for primary markers analyzed by MS/MS, previously reported to be 8% to 10% (Schulze et al, 2003; Wilken et al, 2003), can be improved substantially if the marker is evaluated in context with other metabolites that are screened or when a tiered approach is applied (Frazier et al, 2006; Sahai et al, 2007). Second-tier assays, such as molecular assays for cystic fibrosis or secondary immunoassays for hypothyroidism, are commonly performed to reduce the false-positive rates for primary markers analyzed by immunoassays or enzymatic assays. Nevertheless, false-positive results cannot be entirely eliminated; therefore it is important to reassure the parents that not every abnormal result of newborn screening inevitably implies a disorder and that transient or nonspecific abnormalities are common. Although all infants with an abnormal screening result must undergo repeated testing, the families should be informed that an initial positive result might have no medical implications. This approach can alleviate excessive anxiety and prevent unnecessary diagnostic procedures and treatment. INCREASED DETECTION BY SCREENING For certain disorders screened by MS/MS, many more infants are identified than the expected numbers based on clinical identification. The greatest increases are in the fatty acid oxidation disorders, including short-chain acyl-CoA dehydrogenase deficiency SCADD, MCADD, CHAPTER 27 Newborn Screening VLCADD, and carnitine deficiency, and in three organic acid disorders (glutaric acidemia type I, 3-methylcrotonylCoA carboxylase deficiency, and 3-ketothiolase deficiency; Wilken et al, 2003). Although some of the excess detection by screening could represent cases that were symptomatic but not diagnosed clinically, it is likely that the greater part of the excess represents infants with benign or milder forms of the disorders who would not come to clinical attention. Notably, Spiekerkoetter et al (2003) identified a frequent mutation in asymptomatic patients with VLCADD detected by newborn screening, and Ensenauer et al (2004) found a common, mild, and perhaps asymptomatic mutation in patients with isovaleric acidemia identified by screening. MISSED CASES Infants with congenital hypothyroidism, PKU, intermittent MSUD, glutaric acidemia, tyrosinemia I, and other screened disorders have been undetected by newborn screening. Laboratory or program errors were reported as the most common cause of these missed cases (Holtzman et al, 1974). In some instances, a specimen was never collected, such as when infants were transferred to another hospital. However, in screening for a multitude of disorders, each with its own biomarker that varies with time and physiologic states, an occasional affected neonate may have normal biomarker concentrations in the newborn specimen simply because the timing of collection was not ideal for the particular condition. Therefore physicians must exercise clinical judgment and not fall into the trap of excluding a diagnosis because an infant has presumably been screened. Specific testing for metabolic and endocrine disorders should be performed in any infant or child with symptoms that suggest the presence of such a disorder, regardless of the assumed or actual newborn screening result. THE FUTURE Many factors are impinging on newborn screening, including rapidly advancing technology, new and increasingly available therapeutic approaches to previously untreatable disorders, and family support groups and influential citizens or legislators who are heavily invested in individual disorders that may or may not be ready for inclusion as screening tests. To address these pressures, a committee convened by the American College of Medical Genetics 327 and the Maternal and Child Health Division of the U.S. Department of Health and Human Services meets regularly to study newly proposed newborn screening tests (and those currently recommended) to judge their suitability for inclusion in the screening panels. One new candidate for screening severe combined immunodeficiency has passed through a careful vetting process and has been recommended for inclusion in the standard panel. Other candidate conditions being considered are the lysosomal storage disorders and adrenoleukodystrophy, for which enzyme or early bone marrow therapies are available and are being studied, Fragile X syndrome, and the SmithLemli-Opitz syndrome. These disorders will be judged on the basis of frequency, severity, availability of preemptive therapies, and the cost and robustness of the screening test itself. Some disorders have been included in isolated state panels on the basis of the political influences mentioned above. Finally, there is the issue of retention of blood spots for future study. Despite multiple safeguards to protect the identity and anonymity of individuals, parents and civil libertarians are concerned that retention of these blood spots poses a threat to the privacy of individuals and that the specimens should be destroyed. If this view prevails, a resource of great value in the development of new and more effective tests, and one that is increasingly recognized as the avatar of personalized medicine, will be lost. SUGGESTED READINGS American College of Medical Genetics: Newborn Screening Expert Group. Newborn screening: toward a uniform panel and system, Genet Med, 2006. Clinical and Laboratory Standards Institute: Newborn screening for preterm, low birth weight and sick newborns: approved guideline. CLSI document I/LA31-A, Wayne, Penn, 2009, Clinical and Laboratory Standards Institute. Dobrowolski SF, Banas RA, Naylor EW, et al: DNA microarray technology for neonatal screening, Acta Paediatr Suppl 88:61-64, 1999. Kaye CI, Accurso F, La Franchi S, et al: Newborn screening fact sheets, Pediatrics 118:e934-e963, 2006. Matern D, Tortorelli S, Oglesbee D, et al: Reduction of the false-positive rate in newborn screening by implementation of MS/MS-based second-tier tests: the Mayo Clinic experience (2004-2007), J Inherit Metab Dis 30:585-592, 2007. Pass KA, Neto EC: Update: newborn screening for endocrinopathies, Endocrinol Metab Clin North Am 38:827-837, 2009. Rashed MS, Ozand PT, Bucknall MP, et al: Diagnosis of inborn errors of metabolism from blood spots by acylcarnitines and amino acids profiling using automated electrospray tandem mass spectrometry, Pediatr Res 38:324-331, 1995. Sahai I, Marsden D: Newborn screening, Crit Rev in Cli Lab Sci 46:55-82, 2009. Smith J, Kinney T: Clinical practice guidelines, quick reference guide for clinician: sickle cell disease: screening and management in newborns and infants, Am Fam Physician 48:95-102, 1993. Wilken B, Wiley V, Hammond J, et al: Screening newborns for inborn errors of metabolism by tandem mass spectrometry, N Engl J Med 348:2304-2312, 2003. Complete references used in this text can be found online at www.expertconsult.com C H A P T E R 28 Resuscitation in the Delivery Room Tina A. Leone and Neil N. Finer The transition from fetal to neonatal life is a dramatic and complex process involving extensive physiologic changes that are most obvious at the time of birth. Individuals who care for newly born infants must monitor the progress of the transition and be prepared to intervene when necessary. In the majority of births this transition occurs without a requirement for any significant assistance. However, when the need for intervention arises, the presence of providers skilled in neonatal resuscitation can be life saving. Each year approximately 4 million children are born in the United States (Martin et al, 2008) and more 30-fold as many are born worldwide. It is estimated that approximately 5% to 10% of all births will require some form of resuscitation beyond basic care, making neonatal resuscitation the most frequently practiced form of resuscitation in medical care. Throughout the world approximately 1 million newborn deaths are associated with birth asphyxia (Lawn et al, 2005). Whereas early effective newborn resuscitation will not eliminate all early neonatal mortality, such intervention will save many lives and significantly reduce subsequent morbidities. Attempts at reviving nonbreathing infants immediately after birth have occurred throughout recorded time with references in literature, religion, and early medicine. Although the organization and sophistication have changed, the basic principle and goal of initiating breathing have remained constant. During the last 20 years, attention has focused on the process of neonatal resuscitation. Resuscitation programs in other areas of medicine were initiated in the 1970s in an effort to improve knowledge about effective resuscitation and provide an action plan for early responders. The first such program was focused on adult cardiopulmonary resuscitation (Kattwinkel, 2006). These programs then began increasing in complexity and becoming more specific to different types of resuscitation needs. With the collaboration of the American Heart Association and the American Academy of Pediatrics, the Neonatal Resuscitation Program (NRP) was initiated in 1987 and was designed to address the specific needs of the newly born infant. The NRP textbook (Kattwinkel, 2006) now includes specific recommendations for the preterm infant. Various groups throughout the world also provide resuscitation recommendations that are more specific to the practices in certain regions. An international group of scientists, the International Liaison Committee on Resuscitation meets on a regular basis to review available resuscitation evidence for all the different areas of resuscitation and puts forth a summary of its review (Chamberlain, 2005). The overall goal of the NRP is similar to other resuscitation programs, in that it intends to teach large groups of individuals of varying backgrounds the principles of resuscitation and provide an action plan for providers. Similarly, a satisfactory end result of resuscitation would be common to all forms of resuscitation, namely to provide 328 adequate tissue oxygenation to prevent tissue injury and restore spontaneous cardiopulmonary function. However, when comparing neonatal resuscitation with other forms of resuscitation, several distinctions can be noted. First, the birth of an infant is a more predictable occurrence than most events that require resuscitation in an adult, such as an arrhythmia or a myocardial infarction. Although not every birth will require resuscitation, it is more reasonable to expect that skilled individuals can be present when the need for neonatal resuscitation arises. It is possible to anticipate with some accuracy which newborns will more likely require resuscitation based on perinatal factors and thus allow time for preparation. The second distinction of neonatal resuscitation compared with other forms of resuscitation involves the unique physiology involved in the normal fetal transition to neonatal life. The fetus exists in the protected environment of the uterus where temperature is closely controlled, the lungs are filled with fluid, continuous fetal breathing is not essential, and the gas exchange organ is the placenta. The transition that occurs at birth requires the newborn infant to increase heat production, initiate continuous breathing, replace the lung fluid with air, and significantly increase pulmonary blood flow so that gas exchange can occur in the lungs. The expectations for this transitional process and knowledge of how to effectively assist the process help to guide the current practice of newborn resuscitation. FETAL TRANSITION TO EXTRAUTERINE LIFE The complete transition from fetal to extrauterine life is complex and much more intricate than can be discussed in these few short paragraphs, but a basic knowledge of these processes will contribute to the understanding of the rationale for resuscitation practices. The key elements necessary for a successful transition to extrauterine life involve changes in thermoregulation, respiration, and circulation. In utero the fetal core temperature is approximately 0.5° C greater than the mother’s temperature (Gunn and Gluckman, 1983). Heat is produced by metabolic processes and is lost over this small temperature gradient through the placenta and skin (Gilbert et al, 1985). After birth the temperature gradient between the infant and the environment becomes much greater and heat is lost through the skin by radiation, convection, conduction, and evaporation. The newborn infant must begin producing heat through other mechanisms such as lipolysis of brown adipose tissue (Dawkins and Scopes, 1965). If heat is lost at a pace greater than it is produced, the infant will develop hypothermia. Preterm infants are at particular risk because of increased heat loss through immature skin, a greater surface area– to–body weight ratio, and decreased brown adipose tissue stores. 329 CHAPTER 28 Resuscitation in the Delivery Room The fetus lives in a fluid-filled environment, and the developing alveolar spaces are filled with lung fluid. Lung fluid production decreases in the days before delivery ­(Kitterman et al, 1979), and the remainder of lung fluid is reabsorbed into the pulmonary interstitial spaces after delivery (Bland, 1988). As the infant takes the first breaths after birth, a negative intrathoracic pressure of approximately 50 cm H2O is generated (Vyas et al, 1986). The alveoli become filled with air, and with the help of pulmonary surfactant the lungs retain a small amount of air persisting at the end of exhalation, which is known as the functional residual capacity (FRC). Although the fetus makes breathing movements in utero, these efforts are intermittent and are not required for fetal gas exchange. Continuous spontaneous breathing is maintained after birth by several mechanisms, including the activation of chemoreceptors, the decrease in hormones that inhibit respirations, and the presence of natural environmental stimulation. Spontaneous breathing can be suppressed at birth for several reasons, most critical of which is the presence of acidosis secondary to compromised fetal circulation. The natural history of the physiologic responses to asphyxia and acidosis has been described by researchers evaluating animal models. Geoffrey Dawes described the breathing response to acidosis in different animal species (Dawes, 1968). He noted that when pH was decreased, animals typically had a relatively short period of apnea followed by gasping. The gasping pattern then increased in rate until breathing ceased again for a second period of apnea. The physiologic effects that occur with worsening acidosis are noted in Figure 28-1. Dawes also noted that the first period of primary apnea could be reversed with stimulation, whereas the second period, secondary or terminal apnea, required assisted ventilation to ultimately establish spontaneous breathing. The first sign of improvement was noted to be an increase in heart rate. Further recovery was noted when the newborn begins gasping again. The secondary period of apnea varies in duration depending on the duration of asphyxia and degree of acidosis. In the clinical situation, the exact timing of onset of acidosis is generally unknown; therefore any observed apnea may be either primary or secondary. This finding is the basis of the resuscitation recommendation that stimulation may be attempted in the presence of apnea, but if not quickly successful, assisted ventilation should be initiated promptly. Without the presence of acidosis, a newborn may also develop apnea because of recent exposure to respiratory suppressing medications such as narcotics, anesthetics, and magnesium. These medications cross the placenta when given to the mother and may depress the newborn’s ­respiratory drive, depending on the time of administration and dose. Fetal circulation is unique because gas exchange takes place in the placenta. In the fetal heart, oxygenated blood returning via the umbilical vein is mixed with deoxygenated blood from the superior and inferior venae cavae, and it is distributed differentially throughout the body. The most oxygenated blood is directed toward the brain, and the most deoxygenated blood is directed toward the placenta. Thus blood returning from the placenta to the right atrium is preferentially streamed via the foramen ovale to the left atrium and ventricle and then to the ascending RESPIRATORY EFFORT HEART RATE AORTIC BLOOD PRESSURE CENTRAL VENOUS PRESSURE CARDIAC OUTPUT BLOOD FLOW, HEAD & HEART BLOOD FLOW, BODY BLOOD FLOW, LUNGS BRAIN DAMAGE TIME pH 7.4 7.1 7.0 6.7 ASPHYXIA RESUSCITATION FIGURE 28-1 The sequence of cardiopulmonary changes with asphyxia and resuscitation. Time is on the horizontal axis. Asphyxia progresses from left to right; resuscitation proceeds from right to left. Units of time are not given. If there is complete interruption of respiratory gas exchange, the entire process of asphyxia from extreme left to right could occur in approximately 10 minutes. It could take much longer with an asphyxiating process that only partly interrupts gas exchange or does so completely, but only for repeated brief periods. With resuscitation, the process reverses, beginning at the point to which asphyxia has proceeded. (Adapted from Dawes G: Foetal and neonatal physiology, Chicago, 1968, Year Book; and Avery GN: Neonatology, Philadelphia, 1987, JB Lippincott.) aorta, providing the brain with the most oxygenated blood. Fetal channels, including the ductus arteriosus and foramen ovale, allow most blood flow to bypass the lungs with their intrinsically high vascular resistance. As a result, pulmonary blood flow is approximately 8% of the total cardiac output. In the mature postnatal circulation, the lungs must receive 100% of the cardiac output. When the lowresistance placental circulation is removed after birth, the infant’s systemic vascular resistance increases while the pulmonary vascular resistance begins to fall as a result of pulmonary expansion, increased arterial and alveolar oxygen tension, and local vasodilators. These changes result in a dramatic increase in pulmonary blood flow. The average fetal oxyhemoglobin saturation as measured in fetal lambs is approximately 50% (Nijland et al, 1995), but ranges in different sites within the fetal circulation between values of 20% and 80% (Teitel, 1988). The oxyhemoglobin saturation rises gradually over the first 5 to 15 minutes of life to 90% or greater as the air spaces are cleared of fluid. A diagram of the blood flow patterns in the fetus and normally transitioning newborn is shown in Figures 28-2 and 28-3. In the face of poor transition secondary to asphyxia, 330 PART VII Care of the Healthy Newborn FIGURE 28-2 Fetal circulation. Oxygenated blood leaves the placenta by way of the umbilical vein (vessel without stippling). The blood flows into the portal sinus in the liver (not shown), and a variable portion of it perfuses the liver. The remainder passes from the portal sinus through the ductus venosus into the inferior vena cava, where it joins blood from the viscera (represented by the kidney, gut, and skin). Approximately half of the inferior vena cava flow passes through the foramen ovale to the left atrium, where it mixes with a small amount of pulmonary venous blood. This relatively well-oxygenated blood (light stippling) supplies the heart and brain by way of the ascending aorta. The other half of the inferior vena cava stream mixes with superior vena cava blood and enters the right ventricle (blood in the right atrium and ventricle has little oxygen, which is denoted by heavy stippling). Because the pulmonary arterioles are constricted, most of the blood in the main pulmonary artery flows through the ductus arteriosus (DA), so that the descending aorta’s blood has less oxygen (heavy stippling) than blood in the ascending aorta (light stippling). (From Avery GN: Neonatology, Philadelphia, 1987, JB Lippincott.) meconium aspiration, pneumonia, or extreme prematurity, the lungs might not be able to develop efficient gas exchange; therefore the oxygen saturation might not increase as expected. In addition, in some situations the normal reduction in pulmonary vascular resistance does not fully occur, resulting in persistent pulmonary hypertension and decreased effective pulmonary blood flow with continued right-to-left shunting through the aforementioned fetal channels. Right-to-left shunting will lead to persistent hypoxemia and potentially to significant newborn illness requiring intensive care until the circulatory pattern adjusts to extrauterine life. The circulatory pattern associated with poor transition is noted in Figure 28-4. FIGURE 28-3 Circulation in the normal newborn. After expansion of the lungs and ligation of the umbilical cord, pulmonary blood flow increases and left atrial and systemic arterial pressures increase while pulmonary arterial and right heart pressures decrease. When the left atrial pressure exceeds right atrial pressure, the foramen ovale closes so that all of the inferior and superior vena cava blood leaves the right atrium, enters the right ventricle, and is pumped through the pulmonary artery toward the lung. With the increase in systemic arterial pressure and decrease in pulmonary arterial pressure, flow through the ductus arteriosus becomes left to right, and the ductus constricts and closes. The course of the circulation is the same as in the adult. (From Avery GN: Neonatology, Philadelphia, 1987, JB Lippincott.) ENVIRONMENT AND PREPARATION The environment in which the infant is born should facilitate the transition to neonatal life as much as possible and should be able to readily accommodate the needs of a resuscitation team when necessary. Hospitals vary in the approach to the details of how to prepare for resuscitation. For example, some hospitals have a separate room designated for resuscitation where the infant will be taken after birth, whereas others have the delivery room adjacent to the neonatal intensive care unit (NICU) and the infant is resuscitated in the NICU if necessary. Hospitals may bring all the necessary equipment into the delivery room when resuscitation is expected or have every delivery room equipped for any resuscitation. Wherever the resuscitation will take place, a few key elements must be considered. The room should be warm enough to prevent excessive newborn heat loss, bright enough to assess the infant’s clinical status, and large enough to accommodate the necessary personnel and equipment to care for the baby. When no added risks to the newborn are identified, the term birth frequently occurs without the attendance of a CHAPTER 28 Resuscitation in the Delivery Room 331 TABLE 28-1 Risk Factors for Neonatal Resuscitation Maternal ­Factors Fetal ­Factors Intrapartum Factors Maternal ­hypertension Preterm delivery Opiates in labor Maternal ­infection Breech ­presentation Rupture of membranes >18 hours Multiple ­gestation (­preterm) Shoulder ­dystocia Meconium-stained amniotic fluid Nonreassuring fetal heart rate patterns Emergency cesarean section Prolapsed cord Data from Aziz K et al: Ante- and intra-partum factors that predict increased need for neonatal resuscitation, Resuscitation 79:444-452, 2008. FIGURE 28-4 Circulation in an asphyxiated newborn with incomplete expansion of the lungs. Pulmonary vascular resistance is high, pulmonary blood flow is low (normal number of pulmonary veins), and flow through the ductus arteriosus is high. With little pulmonary arterial flow, left atrial pressure decreases below right atrial pressure, the foramen ovale opens, and vena cava blood flows through the foramen into the left atrium. Partially venous blood goes to the brain via the ascending aorta. The blood of the descending aorta that goes to the viscera has less oxygen than that of the ascending aorta (heavy stippling), because of the reverse flow through the ductus arteriosus. Therefore the circulation is the same as in the fetus, except that there is less well-oxygenated blood in the inferior vena cava and umbilical vein. (From Avery GN: Neonatology, Philadelphia, 1987, JB Lippincott.) specific neonatal resuscitation team. However, it is recommended that one individual be present who is responsible only for the infant and can quickly alert a neonatal resuscitation team if necessary. Even the best neonatal resuscitation triage systems will not anticipate the need for resuscitation in all cases. Using a retrospective risk assessment scoring system, 6% of newborns requiring resuscitation were not be identified based on risk factors (Smith et al, 1985). Similarly, a recent review found that when a riskbased determination of neonatal resuscitation team attendance at deliveries was used, 22% of infants at attended deliveries required at least assisted ventilation (Aziz et al, 2008). These investigators found that the most significant risk factors were preterm birth, emergency cesarean section, and meconium-stained amniotic fluid. Other significant risk factors for the need for resuscitation are listed in Table 28-1. Antenatal determination of risk allows the resuscitation team to be present for the delivery and to be more thoroughly prepared for the situation. The composition of the neonatal resuscitation team will vary tremendously among institutions. Probably the most important factor in how well a team functions is how the group has prepared for the delivery. Preparation involves both the immediate tasks of readying equipment and personnel for an individual situation as well as the more broad institutional preparation of training team members and providing appropriate space and equipment. We believe that when there is a strong suspicion that the newborn infant will be born in a compromised state, a minimum of three team members should be present, including one member with significant previous experience leading neonatal resuscitations. Each team member is assigned tasks that are performed on a regular basis. The leader is expected to ensure that the appropriate interventions are performed and that they are performed well. All team members are encouraged and expected to speak up if a problem is noticed or if they believe an alternate course would be beneficial. It seems logical that teams that regularly work together and divide tasks in a routine manner will have a better chance of functioning smoothly during a critical situation. Institutions can facilitate team readiness with regular review of practices and mock codes or simulator training to practice uncommon scenarios. In our institution, University of California San Diego (UCSD), video-taped resuscitations are reviewed twice monthly with representatives from all disciplines involved in the resuscitation team (Carbine et al, 2000); this is done as a quality-assurance procedure and allows for ongoing identification of areas needing improvement. In addition, this practice provides an opportunity for education and discussion about potential solutions to repetitive problems of newborn resuscitation. UCSD also instituted a supplemental training program for pediatric trainees to obtain experience in a preclinical situation (Garey et al). These training sessions allow adequate time to review scenarios in detail, and trainees are given the opportunity to prepare and operate the equipment and practice procedures on an individualized basis. Others have used simulators to provide additional resuscitation training (Halamek, 2008). All of these training elements help to prepare teams for future resuscitations. 332 PART VII Care of the Healthy Newborn When preparing for a resuscitation that will occur in the immediate short-term period, additional aspects must be addressed. The most important element in preparing for the resuscitation is the discussion that occurs among the team members who will participate. This discussion, known as a prebrief, is an essential part of improving communication during resuscitation and involves reviewing the clinical information, reviewing the available resources and introducing the individuals who will participate, assigning tasks to individuals, and making equipment as readily available as possible. A discussion of the overall plan will help team members know what to expect and how to address problems that could occur. We have formalized this process by creating a prebrief checklist for the team to review while preparing for the delivery. Checklists are considered one of the most important methods of preventing medical errors. A briefing after the resuscitation (i.e., a postbrief) allows team members to address positive and negative aspects of the resuscitation. This briefing is an important element in the preparation for future resuscitations, because team members can immediately identify how to improve the care being provided. An example of our current checklist with prebrief and postbrief documentation is shown in Figure 28-5. To better train the staff in these techniques, UCSD instituted training in crew resource management, a methodology developed for training air crews from the late 1970s that evolved from a careful evaluation of the role of human error in air crashes (Cooper et al, 1980). ASSESSMENT Immediately after birth the infant’s condition is evaluated by general observation and measurement of specific parameters. Typically a healthy newborn will cry vigorously and maintain adequate respirations. The color will transition from blue to pink over the first 2 to 5 minutes, the heart rate will remain in the range of 140 to 160 beats/ min, and the infant will demonstrate adequate muscle tone with some flexion of the extremities. The overall assessment of an infant who is having difficulty with the transition to extrauterine life will often reveal apnea, bradycardia, hypotonia, and cyanosis or pallor. After the initial steps of resuscitation, interventions are based mainly on the evaluation of respiratory effort and heart rate, which need to be assessed continually throughout the resuscitation. Heart rate can be monitored by auscultation or palpation of the cord pulsations, with auscultation being the more reliable method (Owen and Wyllie, 2004). The continuous monitoring of this parameter gives providers an ongoing indication of the effectiveness of interventions. The use of more sophisticated monitoring devices, such as a pulse oximeter or electrocardiographic leads, can be helpful during resuscitation. A pulse oximeter can provide the resuscitation team with a continuous audible and visual indication of the newborn’s heart rate throughout the various steps of resuscitation while allowing all team members to perform other tasks. In addition, the pulse oximeter can be used as a more accurate measure of oxygenation than the evaluation of color alone, which is an unreliable measure of the infant’s oxygen saturation (O’Donnell et al, 2007). Whenever interventions beyond brief positive-pressure ventilation by mask are required, a pulse oximeter should be considered. Electrocardiographic monitoring may be helpful, but at this point has been infrequently evaluated in the delivery room and does not provide information about oxygenation. The overall assessment of a newborn was quantified by Virginia Apgar in the 1950s with the Apgar score (Apgar, 1953). The score describes the infant’s condition at the time it is assigned and consists of a 10-point scale, with a maximum of 2 points assigned for each of the following categories: respirations, heart rate, color, tone, and reflex irritability. Table 28-2 shows the detailed components of the score. Although the components of the score include items that are assessed for determining interventions, the score itself is not used to determine the need for interventions. The score was initially intended to provide a uniform, objective assessment of the infant’s condition and was used as a tool to compare different practices, especially obstetric anesthetic practices. Despite the intent of objectivity, there is often disagreement in score assignment among various practitioners (O’Donnell et al, 2006). Low scores are associated with increased risk of neonatal mortality (Casey et al, 2001), but have not been predictive of neurodevelopmental outcome (Nelson and Ellenberg, 1981). Interpreting the score when interventions are being provided may be difficult, and current recommendations suggest that clinicians should document the interventions used at the time the score is assigned (American Academy of Pediatrics Committee on Fetus and Newborn and American College of Obstetricians and Gynecologists and Committee on Obstetric Practice, 2006). INITIAL STEPS: TEMPERATURE MANAGEMENT AND MAINTAINING THE AIRWAY In the first few seconds after birth, all infants are evaluated for signs of life and a determination of the need for further assistance is made. This evaluation is done both formally as described in the NRP and informally as the initial care providers observe the infant in the first few moments of life. When the determination is made that further assistance and formal resuscitation are necessary, the infant is then placed on a radiant warmer and positioned appropriately for resuscitation to proceed. Appropriate positioning includes placing the infant supine on the warmer in such a way that care providers have easy access. In addition, the head should be in a neutral or “sniffing” position to facilitate maintenance of an open airway. Frequently the oropharynx contains fluid that can be removed by suctioning with a standard bulb syringe. An infant born through meconium-stained amniotic fluid is at risk for aspirating meconium and developing significant pulmonary disease, known as meconium aspiration syndrome, which may also be accompanied by persistent pulmonary hypertension. For many years, routine management of all infants with meconium-stained amniotic fluid included suctioning of the mouth and pharynx once the head was delivered, but before the shoulders were delivered, and endotracheal intubation and tracheal suctioning in an attempt to remove any meconium from the trachea and prevent the development of meconium CHAPTER 28 Resuscitation in the Delivery Room 333 FIGURE 28-5 Delivery room resuscitation checklist form. TABLE 28-2 The Apgar Score Feature Evaluated 0 Points 1 Point 2 Points Heart rate (beats/min) 0 <100 >100 Respiratory effort Apnea Irregular, shallow, or gasping respirations Vigorous and crying Color Pale, blue Pale or blue extremities Pink Muscle tone Absent Weak, passive tone Active ­movement Reflex ­irritability Absent Grimace Active ­avoidance aspiration syndrome. Routine suctioning of the mouth and nose before delivery of the shoulders has been shown to be of no benefit in decreasing the incidence of meconium aspiration syndrome, ventilation for meconium aspiration syndrome, or mortality (Vain et al, 2004). Recognizing that intubation may not be necessary for all infants and that the procedure may be associated with complications, a more selective approach was evaluated in randomized controlled trials (Wiswell et al, 2000). A metaanalysis of studies that have evaluated this question supported the notion that universal endotracheal suctioning does not result in a lower incidence of meconium aspiration syndrome or improve mortality compared with selective endotracheal 334 PART VII Care of the Healthy Newborn suctioning (Halliday, 2000). The likelihood that an infant with meconium-stained amniotic fluid will develop meconium aspiration syndrome is increased in the presence of fetal distress. The selective approach to endotracheal suctioning requires a quick evaluation of the infant after delivery. If the infant is vigorous with good respiratory effort, normal heart rate, and normal tone, the steps of resuscitation proceed as usual. However, if the infant is not vigorous and has poor respiratory effort, a heart rate less than 100 beats/min, or decreased tone, rapid endotracheal intubation for suctioning is currently indicated, although this has not been evaluated in a standardized manner. When endotracheal intubation for suctioning is performed, the resuscitation team must continue to adequately monitor the infant and may need to proceed quickly to providing assisted ventilation if bradycardia is present. Whereas temperature control is important for all newborn infants, it is particularly important for the extremely preterm infant. Preterm infants are commonly admitted to the NICU with core temperatures well below 37° C. In a population-based analysis of all infants younger than 26 weeks’ gestation, more than one third of these preterm infants had an admission temperature of less than 35° C (Costeloe et al, 2000). More disturbing is the fact that these infants with hypothermia on admission survived less often than those with admission temperatures greater than 35° C. Admission temperatures can be improved in preterm infants by immediately covering the infant’s body with polyethylene wrap before drying the infant (Vohra et al, 2004). With this approach the infant’s head is left out of the wrap and is dried, but the body is not dried before wrap application. Other measures for maintaining infant temperatures include performing resuscitations in a room that is kept at an ambient temperature of approximately 27° to 30° C, using radiant warmers with servo-controlled temperature probes placed on the infant within minutes of delivery, and the use of accessory prewarmed mattress or heating pads for the tiniest infants. Hats are routinely used as a method of decreasing heat loss, but they have not been shown to be consistently effective (McCall et al, 2008). It is important to note that as a required safety feature, radiant warmers substantially decrease their power output after 15 minutes of continuous operation at full power. If this decrease in power is unrecognized, the infant will be exposed to much less radiant heat. By applying the temperature probe and using the warmer in servo-controlled mode, the temperature output will adjust as needed and the power will not automatically decrease. ASSISTING VENTILATION As the newborn infant begins breathing and replaces the lung fluid with air, the lungs become inflated and an FRC is established and maintained. With inadequate development of FRC, the infant will not be able to oxygenate and will develop bradycardia if inadequate FRC is prolonged. Providing assisted ventilation when the infant’s spontaneous breathing is inadequate is probably the most important step in newborn resuscitation. The indications for providing assisted ventilation (positive-pressure ventilation) include apnea or bradycardia of less than 100 beats/min. Positive-pressure ventilation can be delivered noninvasively with a pressure delivery device and a face mask or invasively with the same pressure delivery device and an endotracheal tube. Pressure delivery devices can include self-inflating bags, flow-inflating or anesthesia bags, and t-piece resuscitators, each with its own advantages and disadvantages. The self-inflating bag is easy to use for inexperienced personnel and will work in the absence of a gas source. However, the self-inflating bag requires a reservoir to provide nearly 100% oxygen and does not consistently provide adequate positive end expiratory pressure (PEEP). These devices can deliver very high pressure if not used carefully. Although they have pressure blow-off valves, these valves do not always open at the target blow-off pressures (Finer et al, 1986). An anesthesia bag or flow-inflating bag requires a gas source for use, allows the operator to vary delivery pressures continuously based on the felt compliance, but requires significant practice to develop expertise. However, using a test lung and intermittent airway occlusion, experienced anesthesiologists were unable to recognize the increased resistance from an airway obstruction using only their hands (Spears et al, 1991). A t-piece resuscitator is easy to use, requires a gas source, and delivers the most consistent levels of pressure, but requires intentional effort to vary the pressure levels (Hoskyns et al, 1987). The flow-inflating bag and t-piece resuscitator allow the operator to deliver continuous positive airway pressure (CPAP) or PEEP relatively easily (Bennett et al, 2005; Finer et al, 2001). A level of experience is required to perform assisted ventilation using a face mask and resuscitation device, and this is especially true for an infant with extremely low birthweight. It is important to maintain an open airway for pressure to be transmitted to the lungs. The procedure of obtaining and maintaining an open airway includes at minimum clearing fluid with a suction device, holding the head in a neutral position, and sometimes lifting the jaw slightly anteriorly. The face mask must make an adequate seal with the face in order for air to pass to the lungs effectively. No device will adequately inflate the lungs if there is a large leak present between the mask and the face. Wood et al (2008) measured face mask leaks of more than 55% when participants were evaluated, providing positive pressure to manikins at baseline. The amount of the leak was able to be decreased to approximately 30% with specific instruction (Wood et al, 2008). Until recently there were no masks that were small enough to provide an adequate seal over the mouth and nose for the smallest infants. Such masks are now readily available and facilitate bagmask resuscitation of small infants. Signs that the airway is open and air is being delivered to the lungs include visual inspection of chest rise with each breath and improvement in the clinical condition, including heart rate and color. The use of a colorimetric carbon dioxide detector during bagging will allow confirmation that gas exchange is occurring by the observed color change of the device, or it will alert the operator of an obstructed airway by not changing color (Leone et al, 2006). Airway obstruction is common in the preterm infant during positive-pressure ventilation immediately after birth (Finer et al, 2009). It is important to remember that these devices will not change color in the absence of pulmonary blood flow, as occurs with inadequate cardiac output. At times, multiple maneuvers are CHAPTER 28 Resuscitation in the Delivery Room required to achieve an open airway, such as readjusting the head and mask positions, choosing a mask of more appropriate size, and further suctioning of the pharynx. Alternate methods of providing an open airway include the use of a nasopharyngeal tube (Lindner et al, 1999), a laryngeal mask airway device (Grein and Weiner, 2005), or an endotracheal tube. The amount of pressure provided with each breath during assisted ventilation is critical to establishing lung inflation and therefore adequate oxygenation. Although it is important to provide adequate pressure for ventilation, excessive pressure can contribute to lung injury. Achieving the correct balance of these goals is not simple and is an area of resuscitation that requires more study. A single specific level of inspiratory pressure will never be appropriate for every baby. Initial inflation pressures of 25 to 30 cm H2O are probably adequate for most term infants. The NRP textbook (Kattwinkel, 2006) recommends initial pressures of 20 to 25 cm H2O for preterm infants. The first few breaths may require increased pressure if lung fluid has not been cleared, as occurs when the infant does not initiate spontaneous breathing. Newborn infants with specific pulmonary disorders such as pneumonia or pulmonary hypoplasia also frequently require increased inspiratory pressure. It has been shown that using enough pressure to produce visible chest rise is associated with hypocarbia on admission blood gas evaluation (Tracy et al, 2004), and excessive pressure may decrease the effectiveness of surfactant therapy (Bjorklund et al, 1997). It may be possible to establish FRC without increasing peak inspiratory pressures by providing a few prolonged inflations (3 to 5 seconds of inspiration), although the use of prolonged inflations has not been associated with better outcomes than conventional breaths during resuscitation (Lindner et al, 2005). Choosing the actual initial inspiratory pressure is less important than continuously assessing the progress of the intervention. A manometer in the circuit during assisted ventilation provides the clinician with an indication of the administered pressure, although if the airway is blocked this pressure is not delivered to the lungs. The volume of air delivered to the lungs seems to be more important than the absolute pressure delivered in the development of lung injury. Tidal volume can be monitored with respiratory function monitors that are placed in the respiratory circuit (Schmolzer et al, 2010). The most critical component of continued assessment is evaluation of the infant’s response to the intervention. If the condition of the infant does not improve after initiating ventilation, then the ventilation is most likely inadequate. In our experience the two most likely reasons for inadequate ventilation are a blocked airway and insufficient inspiratory pressure. The occluded airway can be noted using a colorimetric carbon dioxide device as described previously, and frequently it can be corrected with changes in position or suctioning while inadequate pressure is corrected by adjusting the ventilating device. The use of continuous pressure throughout the breathing cycle seems to be beneficial for establishing FRC and improving surfactant function (Hartog et al, 2000; Michna et al, 1999; Siew et al, 2009); this is accomplished during assisted ventilation with the use of PEEP or with CPAP when 335 additional inspiratory pressure is not needed. In the absence of PEEP, a lung that has been inflated with assisted inspiratory pressure will lose on expiration most of the volume that had been delivered on inspiration. This pattern of repeated inflation and deflation is frequently thought to be associated with lung injury. In preterm infants, a general approach of using CPAP as a primary mode of respiratory support in NICUs has been associated with a low incidence of chronic lung disease (Ammari et al, 2005; Avery et al, 1987; Van Marter et al, 2000; Vanpee et al, 2007). Recently, te Pas and Walther (2007) evaluated a ventilation strategy that included the use of a I-piece resuscitator and a nasopharyngeal tube to deliver a prolonged breath followed by CPAP, compared with the use of a self-inflating bag to deliver positive pressure ventilation when needed without any CPAP provided until infants reached the NICU. Infants treated with the prolonged breath and CPAP required less endotracheal intubation and had lower rates of bronchopulmonary dysplasia. The use of CPAP compared with endotracheal intubation and mechanical ventilation as the initial mode of respiratory support was evaluated in the Continuous Positive Airway Pressure or Intubation at Birth (COIN) trial (Morley et al, 2008), but the intervention did not begin until after 5 minutes of life and therefore cannot inform the decision about the use of these therapies as immediate resuscitative interventions. The recently completed Surfactant Positive Airway Pressure and Pulse Oximetry Trial (SUPPORT Study Group 2010. Available at http://clinicaltrials.gov/ct2/show/NCT002333 24?term=neonatal+respiratory+support&rank=15) evaluated the use of CPAP versus intubation and surfactant use in the first hour of life as the primary respiratory support for preterm infants. These trials suggest that CPAP is beneficial for improving some short-term outcomes associated with prematurtiy. Although these trials do not evaluate these therapies specifically in the resuscitation period, the respiratory support provided shortly after birth should probably mimic the support provided in the NICU after delivery. If assisted ventilation is necessary for a prolonged period of time or if other resuscitative measures have been unsuccessful, ventilation must be provided by a more secure device such as an endotracheal tube. If it has been difficult to maintain an open airway while ventilating via a face mask, an appropriately placed endotracheal tube will provide a stable airway. This stability will allow more consistent delivery of gas to the lungs and therefore provide for the ability to establish and maintain FRC. At this time intubation is required for administering surfactant, and it can be used to administer other medications such as epinephrine if necessary for resuscitation. Finally, for non-vigorous infants born through meconium-stained amniotic fluid, intubation is performed for suctioning of the airway. The intubation procedure is often critical for successful resuscitation, requires a significant amount of skill and experience to perform reliably, and may be associated with serious complications. The placement of a laryngoscope in the pharynx often produces vagal nerve stimulation, which leads to bradycardia. A photograph of the desired view of the larynx is shown in Figure 28-6. Assisted ventilation must be paused for the procedure, which if prolonged can lead to hypoxemia and bradycardia. Intubation has been shown to increase blood pressure 336 PART VII Care of the Healthy Newborn TRANSITIONAL OXYGENATION AND OXYGEN USE FIGURE 28-6 (See also Color Plate 18.) View of the glottis and vocal cords as the laryngoscope is gently lifted. (From American Heart Association, American Academy of Pediatrics: Kattwinkel J, editor: Neonatal resuscitation textbook, ed 5, Elk Grove Village, Ill., 2006, American Academy of Pediatrics.) and intracranial pressure (Kelly and Finer, 1984). Trauma to the mouth, pharynx, vocal cords, and trachea are all possible complications of intubation. Performing the intubation procedure when the infant already has bradycardia and hypoxemia can lead to further decline in heart rate and oxygenation (O’Donnell et al, 2006). In addition, hypoxia and bradycardia are more likely when intubation attempts are prolonged beyond 30 seconds. Therefore it is most appropriate to make an attempt to stabilize the infant with noninvasive ventilation before performing the procedure, limiting each attempt to 30 seconds or less (Lane et al, 2004), and allow time for the infant to recover with noninvasive ventilation between attempts. If misplacement of the endotracheal tube in the esophagus goes unrecognized, the infant may experience further clinical deterioration. Clinical signs that the endotracheal tube has been correctly placed in the trachea include auscultation of breath sounds over the anterolateral aspects of the lungs (near the axilla), mist visible in the endotracheal tube, chest rise, and clinical improvement in heart rate and color or oxygen saturation. The use of a colorimetric carbon dioxide detector to confirm intubation significantly decreases the amount of time necessary to determine correct placement of the endotracheal tube from approximately 40 seconds to less than 10 seconds (Aziz et al, 1999; Repetto et al, 2001); this is the primary method of determining endotracheal tube placement. Successful placement of the endotracheal tube is not always easy and is sometimes not possible. Airway anomalies may make alternate methods of airway management necessary. A laryngeal mask airway is one such alternative that has been described for use in patients with Pierre Robin sequence or other airway anomalies (Yao et al, 2004). Some practitioners have reported using the laryngeal mask airway for all positive-pressure delivery after birth, with success noted in infants as small as 1.2 kg (Gandini and Brimacombe, 1999). Administration of medications including surfactant and epinephrine through this device has undergone preliminary investigations (Chen et al, 2008; Trevisanuto et al, 2005). It is critical to remember that fetal arterial oxygen levels are much lower than newborn arterial oxygen levels. The transition from fetal to newborn levels does not take place instantaneously. Using pulse oximetry in the immediate newborn period, several investigators have established that this transition with an ultimate oxygen saturation level greater than 90% takes 5 to 15 minutes to occur in infants who do not otherwise require resuscitation. For term and preterm nonresuscitated infants, the median interquartile range (IQR) of oxygen saturation at 3 minutes was 76% (64% to 87%) and at 5 minutes was 90% (79% to 91%; Kamlin et al, 2006). Expected oxygen saturation levels are slightly lower for preterm compared with term and for infants delivered via cesarean section compared with those delivered vaginally (Rabi et al, 2006). Oxygen saturation levels measured in preductal sites are 5% to 10% higher than those measured from postductal sites for approximately 15 minutes of life (Mariani et al, 2007). A great deal of variability occurs in the saturation values among different healthy individuals during the first 5 minutes of life, but a resuscitation team can expect that there be a steady, albeit slow, increase in levels over several minutes. If values are below a threshold at different time points or not progressively increasing, intervention should be considered. The use of pure oxygen for ventilation became routine practice in resuscitation simply because it seemed logical that oxygen would be beneficial. However, the recognition that oxygen could also be toxic led some investigators to question this previously well accepted practice. The toxicity of oxygen is anticipated when the cellular antioxidant capacity is impaired, as occurs during the reperfusion phase following an hypoxic-ischemic insult. After animal studies showed the potential harmful effects of oxygen (Poulsen et al, 1993; Rootwelt et al, 1992), clinical trials were conducted to evaluate the effects of oxygen use during resuscitation of depressed infants. Several worldwide trials have compared the use of pure (100%) oxygen with room (ambient) air (21% oxygen) as the initial ventilating gas for asphyxiated newborns. These trials found that air was as successful as oxygen in achieving resuscitation, and infants resuscitated with air had a shorter time to initiate spontaneous breathing and less evidence of oxidative stress (Ramji et al, 2003; Saugstad et al, 1998; Vento et al, 2001, 2003). Metaanalyses of several of the trials indicated that infants resuscitated with air had a lower risk of mortality than those resuscitated with pure oxygen (Rabi et al, 2007; Tan et al, 2005). These trials have been criticized because they were not all strictly randomized, and some sites were in developing countries; however, metaanalysis of the strictly randomized trials, which were mostly done in European centers, demonstrates significant benefit for survival with the use of air compared with pure oxygen (Saugstad et al, 2008). The preterm infant may be more susceptible to any harmful effects of excessive oxygen exposure, because of decreased antioxidant enzyme capacity. Some of the infants in the previous oxygen trials were preterm, but few weighed less than 1000 g. Review of the outcomes for preterm infants (<37 weeks’ gestation) demonstrated an CHAPTER 28 Resuscitation in the Delivery Room even greater reduction in death for those initially resuscitated with room air compared with oxygen (Saugstad et al, 2005). NICUs generally attempt to reduce oxygen toxicity by limiting the amount of oxygen administered to newborns using an upper limit for oxygen saturation and adjusting supplemented oxygen levels to maintain saturation levels within that limit. The unlimited use of oxygen during resuscitation exposes the preterm infant to higher oxygen saturation levels than would routinely be accepted in the NICU. Several small trials of oxygen use during resuscitation of preterm infants have been performed in the last 3 years. Among 42 preterm infants younger than 32 weeks’ gestation treated with either pure oxygen for 5 minutes or a targeted oxygen strategy beginning with 21% oxygen, all infants were provided some level of supplemental oxygen (Wang et al, 2008). Infants initiated on pure oxygen had higher pulse oximeter saturation (Spo2) levels during transition, but did not have any differences in heart rate, need for intubation, or survival. The infants provided air from the start of resuscitation all required an increase in inspired oxygen to obtain the specified oxygen saturation targets. Using 30% versus 90% oxygen at the start of resuscitation and adjusting the concentration based on the clinical status of the infant, Escrig et al (2008) found that infants initially receiving 30% oxygen had lower overall exposure to oxygen without any adverse effects. The group receiving 30% oxygen received increased oxygen concentrations up to approximately 55% at 5 minutes, but both groups received similar oxygen concentrations after 4 minutes of life, with a level of approximately 35% by 15 minutes of life (Escrig et al, 2008). In a more recent study, these investigators also reported a decrease in the incidence of bronchopulmonary dysplasia in infants initially receiving 30% oxygen compared with 90% oxygen (Vento et al, 2009). Finally, the Room Air versus Oxygen Administration during Resuscitation (ROAR) study was a blinded, randomized controlled trial in 106 infants at 32 weeks’ gestation or less comparing three oxygen strategies: one group received 100% throughout (HOB), one received 100% initially (MOB), and one received 21% (LOB) initially; the last two groups had the oxygen concentration titrated to keep Spo2 at 85% to 92%. Spo2 levels were below the specified target range 61% of the time in the LOB group and were above the target range 49% of the time in the HOB group (p <0.001; Rabi et al, 2008). Larger trials of preterm infants treated with different oxygen strategies are necessary to determine the long-term effects of resuscitation with different oxygen concentrations. Oxygen use in the first minutes of life could affect survival and common neonatal morbidities associated with prematurity and free radical disease, such as neurodevelopmental impairment, retinopathy of prematurity, bronchopulmonary dysplasia, and necrotizing enterocolitis. The use of oxygen concentrations between 21% and 100% requires compressed air and a blender. Different organizations throughout the world have provided differing recommendations on the use of oxygen for newborn resuscitation. The most recent review of evidence for the updated NRP guidelines suggest that resuscitation could begin with either air or blended oxygen; if blended oxygen is unavailable, air is preferred to pure 337 oxygen. The suggested goal for oxygen saturation values is the IQR at each minute (Kattwinkel et al, 2010). Our approach has been to begin with 40% oxygen and adjust slowly, attempting to mimic the gradual transition in Spo2 values that occur in healthy newborns transitioning from fetal oxygenation levels with an expected increase in Spo2 from the fetal level of approximately 50% to a target of 85% to 90% by 7 to 8 minutes, an increase of approximately 5% per minute. Figure 28-7 demonstrates continuous physiologic data during the first 10 minutes of life for an infant resuscitated using a targeted oxygen strategy. ASSISTING CIRCULATION In newborn infants, the need for resuscitative measures beyond assisted ventilation is extremely rare. Additional circulatory assistance can include chest compressions, administration of epinephrine, and volume infusion. In a large urban delivery center with a resuscitation registry, 0.12% of all infants delivered received chest compressions, epinephrine, or both during 1991 to 1993 (Perlman and Risser, 1995), and 0.06% of all infants delivered received epinephrine during 1999 to 2004 (Barber and Wyckoff, 2006). Ventilation remains the most critical priority in neonatal resuscitation; however, if adequate ventilation is provided for 30 seconds, and bradycardia with a heart rate less than 60 beats/min persists, chest compressions are initiated. Further attention to ventilation with the use of increased pressures or intubation may be required. Chest compressions are preferably provided with two thumbs on the sternum while both hands encircle the chest with (Menegazzi et al, 1993). The chest is then compressed in a 3:1 ratio coordinated with ventilation breaths. Further circulatory support may be necessary if adequate chest compressions do not result in an increase in heart rate after 30 seconds. Epinephrine is then indicated as a vasoactive substance, which increases blood pressure by α-receptor agonism, improves coronary perfusion pressure, and increases heart rate by β-receptor agonism. Intravenous administration of epinephrine is more likely to be effective than endotracheal administration. The intravenous dose of 0.01 to 0.03 mg/kg (0.1 to 0.3 mL/kg of a 1:10,000 solution) is currently recommended. Early placement of an umbilical venous catheter is critical to delivering intravenous epinephrine quickly enough to be effective. In order to place an umbilical catheter as quickly as possible, it is necessary to have the equipment readily available and to begin the procedure as soon as possible. This could be accomplished by the lead resuscitator assigning the task of placing the catheter as soon as chest compressions are initiated. If there is any prenatal indication that substantial resuscitation will be required, the necessary equipment for umbilical venous catheter placement should be prepared before delivery as completely as possible. Epinephrine can be given by endotracheal tube, but the efficacy of this delivery method is not as certain; therefore an increased dose (0.05 to 0.1 mg/kg) is recommended. Epinephrine doses can be repeated every 3 minutes if heart rate does not increase. Excessive epinephrine can result in hypertension, which in preterm infants may be a factor in the development of intraventricular hemorrhage. However, the risks PART VII Care of the Healthy Newborn 200.0 A Pulse rate 150.0 100.0 bpm 338 50.0 0.0 102.0 B 65.0 Sat Oxygen saturation 83.5 46.5 50.0 0.0 0.0 Percent FiO2 28.0 100.0 2.5 5.0 7.5 Minutes FIGURE 28-7 Tracing of physiologic data during transition. The pulse rate, oxygen saturation, and oxygen concentration (FiO2) measured within 30 seconds of birth are displayed here. The pulse oximeter is placed on the infant’s right hand as soon as possible. Point A shows when the pulse oximeter begins working. Point B shows the oxygen saturation (SpO2) at 3 minutes (approximately 80%). This infant was initially treated with 40% oxygen, and the concentration was adjusted to achieve SpO2 of 85% to 90% by 5 minutes of life. When the oxygen is weaned down to 21% at approximately 7.5 minutes of life, the SpO2 decreases and oxygen concentration is increased again. are balanced by the benefit of successful resuscitation in an infant who might not otherwise survive. If the infant has not responded to all the prior measures, a trial of increasing intravascular volume should be considered, which involves administering crystalloid or blood. Situations associated with fetal blood loss are also frequently associated with the need for resuscitation. These situation include placental abruption, cord prolapse, and fetal maternal transfusion. Some of these clinical circumstances will have an obvious history associated with blood loss, whereas others might not be readily evident at the time of birth. Signs of hypovolemia in the newborn infant are nonspecific, but include pallor and weak pulses. Volume replacement requires intravenous access, for which emergent placement of an umbilical venous catheter is essential. Any infant who has signs of hypovolemia and has not responded quickly to other resuscitative measures should have an umbilical venous line placed and a volume infusion administered. The most common, and currently recommended, fluid for volume replacement is isotonic saline. A trial volume of 10 mL/kg is given initially and repeated if necessary. If a substantial blood loss has occurred, the infant may require infusion of red blood cells to provide adequate oxygen-carrying capacity. This infusion can be accomplished emergently with noncrossmatched, O-­negative blood, with blood collected from the placenta or with blood drawn from the mother, who will have a compatible antibody profile with her infant at the time of birth. Because not all blood loss is obvious and resuscitation algorithms usually discuss volume replacement as a last resort of a difficult resuscitation, the clinician needs to keep a high index of suspicion for significant hypovolemia so that action can be taken to correct the problem as promptly as possible. Therefore in situations where the possibility for hypovolemia is known before birth, it would be wise to prepare an umbilical catheter and an initial syringe of isotonic saline and discuss with the blood bank and the delivering physicians the possibility that non-crossmatched blood may be required. SPECIFIC PROBLEMS ENCOUNTERED DURING RESUSCITATION NEONATAL RESPONSE TO MATERNAL ANESTHESIA AND ANALGESIA Medications administered to the mother during labor can affect the fetus by transfer across the placenta and direct action on the fetus or by adversely affecting the mother’s condition, and therefore altering uteroplacental circulation and fetal oxygen delivery. The most commonly discussed complication of intrapartum medication exposure is perinatal respiratory depression after maternal opiate administration. The fetus can develop respiratory depression from the direct effect of the drug. Naloxone has been used during neonatal resuscitation as an opiate receptor antagonist to reverse the effects of fetal opiate exposure. Naloxone chloride (0.1 mg/kg per dose, intravenously, intramuscularly, or endotracheally) can be considered if the newly born infant does not develop spontaneous respirations after adequate resuscitation and the mother has received an opiate analgesic during labor, provided CHAPTER 28 Resuscitation in the Delivery Room the opiate exposure is not chronic. Infants of mothers who have a chronic opiate exposure can potentially have a sudden withdrawal syndrome, including seizures if they receive a narcotic antagonist. It is also critical that assisted ventilation be provided as long as spontaneous respirations are inadequate. It should be noted that the administration of a narcotic antagonist is never an acutely required intervention during neonatal resuscitation, because these infants can be treated with assisted ventilation. CONDITIONS COMPLICATING RESUSCITATION When resuscitation has proceeded through the described steps without improvement in the infant’s clinical condition, other problems should be considered. Some of these problems may be modifiable with interventions that could improve the course of the resuscitation. For example, an unrecognized pneumothorax could prevent adequate pulmonary inflation and if under tension could impair cardiac function. If the pneumothorax is recognized and drained, gas exchange and circulation can be improved. Some congenital anomalies that were not diagnosed antenatally make resuscitation more difficult. Congenital diaphragmatic hernia is difficult to recognize on initial inspection of the infant, but can cause significant problems with resuscitation. The abdominal organs are displaced into one hemithorax, and the lungs are unable to develop normally; this will cause ventilation to be difficult. If the intestines are displaced into the thorax and mask ventilation is provided, the intestines will become inflated, making ventilation more difficult. If the congenital diaphragmatic hernia is known before delivery or a presumptive diagnosis is made in the delivery room, the baby should be intubated early to prevent intestinal inflation. An orogastric suction tube should also be placed to decompress the inflated intestines. Many other congenital anomalies that can lead to a difficult resuscitation will be more visibly obvious when the baby is born. For example, hydrops fetalis occurring for any reason can be associated with difficult resuscitation. Although most cases are diagnosed on fetal ultrasound examination before delivery, severe hydrops would be visible on examination with skin edema and abdominal distention. Frequently peritoneal and pleural fluid will need to be drained to achieve adequate ventilation. Abdominal wall defects such as gastroschisis require special attention in the delivery room to ensure that the exposed bowel is covered in plastic, to prevent excessive fluid loss, and is protected from twisting or trauma. An orogastric suction tube is also important to decompress the stomach and limit the chances of further vomiting. If assisted ventilation is necessary, early intubation as opposed to mask ventilation should be considered to prevent gastric distention. Infants with neural tube defects should also have any exposed tissues protected with a covering. In addition, all efforts should be made to avoid pressure on the defect. Fetal congenital high airway obstruction can cause particularly difficult resuscitation. If a significant airway obstruction is diagnosed antenatally, an ex utero intrapartum treatment (i.e., EXIT) procedure can be planned. This procedure allows for establishing a stable airway before 339 clamping the umbilical cord, which maintains placental function until the airway is secure. An airway obstruction may not necessarily be diagnosed before delivery and can cause difficult resuscitation. The therapy will vary depending on the cause of obstruction. An alternate airway (oral or nasopharyngeal) can be helpful if endotracheal intubation is not possible, as can occur with micrognathia. Tracheal suctioning can be attempted if a tracheal plug is suspected. In extreme situations of airway obstruction, an emergency cricothyroidotomy may be attempted. Birth trauma of any kind has been documented to occur in approximately 2.6% of all deliveries in the United States (Moczygemba et al, 2010). The most serious injuries are the variety of head injuries that can occur, including subgaleal hemorrhage and intracranial hemorrhage. Subgaleal hemorrhage is more often associated with vacuum-assisted delivery; it and is important to recognize because of the rapid blood loss that can occur into this soft tissue space. Intracranial hemorrhages such as subdural hematomas can occur, although many such injuries are mild and can be found incidentally after uncomplicated vaginal delivery. Spinal cord injuries after birth are extremely rare, but can be severe with long-lasting functional limitations. Brachial plexus and other peripheral nerve injuries may be noticed in the delivery room and evaluated shortly after birth, but they should not interfere with the resuscitation efforts. Pressure injuries from forceps-assisted deliveries can be noted on examination and evaluated as necessary in accordance with the location of the injury. Fractures and lacerations can occur and need to be evaluated after the newborn infant has transitioned adequately. LIMITS OF VIABILITY Neonatal intensive care has increased survival at lower gestational ages, resulting in changes in the definition of viability over time. There is a variety of opinion among practitioners regarding the lower limit of gestational age when intensive care should be offered. The NRP states that not providing intensive care is appropriate at less than 23 weeks’ gestation or 400 g birthweight. When the outcome is uncertain, parents of the infant are frequently included in the decision-making process. Making an informed decision about providing intensive care should be done with the best information available. One of the most useful resources to determine the likelihood of survival without severe disabilities has been provided by Tyson et al (2008) using the National Institute of Child Health and Human Development Neonatal Research Network data. Using information regarding the gestational age, birthweight, sex, number of fetuses, and antenatal steroid exposure, a calculation of the risk of mortality and neurodevelopmental impairment can be made. These predictive data that were compiled from the participating centers of the Neonatal Research Network, a group of mostly academic centers, may currently be the best national data available for predicting outcome, but they do not necessarily represent the outcome at any single unit or of any individual baby. When there is the possibility of delivery at a very preterm gestational age at a small, inexperienced center, all efforts should be made to transport the mother to an experienced inborn center. 340 PART VII Care of the Healthy Newborn CARE AFTER RESUSCITATION Infants who survive a significant resuscitation require special attention in the hours to days that follow. Frequent complications immediately following resuscitation include hypoglycemia, hypotension, and persistent metabolic acidosis. In addition, infants with evidence of hypoxic ischemic encephalopathy may benefit from mild therapeutic hypothermia (Azzopardi et al, 2009; Jacobs et al, 2007; Shankaran et al, 2005). Mild therapeutic hypothermia, in which the core body temperature is kept at 33.5° C, has been extensively evaluated and has been effective at reducing death or impairment in infants with moderate to severe hypoxic ischemic encephalopathy. The decreased temperature is thought to decrease the secondary injury that occurs after an hypoxic-ischemic insult. Hypothermia can be accomplished with both whole-body and head cooling, although clinical trials of whole-body cooling more effectively achieved a reduction in adverse outcomes. The therapy is most beneficial when initiated as quickly as possible after an insult, with beneficial effects noted when treatment was initiated within 6 hours of birth. The timing of insult in relation to the time of birth is not always obvious, making it difficult to know the actual timing of initiating therapy after the insult. Mild hypothermia therapy should be considered when an infant has required a significant resuscitation after birth, Apgar scores are low (especially a 5-minute score less than 5), fetal or neonatal acidosis are documented on cord or newborn blood gases, and signs of encephalopathy are apparent. In addition, the history of a significant event likely to cause a hypoxic-ischemic insult should trigger a thorough evaluation of the newborn to determine whether hypothermia therapy is indicated. Hypothermia therapy is not currently available at all institutions; however, all delivery services must be able to recognize the indications for therapy, so that a transfer can be initiated as quickly as possible, if necessary. In infants born without a heart rate or any respiratory effort, if resuscitation is performed to the full extent without any response, discontinuation is recommended after 10 minutes. From a review of 13 years of data from a database including 81,603 deliveries, Haddad et al (2000) found that survival with an Apgar score of 0 at 1 minute occurred in 1.26 of 1000 delivered infants without major malformations. Of 33 infants assigned Apgar scores of 0 at both 1 and 5 minutes, 67% died before hospital discharge (Haddad et al, 2000). A recent review of the available literature for infants with Apgar scores of 0 at 10 minutes found that 94% of infants either died or were severely handicapped, whereas 3% were mild or moderately handicapped (Harrington et al, 2007). The transition from fetal to neonatal life is a critical time in an individual’s life and is an opportunity for care providers to have significant effect on the outcome of infants who need assistance. The need for neonatal resuscitation, even when no signs of encephalopathy are recognized, increases the risk that children will have lower scores on intelligence quotient tests at school age (Odd et al, 2009). This risk arises most likely because resuscitation is a marker for a prior insult. However, a well-performed resuscitation could be critical for a successful recovery. Neonatal care providers have an obligation to ensure that this process is performed as well as possible and that the techniques of resuscitation are evaluated in an objective manner to promote continued improvement. SUGGESTED READINGS Apgar V: A proposal for a new method of evaluation of the newborn infant, Curr Res Anesth Analg 32:260-267, 1953. Carbine DN, Finer NN, Knodel E, et al: Video recording as a means of evaluating neonatal resuscitation performance, Pediatrics 106:654-658, 2000. Costeloe K, Hennessy E, Gibson AT, et al: The EPICure study: outcomes to discharge from hospital for infants born at the threshold of viability, Pediatrics 106:659-671, 2000. Kamlin CO, O’Donnell CP, Davis PG, et al: Oxygen saturation in healthy infants immediately after birth, J Pediatr 148:585-589, 2006. Morley CJ, Davis PG, Doyle LW, et al: Nasal CPAP or intubation at birth for very preterm infants, N Engl J Med 358:700-708, 2008. Saugstad OD, Ramji S, Soll RF, et al: Resuscitation of newborn infants with 21% or 100% oxygen: an updated systematic review and meta-analysis, Neonatology 94:176-182, 2008. Shankaran S, Laptook AR, Ehrenkranz RA, et al: Whole-body hypothermia for neonates with hypoxic-ischemic encephalopathy, N Engl J Med 353:1574-1584, 2005. Siew ML, te Pas AB, Wallace MJ, et al: Positive end-expiratory pressure enhances development of a functional residual capacity in preterm rabbits ventilated from birth, J Appl Physiol 106:1487-1493, 2009. te Pas AB: Walther FJ. A randomized, controlled trial of delivery-room respiratory management in very preterm infants, Pediatrics 120:322-329, 2007. Teitel DF: Circulatory adjustments to postnatal life, Semin Perinatol 12:96-103, 1988. Tyson JE, Parikh NA, Langer J, et al: Intensive care for extreme prematurity: moving beyond gestational age, N Engl J Med 358:1672-1681, 2008. Vohra S, Roberts RS, Zhang B, et al: Heat Loss Prevention (HeLP) in the delivery room: a randomized controlled trial of polyethylene occlusive skin wrapping in very preterm infants, J Pediatr 145:750-753, 2004. Wiswell TE, Gannon CM, Jacob J, et al: Delivery room management of the apparently vigorous meconium-stained neonate: results of the multicenter, international collaborative trial, Pediatrics 105(1 Pt 1):1-7, 2000. Complete references and supplemental color images used in this text can be found online at www.expertconsult.com P A R T V  III Care of the High-Risk Infant C H A P T E R 29 Stabilization and Transport of the High-Risk Infant George A. Woodward, Roxanne Kirsch, Michael Stone Trautman, Monica E. Kleinman, Gil Wernovsky, and Bradley S. Marino NEONATAL TRANSPORT MEDICINE INTRODUCTION Regionalization of medical care has improved the ability to centralize resources and improve patient outcomes. For optimal coordinated care, however, adequate medical transport needs to be developed and continually refined to enable delivery of patients to regional centers, and for specialized care to be available for and delivered to patients in distress. For centers that serve as a basic or specialized service, there will be times when subspecialty care is required. For those who deliver subspecialty care, there will also be times for most organizations when transfer to a similar level organization may be required for reasons such as capacity or extraordinary care. For hospitals that do not have birthing centers as part of their facility, the patient population, the acuity of the arriving patients, and potentially the ultimate morbidity and mortality of those patients depend on skilled, efficient, and quality neonatal transport. Although transfers to neonatal centers in the 1960s and early 1970s often occurred in an ad hoc manner, such as in a police car or an ambulance (i.e., a converted hearse) with variably skilled providers, transfer programs today offer a more sophisticated level of care. This level of care, however, is not consistent throughout cities, nationwide, and internationally. Many transport services and centers have grown around individual center needs, without clear attention to coordination and regionalization of services. Competitive systems, often located in similar areas or vying for similar patient populations, have resulted in the duplication of services at the ground and air levels, and at times increased risk and cost to patients and providers as part of the efforts to increase patient volume and revenue. When considering transport of neonatal patients, several situations can occur. Intrafacility transport may be required for specialty services within a particular institution. Interfacility transport between lower and higher levels of service capability can also occur, as well as between relatively equivalent levels of service because of capacity or other issues. Transported patients may be of high acuity, relatively stable, or in various stages of convalescent care. Each type of transport requires anticipatory planning, appropriate staffing, adequate modalities (e.g., transport vehicles), and strong relationships between referring and receiving providers, in addition to skilled, qualified, and certified transport personnel. As noted in the next section, transfer agreements can help to minimize inefficiencies and enable rapid approval and eventually transport of patients. This chapter will review considerations and requirements for neonatal transport; discuss issues involved in transport team operation, including equipment, personnel, mode of transport, and medical-legal issues; and present general and specific topics, including quality improvement opportunities, that might be encountered in a neonatal transport system (American Academy of Pediatrics et al, 2007; Cornette, 2004; Woodward et al, 2007). REGIONALIZATION OF NEONATAL CARE, TRANSFER AGREEMENTS, AND BACK TRANSPORT The concept of regionalization of neonatal care and transport developed from the formation of neonatal stations in the 1920s and 1940s (Oppenheimer, 1996). These stations were located within certain area hospitals, where additional resources were allocated to provide care for premature infants. One consequence of the formation of these stations was the development of equipment and protocols to transport premature infants from other area hospitals to those specialized areas to receive care. In the 1960s and 1970s, as interest in neonatal care grew, so did the number of hospitals offering services for premature infants. To help optimize the care being delivered, the March of Dimes produced Toward Improving the Outcome of Pregnancy in a 1976 report (Committee on Perinatal Health, 1976). The report stratified maternal and neonatal care into levels based on their complexity, and it proposed the referral of high-risk patients to centers with sufficient personnel and resources to provide care. The goal was to create standard definitions so that comparisons of health outcomes, resource utilization, and costs among regional institutions could be made. High-risk maternity patients would be able to actively participate in selecting a delivery service, and businesses would be able to select appropriate health care resources for their employees. The subsequent March of Dimes publication in 1993 Toward Improving the Outcome of Pregnancy: The 90s and Beyond reiterated the importance of regionalized care and further delineated care levels (Committee on Perinatal Health, 1993). The concepts of regional care were adopted and incorporated into the Guidelines for Perinatal Care (American Academy 341 342 PART VIII Care of the High-Risk Infant of Pediatrics Committee on Fetus and Newborn and Bell, 2007; Woodward et al, 2007). Whereas the original driving forces for regionalization in the 1970s were the shortage of trained personnel to care for low birthweight (LBW) infants and the economic expense to maintain these skills, during the late 1980s and 1990s, technology and clinical expertise disseminated outside the regional tertiary centers, resulting in proliferation of the number of intermediate care neonatal intensive care units (NICUs). This proliferation has blurred many of the original distinctions between various care systems. Whether driven by third party payers or other factors, with various interpretations and applications of what “regional care” means, the results have been the creation of a variety of care options (Lainwala et al, 2007). Limited space and cooperative longitudinal care planning in some tertiary care units have created the necessity for patient transport back to a unit with less acuity or fewer resources once their critical condition has resolved or stabilized (Attar et al, 2005; Donovan and Schmitt, 1991; Lynch et al, 1988). Regionalization guidelines should support the return to the community for patients who no longer need the highest care level. Patient selection for back transport (transport back to referring service, hospital, or a hospital closer to the patient’s home) and care in another facility should match the capabilities and expertise of the community hospital (Stark and American Academy of Pediatrics Committee on Fetus and Newborn, 2004). Although third-party payers often drive decision making, transport relationships develop between various institutions either by formalized transfer and preferred provider agreements or by historical and personal relationships (Attar et al, 2005). Transfer agreements can help to define the roles, understanding, and expectations between institutions and the transport service; they also help frequently to detail reimbursement issues. These agreements set the expectations for participating facilities, with the ultimate goal of the timely movement of patients from one facility to another 2007; Woodward et al, 2007). Whether optimal care can be provided in some smaller NICUs or outside of a tertiary unit is frequently debated. Nonetheless, there is a trend in the United States and elsewhere toward the centralization of perinatal care (Howell et al, 2002; Wall et al, 2004). Several investigators have shown that mortality was lowest for deliveries of infants with very LBW that occurred in hospitals with tertiary care NICUs (Chien et al, 2001; Phibbs et al, 2007; Rautava et al, 2007; Robertson et al, 1994). Cifuentes et al (2003) supported the idea that whenever possible, women in early preterm labor should be moved to the regional hospital rather than transferring the infant after birth. The study was insufficiently powered, however, to make a recommendation regarding the difference between regional and large community NICUs (Cifuentes et al, 2003). In trying to discern which LBW infant might be better cared for in tertiary regional care centers, Vieux et al (2006) showed that the risk factors for requiring high levels of intensive care were low gestational age, twin pregnancy, maternal hypertension, antepartum hemorrhage, infection, and male gender, whereas antenatal steroid therapy and premature rupture of membranes were protective factors against requiring intensive care. Regional care concerns can apply to the neonates with extremely LBW and to the late preterm population (Khashu et al, 2009). Investigators trying to describe the optimal neonatal unit caring for very preterm infants in Europe were unable to produce a consensus as to what the ideal neonatal unit size and patient volume should be (Van Reempts et al, 2007). Given the wide variability in levels of neonatal care and the inability to predict premature delivery, neonatal transport will continue to be a dynamic process. TRANSPORT COMMUNICATIONS When developing and maximizing transport capabilities, a key concept is the centralization of communication. Although a telephone call from a referring provider to a receiving provider might be the most efficient way to receive advice or notify a receiving provider of a potential transport, there are more centralized and effective means of communication. First, a centralized, easily recalled, advertised, and monitored (24 hours per day, 7 days per week) communication system should be available. Identifying and publishing a centralized access number for immediate access to the transport system or receiving center personnel are imperative for optimal communication and efficiency (Southard et al, 2005). Anyone who has transported or referred patients to systems without centralized access understands the challenges in working through operators, unit clerks, multiple providers, and often multiple services to enable a singular transport. This process is time consuming and often frustrating for the referring provider, and it is often time that could be better spent in direct assessment and care of the patient, rather than on the telephone with repetitive informational transfer. Ideally a centralized transport communication center would enable a referring provider to make a single call to a single number and immediately receive all the services that they might require. Those services include appreciation and recognition of the need for transfer, identification of appropriate hospital and facility, review of medical issues, and determination of required services (i.e., personnel, equipment, time of response) needed for transport. Also included would be simultaneous awareness of need for transport by the personnel responsible for arrangement of particular modes of transport, verification of bed capacity, and any other logistic items that may need to take place prior to the transport. A telephone call made to an individual provider requesting transport requires a sequential process from data gathering through acceptance for admission and eventual transport service or modality identification and dispatch. Centralized access to a communication center allows all those functions to occur simultaneously, enabling more rapid transport response and appropriate involvement of all those required for the management of the particular patient (American Academy of Pediatrics Committee on Fetus and Newborn and Bell, 2007; Woodward et al, 2007). MEDICAL SUPERVISION A key requirement for any system is to have appropriately skilled and immediately available medical command physicians (American Academy of Pediatrics Committee on Fetus and Newborn and Bell, 2007; Woodward et al, 2007). A medical command physician should be literate CHAPTER 29 Stabilization and Transport of the High-Risk Infant and expert in the medical area of concern. In most cases involving neonatal transport, this provider should be a neonatologist. There may be instances, however, when the referring or receiving physicians may request or desire additional medical expertise. For example, a cyanotic newborn with congenital heart disease may be temporarily improved or stabilized well by the referring neonatologist and have additional stabilization direction provided by the receiving medical command physician; however, invaluable additional management and planning might be added by a partnering cardiac intensive care physician. A communication or transfer center can allow for multiple providers to be linked together during an initial referral call to allow the highest level advice to be presented and discussed among the providers. These telephone calls should also include the transport personnel so that the providers have the background information and care plans delivered directly. MODE OF TRANSPORT Once the transport referral has been made, ideally through a communication center, and discussions have been started with the medical command physician, the transport can take place. Decision on mode of transport is an important component to consider at this juncture. In general, ground transport allows door-to-door service between facilities, enables a well-lighted environment, provides space for several providers as well as the patient and a family member, and is efficient in urban and short-range transfers. As the distance for transfer increases, and the need for decreased out-of-hospital time becomes more important, consideration of air transport becomes important depending on acuity and patient disease process. In general, air transport is most expeditious over approximately 50 miles, with helicopter being favored between 50 and 150 miles and airplane being effective for transports greater than 150 miles. The decision regarding mode of transport is influenced by several issues, which include available modes of transport, staffing and medical expertise of the providers involved in those particular modes, current and projected patient condition (which includes preliminary identification of the underlying disease process), consideration of stability or illness progression during the projected transport timeframe, capabilities of the referring facility and personnel, and the urgency of intervention and definitive placement of the patient. Other issues that influence the process include geographic and weather constraints, distance and duration of transport, and provider perception about the transport process. In addition to the potential risks of ground transport, which include traffic accidents, vibration, noise, and other issues, air transport creates concerns for accident, vibration, gravitational forces during acceleration and deceleration, excessive noise, temperature variations, and decreased humidity with altitude. Air transport also includes issues related to altitude physiology that can affect patients with respiratory issues or air trapping and their air-containing equipment (e.g., endotracheal tube cuffs, laryngeal mask airways; Wilson et al, 2008; Woodward et al, 2002, 2006). Dalton’s Law recognizes that ambient oxygen decreases as altitude increases; therefore there may be a need for pressurization and augmentation with increased FiO2. Boyle’s Law states that 343 as altitude increases, the volume of a gas also increases, and the barometric pressure is inversely related to the volume of the gas. This law is potentially a serious issue for patients with an enclosed gas collection, such as a simple or developing pneumothorax. TRANSPORT PERSONNEL, EDUCATION, AND TEAM COMPOSITION Awareness of the capabilities of the transport system and of the personnel involved in transport is imperative in decisions regarding mode of transport. Although it is ultimately the responsibility of the referring physician to identify the appropriate mode and personnel for transport, per Emergency medical Treatment and Active Labor Act (EMTALA), opportunities exist for tertiary care and referral centers to help to educate the referral providers regarding optimal transport planning and usage (Bolte, 1995; Woodward, 1995). In general, issues influencing transport decisions include current level of care, urgency for a different level of medical capability or equipment, current providers of care, stability of the patient, options available to the provider and patient, and efficiency and quality of the transport process. Ideally, these issues are the determinants of appropriate transfer; however, referring providers are often overwhelmed by the severity or acuity of the patient and their primary desire may be to have the patient removed from their facility as quickly as possible. The providers may decide on a transport or transfer process based solely on speed of transport process rather than quality of care. It is imperative for the receiving and tertiary care centers to educate the referring population on the importance of stabilization, initiation and quality of initial response, transport options, and definitive care to maximize potential patient outcome. When examining the transfer of patients, providers should ask a simple question: “Are we trying to deliver the patient to tertiary care, or are we trying to deliver tertiary care to the patient?” In most high-functioning transport and referral centers, the latter is true. The referring physician should expect to have tertiary care advice and direction delivered at the moment of the referral telephone call and continued throughout the transport process (American Academy of Pediatrics Committee on Fetus and Newborn and Bell, 2007; Woodward, 1995; Woodward et al, 2007). At no time should the care delivered to the patient decrease in sophistication after a referral call is made. If a transfer places the infant in a suboptimal environment or with providers who are not competent to address the infant’s current or potential medical needs, the patient might receive suboptimal care and the referring provider is liable for arranging and providing inappropriate care. When considering the team composition of the transport providers, it is important to consider the quality of the personnel, their expertise and experience, and their ability to work in the transport environment (King and ­Woodward, 2002a; King et al, 2001, 2007). There are many variations of transport teams in the United States and abroad. These teams can be composed of physicians, nurse practitioners, nurses, respiratory therapists, paramedics, and other health care providers. In the United States, medical (pediatric or specialty) trainees are often 344 PART VIII Care of the High-Risk Infant used as primary providers in the transport environment. Regardless of the formal educational background of an individual, there are several criteria that must be met to be optimally effective in the transport environment. First, the provider must have adequate certification, be licensed for the care they deliver, and be able to provide the assessments and interventions that the patient currently or potentially requires during the transport process. For example, a neonatal retrieval service must be able to manage acute and critical airways in the neonatal population, both at a referring hospital and during the transport. It is interesting that those same providers might not be credentialed to provide those skills within their home hospital; however, they have been certified to provide them in the ambulance environment. In general, this must be done under the auspices of a physician’s care, which may be from an accompanying physician or via online medical control (real-time medical advice during the transport process). Transport care also often involves offline medical control, which involves the use of protocols developed by medical personnel. It is important to recognize that the transport timeframe is somewhat limited; therefore the personnel might not need to have the longitudinal or differential diagnosis expertise of a fully trained neonatologist. However, these personnel must have the acute care assessment abilities and intervention skills of an experienced neonatal expert. In addition to training the onsite personnel, it is imperative to have medical command physicians understand the opportunities and limitations of the transport services and environment and to understand the risks and challenges that referring personnel can potentially encounter with situations and patients who exceed their own personal or their facility’s management abilities. The physicians must have significant awareness of the transport environment to understand the limitations of potential interventions. It is also imperative that medical command positions have superb communication skills, not only at the referring physician level but also within the transport team. Medical command physicians will need to be able to effectively and efficiently communicate with providers from different disciplines. QUALITY IMPROVEMENT Transport medicine offers an opportunity to identify areas for potential improvement within the inpatient arena, in the transport system, and at the referring facilities (Browning Carmo et al, 2008; Chen et al, 2005; Lim and Ratnavel, 2008; McPherson et al, 2008; Ramnarayan, 2009). This glimpse into medical sophistication and capability is one that is privileged and should be used to identify educational opportunities rather than prompt judgmental and critical review. Education by referring physicians and transport teams can have a significant effect on the quality and outcome of patient care and the volume of future referrals. Ideally, once a referral call is made, a receiving physician or medical command physician will direct the care so that the job of the transport team is to verify that an appropriate working diagnosis has been made and that adequate stabilization procedures have been performed. The team can verify that the interventions are secure and then transport the patient in a stable fashion. Systems that do not gather adequate information or offer appropriate advice, or in which the referring facilities do not follow that advice or choose not to perform needed interventions, put the patient at risk by having a delay in care of initial interventions, prolonging the transport, and delaying delivery of definitive care. The transport team that has invested several hours at a bedside, stabilizing a newborn with medical or surgical issues, may be spending time in a facility that is not ideal, has a limited number of skilled personnel, and has minimal backup, thus prolonging the transport process and potentially putting that individual patient at risk (Chen et al, 2005; Haji-Michael, 2005). However, the patient and the system are at risk because the valuable resource of specialized neonatal transport personnel is not available for another patient. Ideally, care delivery would be the same at referral and receiving centers; the development of practice guidelines can be helpful in that regard. Guidelines that are evidence based, developed by regional and local experts, and disseminated to referring locations and transport teams will help to standardize and enable consistent care across variable locations. It is necessary, however, to assess and reassess the quality of the guidelines and the competency of their use to ensure optimal results from this process. Even in the best of hands, near-miss or realized adverse events may happen. It is clear that identification of those events, discussion with families (where appropriate), and root cause analysis are imperative. Several studies have examined adverse events in transported patients. Ligtenberg et al (2005) noted that one third of patients had an adverse event, and 50% of those resulted from not following the advice of the medical command physician. Of that group, 70% of events were avoidable and 30% were logistical. In a review of the London Neonatal Transfer ­Service, Lim and Ratnavel (2008) noted that 36% of their patients had greater than or equal to one adverse event, and two thirds of those were due to human error; half of those occurred before the team arrived at the referral center, and their major etiologies included preparation and communication. A neonatal transport service must determine if maternal transport is part of their purview. It is clear that preterm infants who are born at outside hospitals and require transfer to tertiary care centers or who are born at one tertiary care center and require transfer to another have worse outcomes that include increased mortality and morbidities such as intraventricular hemorrhage (IVH) and other medical issues (Baskett and O’Connell, 2009; Janse-Marec and Mairovitz, 2004; Jony and Baskett, 2007; O’Brien et al, 2004; Ohara et al, 2008). Developing appropriate criteria for transporting women in preterm labor may help to direct the optimal use of resources required with maternal transfer. Identifying the appropriate time to transport women in preterm labor, as well as those with other preterm medical issues, will help in developing appropriate criteria and limit resource utilization for maternal transfers. It is important to note that significant work regarding neonatal and maternal transport and regionalization of care in the United States, United Kingdom, Australia, Europe, and other areas has already been accomplished. There are, however, multiple opportunities to improve transport on an individual and regional basis in these developed countries and in other developing countries. CHAPTER 29 Stabilization and Transport of the High-Risk Infant TRANSPORT ADMINISTRATION As a hospital develops and optimizes a neonatal transport program, experts in transport medicine are integral to the success of this program (American Academy of Pediatrics Committee on Fetus and Newborn and Bell, 2007; ­Woodward et al, 2007). A quality medical director and program director, often a nurse or respiratory therapist, are essential to understand the potentially complicated and challenging environment of transport medicine. These leaders should be instrumental in identifying expectations, roles, and responsibilities for the process; this includes access through a communication center and developing and disseminating referral center expectations. These expectations include stabilization and preparation of the patient before transport, making an appropriate decision to transfer, choosing the appropriate transport process and destination, obtaining consent from the family for transport, discussing plans for stabilization and intervention with the medical command physician, and initiating that plan as able. Referral centers must be able to be direct when they believe that the suggested interventions are inappropriate or beyond their scope. The referring team also needs to be available to participate in transition care to the transport team and the receiving service. The receiving and transport team responsibilities include being immediately available for case discussion, having the ability to rapidly accept the patient if capacity is available, offering clear and concise expert recommendations, ensuring preparation of the environment and the staff both for the transport and arrival of the patient, organizing additional diagnostics and interventions, and providing available and accessible neonatal advice throughout the process. Most important, the team needs to ensure that appropriate skills and therapy are available and delivered throughout the process, from referral call through definitive placement, and ensure seamless transition at each point of care. The team needs to communicate well with referring and the patient’s physicians and document their advice, interventions and activities in a clear, concise fashion to enable appropriate patient and provide protection for the transport service. TRANSPORT EXPERTISE As transport medicine has developed over the past 30 to 40 years, there has been an increase in the quality and volume of transport research, as well as organizations that are discipline and process specific. Multiple disciplines have organizations with transport focus and expertise, including the American Academy of Pediatrics Section on Transport Medicine, which includes executive committee representation from neonatology, pediatric critical care, and emergency medicine physicians, as well as a broad membership from other subspecialties and disciplines. The American Academy of Pediatrics Transport Section published its most recent Guidelines for Neonatal and Pediatric Transport in 2007; they also host a listserv at [email protected], which is a national repository of information and a conduit to others in the transport arena for transport-related issues. There are many keys to success in transport medicine. These keys include integrity, professionalism, preparation, 345 anticipation, education, competency, critical evaluation of current and new practices, the ability to critically assess the quality of care being delivered at every point along the way, and looking for opportunities to improve access, efficiency, and delivery of care. An area that is vital for appropriate transport care is the assurance of competency of the providers (American Academy of Pediatric Committee on Fetus and Newborn and Bell, 2007). As noted, providers can come from many different backgrounds and bring different skills from their clinical and training experience. It is imperative for the transport system and leadership to ensure that the skills required are present in their personnel. This responsibility includes initial education, continuing education, and competency assessment. A team may decide to have individuals who are experts in defined areas of care, such as a physician who manages airways and the pneumothorax, a nurse who manages intravenous access and medication delivery, and a respiratory therapist who is responsible for airway and ventilatory support, whereas another team may choose to have all team members competent in all skills. Specific procedures may be useful or needed in transport that are not as common in the hospital environment or to the hospital-based providers (e.g., laryngeal mask airway; Trevisanuto et al, 2005). However the team is structured, there must be clear and concise guidelines for ongoing training and assessment of the personnel. High-fidelity simulation models offer additional opportunities to assess and potentially improve technical and cognitive capabilities (LeFlore and Anderson, 2008). The American Academy of Pediatrics Guidelines for Air and Ground Transport of Neonatal and Pediatric Patients (available at www.aap.org) is an invaluable resource that outlines required education and competencies for each level and discipline. The true quality of a transport service is sometimes difficult to determine, because many teams and services have been developed independently and are not part of larger regional systems. It is difficult to compare and contrast systems in different areas for which different patient populations are transported (Berge et al, 2005; Cornette and Miall, 2006; Craig, 2005; Doyle and Orr, 2002; Khilnani and Chhabra, 2008; Van Reempts et al, 2007). One avenue for potential standardization of the transport environment is by assessment and accreditation by the Commission on Accreditation of Medical Transport Systems, which was initiated in 1990 as a direct response to the number of air medical accidents in the 1980s. This certification is voluntary in some areas and required in others, and it serves to assure providers, stakeholders, and the public of adherence to quality of care and transport safety standards. The Commission on Accreditation of Medical Transport Systems is an independent organization, supported by 16 member organizations, which publishes standards and arranges reviews of interested air and ground transport programs. Further information can be found at http:www.camts.org. TRANSPORT SAFETY Safety of the transport system and for its providers is paramount and must be assessed and ensured before transporting any patient. Vehicles must be safe and meet the standards for air or ground transport, the personnel must be appropriate, licensed, and competent, and the patients 346 PART VIII Care of the High-Risk Infant must be managed in the most appropriate and professional fashion. In addition, the logistics of travel must include a safe environment, including helmets and fire-retardant suits for those who fly in helicopters, three-point restraints, and appropriate ambulance seating arrangements. There should not be an occasion when providers put themselves at risk by being unrestrained or being in an area where unsecured debris or inappropriately placed equipment may damage a provider or a patient. Adherence to rules and regulations of air and ground transport is imperative as well (Clawson, 2002; Greene, 2009; King and Woodward, 2002b; Levick et al, 2006; National Highway Traffic Safety Administration, 2009). It is imperative to recognize that there is risk with both air and ground transport. The air transport industry has seen an acute spike in tragic and fatal air accidents (Greene, 2009; National Transportation Safety Board Accident Database, 2009). This increase has caused the industry, and the U.S. government, to critically investigate these issues and offer suggestions to improve transport safety (Federal Aviation Administration Helicopter Safety Initiative, 2011; 10.5-year U.S. HEMS Safety, 2008). Requirements such as duty hours for pilots, weather restrictions, flight under instrument, flight rules with terrain avoidance equipment, and night vision goggles can to help minimize the risk for these transports. Although ground ambulances are used much more frequently, and the risk of injury and death is evident, the fatality rate is lower in ambulance accidents than it is in aircraft accidents (Becker, 2003; Becker et al, 2003; King and Woodward, 2002b). It is necessary, however, that the vehicles be maintained and operated in a safe manner. Many systems do not allow ambulances to exceed posted speed limits and use lights and sirens only as a way to identify an emergency response, not to enable the vehicle to circumvent or ignore standard traffic laws (Clawson, 2002). Appropriate equipment for ambulances is required as well, and the July 2009 statement by the American College of Surgeons Committee on Trauma regarding appropriate equipment for ambulances (also published as an American Academy of Pediatrics’ policy statement) should be reviewed by all transport systems (American College of Surgeons Committee on Trauma, 2009; American College of Surgeons Committee on Trauma et al, 2009). It is evident from the transport and pediatric literature that patient outcomes are improved with specialty providers. There have been multiple studies to examine this particular issue (Belway et al, 2006; Mullane et al, 2004). Perhaps the most compelling is the study by Orr et al (2009), which examined transports provided by variable providers within the same system. This study compared outcomes in patients whose care was delivered by specialized pediatric critical care teams with those whose care was delivered by general providers. Both teams had the same medical command oversight, equipment, and modalities. Patient outcomes were worse for those whose care was not delivered by specialty teams, and much improved for those who were. One challenge, however, with transport teams is that differentiation of medical resources, such as a neonatal specialty team, likely means that there may be a scarcity and potential need for rationing of those resources. It is possible to develop teams with a variety of personnel with complementary cognitive and procedural skill sets and work toward appropriate triage of transport requests to ensure the proper level of onsite skill provision. There have been multiple attempts to develop triage tools for pediatric and neonatal care providers, including the Mortality Index for Neonatal Transport, the Modified Clinical Risk Index for Babies, and the Risk Score for Transported Patients, which are noted in the bibliography (Broughton et al, 2004a, 2004b; Markakis et al, 2006). FAMILY-CENTERED CARE Transport team research has shown that family-oriented care, as in other areas of the medical systems, is an important component of transport (Woodward and Fleegler, 2000, 2001). Families who have been formally surveyed appreciate the opportunity to participate in the care of their child. In neonatal transport, however, there are times when there are two patients who may require care in two disparate locations. A mother who has had a cesarean section and has delivered an acutely ill child in need of care in a higher facility is one such example. It is required that the mother be in one facility and the child be in another. The father may be conflicted regarding accompanying his new child on the transport or staying with the child’s mother. Transport teams need to be sensitive to the challenges and opportunities for the family and include them in the process when possible. It is evident that when parents attend or accompany transport team members on critical care transports, they are not there to assess the medical skill set of the provider, but to provide support to their child. It is also a great opportunity for the transport team to demonstrate to the family that their patient is in focused, professional, caring, and capable hands. MEDICAL LEGAL ISSUES There are many medical legal issues in transport medicine, as elsewhere in the medical system (Hedges et al, 2006; Williams, 2001; Woodward, 2003). The Health Insurance Portability and Accountability Act (HIPAA) is a required component of transport planning and delivery. Discussion of patients should not take place in a public area or via public communication airways where non–patient-related personnel or bystanders could overhear information about a specific patient. As noted earlier, a requirement of EMTALA is that the referring physician chooses the appropriate mode of transport and ensures that the transport process and receiving hospital are appropriate for the particular patient. Patients cannot be transferred if they are unstable and the ability to further stabilize them is available at the initial site of care. If a patient must be transferred for care while in an unstable condition—a frequent scenario for critically ill patients who need care not available at the referring ­institution—consent must be obtained from the family, which acknowledges their understanding of the potential risks and benefits of the process. In practice, there are often patients in unstable condition who are transferred from lower to higher levels of care, because the level of care that can be provided at the referring or initial center is not optimal for the child. This reason is appropriate for transfer as compared with transferring patients because of financial or other economic drivers. The medical liability for transport CHAPTER 29 Stabilization and Transport of the High-Risk Infant is a shared process. Before the referring center calls regarding the patient and the center of referral has accepted care, the entire medical responsibility lies with the referring provider. Once the receiving team has accepted the patient and offered advice, medical liability is a shared process. The referring physician maintains the majority of the liability, as well as medical control of the patient, throughout the process until the transport team has left the referring hospital. It is important to recognize that most transport teams and personnel do not have privileges at referring hospitals and are working under the guidance and supervision of the referring physician team. Transport teams that act independently, or referring physicians who are not available when the transport team arrives, put both the referring provider and the transport team at risk if there is disagreement or inappropriate care delivered to the patient. There will be times, however, when there is disagreement regarding the optimal care to be delivered. For example, a child with a hypoplastic left heart syndrome and an open ductus arteriosus may be given a low-dose prostaglandin infusion and be in stable condition. A transport team can insist on intubation before a long air transport, whereas the referral physician might believe that intubation is not required and that it poses a risk to that particular patient. This situation can be challenging, and it must be handled appropriately. It is never appropriate to have obvious provider conflict occur at a patient’s bedside, especially in front of family members. The appropriate way to handle a situation that cannot be easily mitigated is to involve the medical command physician with a telephone call to the referring physician in a discussion at a peer-topeer level. Transport teams have been known to comply with the wishes of the referring providers to not perform advanced procedures at the referring hospital, only to perform those procedures in the ambulance, which is a much less desirable location. Ideally, all disagreements and considerations of different therapies are discussed in a collegial fashion in the appropriate environment to enable the safe care and transport of the child to the receiving center. As noted previously, documentation of all information received and advice offered is imperative. If there is future review or challenge to the care delivered during transport, as with any other care delivered in the hospital, clear and appropriate documentation should stand alone as an excellent defense. In addition, many centers use recorded (i.e., digital, tape, other retrievable recording process) for their intake and advice lines; this is another way to review, educate, and ensure that appropriate information is delivered in an effective communication style. The use of recorded lines with frequent review, for educational and quality assurance purposes, can be invaluable. Review with legal advisors can help to define the length of time the recorded materials should be maintained for quality improvement or patient record addendum. PATIENT CARE DURING TRANSPORT EXTREME PREMATURITY The most effective method of transporting premature infants is within the mother as a maternal transport. By delivering at a referral hospital, the premature infant is exposed to all the variability of the extrauterine environment, with the added complexity of having to be 347 transported to a facility capable of providing the level of care the infant needs. Most neonatal transport teams are regarded as extensions of the NICU. The team initiates and provides much of the same level of complex neonatal care as the receiving hospital, but in a changing environment. It is this changing environment that poses unique challenges for both patient and caregiver. These issues would be amplified in the case of a disaster with care and transfer required for premature infants (Gershanik, 2006). Limit of Viability The incidence of premature births has been increasing in the United States. Of the approximate 4.1 million live births occurring annually, 520,000 (12.7%) infants are born premature (Ananth et al, 2009; Martin et al, 2008). Although some premature infants require minimal care to survive, infants with very LBW often test the resources and skills of the most experienced care provider. The longterm outcome for many of these small infants is often unknown for months to years after leaving the hospital. In some instances, the long-term results have not been ideal (­Donohue et al, 2009; Tyson and Saigal, 2005). Several studies of extremely premature infants born outside of tertiary care centers show increases in morbidity and mortality (Cifuentes et al, 2002; EXPRESS Group et al, 2009; Phibbs et al, 2007; Rautava et al, 2007). Human viability is currently limited by the physiology of pulmonary development and its ability to exchange gases. Currently this limit appears to be at approximately 22 to 24 weeks’ gestation (Pignotti and Donzelli, 2008). Unfortunately, for transport teams and the care givers at referring hospitals, it is difficult to determine which infants born at the margins of viability should be resuscitated and provided with aggressive neonatal care and which should be allowed to die (American Academy of Pediatrics Committee on Fetus and Newborn and Bell, 2007; Buchanan, 2009). These decisions are best made collaboratively with the family, transport team members, and the referring and receiving physicians (Ahluwalia et al, 2008; Gunderman and Engle, 2005; Tyson et al, 1996) and may ultimately result in a patient transport, even when the likelihood for survival is minimal. Thermoregulation Problems in neonatal thermoregulation continue to be a major contributor to neonatal morbidity and mortality worldwide and can be especially problematic in neonatal transport (World Health Organization, 1996). During transport, neonates often cross into and out of multiple different environments with wide temperature and humidity variations. Although a normal term infant is capable of a significant homoeothermic response by using their sympathetic nervous system to vasoconstrict peripherally, thus placing their normal layer of insulating white fat between the body’s core and the exposed skin, preterm infants lack the subcutaneous white fat insulation to protect their core temperature. Moreover, term infants use brown fat as a source for nonshivering thermogenesis. Preterm infants, although as sensitive to temperature as term infants, lack sufficient brown fat to sustain a response when exposed to a cold environment (Baumgart, 2008). The consequences 348 PART VIII Care of the High-Risk Infant of hypothermia can include hypoglycemia, metabolic acidosis, intraventricular hemorrhage, persistent pulmonary hypertension, and hypoxemia (Bartels et al, 2005). Silverman et al (1958) demonstrated that using a higher incubator temperature, even without additional humidity resulted in improved premature survival rates. Additional measures such as chemical gel packs and polyethylene occlusive skin wrapping help in maintaining the temperature of an infant with very LBW (Vohra et al, 2004). For all newborns, an equally important condition to avoid is hyperthermia. Although elevated temperatures in neonates occur with increased metabolic rates, prolonged seizures, dehydration, or infection, the most common cause of neonatal hyperthermia is high ambient air temperature and humidity (Baumgart, 2008). In a review of a subset of patients referred for a trial of hypothermia for hypoxic ischemic encephalopathy, patients with an elevated body temperature had poorer neurologic outcomes than those with normal body temperatures (Laptook et al, 2008; Yager et al, 2004). Humidity also has an extremely important role in temperature control of LBW infants, especially those receiving mechanical ventilator support. The importance of delivering humidified gas to neonates receiving mechanical ventilation is widely acknowledged (Sousulski et al, 1983). Ventilation with dry gases affects the airway epithelium in very LBW infants, and it can result in hypothermia secondary to their large surface area–to–body mass ratio and their relatively large respiratory minute volume (Fassassi et al, 2007). There is a linear correlation between incubator temperature and the humidity generated by heated humidity systems (Fassassi et al, 2007). Whereas ventilator complications can be reduced and thermoregulation can be improved by providing exogenous heat and humidity to the gases, active heated humidification systems are used infrequently during neonatal transport. Passive hygroscopic heat and moisture exchangers have been used for short-term conventional mechanical ventilation and with some types of high-frequency ventilation (Fassassi et al, 2007; Schiffmann, 1997; Schiffmann et al, 1999). Humidifying the incubator and the ventilator system helps in thermoregulation while transporting very LBW neonates. Surfactant Surfactant replacement has had a significant effect on newborn intensive care. Numerous systematic reviews have demonstrated the benefit of surfactant administration in reducing oxygen needs and ventilation requirements, as well as ventilator complications such as pneumothorax and pulmonary interstitial emphysema (Horbar et al, 1993; Liechty et al, 1991; Schwartz et al, 1994). Extremely LBW infants benefit from both prophylactic (defined as within 10 to 30 minutes of birth) and rescue surfactant administration (within 12 hours of birth) (Kendig et al, 1991; Soll, 2000). Infants given prophylactic surfactant appear to have fewer complications, including death, pneumothorax, and bronchopulmonary dysplasia (Soll and Morley, 2001), supporting its use during transport (Riek, 2004). Surfactant administration can be complicated by airway obstruction and right mainstem bronchus or esophageal instillation, justifying the American Academy of Pediatrics Committee on Fetus and Newborn recommendations that “preterm and term neonates who are receiving surfactant should be managed by nursery and transport personnel with the technical and clinical expertise to administer surfactant safely and deal with multisystem illness” (Engle, 2008). While investigations are still ongoing as to the optimal manner of surfactant administration, future technology might eventually allow for surfactant to be delivered without airway intubation. Currently, aerosolized surfactant has yet to be shown to be superior to endotracheal tube administration (Berggren et al, 2000; Mazela et al, 2007). Although there are concerns about the composition variability of animal-derived surfactants, use of a synthetic surfactant such as Lucinactant creates other concerns. Lucinactant requires a special warming cradle to convert it from a gel to a liquid before administration, adding a level of difficulty to its administration during neonatal transport (Kattwinkel, 2005; Moya et al, 2005). Smaller community hospitals might not stock surfactant in their pharmacies, making it necessary for the transport team to carry it as part of their medical supplies. Earlier forms of surfactant came as lyophilized powder and could be reconstituted when needed. Currently, the most frequently used surfactant preparations in the United States are derived from either bovine or porcine lung extracts. These surfactants require refrigeration and are usually carried by transport teams in small containers cooled by gel packs. The same precautions used in the delivery room or the NICU apply to its use in transport. Checklists should be used to avoid inadvertently leaving this lifesaving therapy behind when the transport team is dispatched. Before surfactant administration, proper endotracheal tube position must be confirmed either clinically or radiographically. Surfactant administration is generally done using transport team protocols, which usually mirror the package inserts. After administering surfactant, monitoring pulmonary compliance and adjusting the patient’s ventilator support will occasionally extend the time of a transport to avoid clinical complications, such as pneumothorax or endotracheal tube occlusion, during the transfer process. Often by the time the transport team arrives at the receiving hospital, significant clinical improvement has occurred. Surfactant has a role in other disease processes, such as meconium aspiration, sepsis, and pulmonary hemorrhage, during which a patient’s endogenous surfactant is depleted or inactivated, thus creating a condition of secondary surfactant deficiency (Chinese Collaborative Study Group for Neonatal Respiratory Diseases, 2005; Finer, 2004; Wiswell et al, 2002). HYPOXIC RESPIRATORY FAILURE Meconium Aspiration Syndrome, Persistent Pulmonary Hypertension of the Newborn, Inhaled Nitric Oxide on Transport, and Extracorporeal Membrane Oxygenation Referrals Hypoxic respiratory failure describes a heterogeneous group of neonatal disorders that have in common impaired oxygenation and the need for assisted ventilation. The most frequently observed conditions are respiratory CHAPTER 29 Stabilization and Transport of the High-Risk Infant distress syndrome, meconium aspiration syndrome, and persistent pulmonary hypertension of the newborn (PPHN), which can occur individually or in combination. Interfacility transport of the newborn with hypoxic respiratory failure is potentially hazardous, with a high risk for patient deterioration and complications of therapy, such as pneumothorax. Assisted ventilation during neonatal transport can be accomplished with a variety of devices, although not all modes of mechanical ventilators have been modified for use in a mobile setting. Newborns with milder degrees of illness may be managed successfully with continuous positive airway pressure (Murray and Stewart, 2008). Infant transport ventilators range from simple time-cycled, pressure-limited machines to more sophisticated devices with flow sensitivity and the ability to synchronize. Certain high-frequency ventilators have been successfully adapted for the transport environment (see later). Inhaled nitric oxide (iNO) is approved for use in term and near-term newborns with hypoxic respiratory failure with clinical or echocardiographic evidence of pulmonary hypertension. Persistent PPHN can exist in isolation or be secondary to another insult, such as sepsis or meconium aspiration syndrome. Administration of iNO to newborns with PPHN reduces the need for extracorporeal membrane oxygenation (ECMO; Clark et al, 2000; Lowe and Trautwein, 2007). With its availability to community NICUs, iNO is often initiated before transport to a tertiary care center. Once iNO is administered, abrupt cessation of therapy can result in rapid clinical deterioration from rebound pulmonary hypertension, so it is essential to continue iNO if already initiated (Kinsella et al, 1995). Transport teams may also consider initiation of iNO therapy for term and near-term newborns with hypoxic respiratory failure. If the referring facility does not have the capability to perform echocardiography, transport staff should carefully consider the possibility of congenital heart lesions that could be worsened by the use of iNO, including total anomalous venous return and lesions dependent on right-to-left ductal flow such as critical aortic stenosis. The rationale for initiating iNO during transport is to reduce pulmonary vasoreactivity and improve stability during transport. However, there have been no prospective studies to determine whether this practice affects patient outcome. INO during transport has been delivered by a number of different systems, such as the aeroNOX (Aeronox Technology Corporation, Quezon City, ­Philippines) iNOvent, and INOmax DS (INO Therapeutics, New Jersey) transport systems (Kinsella et al, 1995, 2002; Lutman and Petros, 2008; Tung, 2001). The use of iNO in combination with high-frequency ventilation in non-ECMO centers can complicate the transfer process. If the transport team does not have the capability to provide mobile high-frequency ventilation, it is recommended that the referring hospital perform a trial of conventional mechanical ventilation before the arrival of the transport team to establish that the infant can tolerate the transition. Both high-frequency jet ventilation (Bunnell ventilators [Bunnell Inc., Utah]) and high-frequency flow interrupter ventilation (Bird ventilators [Viasys, California]) have been configured and used with nitric oxide during transport (Honey et al, 2007; Mainilli et al, 2007). The 349 high-­frequency oscillator (SensorMedics 3100A [­Viasys, California]) that is commonly used in many NICUs is impractical for ground transport and is not configured for helicopter or fixed wing transport. However, even with these technologies, 30% to 40% of critically ill neonates improve only temporarily with iNO therapy and will ultimately require higher care levels (Fakioglu et al, 2005; Kinsella et al, 2002). In patients with pulmonary or cardiac failure that is unresponsive to maximal medical therapy, conventional ECMO is often used as a bridge therapy to allow either the lungs and heart to recover. Ideally, centers without ECMO capability have prospective criteria to guide the transfer of newborns before the need for ECMO cannulation. In some cases, the patient’s condition is so unstable that conventional transport cannot be conducted safely. Selected programs have the capability to provide mobile ECMO, during which patients are cannulated for ECMO before transport to the referring institution. Mobile ECMO can also benefit patients already receiving ECMO at a tertiary facility who are in need of advanced quaternary therapies, such as heart or heart-lung transplants (Wagner et al, 2008; Wilson et al, 2002). The resources and skill set necessary to safely and consistently perform ECMO in the transport environment have, by their complexity, restricted the number of transport programs with mobile ECMO capabilities (Coppola et al, 2008). NEUROLOGIC ISSUES Perinatal Depression and Therapeutic Hypothermia Hypoxic ischemic encephalopathy affects approximately 1 to 2 in 1000 term neonates (du Plessis and Volpe, 2002). Several early studies from the 1950s and 1960s suggested that hypothermia might be an effective therapy for the asphyxiated neonates, but there was no systemic long-term follow-up for these patients to confirm its effectiveness until a series of animal studies suggested that hypothermia was an effective neuroprotective therapy (Gunn et al, 1997, 1998). Several randomized trials of therapeutic hypothermia for term human infants have demonstrated a reduction in the combined outcome of death and neurodevelopmental disability when cooled, compared with a control population (Gluckman et al, 2005; Shankaran et al, 2005). However, the efficacy of hypothermia may diminish as the time from insult to cooling increases. Although clinical trials and current guidelines suggest that cooling should be initiated within 6 hours of insult, animal studies demonstrated that hypothermia is most effective if it is started as soon as possible (Gunn et al, 1997, 1998). For patients with a delay in referral or who are being transported from a distant center, waiting to start hypothermic therapy until arrival at the receiving center potentially places the patient at a significant disadvantage. The initial larger clinical trials did not address these issues in regard to critical care transport, but several case (Anderson et al, 2007), pilot (Eicher et al, 2005), and single-center studies have demonstrated the feasibility of controlled hypothermia during transport to significantly shorten the time to therapy initiation (Zanelli et al, 2008). Several new 350 PART VIII Care of the High-Risk Infant large-scale hypothermia trials that have been recently completed or are nearing completion include initiation of hypothermia before arrival at the referral center. The Total Body Hypothermia for Neonatal Encephalopathy Trial (the Toby Trial) compared standard intensive care plus total body cooling with standard intensive care for 72 hours. Recruited patients from referring hospital were assessed by the transport team, who performed amplitude integrated electroencephalograms and obtained consent if the patient was eligible and less than 6 hours of age. The patients randomized to cooling had gel packs applied if needed to cool the body to rectal temperature between 33° and 34° C (Azzopardi et al, 2009). The Australian Cooling Trial for Hypoxic-Ischemic Encephalopathy achieves the temperature range for referral study patients by turning off the normal heating systems and applying gel packs (at 10° C) around the infant’s head and chest, until the rectal temperature is reduced to 33° to 34° C (Jacob et al, 2008). While hypothermia is still being evaluated as a therapeutic tool to manage hypoxic ischemic encephalopathy, the current management trend for term and near-term infants being considered for hypothermia is to delay actively warming the patient at a referral hospital until there is a consultation with a receiving hospital that offers therapeutic hypothermia. If criteria are met, further cooling could be achieved by placing wrapped disposable cooling packs next to the trunk and head with continuous monitoring of rectal temperature. Once the rectal temperature stabilizes, the infant is transferred to a transport incubator with the heater turned off, and the transport team applies cooling packs or a blanket as needed to maintain a rectal temperature of 33° to 34° C during transport. Just as careful temperature monitoring during hypothermia is necessary for the safe transport of these patients, observational studies suggest that the avoidance of hyperthermia is critically important. In an observational study, Laptook et al (2008) noted an increase in mortality or disability in asphyxiated neonates who had elevated skin or esophageal temperatures. As additional studies are completed, therapeutic hypothermia may enter common clinical and transport practices. When combined with other therapeutic interventions, hopefully therapeutic hypothermia will have a greater effect on the consequences of perinatal asphyxia. Neonatal Seizures The highest rate for seizures in pediatrics occurs during the neonatal period, with an incidence of 2 to 3.5 in 1000 live births (Cowan, 2002); however, there is considerable variability in treatment regimens among pediatric hospitals in the United States (Blume et al, 2009). Phenobarbital, phenytoin, or midazolam have not been shown to significantly improve seizure management or reduce morbidity and mortality (Booth and Evanse, 2004; Carmo and Barr, 2005; Painter et al, 1999). Despite the evidence of limited efficacy, phenobarbital remains the first line of therapy to treat neonatal seizures (Painter et al, 1999; ­Sankar and Painter, 2005). Before initiating therapy, the transport team needs to consider the diverse causes for neonatal seizures. Metabolic disorders such as hypoglycemia, hypocalcemia, and inborn errors of metabolism, cerebral vascular events including intraventricular hemorrhage and stroke, bacterial and viral meningitis, infectious and developmental abnormalities can all manifest as or with neonatal seizures (Silverstein and Jensen, 2007). Having initiated treatment, the transport team needs to remain vigilant for potential changes in the patient’s clinical status either as a response to therapy or from the primary disease process. Attention to the basics of airway, breathing, and circulation should continue throughout the transport, because the major anticonvulsants used to treat neonatal seizure all have potential side effects, most commonly respiratory depression and hypotension. Controversy remains as to whether to treat all forms of neonatal seizures, although treatment seems prudent in the transport environment (Bartha et al, 2007). In animal models, seizures create neural, biochemical, and structural changes that have long-term cognitive and behavioral consequences (Thibeault-Eybalin et al, 2009). Similar neurologic findings have been suggested in human neonates. Both of the large-scale clinical hypothermia trials for hypoxic ischemic encephalopathy noted poorer prognosis for children with hypoxic ischemic encephalopathy with seizures (Gluckman et al, 2005; Shankaran et al, 2005). Glass et al (2009) compared magnetic resonance imaging findings and the presence of seizures in a group of newborns with suggested hypoxic ischemic encephalopathy. Adjusting for the magnetic resonance imaging results, patients with seizures were more likely to have long-term neurologic issues than those without seizures (Glass et al, 2009). The ideal anticonvulsant agent would reliably stop clinical and electrographic seizures with minimal adverse effects (Thibeault-Eybalin et al, 2009). Several trials of new antiepileptic medications are underway. Topirmante, levetiracetam, bumetanide, and zonisamide have all been used to treat neonatal seizures, but there have been no randomized trial to demonstrate their effectiveness compared with current therapies. Trials with levetiracetam and bumetanide are reportedly ongoing. Further work is necessary to develop more effective and safer antiepileptic drugs so that, along with potentially neuroprotective strategies, the vulnerable and immature brain can be protected. Several trials and studies have been proposed to hopefully elucidate a better treatment strategy (Clancy, 2006). STABILIZATION AND TRANSPORT OF THE NEONATE WITH CONGENITAL HEART DISEASE Transport of the neonate with congenital heart disease (CHD) follows the general guidelines for the transport of any critically ill neonate. However, neonates with complex CHD often need therapeutic intervention, requiring the support of multiple subspecialty consultants, including the pediatric cardiothoracic surgeon and pediatric cardiologist (Allen et al, 2003; Stark et al, 2000). The care of the neonate with CHD generally necessitates transport to a specialized center (Castaneda et al, 1989; Penny and Shekerdemian, 2001). The preoperative care of the patient affects postoperative outcomes and mortality (Mahle and Wernovsky, 2000; Mahle et al, 2000; Robertson et al, 2004; Simsic et al, 2007; Wernovsky et al, 1995, 2000). With the increasing use of fetal ultrasound examination to diagnose CHD antenatally, the ability to intervene immediately at delivery CHAPTER 29 Stabilization and Transport of the High-Risk Infant has become increasingly expected. In a single tertiary care center, 53% of neonates with CHD had a prenatal diagnosis (Dorfman et al, 2008), with an additional 38% obtaining a diagnosis before discharge from the newborn nursery. The aspects of stabilization of the neonate with CHD either diagnosed by prenatal ultrasound or suggested by postnatal clinical examination include initial resuscitation, airway management, vascular access, a judicious use of supplemental oxygen, prostaglandin E1 (PGE1) therapy, inotropic support, and communication with cardiac specialty services for timely transport and intervention. Neonates with Prenatally Diagnosed Congenital Heart Disease Significant advances in the diagnosis and treatment of neonates with critical CHD across the last decade have altered the timing of intervention, the types of interventions available, and the way in which CHD is diagnosed. With increasing numbers of patients with prenatal diagnosis (Berkley et al, 2009; Friedman et al, 2002; Jone and Schowengerdt, 2009; Khoshnood et al, 2005), fewer neonates may have their initial presentation of CHD in the emergency department. In a single-center review, prenatal diagnosis accounted for 53% of presentations of CHD (Dorfman et al, 2008). Another European center reported that 51% of transports with CHD had a prenatal diagnosis (Bouchut, 2008). In addition, a single-center cohort of critical CHD showed 44.3% to have had a prenatal diagnosis (Schultz et al, 2008). Clearly, prenatal diagnosis of CHD has increased across the past decade. Neonates with Suspected Congenital Heart Disease In a single tertiary care center, Dorfman et al (2008) found that 38% of neonates received a diagnosis with CHD in the newborn nursery before discharge, and 8% of patients presented with CHD after initial hospital discharge at 4 to 27 days of life. The most common nursery findings prompting diagnosis of heart disease in the nursery were an isolated murmur, cyanosis, or both (Berkley et al, 2009; Table 29-1). In a single emergency department review of patients younger than 5 months with undiagnosed cardiac disease, the most common presentation was congestive TABLE 29-1 Presenting Symptoms of Patients Diagnosed Postnatally Murmur Cyanosis Respiratory distress Shock Arrhythmia Other Multiple symptoms Diagnosis in Nursery, n = 73 (%) Diagnosis After Discharge, n = 16 (%) 38 32 7 4 3 3 14 25 0 19 38 0 6 13 Modified from Dorfman AT, Marino BS, Wernovsky G, et al: Critical heart disease in the neonate: presentation and outcome at a tertiary care center, Pediatr Crit Care Med 9:193-202, 2008. 351 heart failure, shock, and cyanosis; if restricting the findings to a neonatal population, shock and profound cyanosis were the presenting symptoms (Savitsky et al, 2003). Another study showed that 45% of critical CHD cases (CHD with signs of shock, end organ dysfunction, or in cardiac arrest) were diagnosed postnatally, particularly left-sided obstructive lesions (Schultz et al, 2008). This finding is consistent with other centers showing a preponderance of left-sided obstructive lesions (excluding hypoplastic left heart syndrome) in those diagnosed postnatally (Friedberg et al, 2009). The cyanotic neonate who fails the hyperoxia test without an obvious pulmonary etiology, who has an equivocal result on the hyperoxia test but has other signs or symptoms of CHD, or who is in shock within the first 3 weeks of life is highly likely to have complex CHD. Furthermore, pulse oximeter measurement of the postductal (probe on the foot) oxygen saturation has been found to be specific for the detection of CHD, particularly when combined with findings on the clinical examination (de-Wahl Granelli et al, 2009; Meberg et al, 2008, 2009; Reich et al, 2008; Sendelbach et al, 2008; Thangaratinam et al, 2007; ­Valmari, 2007). Although debate is ongoing as to the appropriateness of routine pulse oximetry as a screening tool (Sendelbach et al, 2008), the detection of saturations of less than 95% has a sensitivity of 72% to 77%, specificity rates over 99%, and false-positive rates of 0.2% to 0.6% in largevolume studies (Meberg et al, 2008; Valmari, 2007). These neonates are likely to have heart lesions that depend on blood flow through a patent ductus arteriosus to contribute to either systemic or pulmonary blood flow, or improvement of intercirculatory mixing. PGE1 given as continuous infusion will reopen the ductus arteriosus. In babies with ductal-dependent pulmonary blood flow, hypoxemia is lessened as the ductus opens, and the resultant metabolic acidosis will resolve. Babies with systemic flow that is dependent on the ductal connection to provide flow to the descending aorta will have congestive heart failure, low cardiac output, or shock, which is unlikely to be treatable by standard measures without reopening the ductus arteriosus. Patients with transposition of the great arteries will have improved intercirculatory mixing with a patent ductus arteriosus (Wernovsky, 2008; Wernovsky and Jonas, 1998). Interventions such as supplemental oxygen, airway management, volume resuscitation, inotropic therapy, and prostaglandin infusion require careful thought and consideration in the child with CHD. Considerations of Therapies in the Neonate with Congenital Heart Disease The neonatal resuscitation algorithm is still applicable in the presence of CHD (Johnson and Ades, 2005), but should be modified in certain circumstances. In presentations with hypoxemia that is unresponsive to supplemental oxygen, congestive heart failure, or shock, simultaneous attention is devoted to the basics of neonatal advanced life support and to assurance of a patent ductus arteriosus. A stable airway must be maintained, allowing for adequate alveolar oxygenation and ventilation. In critically ill neonates with CHD presenting with severe cyanosis or circulatory collapse, intubation should be performed if 352 PART VIII Care of the High-Risk Infant possible after premedication with sedation and neuromuscular blockade (discussed subsequently). Reliable venous access is important, and arterial monitoring is helpful for ongoing assessment of blood pressure, acid-base status, and gas exchange. Volume resuscitation, inotropic support, and correction of metabolic acidosis may be required to maximize cardiac output and tissue perfusion. Blood glucose and ionized calcium should be checked and treated to achieve normal range for age. An evaluation for sepsis is typically performed simultaneously, and empiric antibiotic therapy is initiated while evaluation continues. Supplemental Oxygen Supplemental oxygen is a potent pulmonary vasodilator and systemic vasoconstrictor, and it can adversely affect the physiology in neonates with a single ventricle, as well as those with two ventricles with an unrestrictive ventricular septal defect or great vessel communication (see ­Chapter 55, Congenital Heart Disease). In these babies, the ratio of pulmonary vascular resistance to systemic vascular resistance will determine the proportion of blood flow to each vascular bed. The oxygen-induced pulmonary vasodilation can decrease pulmonary vascular resistance and increase pulmonary blood flow at the expense of systemic blood flow, thus reducing systemic output. Titrating oxygen via nasal cannula or face mask to a target peripheral saturation of 75% to 85% usually corresponds to adequate blood flow in both the pulmonary and systemic systems. In the setting of normal hemoglobin, cardiac output, and oxygen consumption, these oximetry values will provide adequate oxygen delivery. Higher oxygen saturations are typically not necessary and importantly may in fact ultimately result in decreased oxygen delivery to the peripheral tissues. Considerations for intubation and mechanical ventilation are discussed Intubation and Ventilation, later. Prostaglandin E1 Therapy In the instance of antenatally diagnosed CHD, ductal dependency (either for systemic or pulmonary blood flow) is often already determined. Whenever possible, PGE1 infusions should be prepared ahead of delivery and started promptly in a peripheral intravenous catheter. The dose of PGE1 varies according to the timing of the diagnosis and the degree of ductal closure that is present upon discovery (or suspicion) of the cardiac lesion. In most cases, PGE1 at 0.01 to 0.025 μg/kg/min via intravenous infusion is adequate for stabilization. PGE1 may be given via a central venous catheter or peripheral intravenous catheter, or in situations where intravenous access cannot be obtained, via an umbilical arterial catheter or intraosseous line. The dose is usually titrated down once the patent ductus arteriosus has been demonstrated echocardiographically. However, doses as high as 0.1 μg/kg/min may be required in the newborn with significant ductal constriction or in the neonate who at 1 to 2 weeks of life presents with a closing, closed, or severely restrictive ductus arteriosus. As in the prenatally diagnosed neonate, once therapeutic effect is achieved, the dose can generally be decreased without loss of therapeutic effect. Prostaglandin response is often immediate if ductal patency is important to the hemodynamics of the infant. Failure to respond may mean that the initial diagnosis of ductal dependent CHD is incorrect, the ductus is unresponsive to PGE1 (which may occur in older infants), or that there is no ductus arteriosus present. On rare occasion, the neonate may have progressive instability after initiating PGE1 therapy. This important diagnostic finding strongly suggests a congenital heart defect with obstructed blood flow out of the pulmonary veins or the left atrium. These lesions include (1) hypoplastic left heart syndrome with restrictive foramen ovale or intact atrial septum, (2) other variants of mitral atresia with a restrictive foramen ovale, (3) transposition of the great arteries with intact ventricular septum and restrictive foramen ovale, and (4) total anomalous pulmonary venous return with obstruction of the common pulmonary vein. If the neonate clinically deteriorates despite PGE1 therapy, urgent echocardiography with plans for potential interventional cardiac catheterization or cardiac surgery need to be made. Early recognition of this deterioration despite PGE1 requires prompt transfer to a cardiac unit (Penny and Shekerdemian, 2001). Controversy exists as to whether PGE1 should be continued in these rare instances; it has generally been the authors’ practice to continue the infusion at the usual dose. More important than the decision to continue or discontinue PGE1 is that the lack of response to PGE1 in a child with suspected CHD is a marker for rare forms of CHD that do not respond to medical management and require urgent surgical or catheter intervention. Although PGE1 is critical in the management of most neonates with CHD, there are a number of potential adverse effects associated with PGE1 continuous infusion that must be anticipated, particularly in the premature or LBW infant, in which they occur more commonly. The most common adverse effects include hypotension (caused by vasodilation), apnea, rash, and fever (Kramer et al, 1995; Lewis et al, 1981). Other less common side effects include seizures, gastric outlet obstruction, cortical hyperostosis, and leukocytosis (Arav-Boger et al, 2001; Teixeira et al, 1984). Although high-dose PGE1 infusions may occasionally be required, most often neonates are in stable condition receiving a low-dose infusion. The resultant vasodilation and hypotension is less common and usually treatable with fluid resuscitation rather than inotropic support (Kramer et al, 1995). A separate intravenous catheter is typically necessary for the purpose of volume administration and should be considered prior to transport. A 5 to 10 mL/kg bolus of normal saline, lactated Ringer’s solution, or 5% albumin will generally normalize the blood pressure in cases of hypotension related to PGE1. Serum ionized calcium levels should be checked and normalized if low. If hypotension is refractory to fluid administration, an alternative cause of hypotension should be considered (e.g., a restrictive ductus, pericardial effusion, myocardial dysfunction, sepsis), and dopamine (3 to 5 μg/kg/min) may be given to offset the vasodilatory effects of PGE1. If apnea and hypotension occur, they will usually manifest during the first hours of administration, but may occur at any time during the infusion. This occurrence mandates the need for ongoing cardiorespiratory monitoring even after stabilization on PGE1 infusion. Furthermore it has CHAPTER 29 Stabilization and Transport of the High-Risk Infant been shown that, although the side effect profile (particularly apnea) increases with increasing doses and lower weight, the clinical response is not dose dependent. In other words, infants can be maintained safely on a lowdose PGE1 infusion until definitive care is achieved and thereby potentially avoid the significant side effects of apnea and hypotension. Certainly, PGE1 therapy alone does not require mechanical ventilation without the presence of significant or recurrent apnea (Browning et al, 2007; Kramer et al, 1995) or other clinical states requiring mechanical ventilation. The use of low-dose PGE1 (<0.015 μg/kg/min) for duct-dependent lesions is unlikely to cause apnea requiring mechanical ventilation, and it is safe to transport the infant without establishing an artificial airway in most cases (Carmo and Barr, 2005). When starting PGE1 infusion, it is prudent to remeasure arterial blood gases and reassess vital signs and perfusion within 15 to 30 minutes of initiation. Doing so allows effectiveness to be assessed and any adverse events to be addressed promptly. Intubation and Ventilation Although the initiation of PGE1 alone does not require the neonate to be intubated, the presence of profound hypoxemia, respiratory distress, or hemodynamic instability may demand airway intervention. In most cases, intubation should be performed after premedication with sedation (narcotic or benzodiazepine), preferably with neuromuscular blockade. Although intubation can be performed without sedation and neuromuscular blockade, there are physiologic reasons to use these agents for the intubation of a neonate with CHD. First, the catecholamine surge that will occur with a nonsedated intubation may result in significant dysrhythmias in the at-risk myocardium. Second, vagally mediated bradycardia from hypoxemia, hypercapnia, or laryngeal stimulation may lead to asystole in these neonates, who have little reserve. Finally, sedation and neuromuscular blockade will reduce total body oxygen consumption, raising the mixed venous oxygen saturation and improving oxygen delivery. Premedication with atropine (0.02 mg/kg) can blunt the vagal effects of laryngoscopy. Fentanyl (1 to 2 μg/kg) or midazolam (0.05 to 0.1 mg/kg) for sedation may be given, with titration of dose to effect. Chest wall rigidity may occur with low-dose fentanyl (more likely if given rapid push) and may require neuromuscular blockade for adequate ventilation. In most cases, the neonate is preoxygenated with 100% FiO2, which can be down-titrated to achieve an acceptable oxygen saturation given the underlying CHD after intubation. In an effort to maintain a balanced circulation, ventilation should attempt to achieve normocarbia (Pco2, 35 to 40 mm Hg). In the single-ventricle patient (i.e., a patient in whom the systemic and pulmonary circulations are dependent on a single-ventricular pumping chamber) or the neonate with unrestrictive ventricular or great vessel communications, significant adjustments in ventilation will alter pulmonary vascular resistance, and therefore may have an effect on both pulmonary and systemic blood flow. Hyperventilation in the neonate decreases pulmonary vascular resistance, thereby increasing pulmonary blood flow potentially at the expense of systemic blood flow. 353 Although intubation and ventilation may be necessary interventions, it should be noted that preoperative mechanical ventilation is a risk factor for mortality or poor outcome after surgery for CHD in multiple studies ­(Gottlieb et al, 2008; Robertson et al, 2004; Simsic et al, 2007; Tabbutt et al, 2008). Inotropic Therapy The neonate with CHD discharged from the nursery before diagnosis may be experiencing congestive heart failure or circulatory collapse. These infants, while urgently requiring the opening of the ductus arteriosus to provide either systemic or pulmonary blood flow, may also require inotropic therapy to recover from the adverse effects of a myocardium that has decompensated in the face of high afterload (obstructive left-sided lesions requiring ductal patency for systemic blood flow) or hypoxemia (obstructive right-sided lesions requiring ductal patency for pulmonary blood flow). Appropriate volume status should be achieved in conjunction with the institution of inotropic therapy and secure access for its delivery. The choice of inotropic agent remains largely practitioner dependent. Sympathomimetic amines are the most commonly used inotropic agents. They can be endogenous (dopamine and epinephrine) or synthetic (dobutamine and isoproterenol). Dopamine is a norepinephrine precursor and stimulates dopaminergic, β1-adrenergic, and α-adrenergic receptors in a dose-dependent manner. Dopamine improves myocardial contractility, which will increase stroke volume. This increase leads to improved cardiac output and higher mean arterial pressure, with resultant increased urine output. There is a low incidence of side effects at doses less than 10 μg/kg/min. Dobutamine is an analogue of dopamine; it stimulates β1-adrenergic receptors predominantly, with relatively weak β2-adrenergic receptor and α-adrenergic receptor activity. No definitive benefit has been found in the use of dopamine versus dobutamine, although dopamine is more likely to increase the systemic blood pressure in the short term. No differences in outcomes or mortality were found between the two inotropes (Subhedar and Shaw, 2003). Both drugs are generally initiated at 3 to 5 μg/kg/min. Epinephrine may also be used in neonates with hypotension and hemodynamic deterioration. Epinephrine has α1-, α2-, β1-, and β2-adrenergic effects. Few trials have been done to recommend epinephrine over other agents. A randomized controlled trial comparing epinephrine and dopamine in LBW infants found equal efficacy in treating hypotension (Valverde et al, 2006). The recommended starting dose is 0.03 to 0.05 μg/kg/min. Isoproterenol stimulates β1- and β2-adrenergic receptors. It has greater chronotropic effect and stronger vasodilatory effect because of the β2 stimulation. It needs to be started at low dose and titrated to effect, because of the strong chronotropic effect. Chronotropic effects will precede inotropic effects in a responsive heart and can produce tachyarrhythmias. Recommended starting dose is 0.01 to 0.05 μg/kg/min. The adverse effects of the sympathomimetic amines include tachycardia, atrial and ventricular arrhythmias, and peripheral vasoconstriction causing increased afterload. Tachycardia will increase myocardial oxygen 354 PART VIII Care of the High-Risk Infant consumption, whereas arrhythmias and vasoconstriction decrease cardiac output. Vascular Access Umbilical venous catheters (UVCs) and umbilical arterial catheters (UACs) are useful in the stabilization, transport, and preoperative and postoperative management of the neonate with CHD. However, the placement of these lines is not without risk. Beyond the infectious risks, newborns with CHD are at a higher risk for thromboembolic events because of their underdeveloped clotting mechanisms, small vessel lumens, and low flow states. There is also an ongoing risk of systemic embolization of air and particulate matter in babies with an intracardiac right-to-left shunt. There has been much controversy surrounding the optimal placement of a UAC in premature infants, although there are scant data regarding full-term infants with CHD. A Cochrane Database review suggested that “high” catheters were associated with a decreased risk of vascular complications with no significant increase in adverse sequelae (Barrington, 2000; Hermansen et al, 2005). Although there is variation in definition, in general “high” catheters are above the level of the diaphragm and below the level of the left subclavian artery (approximately T6 to T10); “low” catheters are below the renal arteries but above the aortic bifurcation (approximately L3 to L5). Levels between high or low are associated with increased risks of complications, as are placements below the L5 level (Hermansen et al, 2005). Although transient placement of a UAC in this area will generally be well tolerated, long-term placement should be avoided if possible. The UVC should be placed, if possible, at the inferior vena cava/right atrial junction or in the atria. It might not be necessary to obtain ideal placement of either the UAC or UVC for stabilization and transport, and manipulation of lines and reconfirmation can delay transport and stabilization to the tertiary care center. The risk-benefit relation for the use of umbilical lines in the neonate with ductal dependent circulation has not been well delineated. The need for central access should be judged based on the clinical status of the neonate and stability for transport. Transport of the Neonate With Congenital Heart Disease Once the neonate with suspected or known CHD has been sufficiently resuscitated and stabilized, they should be transferred to an institution that provides subspecialty care in pediatric cardiology and pediatric cardiothoracic surgery. In a single tertiary center study of neonates presenting with CHD, 157 of 190 patients required at least one surgical or catheter intervention with a median of 3 days of life at time of surgery (Dorfman et al, 2008). To optimize management, early communication with the specialist center is vital. Successful transport involves two phases: referring hospital staff to the transport team and subspecialists, and transport staff to the accepting hospital staff. The need for this precise and thorough communication between respective teams cannot be overemphasized. Whenever possible, the pediatric cardiologist, neonatologist, or intensivist at the accepting hospital should be included in formulating the transport management plan while the neonate is still at the referring hospital. This procedure will help to guide the timing and urgency of the transport, line placement, and recommendations for airway management and supplemental oxygen. Often the patient with a duct-dependent lesion will improve greatly with the institution of prostaglandin therapy and may not need to be rushed to the cardiac referral center as an emergent case. Similar supports and treatments need to be put into place before transport of any critically ill neonate or infant. Secure vascular access should be obtained, with a port available for volume resuscitation that is not running the inotropic support or PGE1 infusion, to avoid interruption of this medication. Intubated patients should have the airway position recorded and secured and have a nasogastric or orogastric tube for decompression. The intubated infant should remain unfed, and medication administration and fluids should be given intravenously. Appropriate sedation should be used, as well as maintenance of normothermia and avoidance of hyperthermia, to minimize oxygen consumption. Evaluation of acid-base status, oxygen delivery, temperature, serum glucose, and calcium should take place before transport and should be corrected when possible. Neonates with conotruncal anomalies are at particular risk for hypocalcemia, because they have an increased frequency of 22q11 deletion syndrome (Gallot et al, 1998), which may have associated thymic and parathyroid hypoplasia. Furthermore, the neonatal myocardium may be more dependent on calcium for inotropy than the adult. Ionized calcium levels less than 1 mg/dL may have a significant negative effect on contractility and should be considered for treatment with calcium gluconate (50 to 100 mg/kg) or calcium chloride (0.1 mEq/kg of elemental calcium or 0.01 g/kg) intravenously, via a central catheter if possible. In patients with mixing lesions, single ventricle or in those with an unknown diagnosis, supplemental oxygen should be used and titrated to maintain pulse oximeter oxygen saturation of 75% to 85% for “balance” of circulation between pulmonary and systemic systems and to provide adequate oxygen delivery, because higher oxygen saturations may actually be deleterious, as noted previously. The accepting pediatric cardiologist can aid with lesion-specific advice and guidelines to optimize in-transport care. Hypotension is a late finding of shock in neonates. Earlier, more sensitive signs, of impending decompensation include persistent tachycardia despite adequate intravascular volume and temperature control, poor tissue perfusion, and metabolic acidosis. Treatment of shock should occur before transport during the stabilization phase of management, although the patient in profound shock and with significant acidosis at presentation may require significant time for resolution of organ dysfunction and perfusion and clearance of acidosis. Finally, failure to respond to prostaglandin, assuming adequate dose and delivery, should promote prompt transfer to a cardiac unit able to care for lesions with obstruction to pulmonary venous outflow. Before leaving the referring hospital, hemodynamic status (capillary refill, heart rate, systemic blood pressure, and acid-base status) needs to be reassessed and communicated to the accepting hospital. CHAPTER 29 Stabilization and Transport of the High-Risk Infant Summary of Cardiac Transport Stabilization and transport of the neonate with known or suspected critical CHD affect their preoperative condition, potentially contributing to short-term mortality risk and long-term morbidity. One key to stabilization is the recognition of potential or confirmed CHD, appropriate resuscitation, airway management where indicated, stable vascular access, judicious oxygen utilization, PGE1, and inotropic support when necessary. Early consultation with a center specializing in pediatric cardiac care, for advice and timing of transport, will aid the initial care and stability of the infant on transport. Accurate, detailed information must be communicated among the referring hospital, the transport team, and the accepting hospital. SURGICAL EMERGENCIES Because most births occur in hospitals without an NICU or neonatal surgical abilities, the need for surgical evaluation or intervention is a common reason for interfacility transport. Whereas some surgical conditions are relatively not urgent in nature, there are several diagnoses that represent truly life-threatening conditions for which stabilization and transport require expertise and specialized resources. Congenital Diaphragmatic Hernia Advances in ultrasound technology have resulted in the prenatal diagnosis of up to 60% of fetuses affected by congenital diaphragmatic hernia (Gallot et al, 2007). A recent systematic review showed that outcome was improved for newborns with a prenatal diagnosis or born in a tertiary care center (Logan et al, 2007b). Newborns with undiagnosed CDH, especially those delivered at a community hospital, are at high risk for complications. Before transport, critically ill infants with suspected or known CDH should undergo tracheal intubation and placement of a large-bore nasogastric or orogastric tube to continuous suction, because many have received bag-mask ventilations and swallowed air can quickly travel beyond the pylorus and distend the intrathoracic intestinal contents (GrisaruGranovsky et al, 2009; Logan et al, 2007b). Whenever possible, the same principles of management that are used in the tertiary care center should be used during transport: limitation of peak airway pressures and use of low tidal volumes to avoid ventilator-induced lung injury, judicious use of sedation, avoidance of chemical paralysis to permit spontaneous breathing, and maintenance of adequate systemic blood pressure with the use of fluid therapy and inotropes (Logan et al, 2007a). Although there is still controversy over the role of extracorporeal membrane oxygenation for newborns with CDH, transport to a high-volume center with ECMO capabilities should be strongly considered (Grushka et al, 2009; Morini et al, 2006). Abdominal Wall Defects The proper management of a newborn with gastroschisis or omphalocele is critical during the first several hours of life, and delivery in a tertiary care center has been associated with improved outcome (Quirk et al, 1996). In 355 addition to the need for initial resuscitation and cardiorespiratory support, the correct treatment of the exposed bowel or sac may improve the infants’ chances of a successful repair and long-term intestinal function. The mean gestational age for newborns with gastroschisis is 36.6 weeks, and many affected infants are small for gestational age (Baerg et al, 2003; Lausman et al, 2007). As with other mildly premature and growth-restricted infants, patients with gastroschisis are at risk for hypothermia and hypoglycemia. Heat loss is exacerbated by the large surface area of the exposed intestines, which can also serve as a significant source of fluid loss. Prevention of heat and fluid losses can be accomplished by placing the lower part of the infant’s body, including the intestines, into a transport bag (i.e., bowel bag or Lahey bag) before placement of the infant into the heated transport isolette. Significant fluid losses can occur through the exposed mucosa, and the patient may require aggressive fluid replacement (120 to 150 mL/kg/day). The use of antibiotics should be considered if risk factors for sepsis are present and reviewed with the pediatric surgeon. Infants with gastroschisis are at risk for intestinal vascular compromise, because the vascular pedicle containing the arterial supply and venous drainage from the bowel must pass through the relatively small abdominal wall defect. Transport personnel must closely monitor the appearance of the bowel to detect signs of venous congestion or ischemia. Transporting the infant in the lateral position, with support of the exposed intestines to avoid tension or torque, is recommended. The use of intestinal pulse oximetry has been described for monitoring the bowel for ischemia through a transparent silo, but has not been studied as a tool during interfacility transport (Kim et al, 2006). Vascular compromise of the intestine is a surgical emergency, and communication with the receiving facility is essential to coordinate urgent intervention. The transport of an infant with omphalocele has similar considerations, although unless the sac has ruptured there is significantly less risk of heat and fluid loss. Infants with an omphalocele are more likely than those with gastroschisis to have other birth defects (e.g., CHD). Furthermore, infants with giant omphaloceles often have respiratory insufficiency caused by diaphragmatic dysfunction, pulmonary hypoplasia, or both, and they may require ventilatory assistance. All infants with gastroschisis or omphalocele require placement of a large-bore nasogastric or orogastric tube due to functional ileus or intestinal obstruction, as may occur with associated stenoses or atresias. In general, cannulation of the umbilical vessels is not recommended unless other methods for vascular access are not successful. Esophageal Atresia and Tracheo Esophageal Fistula Esophageal atresia, with or without tracheo esophageal fistula, is typically diagnosed within the first day of life because of increased secretions, poor feeding, and respiratory distress. General transport considerations include placement of a large-bore sump-type tube for continuous aspiration of the proximal esophageal pouch, positioning (prone with the head of bed elevated), and respiratory 356 PART VIII Care of the High-Risk Infant support as indicated. Direct aspiration of secretions into the trachea may occur with either a proximal or distal tracheo esophageal fistula. Transport providers should be aware that infants with a distal tracheo esophageal fistula (type C), characterized by the presence of air in the intestinal tract, are at risk for gastric and intestinal insufflation via the fistula when receiving positive-pressure ventilation. Bag-mask ventilation and continuous positive airway pressure should be avoided. If the infant requires endotracheal intubation, the endotracheal tube should be positioned as close to the carina as tolerated in an effort to position the distal tip beyond the fistula and minimize direct inflation of the distal esophageal segment with pressurized gas. In extreme cases, gastric rupture with pneumoperitoneum has been reported, requiring emergency paracentesis, laparotomy, or both (Maoate et al, 1999). Midgut Volvulus Malrotation with midgut volvulus can be a catastrophic event resulting in intestinal ischemia and shock, and it represents a surgical emergency in the neonate. The most common clinical presentation of midgut volvulus is bilious vomiting, which is a nonspecific sign of intestinal obstruction. Expeditious evaluation of the newborn with bilious vomiting is essential to facilitate prompt surgical intervention in the event that midgut volvulus is identified, to prevent progression of vascular insufficiency to actual intestinal necrosis. An upper gastrointestinal series is the radiologic test of choice to diagnose malrotation and midgut volvulus, although some practitioners have reported success with the use of ultrasound examination to identify the relationship of the superior mesenteric vessels (Lampl et al, 2009; Shew, 2009). An infant with suspected midgut volvulus should be rapidly transported to a facility with pediatric radiology and surgical capabilities. Care of the infant with suspected midgut volvulus during interfacility transport is primarily supportive and includes circulatory support with intravenous fluid repletion, correction of metabolic abnormalities, and gastric decompression with a large-bore nasogastric or orogastric tube. Necrotizing Enterocolitis Necrotizing enterocolitis (NEC) is common in premature infants, and approximately 30% of newborns with NEC will require surgical intervention in the form of laparotomy or peritoneal drain placement (Guthrie et al, 2003). The clinical presentation of NEC is typically nonfocal and often mimics signs of systemic sepsis. Clinical and radiologic findings that are specific to NEC include abdominal distention and discoloration of the abdominal wall, dilated or thickened bowel loops, pneumatosis intestinalis, portal venous gas, and free intraperitoneal air. Infants with suspected NEC should be transported to a facility with pediatric surgical capabilities. Care during transport is primarily supportive and includes intravenous fluids, administration of broad-­spectrum antibiotics, correction of metabolic abnormalities, and gastric decompression. Respiratory failure is common because of disordered control of breathing and elevation of the diaphragm from abdominal distension. Meningomyelocele Although most newborns with neural tube defects are diagnosed prenatally because of an elevated AFP and are therefore delivered in a tertiary care center, the unexpected birth of an infant with a meningomyelocele may be an indication for neonatal transport. For purposes of transport, the infant should be placed in the prone position and the spinal defect should be covered with moist sterile dressings as well as some form of plastic wrap to maintain moisture. The lesion can be covered with a moistened Telfa dressing and then loosely encircled with a Kerlix “donut,” with the entire defect covered with a sterile drape. This dressing can be moistened as indicated during the transport process (Jason and Mayock, 1999). Avoid the use of latex gloves during care of these patients. If the skin covering the defect is disrupted, there is an increased risk of infection, and empiric antibiotics should be considered. Infants with meningomyelocele may or may not have accompanying hydrocephalus at birth; approximately 25% of affected patients will require shunting in the immediate newborn period, with up to 85% eventually undergoing shunt placement (Bowman et al, 2001). SUGGESTED READINGS Browning Carmo KA, Barr P, West M, et al: Transporting newborn infants with suspected duct dependent congenital heart disease on low-dose prostaglandin E1 without routine mechanical ventilation, Arch Dis Child Fetal Neonatal Ed 92:F117-F119, 2007. Das UG, Leuthner SR: Preparing the neonate for transport, Pediatr Clin N Am 51:581-598, 2004. Dorfman AT, Marino BS, Wernovsky G, et al: Critical heart disease in the neonate: presentation and outcome at a tertiary care center, Pediatr Crit Care Med 9: 193-202, 2008. Engle WA: American Academy of Pediatrics Committee on Fetus and Newborn: Surfactant-replacement therapy for respiratory distress in the preterm and term neonate, Pediatrics 121:419-432, 2008. King BR, Woodward GA: Procedural training for pediatric and neonatal transport nurses: Part I: - training methods and airway training, Pediatr Emerg Care 17:461-464, 2001. Orr RA, Felmet KA, Han Y, et al: Pediatric specialized transport teams are associated with improved outcomes, Pediatrics 124:40-48, 2009. Phibbs CS, Baker LC, Caughey AB, et al: Level and volume of neonatal intensive care and mortality in very-low-birth-weight infants, N Engl J Med 356: 2165-2175, 2007. Polin RA, Randis TM, Sahni R: Systemic hypothermia to decrease morbidity of hypoxic-ischemic brain injury, J Perinatol 27:S47, 2007. Silverstein FS, Jensen FE: Neurological progress: neonatal seizures, Ann Neurol 62:112, 2007. Thangaratinam S, Daniels J, Ewer AK, et al: Accuracy of pulse oximetry in screening for congenital heart disease in asymptomatic newborns: a systematic review, Arch Dis Child Fetal Neonatal Ed 92:F176-F180, 2007. Woodward GA, Insoft RM, Kleinman ME, editors: Guidelines for air and ground transport of neonatal and pediatric patients, Elk Grove Village, Ill, 2007, American Academy of Pediatrics. Woodward GA, King BR, Garrett AL, et al: Prehospital care and transport medicine. In Fleisher G, Ludwig S, Henretig F, editors: Textbook of pediatric emergency medicine, ed 5, Philadelphia, 2006, Lippincott, Williams, and Wilkins, pp 93-134. Complete references used in this text can be found online at www.expertconsult.com C H A P T E R 30 Temperature Regulation of the Premature Neonate Stephen Baumgart and Sudhish Chandra The human neonate is a homeothermic mammal. Even the smallest premature infants respond adaptively to changes in their environment. Response, however, may be insufficient to maintain core body temperature and can render preterm babies functionally poikilothermic, even in moderately temperate environments. Morbidity (e.g., poor brain and somatic growth) and mortality rates increase when core body temperature is permitted to decline much below 36° C (96.8° F), and moderate to severe hypothermia below 31° C (88.7° F) results in precipitous declines in heart rate and blood pressure. COLD STRESS Evaporative heat loss is widely as regarded the most stressful cooling event upon birth. Severe nonevaporative heat loss is also problematic for several reasons. First, a baby’s exposed body surface area is much larger than an adult’s, relative to metabolically active body mass (Table 30-1). Especially for the extremely low birthweight infant, the heat-dissipating area is fivefold to sixfold greater proportionate to that of the adult. Second, the baby’s small size presents a much smaller heat sink to store thermal reserve. Finally, the radius of curvature of the body is less than in the adult, resulting in a thinner protective boundary layer of warm, still air. Aside from these geometric considerations, characteristics of the premature infant’s skin contribute to the problem of excessive heat loss. The skin and subcutaneous fascia provide little insulation against the flow of heat from the core to the surface. Moreover, the lack of a keratinized epidermal barrier exposes infants to vastly increased evaporative heat loss. Finally, premature infants may not induce effective thermogenesis in response to cold stress. Shivering and nonshivering thermogenesis are compromised by low brown fat stores. Furthermore, the presence of hypoxia (common with preterm birth) seriously reduces nonshivering thermogenesis by reducing mitochondrial oxidative capacity. PHYSICAL ROUTES OF HEAT LOSS CONVECTION Convective heat loss in newborns occurs when ambient air temperature is less than the infant’s skin temperature. Convective heat loss includes natural convection (passage of heat from the skin to the ambient still air) and forced convection, in which mass movement of air over the infant conveys heat away from the skin. The quantity of heat lost is proportional to the difference between air and skin temperatures, and to air speed. The effect of forced convection in disrupting the microenvironment of warm, humid air layered near an infant’s skin usually is not appreciated in the nursery, where drafts, air turbulence, and consequently heat loss can occur within the relatively protective environment of an incubator. EVAPORATION Passive transcutaneous evaporation of water from a newborn’s skin (insensible water loss) results in the dissipation of 0.58 Kcal/mL of latent heat. As shown in Figure 30-1, transcutaneous water loss increases exponentially with decreasing body size and gestation. The tiniest premature baby, who is least able to tolerate cold stress, can incur evaporative heat loss in excess of 4 Kcal/kg/hr (water loss of approximately 7 mL/kg/hr). Evaporation is enhanced by low vapor pressure (i.e., high air temperature and low relative humidity) and air turbulence. The highest evaporative losses occur on the first day of life. During the first week of life in infants born at 25 to 27 weeks’ gestation, evaporative heat losses may be higher than radiant losses (Hammarlund et al, 1986). RADIATION Radiant heat loss constitutes the transfer of heat from an infant’s warm skin, via infrared electromagnetic waves, to the cooler surrounding walls that absorb heat. Radiant heat loss is proportional to the temperature gradient between the skin and surrounding walls. An infant’s posture may affect radiant heat loss by increasing or reducing the exposed radiating surface area. In a moderately humid environment (relative humidity approximately 50%), babies experience an ambient temperature (termed operant temperature) determined 60% by wall temperature and 40% by air temperature. CONDUCTION Conductive heat loss to cooler surfaces in contact with an infant’s skin depends on the conductivity of the surface material and its temperature. Usually babies are nursed on insulating mattresses and blankets that minimize conductive heat loss. PHYSIOLOGY OF COLD RESPONSE AFFERENTS Homeothermic response to a cold environment begins with the sensation of temperature. Traditional physiology identifies two temperature-sensitive sites: the hypothalamus and the skin. Sensation of cold by neonatal skin triggers a cold-adaptive response long before core sensors 357 358 PART VIII Care of the High-Risk Infant 7 TABLE 30-1 Body Surface Area-to-Body Mass Ratio Adult 70 Surface Area (m2) Ratio (cm2/kg) 1.73 250 Premature infant 1.5 0.13 870 Very premature infant 0.5 0.07 1400 in the hypothalamus become chilled. Some investigators conjecture that neonatal cold reception resides primarily in the skin, whereas warm reception resides in the hypothalamus. Both sensors are probably integrated, because cold sensory response is inhibited by core sensor hyperthermia and vice versa. Peripheral skin cold sensation is teleologically important, because early detection of heat loss from the skin aids in the infant’s timely response for maintaining core temperature. CENTRAL REGULATION Integration of multiple skin temperature inputs probably occurs in the hypothalamus; however, no single control temperature seems to exist. Under different environmental conditions, temperature of the skin can fluctuate 8° to 10° C, and temperature of the hypothalamus may vary ±0.5° C. There are also diurnal temperature fluctuations, variations with general sympathetic tone, and blunted regulation with asphyxia, hypoxemia, and other central nervous system defects. Premature infants can regulate core temperature near 37.5° C (99.5° F), whereas term infants may respond by maintaining 36.5° C (97.7° F). Because important thermoregulatory processes are triggered by deviations of as little as 0.5° C at any temperature-sensitive site, environmental temperature homeothermy is important. EFFERENTS The effector limb of the neonatal thermal response is mediated primarily by the sympathetic nervous system, although infant behavior may also be involved. The earliest maturing response is vasoconstriction in deep dermal arterioles, resulting in reduced flow of warm blood from the infant’s core into the exposed periphery. In addition, reducing blood flow effectively places a layer of insulating fat between the warm core tissue compartment and the cooler exposed skin surface in the term infant. Reduced fat content in babies with low birthweight diminishes this effective insulating property. Vasoconstriction nevertheless remains the newborn’s first line of thermally insulating defense, and the response is present even in the most premature infant. Brown fat constitutes a second sympathetic effector organ that provides a metabolic source of nonshivering thermogenesis. (Babies do not shiver like adults to generate heat, because muscle constriction-relaxation fibers are not yet myelinated.) Brown fat located in axillary, mediastinal, perinephric, and other regions of the newborn is particularly enervated and equipped with an abundance of mitochondria to hydrolyze and reesterify triglycerides IL  28.04e1.73(Wt) r   0.90 p 0.001 6 Insensible water loss, ml/kg/hr Body Weight (kg) 5 4 3 2 1 0 0 0.4 0.8 1.2 1.6 2.0 2.4 Weight, kg FIGURE 30-1 Exponential increase in evaporative water loss from skin of infants with very low birthweight who were nurtured under ­radiant warmers. (Adapted from Baumgart S, et al: Fluid, electrolyte and glucose maintenance in the very low birthweight infant, Clin Pediatr 21:199-206, 1982.) and to oxidize free fatty acids. In the term infant, these reactions are exothermic and may increase metabolic rate by twofold or more. Preterm babies, however, have little brown fat and may not be capable of more than a 25% increase in metabolic rate despite the most severe cold stress (Hull, 1966). Finally, recent evidence suggests that control of voluntary muscle tone, posture, and increased motor activity with agitation may serve to augment heat production in skeletal muscle via glycogenolysis and glucose oxidation. Clinical observations of infant posture, behavior, and skin perfusion and measurements of skin and core temperature gradient may ultimately provide the most useful guidelines for assessing infant comfort during incubation. MODERN INCUBATION: INCUBATORS AND RADIANT WARMERS The lifesaving requirement of an appropriate thermal environment was demonstrated conclusively by Day et al (1964) and further defined by Silverman et al (1966). Minor changes in heat balance create an oxygen and energy cost, inducing an increased metabolic rate that can be met only by increased ventilation or increased inspired oxygen and appropriate cardiovascular response to provide oxygen delivery to activated tissues. THERMAL NEUTRAL ZONE The thermal neutral zone is a narrow range of environmental temperatures within which an infant’s metabolic rate is minimal and normal body temperature is maintained. 359 CHAPTER 30 Temperature Regulation of the Premature Neonate °F °C TABLE 30-2 Mean Temperature Needed to Provide Thermal 36 Baby weighing 1 kg at birth Neutral thermal environment 96 95 Neutrality for a Healthy Baby Nursed Naked in Draft‑Free Surroundings of Uniform Temperature and Moderate Humidity during the Days or Weeks after Birth 35 94 93 Operative Environmental Temperature* Birthweight (kg) 34 92 91 90 35° C 34° C 33° C 32° C For 10 days After 10 days After 3 weeks After 5 weeks 1.5 For 10 days After 10 days After 4 weeks 2.0 For 2 days After 2 days After 3 weeks For 2 days After 2 days 1.0 33 32 Baby weighing 2 kg at birth 89 0 5 10 15 20 25 30 35 Age in days FIGURE 30-2 The range of temperature needed to provide neutral environmental conditions for a baby lying naked on a warm mattress in draft‑free surroundings of moderate humidity (50% saturation) when mean radiant temperature is the same as air temperature. The hatched areas show the average neutral temperature range for a healthy baby weighing 1 or 2 kg at birth. Optimal temperature probably ­approximates the lower limit of neutral range as defined here. Approximately 1° C should be added to these operative temperatures to derive the ­appropriate neutral air temperature for a single‑walled incubator when room temperature is less than 27° C (80° F), and more should be added if room temperature is significantly less. ( Adapted from Hey EN, Katz G: The optimum thermal environment for naked babies, Arch Dis Child 45:328-334, 1970.) When thermally neutral, infants regulate temperature through vasomotor tone alone without regulatory changes in metabolic heat production. A range of critical environmental temperatures relevant to modern incubators was clearly articulated by Hey and Katz (1970; Figure 30-2, Table 30-2). Below this range, an increase in the infant’s minimal metabolic rate was observed; therefore the thermal neutral temperature range was defined as the optimal incubator operating (operant) temperature. Several important considerations in regulating incubator temperature were included in these studies: (1) incubator wall temperature was maintained identical to air temperature, (2) relative humidity was controlled near 50%, and (3) the environment was maintained in a steady state, uninterrupted by turbulence or invasion of the incubator’s enclosed perimeter. Rigid application of older air temperature recommendations should be modified. Many modern incubators incorporate a double-walled design that results in lower radiant heat loss to colder incubator walls encountered in single-walled designs. As a result, slightly cooler air temperature may be required. In addition, many nurseries do not humidify incubators artificially, fearing the occurrence of condensation (“rain-out”) resulting in bacterial colonization, particularly around door openings. Finally, it is important to recognize that the incubator’s steady state temperature control is frequently interrupted for nursing and medical procedures that require opening doors to care for the infant. More than 1 hour may be required to recover prior steady state conditions after such procedures; therefore the thermal neutral zone must be redefined in practical terms. Silverman et al (1966) used a modified concept of the thermal neutral zone to simplify >2.5 Data from Hey E: Thermal neutrality, Br Med Bull 31:72, 1975. *To estimate operative temperature in a single‑walled incubator, subtract 1° C from incubator air temperature for every 7° C by which this temperature exceeds room temperature. clinical application. Reasoning that infants sense environmental temperature first on the skin, electronic negativefeedback (servo-controlled) regulation of the incubator heater in response to skin temperature was introduced. These authors demonstrated minimum metabolic expenditure near 36.5° C (97.7° F) abdominal skin temperature as measured by a shielded thermistor in a less rigidly defined incubator environment. The importance of frequently checking core temperatures (axillary or rectal) must be emphasized before delegating the infant’s environment to thermostatic control. In addition, Chessex et al (1988) have demonstrated that incubator temperature can vary by more than 2° C when skin temperature servo control rather than air temperature control is used, obviating homeothermic environmental conditions. Finally, with the modern use of open radiant warmer beds (improving access to the critically ill premature infant without interrupting heat delivery) skin temperature servo control is the only practical method for approximating the thermal neutral zone (Malin and Baumgart, 1987). The extension of infant warming to include extremely low birthweight, critically ill premature babies have generated new problems for determining a universally accepted optimal environment. PARTITIONING INFANT HEAT LOSSES AND HEAT GAINS Wheldon and Rutter (1982) demonstrated the special problems encountered in incubating infants with very low birthweight in a convection-warmed, closed-hood incubator environment (Figure 30-3). Figure 30-3, A, demonstrates the thermal balance achieved by a series of 12 infants (mean weight, 1.58 kg). Heat losses to radiation (R), convection (C), and evaporation (E) are modest, and their sum (Σ) is balanced by the infant’s metabolic heat production (M). Used in this fashion, the incubator reduces physical heat losses such that the infant’s minimal metabolism (larger than any single avenue of heat loss) 360 PART VIII Care of the High-Risk Infant A Incubator, 1.58 kg 40 Heat losses 3.0 Heat gains 2.0 1.0 20 Kcal/kg/hr Heat (watts/m2) 30 10 0 C E  C R M S M 2.0 FIGURE 30-4 Partitional calorimetry in 10 critically ill premature newborns with low birthweight who were nursed under open radiant warmers. C, Convection; E, evaporation; M, infant’s metabolic heat production; R, radiation. (Adapted from Baumgart S: Radiant heat loss vs. radiant heat gain in premature neonates under radiant warmers, Biol Neonate 57:10-20, 1990.) 20 B Incubator, 1.08 kg 40 30 Heat (watts/m2) E 1.0 R 10 20 10 0 C C R E  M 10 20 FIGURE 30-3 A, Partition of heat losses and gains in 12 premature infants nursed within incubators. B, Partition in an infant weighing 1.08 kg at birth. Heat losses to radiation (R), convection (C ), and evaporation (E ) are modest, and their sum (Σ ) is balanced by the infant’s metabolic heat production (M ). (Adapted from Wheldon AE, Rutter N: The heat balance of small babies nursed in incubators and under radiant warmers, Early Hum Dev 6:131-143, 1982.) delicately balances minimal physical heat losses within the controlled thermal environment. In contrast, a subject with very low birthweight (1.08 kg) is evaluated in Figure 30-3, B. As the incubator servo control increases warming power to accommodate massive evaporative heat loss, convective loss becomes a net gain (negative histogram bar). Radiant loss is diminished by warm walls inside the incubator. These conditions differ strikingly from those discussed previously: the incubator truly warms the infant rather than modestly attenuating convective heat loss, and evaporative heat loss vastly exceeds the infant’s metabolism. The small infant’s body temperature is balanced, therefore, between opposing physical parameters of evaporative and convective heat transfer. Metabolism plays a secondary role. The modern use of radiant warmers that are servo controlled to maintain infant abdominal skin temperature between 36.5° and 37° C (97.7° to 98.6° F) also demonstrates the opposition of physical forces described earlier. Figure 30-4 demonstrates the heat balance partition for 10 critically ill premature infants (mean weight, 1.39 kg) nursed on open radiant warmer beds (Baumgart, 1990). Because ambient room air temperature is 5° to 10° C cooler than air inside an incubator, convective heat loss is nearly double the infant’s metabolic heat production. Evaporation adds to the net physical heat loss. In addition, small amounts of heat are lost to conduction and radiation (to cooler room walls). The infant’s metabolism provides only one third of the energy required to maintain body temperature, with the majority of heat supplied by the servo-controlled radiant heat source injecting heat transdermally through skin blood flow. In this instance, radiant warming (not convection as in the incubator discussed earlier) delicately balances the infant’s physical temperature environment. Wheldon and Rutter (1982) demonstrated similar results in their studies. HYBRID INCUBATOR-RADIANT WARMER TECHNOLOGY A more promising new development in commercially available incubators is a hybrid design, combining these two separate warming modes in one device. Manufacturers in the United States and Germany (Air-Shields [Hatboro, Pennsylvania], ­Hill-Rom [Batesville, Indiana], Drager [Lubeck, Germany], and General Electric, Ohmeda [Laurel, Maryland]) have launched such products. In the incubator mode, movable plastic walls enclose the tiny premature infant, providing servo-controlled air warming, while the overhead radiant warmer is incorporated into the roof of this design but remains off. In the radiant warmer mode, the plastic walls are dropped down, and the radiant warmer rises overhead on a motorized pylon and rapidly turns on to maintain servo-controlled skin temperature during infant care procedures. The infant is not moved, and no plastic barrier is placed between the infant and the radiant heat source when turned on in this mode. The CHAPTER 30 Temperature Regulation of the Premature Neonate servo control algorithms, and the integrity of the plastic enclosure in incubator mode, are critical to the performance of such devices, and the utility of these products (e.g., survival, quality of life) is still unproved. Published data are reviewed in the following sections. VERSALET INCUWARMER Greenspan et al (2001) tested nine premature lambs, randomized at delivery, to receive incubation from a conventional radiant warming bed (Resuscitaire) with subsequent transfer into an incubator C550 Isolette or from the hybrid Versalet 7700 Care System (Air-Shields, ­Hatboro, ­Pennsylvania) in both the warmer bed and the incubator modes. Deep central and surface temperatures, heart rate, blood pressure, and blood gas determinations were measured during warming in the radiant warmer bed mode (Versalet) or on the radiant warmer bed Resuscitaire and then during transition to the incubator mode (Versalet or Isolette), and then back to the warmer bed, and bed mode. The animals conditions all remained clinically stable throughout the entire transfer protocol on both arms of the study. Despite careful planning, loss of temperature probe data occurred when probes became unattached in the control group during transfers from one device to the other. There were no significant differences in recorded temperatures or in pH and blood gasses in either group. Compared with the standard warming techniques currently used in neonatal intensive care units (NICUs; separate warmer bed for resuscitation and stabilization with transfer as soon as possible into an incubator device), the Versalet provided similar thermal and cardiovascular stability without adverse physiologic events during transition to different modes of warming. The authors state that the contribution of this hybrid device to the ease of management and improved outcomes in humans needs to be evaluated in a clinical trial (Greenspan et al, 2001; Sherman et al, 2006). No such trial has yet been conducted, and the Versalet is currently not in production (Jay Greenspan, personal communication). GIRAFFE OMNIBED As a radiant warmer, the Giraffe OmniBed (General Electric, Ohmeda Division, Laurel, Maryland) evenly heats the infant’s mattress with a curved reflector surface designed to distribute heat to the baby and its bed surface without overwarming bedside caregivers. Babies are warmed uniformly, regardless of their position on the bed platform surface, which can be rotated 360 degrees, accommodating intravenous and ventilator tubing and attached wire leads. The same platform tilts up to 12 degrees in Trendelenburg or reverse Trendelenburg positions. Three-sided access from drop-down walls in the radiant warmer mode facilitates procedures such as diaper and bedding changes, blood sampling, starting intravenous lines, performing tracheal intubation, administering medications, creating radiographs, and conducting ultrasound examinations without interrupting warming. A stable thermal environment in either the incubator or radiant warmer mode eliminates the stress of moving premature babies, such as when performing chest tube insertion or other surgical procedures 361 performed in the NICU and outside the operating room. In one of several industry sponsored studies, Leef et al (2001) reported that infants were handled significantly less with the Giraffe OmniBed, especially when converted to incubator mode (from 6.9 handling events per hour maximum on a standard radiant warmer bed device to 1.6 times per hour on an OmniBed-closed). These authors concluded that the OmniBed is conducive to providing developmentally appropriate care—that is, medically fragile newborns are not exposed to a variety of visual, auditory, and tactile stimuli that would not occur otherwise within the mother’s womb. Consequences of such stimulation are unknown, although it seems reasonable to avoid excessive handling and inappropriate touches because of documented physiologic effects of procedural handling (Gressens et al, 2002). In a second industry-sponsored study (Gaylord et al, 2001), there were no differences found in mean skin temperature among the four tested conditions in premature neonates (R = radiant warmer configuration of OmniBed; transition R to C = convection-warmed closed OmniBed and transition C to R). Mean heart rate, respiratory rate, blood pressure, and oxygen saturation were not statistically different among the four test conditions. These authors conclude that the Giraffe OmniBed provided thermal and physiologic stability across bed states, eliminating the risk of infant mishap as a result of bed transfer. When transforming the Giraffe OmniBed from incubator to warmer bed and back, the closed-convection heat partition adapts to form a uniform open-radiant heating configuration with sequential alterations of air warming temperature, fan power, and radiant heat delivered while displaying all equipment and baby parameters in one control panel. For example, when returning to the closedconvection mode, the retracting radiant warmer pylon immediately disconnects electrical power to the warming element and opens a mechanical air vent to cool the reflector hood, avoiding overheating the infant upon descent. In closed-convection configuration, bidirectional airflow through a double wall construction provides a stably enclosed thermal environment. When either door port is opened, an air curtain minimizes infant heat loss. Light and sound levels are carefully controlled within the OmniBed to promote infant health and development (Lynam, 2003). An alarm light easily visible to caregivers remains out of infant’s view. The WhisperQuiet mode limits sound to create the most quiet and soothing environment possible. Alarm speakers are deflected to minimize any noise experienced by the baby. An in-bed scale further reduces infant handling. In addition, servo-regulated humidification is supplied within the closed-incubator condition and can be set to a determined relative humidity between 70% and 80%, which is optimal to avoid excessive insensible water loss and electrolyte disturbances often experienced by premature neonates with extremely low birthweight in the first week of life when incubated dry. One recent non–industry-sponsored report of a clinical series compared the use of initial stabilization of babies with extremely low birthweight (<1000 g) under a radiant warmer followed by conventional incubation—dry versus use of humidity control in OmniBeds. The authors demonstrated that humidification improved care by decreasing fluid intake, with more stable electrolyte 362 PART VIII Care of the High-Risk Infant balance, and growth velocity (Kim et al, 2010). The authors did not address the risk-benefit issue of humidification and infection. The Giraffe Humidifier immerses a heating element in a reservoir of sterile, distilled water. Water temperature at equilibrium ranges 52° to 58° C, which is bactericidal to most organisms thriving at temperatures of 20 to 45° C (most human pathogens). As an added safety measure against reservoir contamination, water is boiled off the immersion element as the humidified air is passed inside the infant’s compartment. Sterile humidity is created in a vapor state, with no airborne droplets. In a third industrysponsored study by Lynam and Biagotti (2002), humidified OmniBeds (in vitro, air control mode at 35° C, and humidified to 65% relative humidity) were cultured after investigator inoculation with reservoir contamination with four waterborne pathogens over a 4-week incubating period. No infant environment culture revealed growth of any pathogen. The authors concluded that there is no concern for an increased risk of infection to an infant when the reservoir is filled daily with sterile distilled water and the bed is routinely cleaned, according to their protocol. LESS CONVENTIONAL TECHNIQUES PLASTIC HOODS Plastic hoods comprise rigid body shields sometimes used as miniature incubator hoods placed over infants on open radiant beds. Hoods are usually made of plastic (1 to 3 mm thick) and can obstruct the delivery of radiant heat to the baby’s skin when placed between an infant and the radiant warming element. The plastic can form a heat sink, warming to 42° to 44° C, but disrupts the radiant warmer’s servo control mechanism and is a poor strategy. In addition, such devices might not diminish insensible water loss, especially when used without humidity. Use with ­humidity—a ­technique often referred to as swamping, which has never been validated—encourages bacterial colonization with water born pathogens. PLASTIC BLANKETS Alternatively, a thin, flexible plastic wrap or polyethelene ­plastic “blanket” was introduced in 1968 for covering premature infants under radiant warmers (Baumgart, 1984; Baumgart et al, 1981, 1982a, 1982b). The blanket prevents convective heat loss and is thin enough to transmit almost completely radiant heat from the warmer. The flexible plastic molds closely around the infant’s body (with care taken to avoid skin adhesion), conserves evaporation, and forms a microenvironment near the baby’s body. The plastic acts as a mechanical barrier to prevent convective turbulence from disrupting this microenvironment. A two-thirds reduction in insensible water (and evaporative heat) loss under radiant warmers is prevented, resulting in less servo-controlled radiant heat delivery required to maintain body temperature in even the smallest infants. By moderating heat losses, oxygen consumption is reduced by approximately 10%. This reduction exactly matches the oxygen consumption cost of using radiant warmers reported by LeBlanc (1982). Risks of plastic blankets include sticking to immature skin, causing masceration and accidental airway obstruction in patients without intubation. In rare cases, partial detachment of the skin thermistor probe can result in life-threatening hyperthermia. Diligence is required to avoid these complications. The risk for abnormal bacterial colonization has not been evaluated. A polyethylene plastic bag was evaluated by Vohra et al (1999) to promote temperature maintenance during delivery room resuscitation under radiant warmers and during subsequent transport of infants to the NICU. A significantly higher admission rectal temperature and survival was reported for infants at less than 28 weeks’ gestation. A larger, randomized trial of this technique was performed, confirming the prevention of heat loss with a polyethylene bag but not the improvement in mortality (Vohra et al, 2004). Recently another large, randomized trial was performed to compare polyethylene caps with polyethylene bags; it found that both methods were more effective than conventional treatment in improving NICU admission temperature for infants at less than 29 weeks’ gestation (Trevisanuto et al, 2010). SEMIPERMEABLE POLYURETHANE SKIN A semipermeable, semiocclusive polyurethane dressing (Tegederm [3m Corporation, Maplewood, Minnesota] or Opsite [Smith & Nephew, London, United Kingdom]) can be applied like an artificial skin to shield premature infants with very low birthweight. These breathing polyurethane plastics do not cause skin breakdown and are often used to dress central venous catheter sites. Temperature probe detachment is less likely to occur because of skin adherence. These dressings can also be tailored to avoid airway obstruction. Insensible water loss may be reduced more than 30% with these dressings, and after carefully removing them the skin’s naturally developing keratin moisture barrier may be preserved. Porat and Brodsky (1993) and Bhandari et al (2005) demonstrated that an adherent polyurethane layer over the torso and extremities of infants with very low birthweight improved fluid and electrolyte balance, reduced the occurrences of patent ductus arteriosus and intraventricular hemorrhage, and improved survival. Further study is required to demonstrate the safety of this technique. PETROLEUM OINTMENTS Another occlusive dressing for the skin of preterm infants with low birthweight during the first week of life is the application of Aquaphor (Biersdorf Co, Hamburg, ­Germany). Although early results were encouraging for preserving skin integrity and perhaps controlling excessive dehydration in dry incubators, the findings of a multicenter randomized trial were not so laudable (Edwards et al, 2004). Anecdotal reports suggest that bacterial infections may actually be more common with use of this technique. Clinicians should exercise critical judgment before adopting this practice. KANGAROO CARE Kangaroo care connotes warming by skin-to-skin contact with the mother or father, where premature infants are held naked between the axilla or the breasts, mimicking a kangaroo’s pouch. First implemented in Bogota, CHAPTER 30 Temperature Regulation of the Premature Neonate Columbia, this method was popularized for nonintubated premature infants in Scandinavian and European countries in the 1980s. Early studies suggested a significant reduction in early mortality and morbidity in premature infants weighing less than 1.5 kg and who are nursed by kangaroo care. Behavioral studies demonstrated more stable sleep patterns, less irritability at 6 months of age, and more eye contact with caregivers in infants nursed with kangaroo care. Kangaroo care has been shown to promote a thermal-neutral metabolic response and temperature stability in stably growing premature babies. Moreover during kangaroo care, infants with bronchopulmonary dysplasia have better oxygenation, and other infants show less periodic breathing and reduced apnea. In the modern nursery, kangaroo care be initiated during mechanical ventilation in uncomplicated patients. The infant should be placed between breasts with maximum skin contact and should be covered with a blanket to avoid outward convective and evaporative heat losses. After initial sessions of 30 minutes to 1 hour with careful intermittent temperature monitoring, periods up to 4 hours may be achieved successively. Kangaroo care integrates the family into the neonatal intensive care team. No adverse reports have been published, and use in many nurseries is on the rise. SPECIAL CASES EXTREMELY LOW-BIRTHWEIGHT INFANTS Case Presentation A baby boy (234⁄7 weeks’ gestation, weighing 510 g) was born limp, cyanotic, with no respiratory effort, and a heart rate less than 100 beats/min. The baby was resuscitated under a radiant warmer and dried, and a stocking cap was placed over the head. The warmer’s temperature probe recorded a low skin temperature because it did not attach well to skin. The radiant warmer was on full power to compensate for this low probe reading. After resuscitation, the baby was transported to the NICU in an incubator preheated to a temperature of 35° C. On arrival, the baby was weighed quickly and placed onto a preheated radiant warmer bed. Skin probe (servo control) temperature registered 33° C, with a rectal temperature of 35° C and axillary temperature of 34° C. The radiant warmer set point was targeted at 37° C, and temperatures were monitored every 15 minutes. A plastic blanket was used to minimize heat loss. Heated and 100% humidified air from a respiratory pack was run underneath the blanket. next, the baby was prepared with iodine solution and draped completely for umbilical catheterization (a 30-minute procedure). Over the first 2 hours of life, rectal temperature increased only to 34.5° C. The baby was transferred into a preheated and humidified incubator regulated by a skin probe set at 37° C. Two light bulbs (250 W) were positioned over the incubator hood for supplemental heat and for warming during any procedure requiring opening the incubator (e.g., serial weight determinations). Over the next 12 hours the baby had difficulty maintaining temperature despite incubator air temperature running near a set temperature of 37° C, and sometimes greater than 38.5° C. 363 Discussion Temperature maintenance of an extremely premature infant should be part of resuscitation from the time of delivery. Despite radiant warming in the delivery room and convective incubation during transport, this baby had hypothermia on admission to the NICU. Low admission temperature correlates with increased mortality rates in these infants. They are born wet and prone to excessive transepidermal evaporative and convective heat losses. In addition, heat loss can be exacerbated by suboptimal radiant heating; blowing noncontrolled warm, humidified air under plastic blankets, which is probably not as effective as the still air envelope conserved by the blanket; and the pressure for performing procedures with surgical drapes that block radiant heat delivery. Quick drying, proper placement directly under the radiant heater at birth, and covering the head are small but important steps of temperature resuscitation. Other techniques might be considered from birth, such a use of a plastic bag described by Vohra et al (1999) or a plastic blanket or polyurethane drape during umbilical catheterization. Finally, transferring infants into incubators before adequately rewarming them under a radiant warmer can prolong thermal recovery in hypoxic subjects who are incapable of generating enough metabolic heat to recover. The use of light bulbs to provide supplemental radiant heat in incubators attests to the incubator’s intrinsic operating power deficiency (i.e., seeking to reduce heat loss, rather than rewarm cold babies). Moreover, such supplemental radiant heat sources are not controlled, produce more radiation in the visible light range, and are inefficient and potentially dangerous for promoting burns if placed too close to the infant’s skin. There are few data describing the best rate of rewarming. Cold-stressed babies should probably be nursed under a servo-controlled radiant heat source, and heat loss should be minimized. During rewarming the infants should be closely monitored, and heat delivery should be servo controlled. A rate of approximately 0.5° to 1° C per hour seems reasonable, but it might not be achievable. Profound cardiovascular and electrolyte disturbances characterize rewarming. Alternatively, a preheated and humidified incubator can be used to rewarm infants, although the rate may be slower. The incubator air or skin servo control temperature can be set 1° C higher than the baby’s temperature and gradually increased until a normal core temperature is achieved. Hybrid incubators that can also be used as radiant warmer beds are available and are helpful in rewarming scenarios. NEONATAL FEVER Problem of Temperature Elevation There is no universally accepted definition of neonatal fever. Craig (1963) defined neonatal pyrexia as a rectal (core body) temperature greater than 37.4° C; however, other investigators accept temperatures up to 37.8°C as normal. Between 1% and 2.5% of all newborns admitted to the nursery develop fever, judged by rectal or axillary temperatures and depending on the limits chosen. Fever is an inconsistent and infrequent sign of sepsis (fewer than 10% 364 PART VIII Care of the High-Risk Infant of febrile neonates have culture-proven sepsis), and temperature elevation may be seen with several other clinical entities. Mechanisms Producing Neonatal Fever Mechanisms producing neonatal fever are understood incompletely and result from disturbances in the complex interactions between heat conservation and heat dissipation mechanisms. Fever can occur when immunogenic pyrogens (commonly prostagladin [PG E2]) leads to upward displacement of the normal thermal set-point in the hypothalamus, leading to activation of heat conservation and physiologic heat-generating responses. Generally heat conservation starts with peripheral vasoconstriction and is followed by thermogenesis (generally nonshivering in neonates) while the new set-point is achieved. It is important to note that newborn infants of different animal species react in peculiar ways to different known pyrogens. Human newborns can have severe bacterial infections without increased body temperature. In addition to pyogenic mechanisms of febrile response in newborns, other phenomena can lead to elevated of body temperature. Newborn infants have poor heat dissipation mechanisms (absence of sweating); therefore exposure to excess heat or insulation (e.g., excessive swaddling) can quickly increase their core temperature. Such overheating commonly occurs when term babies are nursed in uncontrolled incubators or under radiant warmers. Temperature elevation can also occur with increased infant metabolic rate such as that seen with skeletal muscle rigidity and status epilepticus. Another cause of temperature elevation is occasionally observed in healthy, breast-feeding newborn infants on the third to fourth day of life, and it is believed to result from dehydration caused by insufficient milk production. Finally, there are more recent reports of an increased incidence of neonatal fever in infants of mothers receiving epidural analgesia. The mechanisms for temperature elevation in these latter instances are unknown. Determining the Cause of Fever Sepsis is an uncommon cause of fever. Paradoxically, neonates with sepsis more frequently have hypothermia as well. However, sepsis is probably the most treatable lifethreatening illness occurring in febrile newborn infants, especially those with temperature elevations to greater than 38° to 39° C, who are more likely to have bacteremia, purulant menigitis, and pnemonia. Most neonatal febrile episodes are noted in first day of life (54% in one series); however, any fever occurring on the third day of life and body temperature greater than 39° C have both been correlated with a significantly higher chance of bacterial disease. Severe temperature elevation is also associated with viral disease, particularly herpes simplex encephalitis; therefore work-ups for sepsis in these infants should include lumbar puncture. Hyperthermia has been reported in tiny premature infants as a complication of improper use of shielding devices under either convection-warmed incubator or radiant warmer conditions. When incubated, babies should always have a strictly monitored and controlled source of heat. Fever secondary to overheating, particularly associated with incubators, is more common in equatorial and tropical countries. Dehydration is an infrequently recognized cause of fever in the newborn period. Dehydration occurring in healthy term infants between the third and fourth day of life was noted previously, and it is probably the result of inadequate milk intake. Dehydration fever is commonly seen in large breastfed babies whose milk intake is poor and who may be exposed to high environmental temperatures during the summertime or in tropical areas. Body temperature can range between 37.8° and 40° C. Rehydration leads to resolution of fever and is key to the diagnosis of dehydration fever. In two reports (Lieberman et al, 1997; Pleasure and Stahl, 1990), fever was more common in neonates born to mothers receiving epidural analgesia during labor when compared with those without analgesia (7.5% versus 2.5% and 14.5% versus 1%, respectively). One of these reports (Lieberman et al, 1977) observed more frequent sepsis evaluations and antibiotic use in the offspring of women receiving epidural analgesia. With the increasing use of epidural analgesia during labor, recognizing epidural neonatal fever is an important consideration when evaluating a febrile neonate. Unusual and uncommon causes of neonatal fever include neonatal typhoid fever and congenital malaria, which should be considered in immigrant populations or in third-world countries. An increase in unexplained neonatal fevers was associated with the introduction of routine hepatitis B vaccination versus historical controls (Lewis et al, 2001), but this was not confirmed in a subsequent, large prospective clinical study (Lewis et al, 2001). In addition, temperature elevations may be seen with hypothalamic or other central nervous system malformations or masses. Subarachnoid or other intracranial hemorrhages may also be associated with temperature elevation. On rare occasions, neonatal spinal neurenteric cyst can manifest with long-lasting neonatal fever and should be considered in the differential diagnosis of acute myelopathy with persistent fever in infancy (greater than 3 weeks’ duration). The presence of myelopathy will help to establish this diagnosis. Management The clinical problem is that fever may be the only indication of severe bacterial disease. The relevant perinatal history should be evaluated for risk factors mitigating a laboratory evaluation, presumptive treatment for infection, or both. Furthermore, signs that are suggestive of sepsis (e.g., diminished activity, irritability, seizures) should be considered. All neonates with fever should be evaluated for hydration, weight loss, and foci of infection (i.e., cellulitis, septic arthritis or osteomyelitis, omphalitis, and the presence of colonized foreign bodies, such as a central venous line). However, febrile neonates without clinical history or any signs of infection present a challenge with insufficient data in the literature regarding appropriate management. An infant’s environment should be examined for overheating. In breastfeeding, infants with fever at 3 to 4 days of age and excessive weight loss, dehydration fever should be considered and treated to establish this diagnosis. Mothers CHAPTER 30 Temperature Regulation of the Premature Neonate receiving epidural analgesia often manifest shivering with their temperature rise, and they experience a rapid defervescence after discontinuing the epidural infusion. Recognition of this pattern may avoid unnecessary sepsis evaluations in neonates with early fever. THERAPEUTIC HYPOTHERMIA FOR BRAIN PROTECTION FOLLOWING NEONATAL ASPHYXIA Since 2005, therapeutic hypothermia has been considered the standard of care for a highly selected population of near-term and term infants suffering hypoxic-ischemic events after birth, or after medically witnessed cardiopulmonary arrest and resuscitation. Two randomized studies suggest that cerebral cooling with either whole body cooling to a core temperature of 33.5° C (considered moderate hypothermia) or selective head cooling to approxmately 10° C (with mild whole-body cooling to 34° C) reduces the risk for death or moderate to severe neurologic injury from approxiately two thirds, to less than half (Gluckman et al, 2005; Shankaran et al, 2005). In the United States, the National Institutes of Health Institute of Child Health and Human Development Experts Panel Workshop held in May 2005 emphasized using standardized protocols adapted from these randomized trials for hypothermia treatment, and it recommended continual follow-up until school age to develop and better refine therapy for treating moderate to severe clinical neonatal encephalopathy observed in term or near-term neonates (Higgins et al, 2006). The Children’s National Medical Center (CNMC) has adapted the Neonatal Network’s whole-body hypothermia protocol (Shankaran et al, 2005). In addition, CNMC provides continuous electroencephalogram (EEG) neurologic monitoring as part of a protocol, without consideration of the EEG as a criteria for initiating cooling. A full montage video-EEG (using a modified International 10-20 system accepted for neonates) is recorded by computer for 120 hours after birth to include cooling and rewarming recovery. EEG data are displayed on a centralized monitoring station and reviewed at least daily for every patient treated for hypothermia. Amplitude-integrated EEG is also reviewed for characterization of background pattern and for detection of seizures that can contribute to evolving brain injury and are treated aggressively. Automated seizure detection software is confirmed on raw EEG signals by an electrophysiology neonatal neurologist who is integrated into the program. To receive cooling therapy, infants must be at 34 weeks’ gestation or greater, arrive for treatment within 6 hours of birth, and have experienced a hypoxic-ischemic event. Evidence of hypoxic-ischemic injury includes: resuscitation being required at delivery, an umbilical vessel blood gas pH of 7.00 or less, or having a significant base deficit of at least −16. Also qualifying for therapy are blood gas disturbances within the first hour of life, with a pH of 7.01 to 7.15 and a base deficit of −10 to −15.9, along with an ominous perinatal history (e.g., fetal heart rate decelerations, umbilical cord prolapse or rupture, uterine rupture, severe maternal trauma preceding birth, abruption of the placenta, maternal life-threatening event requiring cardiopulmonary resuscitation). Signs of 365 moderate to severe neonatal encephalopathy must also be present to qualify for hypothermia intervention: lethargy, complete stupor, diminished or completely absent spontaneous activity and/or aberrant muscle tone, weak or absent sucking and Moro reflexes with pupils fixed and constricted or unresponsive and dilated, fixed flexion or extension posturing of extremities, or a clinically observed seizure. Infants with such symptoms persisting for several days have an approximately 60% risk for death before hospital discharge, whereas a majority of survivors experience moderate to severe life-long neurodevelopmental disabilities (e.g., cerebral palsy, deafness, blindness, mental retardation, or recurrent seizures as in epilepsy). Infants meeting these criteria are quickly examined at admission and then placed supine onto a water-filled cooling blanket that is precooled to 5° C (41° F). CNMC uses a Blanketrol II device (Cincinnati Sub-Zero, Cincinnati, Ohio) that is used commonly in the emergency department to reduce high fevers and in the operating room to promote hypothermia before and during reconstructive heart surgery. An esophageal temperature probe is placed into the distal third of the esophagus to monitor the infant’s core temperature, and the thermostatic controller in the water mattress’ cooling unit is set to 33.5° C. A second, larger pediatric-size blanket is also attached in parallel to the cooling system and is suspended at the bedside as a “sail” heat capacitor. Water circulates through both blankets exposed to both the baby and the room air temperature to diminish continuously monitored temperature fluctuations in esophageal temperature (less than ±0.5° C). Although CNMC uses a warmer bed platform as a crib, the overhead warmer is not turned on during the cooling period. Abdominal wall skin temperature is monitored with a surface probe that is available with the warmer bed (in monitor-only mode). Temperatures of the esophagus, skin, and axilla are thus monitored and recorded every 15 minutes for the first 4 hours of cooling, every hour for the next 8 hours, and every 4 hours during the remaining 72-hour period of hypothermia. CNMC’s electronic medical record (Cerner) provides hourly esophageal temperature recording. After 72 hours, the set-point of the controller on the cooling system is increased by 0.5° C increments per hour to promote gradual rewarming. The Neonatal Network reported febrile rebound and seizures during rewarming (Shankaran et al, 2005); however, CNMC has not experienced this with continuous EEG monitoring over 24 to 48 hours throughout rewarming and a subsequent recovery period. After 6 hours, the esophageal probe and cooling blankets are removed, and anterior abdominal wall skin temperature is then regulated using the radiant warmer’s servomechanism set at 36° to 36.5° C (i.e., warmer is turned on). The purpose of rewarming slowly is to avoid rapid shifts in electrolytes (calcium and potassium in particular), cardiac arrhythmias, and in rewarming hyperthermia, because fever promotes further brain injury. Infants otherwise receive routine neonatal intensive care, with continuous monitoring of vital signs. Mild sinus bradycardia (80 to 90 beats/min) and small decreases in mean arterial blood pressure (less than 20 mm Hg) are commonly observed and treated easily with volume and pressor infusions, according to CNMC’s standard institutional guidelines for babies developmentally. 366 PART VIII Care of the High-Risk Infant Frequent blood samples are monitored for glucose regulation (infants are restricted to 4 to 6 mg/kg/min of dextrose infusion), coagulopathy (in particular treating fibrinogen levels less than 150 mg/dL and platelet levels less than 80,000 per millimeter), and major organ failure (e.g., electrolytes, blood urea nitrogen, creatinine, liver enzymes). CNMC’s clinical experience over the past 3 years in more than 98 babies meeting strict criteria for therapeutic hypothermia has been commensurate with that reported from the Neonatal Network (Shankaran et al, 2005). Target esophageal temperature was achieved at 33.5 ± 0.5° C (range 33° to 34° C) in 30 to 60 minutes without a major circulatory mishap at CNMC. Continuous EEG monitoring has been instructive for intervening seizure activity (observed on EEG in approximately one quarter of infants shortly after admission). CNMC electrophysiologists observed general improvements in background voltages and patterns on EEG recordings during and after hypothermia. Specifically, improvement of background activity, appearance of sleep wake cycling, and disappearance of seizures has been observed at the time of rewarming (El-Dib et al, 2007). The Neonatal Network observed seizures emerging frequently during rewarming, hence the precaution in re-warming more slowly. CNMC is presently acquiring and reviewing infant neurologic and neuron-developmental follow-up data on a 6-month schedule through the first 2 years of life. CNMC has also made the clinical observation that the water mattress felt warm to the touch (i.e., warmer than the examiner’s hand) during most of the cooling period (Baumgart et al, 2007). Median ambient temperature in CNMC’s NICU during October 2007 (23.1° C; range, 20.9° to 25.4° C) was usually less than both the blanket water and baby temperatures. No infant had acidemia during cooling. Temperature gradients suggest that whole-body cooling is achieved through surface cooling from skin exposed to the ambient environment, and not actually by heat loss into the water-blanket. CNMC observers believe that the water matress simply incubates infants at a lower temperature. Except during the first 30 minutes, the blanket more often provided warmth to maintain esophageal temperature at 33.5° C. Whole-body cooling might also be provided by regulating an incubator air temperature or a radiant heater to maintain esophageal temperature. It has been important to CNMC’s outreach education program to emphasize that referring (i.e., resuscitating) institutions should perform cardio pulmonary resuscitation under warm conditions to facilitate cardiac response according to Neonatal Resuscitation Program guidelines. Cold resuscitation is not advocated until after careful clinical evaluation of a stabilized patient has been performed at CNMC to meet therapeutic cooling criteria. SUMMARY The premature newborn with very low birthweight is extremely vulnerable to harsh fluctuations in physical environment. These infants require frequent assessments of skin, core, and air temperature and relative humidity to design an optimal strategy for thermal regulation. In caring for smaller babies, heat replacement is often required, and refinement of techniques to accomplish replacement without inducing hyperthermia is needed. The use of therapeutic hypothermia in highly selective cases at risk for brain injury is no longer considered novel. SUGGESTED READINGS Baumgart S: Partitioning of heat losses and gains in premature newborn infants under radiant warmers, Pediatrics 75:89-99, 1985. Bell EF, Rios GR: Air versus skin temperature servo control of infant incubators, J Pediatr 103:954, 1983. Bell EF, Rios GR: A double‑walled incubator alters the partition of body heat loss of premature infants, Pediatr Res 17:135-140, 1983. Bell EF, Weinstein MR, Oh W: Heat balance in premature infants: Comparative effects of convectively heated incubator and radiant warmer, with and without plastic heat shield, J Pediatr 96:460-465, 1980. Bruck K: Heat production and temperature regulation. In Stave U, editor: Perinatal physiology, New York, 1978, Plenum Publishing, pp 455-498. Dawkins MJ, Hull D: The production of heat by fat, Sci Am 213:62-67, 1965. Day RL, Caliguiri L, Kaminski C, et al: Body temperature and survival of premature infants, Pediatrics 34:171-181, 1964. Knauth A, Gordin M, McNelis W, et al: Semipermeable polyurethane membrane as an artificial skin for the premature neonate, Pediatrics 83:945-950, 1988. Marks KH, Lee CA, Bolan CD, et al: Oxygen consumption and temperature control of premature infants in a double‑wall incubator, Pediatrics 68:93-98, 1981. Mayfield SR, Bhatia J, Nakamura KT, et al: Temperature measurement in term and preterm neonates, J Pediatr 104:271-275, 1984. Okken A, Blijham C, Franz W, et al: Effects of forced convection of heated air on insensible water loss and heat loss in preterm infants in incubators, J Pediatr 101:108-112, 1982. Scopes JW: Thermoregulation in the newborn. In Avery CB, editor: Neonatology, pathophysiology and management of the newborn, ed 2, Philadelphia, 1981, JB Lippincott, pp 171-181. Yager JY, Armstrong EA, Jaharus C, et al: Preventing hyperthermia decreases brain damage following neonatal hypoxic-ischemic seizures, Brain Res 1011:48-57, 2004. Complete references used in this text can be found online at www.expertconsult.com C H A P T E R 31 Acid-Base, Fluid, and Electrolyte Management Michael A. Posencheg and Jacquelyn R. Evans FLUID AND ELECTROLYTE BALANCE Maintenance of fluid and electrolyte balance is essential for normal cell and organ function during intrauterine development and throughout extrauterine life. Pathologic conditions in the newborn often lead to disruption of the complex regulatory mechanisms of fluid and electrolyte homeostasis. Therefore a thorough understanding of the physiologic changes in neonatal fluid and electrolyte homeostasis and the provision of appropriate fluid and electrolyte therapy based on the principles of developmental fluid and electrolyte physiology are among the cornerstones of modern neonatal intensive care. DEVELOPMENTAL CHANGES AFFECTING FLUID AND ELECTROLYTE BALANCE IN THE FETUS AND NEONATE Developmental Changes in Body Composition and Fluid Compartments Dynamic changes occur in body composition and fluid distribution during intrauterine life, labor and delivery, and the early postnatal period. Thereafter the rate of change in body composition and fluid distribution gradually decreases, with more subtle changes taking place especially after the first year of life (Friis-Hansen, 1961). Changes during Intrauterine Development In early gestation, body composition is characterized by a high proportion of total body water (TBW) and a large extracellular compartment (Brans, 1986; Friis-Hansen, 1983). As gestation advances, rapid cellular growth, accretion of body solids, and fat deposition result in gradual reductions in TBW content and extracellular fluid volume while the intracellular fluid compartment increases (Figure 31-1) (Friis-Hansen, 1983). In the 16-week-old fetus, TBW represents approximately 94% of total body weight, and approximately two thirds of the TBW is distributed in the extracellular and one third in the intracellular compartment. By term, TBW contributes only 75% of body weight, and almost half of this volume is located in the intracellular compartment. Therefore infants born prematurely have TBW excess and extracellular volume expansion compared with their term counterparts, with the majority of the expanded extracellular volume being distributed in the interstitium (Brace, 1992). Changes during Labor and Delivery Additional and more acute changes in TBW and its distribution occur during labor and delivery. Arterial blood pressure rises several days before delivery in response to increases in catecholamine, vasopressin, and cortisol plasma concentrations and translocation of blood from the placenta into the fetus. This rise in arterial blood pressure, along with changes in the fetal hormonal milieu and an intrapartum hypoxia-induced increase in capillary permeability, results in a shift of fluid from the intravascular to the interstitial compartment. This fluid shift results in an approximately 25% reduction in circulating plasma volume in the human fetus during labor and delivery (Brace, 1992). The postnatal increase in oxygenation and changes in vasoactive hormone production then restore capillary membrane integrity and favor absorption of interstitial fluid into the intravascular compartment. The ensuing gradual movement of fluid from the expanded interstitial space into the bloodstream aids in maintaining intravascular volume during the first 24 to 48 hours postnatally, when oral fluid intake may be limited. However, prematurity, pathologic conditions, or both can disrupt this delicate process and interfere with the physiologic contraction of the extracellular fluid (ECF) compartment in the immediate postnatal period. In the fetus, body composition and fluid balance depend on the electrolyte and water exchange between mother, fetus, and amniotic space (Brace, 1986). Antenatal events can have significant effects on postnatal fluid balance. Maternal indomethacin treatment or excessive administration of intravenous fluids during labor can result in neonatal hyponatremia with expanded extracellular water content (Rojas et al, 1984; vd Heijden et al, 1988). Placental insufficiency or maternal diuretic therapy can impair fetal hydration, leading to decreases in extracellular volume, urine output, and amniotic fluid volume (Van Otterlo et al, 1977). The timing of cord clamping after delivery is another important factor significantly affecting total circulating blood volume and extracellular volume in the neonate. Immediate cord clamping does not allow for placental transfusion, but if the cord clamping is delayed only 3 to 4 minutes after delivery with the newborn positioned at or below the level of the placenta, up to 25 to 50 mL/kg of blood is transfused into the neonate, representing an approximately 25% to 50% increase in the total blood volume (Linderkamp, 1982; Yao and Lind, 1974). The onset of infant respiratory effort after birth also increases placental transfusion regardless of the infant’s position (Philip and Teng, 1977). Changes in the Postnatal Period In the first few days and weeks after birth, the most important effects on the pace of further changes in body composition, TBW content, and its distribution are exerted by gestational and postnatal ages, the presence or absence of pathologic conditions, the immediate environment, and the type of nutrition. In normal conditions in the first few 367 368 PART VIII Care of the High-Risk Infant FIGURE 31-1 Total body water (TBW) content and its distribution between the extracellular fluid (ECF) and intracellular fluid (ICF) compartments in the human fetus, newborn, and infant from conception until 9 months of age. (Data represent average values from Friis-Hansen B: Body water compartments in children: changes during growth and related changes in body composition, Pediatrics 28:169-181, 1961.) days after birth, the postnatal increase in capillary membrane integrity favors absorption of the interstitial fluid into the intravascular compartment. The ensuing rise in circulating blood volume stimulates the release of atrial natriuretic peptide from the heart, which in turn enhances renal sodium and water excretion (Sagnella and MacGregor, 1984) resulting in an abrupt decrease in TBW and attendant weight loss. Although it is generally accepted that this postnatal weight loss is primarily due to the contraction of the expanded ECF compartment (Cheek et al, 1961), some water loss from the intracellular compartment can also occur, particularly in infants with extremely low birthweight (ELBW) and increased transepidermal water losses (Costarino and Baumgart, 1991; Sedin, 1995). Healthy term newborns lose an average of 5% to 10% of their birthweight during the first 4 to 7 days of life (Brace, 1992); thereafter they establish a pattern of steady weight gain. Because preterm infants have an increased TBW content and extracellular volume, they lose an average of 15% of their birthweight during transition (Shaffer and Meade, 1989) and, depending on the degree of prematurity and associated pathologic conditions, these neonates only regain their birthweight by 10 to 20 days after birth (Figure 31-2). Physiology of the Regulation of Body Composition and Fluid Compartments Although human cells have the ability to adjust their intracellular composition, ultimate regulation of the intracellular volume and osmolality relies on the control of the extracellular compartment. Therefore the human body must be able to monitor the volume and osmolality of the extracellular compartment and to correct the changes resulting from its interaction with the environment. Regulation of the Intracellular Solute and Water Compartment The major intracellular solutes are the cellular proteins necessary for cell function, the organic phosphates associated with cellular energy production and storage, and FIGURE 31-2 Postnatal changes in body weight (expressed in percent of birthweight), extracellular fluid volume (estimated by the bromide dilution method), and sodium balance (defined as the difference between sodium intake and urinary sodium excretion). (From Shaffer SG, Weismann DN: Fluid requirements in the preterm infant, Clin Perinatol 19:233-250, 1992.) the equivalent cations balancing the phosphate and protein anions (MacKnight and Leaf, 1977). Potassium is the major intracellular cation, and sodium is the major extracellular cation. The energy derived from the concentration differences for sodium and potassium between the intracellular and extracellular compartments is used for cellular work. Because changes in osmolality of the extracellular compartment are reflected as net movements of water in or out of the cell, regulation of ECF concentration ultimately controls the osmolality and size of the intracellular compartment (MacKnight and Leaf, 1977). This physiologic principle must be kept in mind by the neonatologist managing sick term and preterm neonates with disturbances of sodium homeostasis. Rapid changes in serum sodium concentration and thus in extracellular osmolality directly CHAPTER 31 Acid-Base, Fluid, and Electrolyte Management 369 affect the osmolality and size of the intracellular compartment and can lead to irreversible cell damage, especially in the central nervous system. Regulation of the Intracellular-Extracellular Interface: The Interstitial Compartment In the healthy term neonate, hydrostatic and oncotic pressures are well balanced, with both being approximately half those in the adult (Sola and Gregory, 1981). In normal physiologic conditions, movement of fluid across the capillary is determined by the direction of the net driving pressure ([PC − PT] − [πP − πT]) and the water and protein permeability characteristics of the capillary wall (Figure 31-3). At the arterial end of the capillary, intracapillary hydrostatic pressure (PC) is high, and plasma oncotic pressure (πP) is relatively low, resulting in a net movement of fluid out of the capillary. As filtration of relatively protein-poor fluid continues along the capillary, plasma oncotic pressure rises and intracapillary hydrostatic pressure drops; therefore on the venous side, fluid moves from the interstitium into the capillary, so that much of the filtered fluid is reabsorbed at the end of the capillary bed. The fluid remaining in the interstitium (arterial-venous side of the capillary) is drained by the lymphatic system. Interstitial hydrostatic (PT) and oncotic (πT) pressures remain virtually unchanged along the capillary bed. However, pathologic conditions readily disturb the delicate balance between the hydrostatic and oncotic forces, leading to an expansion of the interstitial compartment at the expense of the intravascular compartment. The increased interstitial fluid volume (edema) then further affects tissue perfusion by altering the normal function of the extracellular-intracellular interface. Box 31-1 summarizes the mechanisms for conditions resulting in interstitial edema formation in the neonate. There are also some important developmentally regulated differences between the newborn and adult relating to the pathogenesis of edema formation. Capillary permeability to proteins is increased during the early stages of development (Brace, 1992; Gold and Brace, 1988). Because neonatal capillary permeability is further increased under pathologic conditions (see Box 31-1), protein concentration in the interstitial compartment may approach that of the intravascular space, favoring further intravascular volume depletion and interstitial volume expansion. When sick neonates are treated with frequent albumin boluses, much of the infused albumin leaks into the interstitium, creating a vicious cycle of intravascular volume depletion and edema formation. If the cycle is not interrupted, it can result in the formation of anasarca, which has an extremely poor prognosis. In summary, the sick neonate has a limited capacity to maintain appropriate intravascular volume and to regulate the volume and composition of the interstitium. The ensuing intravascular hypovolemia and edema formation result in vasoconstriction and disturbances in tissue perfusion and cellular function with further impairments in the regulation of extracellular volume distribution. Regulation of the Extracellular Solute and Water Compartment The regulation of the volume and osmolality of the extracellular compartment ensures the integrity of the circulation and maintains the osmolality of the extracellular FIGURE 31-3 Filtration and reabsorption of fluid along the capillary under physiologic conditions. BOX 31-1 M  echanisms for Conditions Causing Interstitial Edema in the Neonate CONDTIONS FAVORING FLUID ACCUMULATION IN THE INTERSTITIAL SPACE BY CAUSING A DYSEQUILIBRIUM BETWEEN FILTRATION AND REABSORPTION OF FLUID BY THE CAPILLARIES Increased hydrostatic pressure Elevated capillary hydrostatic pressure Increased cardiac output Venous obstruction Decreased tissue hydrostatic pressure Conditions associated with changes in the properties of the interstitial gel (edematous states, effects of hormones including prolactin) Decreased oncotic pressure gradient Decreased capillary oncotic pressure Prematurity, hyaline membrane disease Malnutrition, liver dysfunction Nephrotic syndrome Increased interstitial oncotic pressure is usually the result of increased capillary permeability Elevation of the filtration coefficient Increased capillary permeability Organs with large-pore capillary endothelium (liver, spleen) State of maturity (preterm infants > term newborns > adults) Production of pro-inflammatory cytokines (sepsis, anaphylaxis, hypoxic tissue injury, tissue ischemia, ischemia-reperfusion, soft tissue trauma, extracorporeal membrane oxygenation) Increased capillary surface area Vasodilation CONDITIONS ASSOCIATED WITH DECREASED LYMPHATIC DRAINAGE Decreased muscle movement Neuromuscular blockade and/or heavy sedation Central and/or peripheral nervous system disease Obstruction of lymphatic flow Increased central venous pressure Scar tissue formation (bronchopulmonary dysplasia) Mechanical obstruction (dressings, high mean airway pressure in mechanically ventilated newborns) Modified from Costarino AT, Baumgard S: Neonatal water metabolism. In Cowett RM, editor: Principles of perinatal/neonatal metabolism, New York, 1991, Springer Verlag, pp 623-649. 370 PART VIII Care of the High-Risk Infant compartment within 2% of the osmolar set point between 275 and 290 mOsm (Robertson and Berl, 1986). Blood pressure and serum sodium concentration (i.e., osmolality) are monitored by baroreceptors and osmoreceptors, respectively. The effector limb of the regulatory system consists of the heart, vascular bed, kidneys, and intake of fluid in response to thirst. The latter part of the effector system is inactive in critically ill term and preterm neonates whose fluid intake is completely controlled by caregivers. By regulating the function of the effector organs, several hormones have a role in the control of the extracellular compartment, including the renin-angiotensin-aldosterone system, vasopressin, atrial natriuretic peptide, brain (B-type) natriuretic peptide, bradykinin, prostaglandins, and catecholamines. Because volume changes in the extracellular compartment must be reflected by similar changes in the intravascular volume for effective operation of the regulatory mechanisms, regulation of the extracellular compartment relies on intact cardiovascular function and on the integrity of the capillary endothelium (Robertson and Berl, 1986). For example, under physiologic conditions, an increase in the extracellular volume is reflected by an increase in the circulating plasma volume, leading to rises in blood pressure and renal blood flow. The ensuing increase in glomerular filtration and urine output returns the extracellular volume to normal. In critically ill neonates, however, the capillary leak and reduced myocardial responsiveness resulting from immaturity and underlying pathologic conditions limit the increase in the circulating blood volume when extracellular volume expands. Thus, especially in sick preterm infants, blood pressure may rise only transiently, and renal blood flow may remain low after volume boluses as fluid rapidly leaks into the interstitium. Inappropriate central regulation of vascular tone results in vasodilatation, further decreasing effective circulating blood volume and compromising tissue perfusion; this leads to impaired gas exchange in the lungs, resulting in hypoxia with further increases in capillary leak. Unless interrupted by appropriate therapeutic measures, a vicious cycle with further deterioration readily occurs in the sick neonate. Maturation of Organs Regulating Body Composition and Fluid Compartments The heart, kidneys, skin, and endocrine system play the most important roles in the regulation of extracellular (and thus intracellular) fluid and electrolyte balance in the neonate. Immaturity of these organ systems, especially in infants with very low birthweight (VLBW), results in a compromised regulatory capacity, which must be noted when estimating daily fluid and electrolyte requirements in these patients. Maturation of the Cardiovascular System There is a direct relationship between gestational maturity and the ability of the neonatal heart to respond to acute volume loading (Baylen et al, 1986). The blunted Starling response of the immature myocardium results from its lower content of contractile elements and incomplete sympathetic innervations (Mahony, 1995). Because central vasoregulation and endothelial integrity are also developmentally regulated (Brace, 1992; Gold and Brace, 1988), an appropriate effective intravascular volume is seldom maintained in a critically ill preterm infant. Since regulation of the extracellular volume requires the maintenance of an adequate effective circulating blood volume, the immaturity of the cardiovascular system contributes to the limited capacity of sick preterm infants to effectively regulate the total volume of their extracellular compartment. Maturation of Renal Function The kidney has a crucial role in the physiologic control of fluid and electrolyte balance. It regulates extracellular volume and osmolality through the selective reabsorption of sodium and water, respectively. Immaturity of renal function renders preterm infants susceptible to both excessive sodium and bicarbonate losses (El-Dahr and Chevalier, 1990). In addition, the inability of the preterm infant to respond promptly to a sodium or volume load results in a tendency toward extracellular volume expansion with edema formation. Because prenatal steroid administration accelerates maturation of renal function (van den Anker et al, 1994), preterm infants treated with steroids in utero have a better capacity to regulate their postnatal ECF contractions. During the first few weeks of life, hemodynamically stable but extremely immature infants produce dilute urine and may develop polyuria because of their renal tubular immaturity. As tubular functions mature, their concentrating capacity gradually improves from the 2nd to 4th week of life. However, it takes years for the developing kidney to reach the concentrating capacity of the adult kidney. Maturation of the Skin Although term infants have a well-developed cornified layer of the epidermis, extremely immature neonates have only two or three cell layers in the epidermis (CartridgePatrick and Rutter, 1992). Because of the lack of an effective barrier to diffusion of water through the immature skin, transepidermal free water losses in the immature infant may be extremely high during the first few days postnatally. Gestational age, postnatal age, the pattern of intrauterine growth, and environmental factors have crucial roles in the magnitude of transepidermal free water losses (Figure 31-4). Although skin cornification rapidly increases, even in the extremely immature infant during the first few days after birth, full maturation of the epidermis does not occur for more than 28 days of age (Sedin, 1995). Chronic intrauterine stress (Hammarlund et al, 1983) and prenatal steroid treatment (Aszterbaum et al, 1993) also enhance maturation of the skin. Increases in free water losses through the immature skin of the VLBW infant can result in early postnatal hypertonic dehydration with rapid changes in intracellular volume and osmolality. In many organs, especially the brain, these abrupt changes in intracellular volume and osmolality can lead to cellular dysfunction and ultimately cell death. Maturation of End-Organ Responsiveness to Hormones Involved in the Regulation of Fluid and Electrolyte Balance Several hormones, including but not limited to the reninangiotensin-aldosterone system, vasopressin, atrial natriuretic peptide, and brain (B-type) natriuretic peptide, directly regulate the volume or composition of the extracellular compartment. These hormones exert their effects Transepidermal Water Loss (g/m2/h) CHAPTER 31 Acid-Base, Fluid, and Electrolyte Management 60 50 40 3 30 5 20 14 10 0 28 21 7 s) ay (d lA a at n st Po ge 0 1 26 28 30 32 34 36 38 40 Gestational Age (weeks) FIGURE 31-4 Transepidermal water loss in relation to gestational age during the first 28 postnatal days in infants who are appropriate for gestational age. There is an exponential relationship between trans­epidermal water loss and gestational age, the water loss being higher in preterm infants than in term infants. Transepidermal water loss is also significantly affected by postnatal age, especially in the immature preterm infant. The measurements were performed at an ambient air humidity of 50% and with the infants calm and quiet. (From Hammarlund K, Sedin G, Stromberg B: Transepidermal water loss in newborn infants. VIII: Relation to gestational age and postnatal age in appropriate and small for gestational age infants, Acta Paediatr Scand 72:721-728, 1983.) mainly by altering renal sodium and water excretion and by inducing changes in systemic vascular resistance and myocardial contractility. Other hormones, including the prostaglandins, bradykinin, and prolactin, modulate the actions of many of the regulatory hormones. Renin-Angiotensin-Aldosterone System Decreases in renal capillary blood flow stimulate renin secretion, which in turn initiates the production of angiotensin. Angiotensin induces vasoconstriction, increased tubular sodium and water reabsorption, and the release of aldosterone (Riordan, 1995). Aldosterone increases potassium secretion and further enhances sodium reabsorption in the distal tubule; therefore the primary function of this system is to protect the volume of the extracellular compartment and maintain adequate tissue perfusion (Bailie, 1992). However, its effectiveness in the neonate is somewhat limited by the decreased responsiveness of the immature kidney to the sodium- and water-retaining effects of these hormones (Sulyok et al, 1985). Vasodilatory and natriuretic prostaglandins generated in the kidney (Gleason, 1987) are the main counterregulatory hormones balancing the renal actions of the renin-angiotensin-aldosterone system. Therefore, when prostaglandin production is inhibited by indomethacin, the unopposed vasoconstrictive and sodium-retentive actions of the activated renin-angiotensin-aldosterone system contribute to the development of the drug-induced renal failure in the preterm infant (Gleason, 1987; Seri, 1995; Seri et al, 2002). Vasopressin (Antidiuretic Hormone) Vasopressin has its major effect in maintaining the osmolality of the extracellular compartment. Vasopressin selectively raises free water reabsorption in the kidneys and 371 results in blood pressure elevation (Elliot et al, 1996). Plasma levels of vasopressin are markedly elevated in the neonate, especially after vaginal delivery, and its cardiovascular actions facilitate neonatal adaptation (Pohjavuori and Raivio, 1985). The high vasopressin levels are in part also responsible for the diminished urine output of the healthy term neonate during the first day of life. Under certain pathologic conditions, the dysregulated release of, or the end-organ unresponsiveness to, vasopressin significantly affects renal and cardiovascular functions and electrolyte and fluid status in the sick preterm and term infant. In the syndrome of inappropriate secretion of antidiuretic hormone (SIADH), an uncontrolled release of vasopressin occurs in sick preterm and term infants, with resulting water retention, hyponatremia, and oliguria. In the syndrome of diabetes insipidus (DI), the lack of pituitary production of vasopressin or renal unresponsiveness to vasopressin results in polyuria, with increased thirst and hypernatremia. Atrial Natriuretic Peptide Via its direct vasodilatory and renal natriuretic actions, atrial natriuretic peptide (ANP) regulates the volume of the extracellular compartment in the fetus and neonate in a fashion opposite to that of the renin-angiotensin-­ aldosterone system (Iwamoto, 1992; Needleman and ­Greenwald, 1986; Seymour, 1985). ANP has a direct inhibitory effect on renin production and aldosterone release (Christensen, 1993). The stretch of the atrial wall caused by an increase in the circulating blood volume is the most potent stimulus for the release of this hormone. Plasma levels of this hormone are high in the fetus (Claycomb, 1988). There are a few specific conditions in which the actions of atrial natriuretic peptide are directly relevant for the neonatologist. For example, the hormone is involved in the regulation of both the fluid shifts during labor (Brace, 1992) and the extracellular volume contraction during postnatal transition (Kojima et al, 1987; Ronconi et al, 1995; Rozycki and Baumgart, 1991; Tulassay et al, 1987). Furthermore, the oliguric effects of positive end-expiratory pressure ventilation are due in part to a decrease in atrial natriuretic peptide secretion (Christensen, 1993) along with the enhanced release of vasopressin (El-Dahr and Chevalier, 1990). Brain (or B-type) Natriuretic Peptide Similar to ANP, brain natriuretic peptide (BNP) is a hormone secreted by cardiac myocytes. Specifically, BNP is thought to be secreted by cardiac ventricular myocytes in response to increases in wall tension. Both of these peptides cause natriuresis, diuresis, and vasodilation while inhibiting the renin-angiotensin-aldosterone system (Gemelli et al, 1991; Holmes et al, 1993; Kojima et al, 1989). BNP levels increase rapidly after birth, with levels on the first day of life up to 20-fold higher than those at birth (Mir et al, 2003; Yoshibayashi et al, 1995), and correlate with the downward trend in pulmonary arterial pressure in the days after birth, unlike ANP (Ikemoto et al, 1996). BNP levels continue to fall during the first week of life (Koch and Singer, 2003; Mir et al, 2003). Infants and children with congestive heart failure have an elevated 372 PART VIII Care of the High-Risk Infant BNP level, which is associated with decreased ejection fraction and increased heart failure score (Mir et al, 2002). In this way, BNP can be used as a surrogate for changes in ventricular volume from either pulmonary hypertension or ventricular overload. Clinical researchers have begun to investigate the utility of this peptide in diagnosing and managing cardiac disorders in neonates. The best-studied association to date is the relationship between BNP and a patent ductus arteriosus (PDA). Multiple studies have described a positive correlation between BNP level and a hemodynamically significant PDA (Choi et al, 2005; Czernik et al, 2008; Flynn et al, 2005; Sanjeev et al, 2005). BNP levels were associated with larger ductal size and degree of shunting, especially in infants older than 2 days (Flynn et al, 2005). Furthermore, significant correlations have been seen between BNP level and left atrial to aortic root diameter (Choi et al, 2005; Czernik et al, 2008) and diastolic flow velocity in the left pulmonary artery (Choi et al, 2005). BNP levels decline significantly with either medical or surgical therapy of a hemodynamically significant PDA (Choi et al, 2005; Sanjeev et al, 2005). The biologically inactive fragment N-terminal pro-B-type natriuretic peptide (NT-proBNP) has also been studied regarding its association with PDA with similar results (Farombi et al, 2008). NT-proBNP has a longer halflife (60 versus 20 minutes) than BNP and has also been associated with sepsis (Farombi et al, 2008). BNP has also been associated with other forms of cardiac dysfunction in neonates. Increases in BNP levels have been described in infants with persistent pulmonary hypertension of the newborn over the first 4 days of life, correlating with the degree of tricuspid regurgitation on an echocardiogram (Reynolds et al, 2004). BNP measurements can also be useful in differentiating infants in respiratory distress who have a cardiac etiology for their disease from those who do not (Ko et al, 2008; Reynolds et al, 2004). The usefulness of this test is limited by the variability of levels in the first few days of life as well as the variety of assays available to measure BNP levels. Its utility may ultimately lie in repeated measures in the same patient over time using the same assay and following trends. Prostaglandins Prostaglandins have a well-documented, counterregulatory role for the renal vascular and tubular effects of renin-angiotensin-aldosterone and vasopressin (Bonvalet et al, 1987). The inhibition of these actions of prostaglandins by indomethacin results in clinically important and sometimes detrimental renal vascular and tubular effects in the preterm infant. The actions of prostaglandins modulating the effects of the other regulatory hormones of neonatal fluid and electrolyte homeostasis are less well studied. Prolactin Prolactin plays a permissive role in the regulation of fetal and neonatal water homeostasis (Coulter, 1983; Pullano et al, 1989). High fetal plasma prolactin levels contribute to the increased tissue water content of the fetus. Interestingly, postnatal prolactin levels remain high in the preterm neonate until approximately the 40th postconceptional week (Perlman et al, 1978). MANAGEMENT OF FLUID AND ELECTROLYTE HOMEOSTASIS General Principles of Fluid and Electrolyte Management Fluid and electrolyte management is the cornerstone of neonatal intensive care, and appropriate management requires an understanding of the previously outlined physiologic principles and careful monitoring of key clinical data. Requirements vary substantially from infant to infant and in the same infant over time; therefore intakes must be individualized and frequently reassessed. The primary goals are to maintain the appropriate ECF volume, ECF and intracellular fluid osmolality, and ionic concentrations. Assessment of Fluid and Electrolyte Status Maternal conditions during pregnancy, drugs and fluids administered to the mother during labor and delivery, and specific fetal and neonatal conditions all affect early fluid and electrolyte balance. Excessive administration of free water or oxytocin use in the mother can result in hyponatremia in the neonate. Maternal therapy with indomethacin, angiotensin-converting enzyme inhibitors, furosemide, and aminoglycosides can all adversely affect neonatal renal function. A newborn’s history of oligohydramnios or birth asphyxia may also alert the clinician to the possibility of abnormal renal function. In young infants, altered skin turgor, sunken anterior fontanel, and dry mucous membrane are not sensitive indicators of dehydration, but tachycardia, hypotension, and metabolic acidosis may be seen when intravascular volume is moderately to severely affected. In addition, edema usually occurs early when there is volume overload. Serial measurements of body weight, intake and output, and serum electrolytes will usually provide the most precise and accurate information regarding overall fluid status. Appropriate fluid balance in the first few days after birth is associated with a urine output of 1 to 3 mL/kg per hour and a weight loss of 5% to 10% in term infants and 10% to 15% in preterm infants. In critically ill infants and in situations of altered homeostasis, additional clinical data that may help in diagnosis and management include blood urea nitrogen, serum and urine osmolarity or specific gravity, urine electrolytes and serum bicarbonate, along with close monitoring of blood pressure and heart rate. The frequency of monitoring depends on the extent of immaturity and the severity of the fluid and electrolyte disturbance and of the underlying pathologic condition. Water Homeostasis and Management Water Losses Free water losses occurring through the skin and the respiratory tract are considered insensible losses, whereas the sensible water losses are composed of the amounts lost through urine and feces. Urine output is the most important source of sensible water loss. Extremely preterm infants without systemic hypotension or renal failure usually lose 30 to 40 mL/kg/day of water in the urine on the first postnatal day and approximately 120 mL/ kg/day by the third day. In stable, more mature preterm infants born after the 28 weeks’ gestation, urinary water 373 CHAPTER 31 Acid-Base, Fluid, and Electrolyte Management loss is approximately 90 mL/kg/day on the first postnatal day and 150 mL/kg/day by the third day (Coulthard and Hey, 1985). Because of their renal immaturity, preterm neonates have a tendency to produce dilute urine, thereby increasing their obligatory free water losses. Normal water losses in the stool are less significant, amounting to approximately 10 mL/kg/day in term infants and 7 mL/kg/day in preterm infants during the first postnatal week (Sedin, 1995). Water losses in the stool increase thereafter and are influenced by the type of feeding and the frequency of stooling. Gestational age, postnatal age, and environmental factors determine the amount of daily insensible water losses through the skin (see Figure 31-4). During the first few postnatal days, transepidermal water losses may be 15-fold higher in extremely premature infants born at 23 to 26 weeks’ gestation than in term neonates (Sedin, 1995). Although the skin matures rapidly after birth, even in extremely immature infants, insensible losses are still somewhat higher at the end of the first month than in the term counterparts. Prenatal steroid exposure is associated with substantially less insensible water loss (IWL) in premature infants (Aszterbaum et al, 1993; Omar et al, 1999; Sedin, 1995). Among environmental factors, ambient humidity has the greatest effect on transepidermal water loss. In extremely immature neonates, a rise in the ambient humidity of the incubator from 20% to 80% decreases the transepidermal water loss by approximately 75% (Sedin, 1995). However, the use of an open radiant warmer more than doubles transepidermal water losses (Flenady and Woodgate, 2003). Applying a plastic heat shield while the infant is under the warmer can decrease transepidermal water loss by 30% to 50% (Costarino et al, 1992). At low ambient humidity, phototherapy increases transepidermal water losses by approximately 30%. However, one study showed that if the ambient humidity is at least 50% in the incubator, regular phototherapy lamps will not increase the transepidermal water loss in infants older than 28 weeks’ gestation (Sedin, 1995). On the other hand, halogen spotlight phototherapy increases transepidermal water loss in premature infants by 20% despite constant skin temperature and relative humidity (Grunhagen et al, 2002). Newer light-­emitting diode phototherapy devices are not associated with significant transepidermal water loss or changes in cerebral hemodynamics (Bertini et al, 2008). Other factors that increase IWL include activity, airflow, elevated body, and environmental temperature as well as skin breakdown and skin or mucosal defects (e.g., gastroschisis, epidermolysis bullosa). Insensible water losses from the respiratory tract depend mainly on the temperature and humidity of the inspired gas mixture and on the respiratory rate, tidal volume, and dead space ventilation. In a healthy term newborn, the water loss through the respiratory tract is approximately half the total IWL if the ambient air temperature is 32.5°C and the humidity is 50% (Sedin, 1995). However, in infants undergoing mechanical ventilation there will be no insensible losses through the respiratory tract if the ventilator gas mixture is humidified at body temperature. Extraordinary losses are also seen in the neonate requiring intensive care. The most commonly encountered extraordinary water losses occur when a nasogastric tube is placed under continuous suction (to be discussed later in this chapter). Large losses may also occur in association with chest tubes, other drains, ostomies, and fistulas as well as with emesis or diarrhea. Management of Water Requirements Appropriate management requires estimating any existing deficits or surpluses, calculating ongoing maintenance needs because of usual sensible and insensible losses and growth and additional needs as a result of extraordinary losses. The infant’s prenatal history, birthweight, gestational age, and need for mechanical ventilation and the environment in which the infant is to be cared for should be considered when determining initial fluid and electrolyte needs. Frequent reevaluations are necessary. The most useful parameter for monitoring fluid balance is the weight of the baby, as rapid changes in weight will reflect changes in water balance. Serial weights can be used to estimate the IWL using the following formula: IWL = Fluid intake − Urine output + Weight loss or IWL = Fluid intake − Urine output − Weight gain It is reasonable to initiate fluid volume based on the sum of an allowance for sensible water loss of 30 to 60 mL/ kg/day and the estimated IWL. Figure 31-4 shows usual IWL ranges by gestational and postnatal age. Factors previously outlined that predictably affect IWL should be considered when prescribing fluids. Prevention of excessive IWL rather than replacement of increased IWL is associated with fewer complications in the preterm neonate and usually can be achieved by modifying the infant’s environment. See Table 31-1 for usual maintenance fluid administration based on birthweight. These numbers are guidelines for initial management only; the approach must subsequently be individualized based on laboratory values and other clinical data. It is important to remember that the TBW excess and extracellular volume expansion of preterm infants implies that their negative water and sodium balance during the first 5 to 10 postnatal days (see Figure 31-2; Shaffer and Meade, 1989) represents an appropriate adaptation to extrauterine life and should not be compensated for by increased fluid administration and sodium supplementation. If this principle is not followed and a positive fluid balance (i.e., weight gain) is achieved during the transitional period, preterm infants are at higher risk of a more severe course of respiratory distress syndrome (Shaffer and Weismann, 1992) and a higher incidence of patent ductus arteriosus (Bell et al, 1980), congestive heart failure (Bell et al, 1980), TABLE 31-1 Estimated Maintenance Fluid Requirements Fluid Requirements (mL/kg/day) Birthweight (g) <750 750-1000 1000-1500 >1500 Day 1 Day 2 Day 3-6 ≥ Day 7 100-140 100-120 80-100 60-80 120-160 100-140 100-120 80-120 140-200 130-180 120-160 120-160 140-160 140-160 150 150 374 PART VIII Care of the High-Risk Infant pulmonary edema (Shaffer and Weismann, 1992), necrotizing enterocolitis (Bell et al, 1979), and bronchopulmonary dysplasia (Oh et al, 2005; Van Marter et al, 1990). Infants with ELBW and others with anticipated fluid problems should be weighed daily or twice daily. Serum sodium levels should be measured every 4 to 8 hours ­untilstabilized, usually by 3 to 4 days after birth, and urine output should be recorded and reviewed every 6 to 8 hours. Once data are available, fluids should be increased if weight loss is greater than 1% to 2% per day in term infants and 2% to 3% per day in preterm infants, if urine output is low, if urine specific gravity is rising, or if serum sodium concentration is rising. Overall, expected and appropriate weight loss in the first week of life is up to 10% in term infants and up to 20% in preterm infants. Conversely, fluids should be decreased if weight is not falling appropriately and serum sodium is decreasing. The goal is to reach 140 to 160 mL/kg/day of fluids by 7 to 10 days to allow for adequate caloric intake. Treatment of Fluid Overload Fluid overload commonly occurs in sick neonates, often because of the use of fluid bolus administration for hypotension. The diagnosis is based on weight gain, edema, and often hyponatremia. Overhydration can sometimes be prevented by the use of blood transfusions or dopamine instead of colloid or crystalloid, if appropriate, for blood pressure support. In addition to reducing the need for volume boluses, dopamine may facilitate the process of extracellular volume contraction via its renal and hormonal effects (Seri, 1995). Once overhydration has occurred, management is usually effected by 10% to 20% decrements of total daily fluid intake while carefully monitoring clinical and laboratory signs to ensure maintenance of adequate intravascular volume as well as normal glucose and electrolyte status while the ECF contraction occurs. Treatment of Dehydration Dehydration may be suspected based on clinical signs. The total water deficit may be estimated by using weight changes, calculating total inputs and outputs, and following serial sodium levels. Free water deficit (or excess) can be calculated as: H2 O deficit (or excess) (L) ={[0.7 × BW (kg) ] × } [Na (mEq/L) ] current −1 [Na (mEq/L) ] desired In this formula, 0.7 × BW is the estimation of TBW. When dehydration is diagnosed, correction should generally occur over 24 hours, with half correction over 8 hours and the remainder over the next 16 hours. Longer correction times are indicated when dehydration is accompanied by moderate to severe hypernatremia. Treatment is best approached by considering separately the fluid resuscitation requirements, fluid to replace current deficits and ongoing losses, and maintenance requirements, because the volume of the fluid, the composition, and the rate of replacement differ for each. Resuscitation of the intravascular volume to restore blood pressure and perfusion should be provided with boluses of isotonic saline (0.9% saline), and this volume is included in the initial half correction. The remaining deficit replacement volume should TABLE 31-2 Free Water Content (as Volume %) of Common Intravenous Solutions at Normal and High Serum Sodium Concentrations* Serum Sodium Concentration 145 mEq/L Intravenous Fluid D5W 0.2% saline 0.45% saline 0.9% saline Lactated Ringer’s ­solution 195 mEq/L Isotonic (%) Water (%) Isotonic (%) Water (%) 0 22 50 100 86 100 78 50 0 14 0 17 39 79 68 100 83 61 21 32 Modified from Molteni KH: Initial management of hypernatremic dehydration in the breastfed infant, Clin Pediatr 33:731-740, 1994. *Note that isotonic saline provides 21% free water when given to a patient with a serum sodium concentration of 195 mEq/L and therefore will induce undesirable decreases in serum sodium concentration when used for volume resuscitation in the severely dehydrated hypernatremic neonate. TABLE 31-3 Approximate Electrolyte Composition of Body Fluids (mEq/L) Body Fluid Sodium Potassium Chloride Gastric Small intestine Bile Ileostomy Diarrhea 20-80 100-140 120-140 45-135 10-90 5-20 5-15 5-15 3-15 10-80 100-150 90-130 80-120 20-115 10-110 be with fluid of appropriate sodium content based on the serum sodium (see Sodium and Potassium Homeostasis and Management, later), usually 0.2% or 0.45% saline. Free water contents of the common intravenous fluids are listed in Table 31-2. Usual maintenance fluids and electrolytes must also be provided. Ongoing urine losses should be replaced volume for volume every 4 to 6 hours with a solution tailored to the urine’s electrolyte concentration (usually 0.45% normal saline). Extraordinary losses caused by tubes, drains, emesis, diarrhea, and ostomies should always be sought in the dehydrated infant and also accounted for in fluid management. The composition of this latter replacement solution depends on the electrolyte concentration of the fluid loss. The most common extraordinary loss, gastric fluid, contains significant sodium and chloride. See Table 31-3 for approximate electrolyte compositions of body fluids. Potassium replacement (usually by adding 20 to 40 mEq potassium per liter of replacement fluid) should not begin until adequate urine output is established. For additional details of fluid correction and sodium management, see the discussion under Management of Hypernatremia later in this chapter. Sodium and Potassium Homeostasis and Management Serum sodium values should be kept between 135 and 145 mEq/L. Sodium chloride supplementation at 1 to 2 mEq/ kg/day should be started in preterm and sick term neonates CHAPTER 31 Acid-Base, Fluid, and Electrolyte Management only after completion of the postnatal extracellular volume contraction, usually after the first few days of life or after more than 5% of birthweight is lost (Hartnoll et al, 2001). In general, as long as the infant’s fluid balance is stable, maintenance sodium requirements do not exceed 3 to 4 mEq/kg/day, and providing this amount usually ensures the positive sodium balance necessary for adequate growth. Extreme prematurity and pathologic conditions associated with delayed transition or disturbance of fluid and electrolyte balance may significantly alter the infant’s daily sodium requirement. For example, although the preterm infant has a limited ability to excrete a sodium load (Hartnoll, 2003), some of the most immature infants may have sodium requirements of as much as 6 to 8 mEq/kg/ day because of the decreased capacity of their kidneys to retain sodium. Neonates recovering from an acute renal insult and preterm infants with immature proximal tubule functions who are in a state of extracellular volume expansion (Ramiro-Tolentino et al, 1996) may need daily sodium bicarbonate supplementation to compensate for their greater renal bicarbonate losses. Hyponatremia Hyponatremia (serum sodium <130 mEq/L) represents a deficit of sodium in relation to body water content and may be caused by either total body sodium deficit or free water excess. In both situations, total body water may be decreased (hyponatremia with volume contraction), normal, or increased (hyponatremia with volume expansion). Hyponatremia may be hypotonic, isotonic, or hypertonic. To initiate effective treatment, it is important to attempt to determine the primary cause of the hyponatremia and whether there is associated volume expansion or contraction. The most common cause of hyponatremia in the sick neonate is excessive administration or retention of free water. In these situations the total body sodium content is normal, and the appropriate treatment is restriction of free water intake and not administration of sodium. In situations of true sodium deficit, the deficits can be estimated by assuming 70% of total body weight as the distribution space of sodium. The formula for calculating sodium ­deficit is: Na+ deficit (or excess) (mEq) ≈ 0.7 × kg × [ (Na+ ) desired − [ (Na+ ) actual ] In most situations of depletional hyponatremia, the sodium deficit should be replaced on a schedule that provides two thirds replacement in the first 24 hours and the remainder in the next 24 hours. Frequent measurements of serum electrolytes are needed to ensure that the correction is occurring appropriately. If the serum sodium concentration is less than 120 mEq/L, regardless of whether the hyponatremia is due to free water overload or total body sodium deficit, then correction of the serum sodium concentration up to 120 mEq/L is recommended with administration of 3% saline solution (513 mEq of sodium per liter). This correction should be done over 4 to 6 hours, depending on the severity of hyponatremia (Avner, 1995) and using the above formula. Although rapid intravenous bolus administration of 4 to 6 mL/kg of 3% saline solution has been effective in children with seizures or 375 coma (Sarnaik et al, 1991), rapid and complete correction of low serum sodium concentration in adults with chronic hyponatremia has been shown to be associated with pontine and extrapontine myelinolysis. Once the risk of acute central nervous system symptoms has been minimized and serum sodium concentration has reached 120 mEq/L, complete correction of hyponatremia should be performed more slowly over the next 48 hours. In patients with asymptomatic hyponatremia whose serum sodium concentration exceeds 120 mEq/L, hypertonic infusions are not indicated. Additional therapy should be directed at fluid restriction if the hyponatremia is dilutional or sodium repletion if the hyponatremia is depletional. The use of 5% dextrose in water with 0.45% to 0.9% saline is a reasonable replacement fluid for depletional hyponatremia once the sodium is above 120 mEq/L. More stable infants with chronic sodium losses can also be corrected with enteral sodium chloride. Figure 31-5 summarizes the clinical evaluation and therapy of neonates with hyponatremia. Hypernatremia Hypernatremia (serum sodium >145 mEq/L) reflects a deficiency of water relative to total body sodium and is most often a disorder of water rather than sodium homeostasis. Hypernatremia does not reflect the total body sodium content, which can be high, normal, or low depending on the cause of the condition. Hypernatremia can also be associated with hypovolemia, normovolemia, or hypervolemia (Box 31-2). If hypernatremia is primarily due to changes in sodium balance, it can result from pure sodium gain or, more commonly, sodium gain coupled with a lesser degree of water accumulation or, rarely, water loss. The hypernatremia-induced hypertonicity causes water to shift from the intracellular to the extracellular compartment, resulting in intracellular dehydration and the relative preservation of the extracellular compartment. This shift is the main reason that neonates with chronic hypernatremic dehydration often do not demonstrate overt clinical signs of intravascular depletion and dehydration until late in the course of the condition. Compared with other organs, the central nervous system has a unique adaptive capacity to respond to the hypernatremia-induced hypertonicity, leading to a relative preservation of neuronal cell volume. The shrinkage of the brain stimulates the uptake of electrolytes such as sodium, potassium, and chloride (immediate effect) as well as the synthesis of osmoprotective amino acids and organic solutes (more delayed response). These idiogenic osmols aid in maintaining normal brain cell volume during longer periods of hyperosmolar stress (Trachtman, 1991). As long as hypernatremia develops rapidly (within hours), as in accidental sodium loading, a relatively rapid correction of the condition improves the prognosis without raising the risk of cerebral edema formation. Intracellular fluid accumulation does not occur because the accumulated electrolytes (sodium, potassium, and chloride) are rapidly extruded from the brain cells, and cerebral edema is unlikely. In these cases, reducing serum sodium concentration by 1 mEq/L per hour (24 mEq/L per day) is appropriate (Adrogue and Madias, 2000). However, because of the slow dissipation of idiogenic osmols over a period of several days (Adrogue and Madias, 376 PART VIII Care of the High-Risk Infant HYPONATREMIA 1. Exclude “pseudohyponatremia” (hyperlipidemia, hyperproteinemia) 2. Exclude hypertonic hyponatremia (↑ ECF osmolality due primarily to hyperglycosemia) 3. Evaluate ECF volume status (clinical and laboratory indicators) Hypovolemic ↓↓ Total body sodium ↓ Total body water Urine [Na] <20 mEq/L Urine [Na] >20 mEq/L Euvolemic ± Total body sodium ↑ Total body water Hypervolemic ↑ Total body sodium ↑↑ Total body water Urine [Na] usually >20 mEq /L Urine [Na] <20 mEq/L Urine [Na] >20 mEq/L Extrarenal losses: 1. Gastrointestinal (vomiting, diarrhea, drainage tubes, fistulas) 2. Pleural effusions, ascites 3. Ileus 4. Necrotizing enterocolitis Renal losses: 1. Diuretics 2. Osmotic diuresis 3. Contraction alkalosis 4. Mineralocorticoid deficiency 5. Mineralocorticoid unresponsiveness 6. Fanconi syndrome 7. Bartter’s syndrome 8. Obstructive uropathy 1. Glucocorticoid, thyroid 2. Excess ADH Edema-forming states: 1. Congestive heart failure 2. Liver failure/cirrhosis 3. Nephrosis syndrome 4. Indomethacin therapy Renal failure: 1. Acute 2. Chronic Volume expansion Volume expansion Water restriction Sodium and water restriction Sodium and water restriction FIGURE 31-5 Flow diagram for the clinical evaluation and therapy of neonates with hyponatremia. ADH, Antidiuretic hormone; ECF, extracellular fluid. (Modified from Avner ED: Clinical disorders of water metabolism: hyponatremia and hypernatremia, Pediatr Ann 24:23-30, 1995.) 2000), in cases of chronic hypernatremia or in cases in which the time frame is unknown, the hypernatremia should be corrected more slowly, at a maximum rate of 0.5 mEq/L per hour (12 mEq/L per day). If correction is performed more rapidly in these cases, the abrupt fall in the extracellular tonicity results in the movement of water into the brain cells, which have a relatively fixed hypertonicity because of the presence of the osmoprotective molecules. The result is the development of brain edema with deleterious consequences (Adrogue and Madias, 2000; Molteni, 1994). In the breastfed term neonate, hypernatremia most commonly develops because of dehydration caused by inadequate breast milk intake (Molteni, 1994), but may also be caused by high sodium levels in maternal breast milk. Reduction in breastfeeding frequency has been shown to be associated with a marked rise in the sodium concentration of breast milk (Neville et al, 1991). A vicious cycle can ensue in which the infant sucks poorly, breast milk production drops, sodium concentration rises, and the infant becomes increasingly dehydrated, hypernatremic, and lethargic. Recognition may be delayed because these infants may appear quiet and content initially. Because this process is chronic, usually occurring over 7 to 14 days, signs of extracellular volume contraction are less prominent until the development of the full clinical presentation consisting of lethargy, irritability, abnormal muscle tone with or without seizures, and cardiovascular collapse BOX 31-2 C  onditions Causing ­Hypernatremia HYPOVOLEMIC HYPERNATREMIA Inadequate breast milk intake Diarrhea Radiant warmers Excessive sweating Renal dysplasia Osmotic diuresis EUVOLEMIC HYPERNATREMIA Decreased Production of Antidiuretic Hormone Central diabetes insipidus, head trauma, central nervous system tumors (craniopharyngioma), meningitis, or encephalitis Decrease or Absence of Renal Responsiveness Nephrogenic diabetes insipidus, extreme immaturity, renal insult, and medications such as amphotericin, hydantoin, aminoglycosides HYPERVOLEMIC HYPERNATREMIA Improperly mixed formula NaHCO3 administration NaCl administration Primary hyperaldosteronism CHAPTER 31 Acid-Base, Fluid, and Electrolyte Management 377 with renal failure. This presentation can be associated with serious central nervous system morbidity from both the hypertonicity (sagital or other venous sinus thrombosis, subdural capillary hemorrhage, white matter injury) and inappropriately rapid rehydration therapy (brain edema, myelinolysis). In the extremely immature neonate, hypernatremia most commonly occurs from excessive transepidermal free water losses. The condition usually develops rapidly, within 24 to 48 hours after birth. The diagnosis is based on the attendant decrease in body weight and clinical signs of extracellular volume contraction. Prevention of this condition has been successful in the majority of immature neonates by frequent monitoring of serum electrolyte levels, appropriate adjustments of free water intake, and the use of humidified incubators (Modi, 2004). The central and nephrogenic forms of diabetes insipidus are much less commonly encountered and result in hypernatremia because of the lack of production of and renal responsiveness to ADH, respectively. Hypernatremia can also develop in response to excessive sodium supplementation, mainly in the sick neonate receiving repeated volume boluses for cardiovascular support. In these cases, clinical signs of edema, increased body weight, and the history of volume boluses help to establish the diagnosis. And the amount of free water required to decrease serum sodium by 12 mEq/L over a 24-hour period when hypernatremia is severe is calculated as: Treatment of Hypernatremia Percentage of free water content = 1 − (IV fluid sodium / Serum sodium) Thorough analysis of the medical history and the changes in clinical signs, laboratory findings, and body weight usually aid in determining the major etiologic factor in hypernatremia and thus the appropriate treatment. In the critically ill infant, the cause of the serum sodium abnormality may be multifactorial however, and the treatment less straightforward. Although some cases of hypernatremia are a result of sodium excess with normal or high TBW, most cases in neonates are due to hypernatremic dehydration. Treatment of this condition is generally divided into two phases: the emergent phase where the intravascular volume is restored, usually by administration of 10 to 20 mL/kg of isotonic saline, and the rehydration phase, where the sum of the remaining free water deficit and usual maintenance needs are administered evenly over at least 48 hours. The free water deficit can be calculated as: H2 O deficit (or excess) (L) = [0.7 × BW (kg) ] × { } 1 −[NA+ (mEq/L) ] current /[NA+ (mEq/L)] desired In this formula, (0.7 × BW) is the estimation of TBW. It is important to note that the amount of free water required to decrease the serum sodium by 1 mEq/L is 4 mL/kg with moderate hypernatremia, but only 3 mL/kg when the hypernatremia is as high as 195 mEq/L (Molteni, 1994). Therefore the amount of free water required to decrease serum sodium by 12 mEq/L over a 24-hour period when hypernatremia is moderate is cal­ culated as: Free water required = Current weight (kg) × 4mL/kg × 12mEq/L or Free water required = Current weight × 48 mL/kg /day Free water required = Current weight × 36 mL / kg / day The free water contents of the common intravenous fluids are listed in Table 31-2. In most mild to moderate hypernatremic states, during the rehydration phase, replacement fluids of 5% dextrose in 0.2% normal saline (31 mEq/L) or 0.45% normal saline (77 mEq/L) are appropriate. Infants with serum sodium levels greater than 165 mEq/L should initially be given 0.9% saline to avoid sudden drops in serum sodium concentration. When the serum sodium is greater than 175 mEq/L, however, normal saline will be hypotonic compared with the patient’s serum. In these instances of severe hypernatremia, an appropriate amount of 3% saline (513 mEq/L) should be added to the intravenous (IV) fluid so that the sodium concentration is approximately 10 to 15 mEq/L less than the serum sodium level (Rand and Kolberg, 2001). The relative free water content of an IV solution for a specific patient with sodium perturbations can be calculated using the formula: Serum electrolytes should be monitored every 2 to 4 hours until the desired rate of decline in serum sodium concentration is established. At this point, the frequency of the laboratory measurements can be relaxed to every 4 to 6 hours until the serum sodium concentration is less than 150 mEq/L. The speed of correction of hypernatremia depends on the rate of its development. This approach provides a reasonable chance that the serum sodium concentration will gradually decrease to the normal range over 2 to 4 days. Except in cases of acute massive sodium overload, the goal should be to drop the serum sodium concentration at a rate no greater than 1 mEq/L/h. A slower pace of correction of 0.5 mEq/L/h is prudent in patients with hypernatremia of chronic or unknown duration to avoid iatrogenic central nervous system sequelae. Once serum sodium concentration, urine output, and renal function are normal, the patient should receive standard maintenance fluids, either intravenously or orally, depending on his or her condition. At this time, electrolyte status must still be monitored for an additional 24 hours to ensure that complete recovery has occurred. Hyperglycemia and hypocalcemia commonly accompany hypernatremia. The use of insulin to treat the hyperglycemia is not recommended, because it can increase brain idiogenic osmol content. Hypocalcemia should be corrected with appropriate calcium supplementation. Potassium Homeostasis and Management Serum potassium should be kept between 3.5 and 5 mEq/L. In the early postnatal period, neonates, especially immature preterm infants, have higher serum potassium concentrations than older persons. The etiology of the relative hyperkalemia of the newborn is multifactorial and 378 PART VIII Care of the High-Risk Infant involves developmentally regulated differences in renal function, Na+,K+-ATPase activity (Vasarhelyi et al, 2000), and hormonal milieu. Exposure to prenatal steroids in premature infants is associated with a decreased incidence of hyperkalemia, believed to be due to improved renal function (Omar et al, 2000). In general, potassium supplementation should be started only after urine output has been well established, usually by the third postnatal day. Supplementation should be started at 1 to 2 mEq/kg/day and increased over 1 to 2 days to the usual maintenance requirement of 2 to 3 mEq/ kg/day. Some preterm infants may need more potassium supplementation after the completion of their postnatal volume contraction, because of their increased plasma aldosterone concentrations, prostaglandin excretion, and disproportionately high urine flow rates. Most term and preterm neonates will require potassium supplementation if they are receiving diuretics. Hypokalemia Hypokalemia in the neonate is usually defined as a serum potassium level of less than 3.5 mEq/L. Hypokalemia can occur from potassium loss due to diuretics, diarrhea, renal dysfunction, or nasogastric drainage from inadequate potassium intake or from intracellular movement of potassium in the presence of alkalosis. Except in patients receiving digoxin, hypokalemia is rarely symptomatic until the serum potassium concentration is less than 2.5 mEq/L. Electrocardiogram (ECG) manifestations of hypokalemia include flattened T waves, prolongation of the QT interval, or the appearance of U waves. Severe hypokalemia can result in cardiac arrhythmias, ileus, and lethargy. Treatment of Hypokalemia Hypokalemia is treated by slowly replacing potassium either intravenously or orally, usually in the daily fluids. Rapid administration of potassium chloride is not recommended, because it is associated with life-threatening cardiac dysfunction. In extreme emergencies, potassium can be given as an infusion over 30 to 60 minutes of not more than 0.3 mEq/kg potassium chloride. If hypokalemia is secondary to alkalosis, the alkalosis should be corrected before considering increasing the potassium intake. Hyperkalemia Hyperkalemia in the neonate is defined as a serum potassium level greater than 6 mEq/L in a nonhemolyzed specimen. It is important to understand that most of the body’s potassium is contained within cells; therefore serum potassium levels do not accurately reflect total body stores. However, a serum potassium greater than 6.5 to 7 mEq/L can be life threatening, even if stores are normal or low, because of its effect on cardiac rhythm. ECG manifestations of hyperkalemia include peaked T waves (the earliest sign), a widened QRS configuration, bradycardia, tachycardia, supraventricular tachycardia, ventricular tachycardia, and ventricular fibrillation. Because pH affects the distribution of potassium between the intracellular and the extracellular space, serum potassium levels rise during acidosis, which may occur acutely. The clinician should be aware of the potential for life-threatening arrhythmias to occur in infants with chronic lung disease on diuretics and potassium supplements who develop a sudden respiratory deterioration with acidosis. Another common cause of hyperkalemia is renal dysfunction, of particular concern in very preterm and asphyxiated infants. In addition, infants who have suffered intraventricular hemorrhage or tissue trauma and those with intravascular hemolysis often have hyperkalemia caused by the release of potassium during breakdown of red blood cells. Finally, hyperkalemia may be one of the earliest manifestations of congenital adrenal hyperplasia. Treatment of Hyperkalemia When hyperkalemia is diagnosed, all potassium intake should be discontinued and the ECG should be monitored. Table 31-4 presents medications used in management of significant hyperkalemia. Calcium gluconate stabilizes cardiac membranes and alkali therapy, insulin, and inhaled albuterol (Singh et al, 2002) all rapidly enhance cellular uptake of potassium and can acutely drop serum potassium levels, but not decrease total body potassium. Intravenous furosemide and rectally administered sodium polystyrene sulfonate (Kayexalate) enhance potassium excretion and will lower total body stores, but they require several hours to take effect. Dialysis or exchange transfusion may be used when the hyperkalemia is life threatening and these measures do not result in improvement. CLINICAL CONDITIONS ASSOCIATED WITH FLUID AND ELECTROLYTE DISTURBANCES Extreme Prematurity Infants born between 23 and 27 weeks’ gestation, or with a birthweight of <1000 g (i.e., ELBW), are at particular risk for acute abnormalities of both fluid and electrolyte status in the immediate postnatal period. Their transepidermal water loss is much higher than that in more mature preterm neonates (see Figure 31-4), and it is difficult for their water balance to be maintained unless the excessive losses are prevented. It should be kept in mind that such an infant, when cared for in an open warmer without the use of a plastic heat shield, may lose up to 150 to 300 mL/ kg/day of free water through the skin during the first 3 to 5 days of life. Infants whose mothers received antenatal glucocorticoids often have fewer problems because prenatal glucocorticoids enhance maturation of the epidermis and result in increases in urine output and fractional excretion of sodium (Ali et al, 2000; Omar et al, 1999). Although the IWL primarily affects the extracellular volume, the intracellular compartment ultimately shares the loss of free water as osmotic pressure in the extracellular compartment rises. As water leaves the cells, intracellular osmolality rises and cell volume diminishes. In the central nervous system, as described earlier, these changes stimulate the generation of idiogenic osmols, resulting in selective increases in intracellular osmolality and a tendency toward normalization of neuronal cell volumes. This protective mechanism has significant clinical implications for the rate at which hypernatremia should be corrected, and the decrease in serum sodium concentration should not exceed 12 mEq/L/day, especially in the infant in whom the hypernatremia has been chronic (>12 hours). CHAPTER 31 Acid-Base, Fluid, and Electrolyte Management 379 TABLE 31-4 Medications Used for Treatment of Hyperkalemia Medication Dosage Onset Length of Effects Calcium ­gluconate 100 mg/kg IV over 2-5 min Immediate 30 min Protects the myocardium from toxic effects of potassium; no effect on total body potassium Can worsen digoxin toxicity Sodium ­bicarbonate 1-2 mEq/kg Immediate Variable Shifts potassium intracellularly; no effect on total body potassium Maximum infusion: mEq/min in emergency situations Tromethamine 3-5 mL/kg Immediate Variable Shifts potassium intracellularly; no effect on total body potassium — Insulin plus dextrose Insulin 0.1-0.15 U/ kg IV plus dextrose 0.5 g/kg IV 15 to 30 min 2 to 6 h Shifts potassium intracellularly; no effect on total body potassium Monitor for ­hypoglycemia Albuterol* 0.15 mg/kg every 20 min for three doses then 0.15-0.3 mg/kg 15 to 30 min 2 to 3 h Shifts potassium intracellularly; no effect on total body potassium Minimum dose, 2.5 mg Furosemide PO: 1-4 mg/kg/dose 1-2 times/day IV: 1-2 mg/kg/dose given every 12-24 hours 15 min to 1 h 4h Increases renal excretion of potassium — Kayexalate 1 g/kg PR every 6 hours 1 to 2 h (rectal route faster) 4 to 6 h Removes potassium from the gut in exchange for sodium Use with extreme caution in neonates, especialy preterm; contains sorbital; may be associated with bowel necrosis and sodium retention Mechanism of Action Comments and ­Cautions IV, Intravenous; PO, by mouth or gastric tube; PR, per rectum. *From Singh BS, Sadiq HF, Noguchi A, et al: Efficacy of albuterol inhalation in treatment of hyperkalemia in premature neonates, J Pediatr 141:16, 2002. Because serum sodium concentration is a reliable clinical indicator of extracellular tonicity, monitoring of this parameter every 6 to 12 hours during the first 2 to 3 postnatal days coupled with daily (or twice daily) measurements of body weight provide valuable information and appropriate guidance for the fluid and electrolyte management of the extremely immature preterm neonate. This is especially true in the absence of prenatal steroid exposure and if the infant is not cared for in a humidified incubator. Serum osmolality should be directly measured in patients in whom calculated serum osmolality is more than 300 to 320 mOsm/L. Table 31-1 shows suggested maintenance fluid requirements by birthweight and postnatal day of life. Because immature neonates in an incubator with an ambient air humidity of 50% to 80% require significantly less free water and less frequent measurements of serum electrolyte and osmolality (Sedin, 1995), open radiant warmers should be used only for critically ill, extremely labile preterm infants requiring frequent hands-on medical management. In these cases, the use of a protective plastic heat shield can help decrease excessive evaporative losses, and total daily fluid intake, similar to that used for infants in incubators, may be started at approximately 100 mL/kg/day containing 5% to 10% dextrose in water with close monitoring for dehydration. Daily fluid intake is then increased by 10 to 30 mL/kg/day every 6 to 12 hours if the serum sodium concentration rises from the baseline, the goal being to keep serum sodium concentration below 145 to 150 mEq/L. As skin integrity improves during the course of the 2nd to 3rd days, serum sodium concentration starts to fall. At this time, a significant stepwise limitation of total fluid intake is obligatory to allow a complete contraction of the extracellular volume to occur and to minimize the possibility of free water overload with its attendant risks for the development of ductal patency, pulmonary edema, and worsening lung disease. Potassium chloride supplementation may be started as soon as urine output has been established and the serum potassium concentration is less than 5 mEq/L. Extremely premature infants are at risk for the development of both oliguric and nonoliguric hyperkalemia, so the serum potassium concentration should be monitored closely, and supplementation should be discontinued if warranted by changes in serum potassium values or in renal function. Critically ill, extremely immature neonates often receive excess sodium with volume boluses, medications, and the maintenance infusion of their arterial lines (Costarino et al, 1992). Therefore extra sodium supplementation usually should not be started during the first few postnatal days, to prevent a rise in total body sodium concentration and thus in extracellular volume, which will hinder the appropriate postnatal diuresis. Many critically ill preterm infants retain their originally high extracellular volumes, even when sodium and water intakes are restricted, and such neonates also tend to lose more bicarbonate in the urine. Interestingly, proximal tubular bicarbonate reabsorption may be appropriate even in the VLBW infant despite the immaturity of their renal functions, as long as extracellular volume contraction 380 PART VIII Care of the High-Risk Infant takes place (Ramiro-Tolentino et al, 1996). Therefore the presence of the extracellular volume expansion appears to be an important factor in the renal bicarbonate wasting in these infants. The diagnosis of functional proximal tubular acidosis in such cases should not rely solely on the finding of an alkaline urine pH, because the distal tubular function is usually mature enough to acidify the urine once serum bicarbonate has decreased to its new threshold. Provided that liver function is normal, daily supplementation of bicarbonate in the form of sodium acetate, potassium acetate, or both normalizes blood pH and serum bicarbonate in these infants and also increases urine pH, aiding in the diagnosis. Once extracellular volume contraction occurs, these neonates generally achieve a positive bicarbonate balance (Ramiro-Tolentino et al, 1996), and supplementation becomes unnecessary. Other general guidelines in the fluid and electrolyte management of the immature preterm infant during the 1st week of life are (1) daily calculation of fluid balance and estimation of sodium balance; (2) daily measurements of body weight, serum electrolytes, blood urea nitrogen, creatinine, and plasma glucose; and (3) testing of urine samples for glucose, and osmolality or specific gravity. The frequency of testing and the addition of other tests, including the measurement of serum albumin concentration and osmolality, depend on the clinical status, severity of underlying disease, and fluid and electrolyte disturbance of the individual patient. Respiratory Distress Syndrome There is a well-established relationship between fluid and electrolyte imbalance and respiratory distress syndrome. Surfactant deficiency results in pulmonary atelectasis, elevated pulmonary vascular resistance, poor lung compliance, and decreased lymphatic drainage. In addition, preterm infants have low plasma oncotic and critical pulmonary capillary pressures and suffer pulmonary capillary endothelial injury from mechanical ventilation, oxygen administration, and perinatal hypoxia (Dudek and Garcia, 2001; Sola and Gregory, 1981). These abnormalities alter the balance of the Starling forces in the pulmonary microcirculation, leading to interstitial edema formation with further impairment in pulmonary functions. In the presurfactant era, an improvement in pulmonary function occurred only during the 3rd to 4th postnatal day. This improvement was usually preceded by a period of brisk diuresis characterized by small increases in glomerular filtration rate and sodium clearance and a larger rise in free water clearance (Costarino and Baumgart, 1991). Although the exact mechanism for this diuresis is not known, it is likely that improving endogenous surfactant production and capillary integrity promoted the recovery of the pulmonary capillary endothelium and lymphatic drainage. The ensuing changes in Starling forces then favored reabsorption of the hypotonic interstitial lung fluid into the circulation, and a delayed physiologic diuresis took place. Currently, with the routine use of surfactant and antenatal steroids, the pulmonary compromise and its consequences are less severe. However, because significant improvements in lung function take place only after the majority of the excess free water is excreted (Costarino and Baumgart, 1991), daily fluid intake should still be restricted to allow the extracellular volume contraction to take place. If this principle is not followed and a positive fluid balance occurs, preterm infants with respiratory distress syndrome are at higher risk for a more severe course of acute lung disease and have a higher incidence of patent ductus arteriosus, congestive heart failure, and necrotizing enterocolitis as well as a greater severity of the ensuing bronchopulmonary dysplasia. Antenatal administration of steroids and postnatal use of surfactants have clearly altered the course and clinical presentation of respiratory distress syndrome (Ballard and Ballard, 1995; Kari et al, 1994). Antenatal steroid administration accelerates maturation of organs including those involved in the regulation of fluid and electrolyte balance (Ballard and Ballard, 1995), whereas the use of exogenous surfactant decreases pulmonary capillary leak and edema formation (Carlton et al, 1995). Furthermore, surfactant administration does not alter the rate and timing of ductal closure (Reller et al, 1993), although it may affect the pattern of shunting through the ductus arteriosus in the acute period (Kaapa et al, 1993; Kluckow and Evans, 2000). Thus, these interventions generally enhance extracellular volume contraction and aid in the stabilization of fluid and electrolyte homeostasis in preterm neonates with respiratory distress syndrome. However, maintenance of a negative water and sodium balance during the first few days of life remains the cornerstone of fluid and electrolyte management in these infants (Tammela, 1995; Van Marter et al, 1990). On the basis of the events in the pathophysiology of pulmonary edema formation in these infants, the use of furosemide has long been suggested to promote a negative fluid balance and to directly inhibit pulmonary epithelial transport processes involved in edema formation in the lungs (Green et al, 1988; Yeh et al, 1984). However, furosemide induces only short-term improvements in pulmonary function in these patients, and no beneficial effects on long-term morbidity or mortality have been documented. Moreover, prophylactic use of the drug during the first postnatal days can lead to intravascular volume depletion with hypotension, tachycardia, and decreased peripheral perfusion as well as to acute and chronic disturbances in serum electrolytes and thus osmolality (Green et al, 1988; Shaffer and Weismann, 1992; Yeh et al, 1984). Furthermore, continued administration may be associated with an increased incidence of patent ductus arteriosus (Green et al, 1983). Therefore the use of furosemide during this period should be restricted to patients with oliguria of renal origin whose intravascular volume appears to be adequate. Chronic Lung Disease or Bronchopulmonary Dysplasia Low gestational age and birthweight, lack of antenatal steroid administration, severe respiratory distress syndrome with oxygen toxicity, volutrauma and barotrauma, air leak, inflammation, patent ductus arteriosus, and insufficient nutrition are among the known etiologic factors for the development of bronchopulmonary dysplasia. In addition, a high fluid and salt intake during the first weeks CHAPTER 31 Acid-Base, Fluid, and Electrolyte Management of life has been shown to increase the incidence and severity of chronic lung disease. Specifically, higher fluid intake and lack of appropriate weight loss in the first 10 days of life are associated with significantly higher risk for bronchopulmonary dysplasia, even after controlling for other known risk factors such as those listed previously (Oh et al, 2005). Therefore careful fluid and electrolyte ­management during the first weeks of life, allowing for the appropriate degree of weight loss, is of great importance in decreasing the incidence and severity of this condition (see Chapter 48). Patent Ductus Arteriosus A PDA has been associated with an increase in morbidity and mortality in epidemiological studies, especially in the infant with ELBW. However, trials aimed at prophylactic or therapeutic closure of the ductus have not demonstrated improvements in these outcomes. Several conditions, including hypoxemia, unstable cardiovascular status, metabolic acidosis, increases in extracellular volume, inflammatory mediators, and ductal prostaglandin synthesis, have been recognized to prolong patency of the ductus arteriosus (Hammerman, 1995; see Chapter 54). Accordingly, clinical management aimed at preventing the occurrence of ductal patency involves interventions that keep the cardiovascular status and oxygenation stable, minimize inflammation, restrict fluid intake, and maintain low levels of local prostaglandin synthesis by the administration of indomethacin (Clyman, 1996). Pharmacologic ductal closure with indomethacin is normally indicated during the first 2 postnatal weeks, because ductal sensitivity to prostaglandins rapidly diminishes thereafter (Clyman, 1996; Van Overmeire et al, 2000). Ibuprofen has been offered as an alternative inhibitor of cyclooxygenase, the enzyme responsible for prostaglandin synthesis, because of its equivalent efficacy in closing the symptomatic PDA with fewer adverse effects (Ohlsson et al, 2008). Specifically, ibuprofen may cause less renal dysfunction (Van Overmeire et al, 2000) and cerebral vasoconstriction (Mosca et al, 1997; Patel et al, 2000) when used in this manner. However, ibuprofen is not recommended for prophylactic use because it has an adverse effect on renal function in this setting and does not reduce grade III and IV IVH as indomethacin has been proven to do (Shah and Ohlssen, 2006). For more details, see Chapter 61. Under physiologic circumstances in the immediate postnatal period, renal prostaglandin production is increased to counterbalance the renal actions of vasoconstrictor and sodium- and water-retaining hormones released during labor and delivery (Bonvalet et al, 1987; Gleason, 1987). Compared with the renal function of the adult kidney in euvolemia, the neonatal kidney is more dependent on the increased production of vasodilatory and natriuretic prostaglandins, rendering it more sensitive to the vasoconstrictive and sodium- and water-retaining actions of cyclooxygenase inhibition. In the preterm infant, indomethacin administration has been shown to have clinically significant, although mostly transient, renal side effects because of decreased prostaglandin production through inhibition of cyclooxygenase. In the indomethacin-treated neonate, the unopposed renal vasoconstriction and sodium 381 and water reabsorption leads to decreases in renal blood flow and glomerular filtration rate and to increases in sodium and free water reabsorption. These side effects occur despite the diminishing left-to-right shunt through the closing ductus. Characteristic clinical findings include a rise in serum creatinine level, oliguria, and hyponatremia (Cifuentes et al, 1979). Hyponatremia occurs because the free water retention caused by the unopposed renal actions of high plasma vasopressin levels is out of proportion to the sodium retention induced by angiotensin and noradrenaline. This pattern of renal response can most likely be explained by the fact that in a preterm infant, the function of the distal tubule is more mature than that of the proximal tubule (Lumbers et al, 1988), leading to an expanded but somewhat hypotonic extracellular space. Therefore fluid management of the preterm infant receiving indomethacin must focus on maintaining an appropriately restricted fluid intake and avoiding extra sodium supplementation. As the prostaglandin inhibitory effects of indomethacin diminish following the last dose, renal prostaglandin production returns to normal, and the retained sodium and excess free water are usually rapidly excreted, especially with the improvement in the cardiovascular status as the ductal shunt decreases. Because furosemide increases prostaglandin production, the drug has been hypothesized to attenuate the renal side effects of indomethacin if the intravascular volume is judged to be adequate (Yeh et al, 1982). However, furosemide administration can increase the incidence of ductal patency (Green et al, 1983), so routine use of the drug might not be prudent in preterm infants treated with indomethacin, because of the theoretical risk of reopening the ductus arteriosus (Seri, 1995). Furthermore, the concomitant use of furosemide and indomethacin has been shown to worsen renal function, as evidenced by increases in serum creatinine and worsening hyponatremia without increasing overall urine output (Andriessen et al, 2009). As a result, furosemide should not be used early in life when the PDA is also being actively managed. In cases of hemodynamic instability or to avoid the potential effect of furosemide on ductal closure, dopamine infusion can be used to support the cardiovascular status and attenuate the indomethacin-induced oliguria (Cochran et al, 1989; Seri et al, 1984, 1993). However, there is controversy regarding the efficacy and clinical significance of this intervention (Baenziger et al, 1999; Fajardo et al, 1992; Seri, 1995). Growing Premature Infant With Negative Sodium Balance The 2- to 6-week-old growing preterm neonate who does not have significant chronic lung disease and is not undergoing diuretic treatment may have hyponatremia (serum sodium concentration 125 to 129 mEq/L) because of a relative sodium deficiency (Sulyok et al, 1979, 1985). Despite the low total body sodium and high activity of sodium retaining hormones, these infants continue to lose sodium in the urine mainly because of immature renal function. Although the infant is usually in a positive sodium balance, it is insufficient to compete with the increased sodium demand because of growth. The treatment of this condition is to provide extra sodium supplementation in the 382 PART VIII Care of the High-Risk Infant form of sodium chloride (usually 2 to 4 mEq/kg/day) to keep serum sodium values greater than 130 mEq/L. Shock and Edema In the uncompensated phase of shock, blood pressure is low; cardiac output may be low, normal, or high; effective circulating blood volume is usually decreased; transcapillary hydrostatic pressure is elevated; and capillary integrity and lymphatic drainage are impaired, resulting in edema formation and increased interstitial compliance. The latter condition further enhances fluid accumulation in the interstitium. The changes in the effective circulating blood volume also trigger the release of antidiuretic hormones, including catecholamines, renin-angiotensinaldosterone, and vasopressin, resulting in the retention of sodium and free water. The specific cause of shock (e.g., infection, asphyxia, myocardial insufficiency, hypovolemia) may independently contribute to this chain of events, further compromising fluid and electrolyte balance. In affected infants, treatment is directed at normalizing tissue perfusion and oxygen delivery by restoring effective intravascular volume, cardiac output, and renal function with the use of vasopressor and inotropic support, as well as with the judicious use of volume expanders while monitoring blood pressure, cardiac output, and changes in organ blood flow (see Chapter 51). In shock refractory to these therapies, early initiation of low-dose glucocorticoid and mineralocorticoid replacement may help to break the vicious cycle by improving capillary integrity and thus effective circulating blood volume, and by potentiating the cardiovascular response to vasopressors and inotropic agents (Seri et al, 2000). Syndrome of Inappropriate Anti-Diuretic Hormone Secretion SIADH may be associated with birth asphyxia, intracerebral hemorrhage, respiratory distress syndrome, pneumothorax, and the use of continuous positive-­pressure ventilation (El-Dahr and Chevalier, 1990; Leake, 1992). The syndrome is characterized by oliguria, free water retention, decreased serum sodium concentration and serum osmolality, increased urine concentration, and weight gain caused by edema formation. However, because the urinary concentrating capacity of the newborn is limited, a less than maximally diluted urine satisfies the diagnosis of SIADH in the presence of the other symptoms. The treatment is based on fluid and sodium restriction despite the oliguria and hyponatremia, as well as on appropriate circulatory and ventilatory support. The clinician must remember that total body sodium is normal, but TBW is elevated in such an infant, and that it is particularly dangerous to treat the hyponatremia caused by free water retention with large amounts of sodium. Because of their more immature renal functions, infants with extreme LBW during the first few weeks of life usually do not exhibit the full-blown syndrome despite their sometimes excessively high plasma vasopressin levels (Aperia et al, 1983). Diminished vasopressin secretion or complete unresponsiveness of the renal tubules to vasopressin results in polyuria, dilute urine production, and increased serum osmolality (Leake, 1992), otherwise known as diabetes insipidus (DI). This condition is not common in neonates, but can occur in association with central nervous system injury or disease, such as in meningitis or cerebral hemorrhage affecting the pituitary gland (central DI) or in an inherited form (nephrogenic DI). The treatment of neonates with this condition consists of allowing for adequate free water intake and the use of desmopressin. Surgical Conditions Surgery has a major effect on metabolism, fluid balance, and electrolyte balance in the newborn. Preterm infants with acute or chronic lung disease are especially sensitive and respond to the procedure with significant catabolic responses, increases in capillary permeability with the attendant shift of fluid into the interstitial space, and retention of sodium and free water (John et al, 1989). The retention of sodium and free water is secondary to the decrease in effective circulating blood volume and to the increased plasma levels of sodium- and water-retaining hormones, including catecholamines, renin-angiotensin-aldosterone, and vasopressin. Preoperative management has a significant effect on outcome and should be aimed at maintaining adequate effective circulating blood volume as well as cardiovascular and renal function. In preterm infants who have evidence of absolute or relative adrenal insufficiency (Watterberg, 2002), the provision of stress doses of steroids may be necessary. In the postoperative period, maintenance of the integrity of the cardiovascular system through the judicious use of volume expanders and pressor support, meticulous replacement of ongoing surgical and nonsurgical fluid and electrolyte losses, close monitoring, and intense and effective communication between the neonatal and surgical teams are essential to ensure a successful outcome. As capillary integrity improves, reabsorption and excretion of the expanded interstitial fluid volume occurs, with normalization in the secretion of hormones regulating fluid and electrolyte balance. At this time, the provision of maximized nutritional support becomes essential to restore the anabolic state and growth of the infant. The most commonly encountered surgical water losses occur when a nasogastric tube is placed under continuous suction to provide relief for the gastrointestinal tract in conditions such as necrotizing enterocolitis and postoperative management after abdominal surgery. Because these losses may be substantial, they should be monitored and a portion of it should be replaced every 6 to 12 hours to maintain appropriate water and electrolyte balance. However, free water retention often develops after surgery; therefore full replacement of the nasogastric free water loss is not usually recommended. The composition of the replacement solution depends on the electrolyte concentration of the fluid loss. Gastric fluid usually contains 50 to 60 mEq/L of sodium chloride, and therefore, 0.45% sodium chloride with potassium is normally used as the fluid of choice for replacement. See Table 31-3 for estimated electrolyte compositions of body fluids. CHAPTER 31 Acid-Base, Fluid, and Electrolyte Management ACID-BASE BALANCE PHYSIOLOGY OF ACID-BASE BALANCE REGULATION Like adults, newborns must maintain their extracellular pH, or hydrogen ion concentration, within a narrow range. A normal pH is essential for intact functioning of all enzymatic processes and, therefore, the intact functioning of all organ systems of the body. Newborns are subjected to many stresses that can affect their acid-base balance. In addition, neonates, especially if they are premature, have a limited ability to compensate for acid-base alterations; therefore acid-base disturbances are common in the neonatal period. An understanding of the principles of acid-base regulation is essential for proper diagnosis and treatment of these disturbances. In healthy humans, the normal range of ECF hydrogen ion concentration is 35 to 45 mEq/L. Because pH is defined as the negative logarithm of hydrogen ion concentration (pH = −log[H+]), these values of hydrogen ion concentration correspond to a pH range of 7.35 to 7.45. Acidosis is a downward shift in pH to less than 7.35, and alkalosis is an upward shift in pH to more than 7.45. Alterations in normal pH are resisted by complex physiologic regulatory mechanisms. The main systems that maintain pH are the body’s buffer systems, the respiratory system, and the kidneys. Some of these systems respond immediately to sudden alterations in hydrogen ion concentration, whereas others respond more slowly to changes but maintain the overall balance between acid and base production, intake, metabolism, and excretion over the long term. The physiologic regulatory systems that respond immediately to changes in acid-base balance include the various intracellular and extracellular buffers as well as the lungs. A buffer is a substance that can minimize changes in pH when acid or base is added to the system. The extracellular buffers, which include the bicarbonate-carbonic acid system, phosphates, and plasma proteins, act rapidly to return the extracellular pH toward normal. The intracellular buffers, which include hemoglobin, organic phosphates, and bone apatite, act more slowly and require several hours to reach maximal capacity. The most important extracellular buffer is the plasma bicarbonate–carbonic acid buffer system, in which the acid component (carbonic acid [H2CO3]) is regulated by the lungs, and the base component (bicarbonate [HCO3−]) is regulated by the kidneys. The buffer equation is: H + + HCO3− ↔ H2 CO3 ↔ H2 O + (CO2 )d where (CO2)d represents the dissolved carbon dioxide. At equilibrium, the amount of (CO2)d exceeds that of H2CO3 by a factor of 800:1; therefore for practical purposes, (CO2)d and H2CO3 can be treated interchangeably. The fact that CO2 excretion can be controlled by the respiratory system markedly improves the efficiency of this buffer system at physiologic pH. The enzyme carbonic anhydrase allows rapid interconversion of H2CO3 to H2O and CO2. If the hydrogen ion (H+) concentration increases for any reason, hydrogen combines with HCO3−, driving the buffer reaction toward greater production of H2CO3 and 383 CO2. Carbon dioxide crosses the blood-brain barrier and stimulates central nervous system chemoreceptors, leading to increased alveolar ventilation and decreased concentration of extracellular CO2. This respiratory compensation begins within minutes after a pH change and is complete within 12 to 24 hours. A similar compensation occurs in response to a decrease in H+ concentration, leading to decreased alveolar ventilation and a resultant increase in extracellular CO2. The relationship of the two components of the bicarbonate–carbonic acid buffer system to pH is expressed by the Henderson-Hasselbalch equation: pH = pK + log [HCO3− ] [H2 CO3 ] Because H2CO3 is in equilibrium with the dissolved CO2 in the plasma, and because the amount of dissolved CO2 depends on the partial pressure of CO2, the equation can be modified as: pH = pK + log [HCO3− ] × PaCO2 0.03 Both the original equation and the modified equation are clinically difficult to use; therefore the modified Henderson-Hasselbalch equation can be rewritten as the Henderson equation without logarithms for easier clinical use. [H+ ] = 24 × PaCO2 [HCO3− ] This last equation clearly indicates the clinically most important aspect of acid-base regulation by the bicarbonate-carbonic acid buffer system, that the change in the ratio of Paco2 to HCO3− concentration, and not in their absolute values, determines the direction of change in H+ concentration and thus in pH. The status of the plasma bicarbonate-carbonic acid buffer system can be monitored easily by serial blood gas measurements, making understanding of this buffer system important in clinical care. The physiologic regulation system that responds more slowly to changes in acid-base balance is the renal system. There must be a long-term balance between net acid increase caused by intake and production and net acid decrease caused by excretion and metabolism. Although infant formula and protein-containing intravenous fluids have small amounts of preformed acid, most of the daily acid load is derived from metabolism. A large amount of the acid produced is in the form of the volatile H2CO3 that can be excreted in the lungs. Nonvolatile or fixed acids are also produced, which must be excreted through the kidneys. Nonvolatile acids normally are sulfuric acid produced in the metabolism of the amino acids methionine and cysteine as well as smaller contributions from phosphoric acid, lactic acid, hydrochloric acid, and incompletely oxidized organic acids. In addition to the excretion of nonvolatile acids, however, the kidneys have a role in long-term acid-base regulation by controlling renal HCO3− excretion. 384 PART VIII Care of the High-Risk Infant Two regions of the kidney act to achieve urinary acidification—the proximal tubule and the collecting tubule. The proximal tubule acidifies the urine by two mechanisms. The first mechanism is by the reabsorption of any HCO3− already present in the blood that is being constantly filtered through the glomeruli. The proximal tubule reabsorbs 60% to 80% of all filtered HCO3− and performs this role through an exchange of Na+ for H+ across the luminal membrane of the proximal tubular cells via the Na+/ H+ exchanger. The excreted H+ combines with filtered HCO3−, producing H2CO3 through the activity of carbonic anhydrase in the cellular brush border. The H2CO3 is then quickly converted to CO2, which crosses into the tubular cell, where HCO3− is regenerated and reabsorbed back into the blood stream, probably in exchange for chloride (Cl–). The regenerated H+ ion reenters the cycle at the Na+/H+ exchanger. The second mechanism by which the proximal tubule acidifies urine is by the production of ammonia (NH3). Inside the tubular cell, NH3 is produced by the deamination of glutamine. The NH3 is secreted into the tubular lumen, where it combines with and traps free H+ to form ammonium (NH4+). The remaining urinary acidification occurs mostly in the collecting tubule. H+ secretion in this region of the kidney is sufficient to combine with or titrate any remaining filtered HCO3− or any filtered anions, such as phosphate and sulfate. Hydrogenated phosphate and sulfate anions produce the titratable acid of the urine. The collecting tubule also takes up NH3 from the medullary interstitium and secretes it into the urine, where again it can combine with and trap H+ as NH4+. This urinary NH4+ can act as a cation and be excreted with urinary anions such as Cl−, PO4−, and SO4−, thereby preventing loss of cations such as Na+, Ca++, and K+. Total acid secretion in the kidney can be represented by Titratable acid + NH4 + − HCO3− and under normal conditions should equal the net production of acid from diet and metabolism that is not excreted in the form of CO2 through the lungs. In adults, the steady state for renal compensation for respiratory alkalosis is reached within 1 to 2 days and for respiratory acidosis within 3 to 5 days. Newborns are able to compensate for acidemia through the previously described renal mechanisms, although the renal response to acid loads is limited, especially in premature infants born before 34 weeks’ gestation. Reabsorption of HCO3− in the proximal tubule and distal tubular acidification are also decreased, with a fairly rapid gestational age-­ dependent maturation of these functions after birth (Jones and Chesney, 1992). To accomplish the tight regulation of pH necessary for survival, H+ ions generated in the form of the volatile acid H2CO3 are excreted by the lungs as CO2. H+ ions generated in the form of nonvolatile acids are buffered rapidly by extracellular HCO3− and more slowly by intracellular buffers. HCO3− is then replenished by the kidneys via the reabsorption of much of the filtered HCO3− and by the excretion of H+ in the urine as NH4+ and titratable acids. DISTURBANCES OF ACID-BASE BALANCE IN THE NEWBORN General Principles The evaluation of the acid-base status in a newborn is one of the most common laboratory assessments made in the neonatal intensive care unit. The status of this system can be monitored with blood gas measurements and should be the starting point for the evaluation of any acid-base disorder. In the blood gas measurement, the pH and Paco2 levels are directly measured; from these, the HCO3− level and base excess or deficit are calculated. The whole blood buffer base, defined as the sum of the HCO3− and non-HCO3− buffer systems, is another important blood gas value used in evaluating acid-base disturbances. The difference between the observed whole blood buffer base of any blood gas sample and the expected normal buffer base of that sample is called the base excess or base deficit. The base excess and base deficit give an accurate measure of the amount of strong acid or base, respectively, that would be needed to titrate the pH back to normal once the respiratory contribution of the acid-base disturbance is also corrected. For example, a base excess of 10 mEq/L indicates that there is an additional 10 mEq of base per liter (or loss of 10 mEq of H+ per liter) that is contributing to the acid-base abnormality. Conversely, a base deficit of 10 mEq/L indicates there is relatively more acid (or less base) in the ECF than expected after accounting for the effect of Pco2 on pH. Acid-base disorders are classified according to their cause as being either metabolic or respiratory. Metabolic acidosis occurs as a result of the accumulation of increased amounts of nonvolatile acid or decreased amounts of HCO3− in the ECF. Metabolic alkalosis occurs as a result of increased amounts of HCO3− in the ECF. Respiratory acidosis is caused by hypoventilation and decreased excretion of volatile acid (CO2), whereas respiratory alkalosis is caused by hyperventilation and increased excretion of volatile acid (CO2). Acid-base disorders are also classified according to the number of conditions causing the disorder. When only one primary acid-base abnormality and its compensatory mechanisms occur, the disorder is classified as a simple acid-base disorder. When a combination of simple acid-base disturbances occurs, the patient has a mixed (or complex) acid-base disorder. Because secondary physiologic regulatory mechanisms often compensate for the alteration in pH caused by primary disturbances, it is sometimes difficult to differentiate simple from mixed disorders or even a simple disorder from its resulting compensation. One important principle that allows the determination of primary acid-base disturbance is that the compensatory regulatory mechanisms do not completely normalize the pH. Nomograms, such as the one shown in Figure 31-6, can help in the diagnosis of the primary disturbance. The nomogram describes the 95% confidence limits of the expected compensatory response to a primary abnormality in either Paco2 or HCO3−. Table 31-5 summarizes the expected respiratory and metabolic compensatory mechanisms for primary acid-base disorders (Brewer, 1990). If the compensation in a given patient differs from CHAPTER 31 Acid-Base, Fluid, and Electrolyte Management that predicted in Figure 31-6 or Table 31-5, the patient either has not had enough time to compensate for a simple acid-base disturbance or has a mixed acid-base disorder. Furthermore, the complete correction of an acid-base disturbance occurs only when the underlying process responsible for the abnormality has been treated effectively. For identification of the primary disturbance, the analysis of blood gas values must be considered in light of the patient’s history and physical findings and with an understanding of expected compensatory responses. Further laboratory evaluation is indicated if the problem is not immediately obvious or if the response to therapy is not as expected. The evaluation of the acid-base disturbance should always involve efforts to determine the underlying cause of the disturbance, because adequate treatment requires correction of the underlying disorder, if possible. Transitional Physiology after Birth Arterial Plasma Bicarbonate (mmol/Liter) As part of a discussion of normal physiology, it is important to understand the in utero environment just before delivery of the newborn and its effects on neonatal 60 56 52 48 44 40 36 32 28 24 20 16 12 8 4 0 100 90 120 110 80 70 60 50 40 35 Metabolic Alkalosis Chronic Respiratory Acidosis 25 7.2 Metabolic Acidosis Metabolic acidosis is a common problem, particularly in the critically ill newborn. Metabolic acidosis occurs when the drop in pH is caused by the accumulation of acid other than H2CO3 by the ECF, resulting in loss of available HCO3−, or by the direct loss of HCO3− from body fluids. Patients who have metabolic acidosis are divided into those with an elevated anion gap and those with a normal anion gap. The anion gap reflects the unaccounted acidic anions and certain cations in the ECF. The unmeasured anions normally include the serum proteins, phosphates, sulfates, and organic acids, whereas the unaccounted cations are the serum potassium, calcium, and magnesium. Thus, in clinical practice, the anion gap is estimated using the formula − Anion gap =[Na + ] serum − ( [Cl ] serum +[HCO3− ] serum ) Normal 15 Acute Respiratory Alkalosis 7.3 10 Chronic Resp. Alkalosis Metabolic Acidosis 7.1 acid-base analysis shortly after birth. Hyperventilation of pregnancy is a known phenomenon with corresponding maternal PCO2 levels of approximately 31 to 34 mm Hg (Thorp and Rushing, 1999). This relative respiratory alkalosis in the mother is compensated for by a corresponding metabolic acidosis in the mother and, ­therefore, fetus. As a result, umbilical arterial blood gases have a normal pH range of 7.20 to 7.28 with a corresponding base deficit ranging from 2.7 ± 2.8 mEq/L (SD) to 8.3 ± 4.0 mEq/L (SD) (Riley and Johnson, 1993; Sykes et al, 1982). In other words, a mild metabolic acidosis in the newborn shortly after birth can be expected and explained by normal physiology. 20 Acute Respiratory Acidosis 7 30 7.4 7.5 PCO (mm Hg) 2 7.6 7.7 7.8 Arterial Blood pH FIGURE 31-6 Acid-base nomogram illustrating the 95% confidence limits for compensatory responses to primary acid-base disorders. (From Cogan MG, Rector FC Jr: Acid-base disorders. In Brenner BM, Rector FC Jr, editors: The kidney, Philadelphia, 1986, WB Saunders.) The normal range of the serum anion gap in newborns is 8 to 16 mEq/L, with slightly higher values in very premature newborns. Accumulation of strong acids because of increased intake, increased production, or decreased excretion results in an increased anion gap acidosis, whereas loss of HCO3− or accumulation of H+ results in a normal anion gap acidosis. A decrease in serum potassium, calcium, and magnesium concentrations, an increase in serum protein concentration, or a falsely elevated serum sodium concentration can also result in an increased anion gap in the absence of metabolic acidosis. In clinical practice, although TABLE 31-5 Expected Compensatory Mechanisms Operating in Primary Acid-Base Disorders Acid-Base Disorder Primary Event Compensation Rate of Compensation ↓ Pco2 For 1 mEq/L ↓ [HCO3–], Pco2 ↓ by 1-1.5 mm Hg Metabolic acidosis Normal anion gap ↓ [HCO3–] For 1 mEq/L ↓ [HCO3–], Pco2 ↓ by 1-1.5 mm Hg ↑ Acid production ↑ Acid intake ↓ Pco2 Respiratory acidosis Acute (<12-24 h) Chronic (3-5 days) ↑ Pco2 ↑ Pco2 ↑ [HCO3 ↑ [HCO3–] For 10 mm Hg ↑ Pco2, [HCO3–] ↑ by 1 mEq/L For 10 mm Hg ↑ Pco2, [HCO3–] ↑ by 4 mEq/L Respiratory alkalosis Acute (<12 h) Chronic (1-2 days) ↓ Pco2 ↓ Pco2 ↓ [HCO3–] ↓ [HCO3–] For 10 mm Hg ↑ Pco2, [HCO3–] ↑ by 1-3 mEq/L For 10 mm Hg ↓ Pco2, [HCO3–] ↓ by 2-5 mEq/L Increased anion gap Metabolic alkalosis 385 ↑ [HCO3–] ↑ Pco2 –] Modified from Brewer ED: Disorders of acid-base balance, Pediatr Clin North Am 37:430-447, 1990. For 1 mEq/L ↑ [HCO3–], Pco2 ↑ by 0.5-1 mm Hg 386 PART VIII Care of the High-Risk Infant BOX 31-3 C  ommon Causes of Metabolic Acidosis INCREASED ANION GAP Lactic acidosis caused by tissue hypoxia ll Asphyxia, hypothermia, shock ll Sepsis, respiratory distress syndrome ll Inborn errors of metabolism ll Congenital lactic acidosis ll Organic acidosis ll Renal failure ll Late metabolic acidosis ll Toxins (e.g., benzyl alcohol) ll NORMAL ANION GAP Renal bicarbonate loss ll Bicarbonate wasting caused by immaturity ll Renal tubular acidosis ll Carbonic anhydrase inhibitors ll Gastrointestinal bicarbonate loss ll Small bowel drainage: ileostomy, fistula ll Diarrhea ll Extracellular volume expansion with bicarbonate dilution ll Aldosterone deficiency ll Excessive chloride in intravenous fluids ll BOX 31-4 Inborn Errors of Metabolism ­Associated With Metabolic ­Acidosis ll ll ll ll Primary lactic acidosis Organic acidemias Pyruvate carboxylase deficiency Pyruvate hydroxylase deficiency ll ll ll Galactosemia Hereditary fructose intolerance Type I glycogen storage disease a serum anion gap value greater than 16 mEq/L is highly predictive of the presence of lactic acidosis and a value less than 8 mEq/L is highly predictive of the absence of lactic acidosis, an anion gap value between 8 and 16 mEq/L cannot be used to differentiate between lactic and nonlactic acidosis in the critically ill newborn (Lorenz et al, 1999). Therefore if the anion gap is within this high normal range and lactic acidosis is suggested, measuring serum lactate is indicated. An increased anion gap metabolic acidosis in the newborn is most commonly caused by lactic acidosis secondary to tissue hypoxia, as seen in asphyxia, hypothermia, severe respiratory distress, sepsis, and other severe neonatal illnesses. Other important but much less common causes of an increased anion gap metabolic acidosis in the neonatal period are inborn errors of metabolism, renal failure, and intake of toxins (Box 31-3). Box 31-4 lists inborn errors of metabolism that can manifest as increased anion gap metabolic acidosis in the newborn period. In the syndrome of late metabolic acidosis of prematurity, first described in the 1960s, otherwise healthy premature infants at several weeks of age demonstrated mild to moderate increased anion gap acidosis and decreased growth. All the infants were receiving high-protein cow’s milk formula, and they demonstrated higher net acid excretion compared with controls. This type of late metabolic acidosis is rarely seen, probably because of the use of special premature infant formulas and changes in regular formulas with decreased casein-to-whey ratios and lower fixed acid loads. A normal anion gap metabolic acidosis most commonly occurs in the newborn as a result of HCO3− loss from the extracellular space through the kidneys or the gastrointestinal tract. Hyperchloremia develops with the HCO3− loss because a proportionate rise in serum chloride concentration must occur to maintain the ionic balance or to correct the volume depletion in the extracellular compartment. The most common cause of normal anion gap metabolic acidosis in the preterm newborn is a mild, developmentally regulated, proximal renal tubular acidosis with renal HCO3− wasting. In infants with this disorder, the serum HCO3− usually stabilizes at 14 to 18 mEq/L in the early postnatal period. The urinary pH is normal once the serum HCO3− falls to this level, because the impairment in proximal tubular HCO3− reabsorption is not associated with an impaired distal tubular acidification of similar magnitude (Jones and Chesney, 1992). The diagnosis of this temporary cause of acidosis can be established by the recurrence of a urinary alkaline pH when serum HCO3− is raised above the threshold after HCO3− or acetate supplementation. Even term newborns have a lower renal threshold for HCO3−, with normal plasma HCO3− levels in the range of 17 to 21 mEq/L. In most infants, plasma HCO3− increases to adult levels over the first year as the proximal tubule matures. Other common causes of normal anion gap metabolic acidosis seen in neonatal intensive care units are gastrointestinal HCO3− losses, often caused by increased ileostomy drainage, diuretic treatment with carbonic anhydrase inhibitors, and dilutional acidosis with rapid expansion of the extracellular space through the use of non-HCO3− solutions in the hypovolemic newborn. The presence of metabolic acidosis in the newborn should be suggested by the clinical presentation and the history of predisposing conditions, including perinatal depression, respiratory distress, blood or volume loss, sepsis, and congenital heart disease associated with poor systemic perfusion or cyanosis. Metabolic acidosis is confirmed by blood gas measurements. The cause of metabolic acidosis is often readily discernible from the history and physical examination. Specific laboratory evaluation of electrolytes, renal function, lactate, and serum and urine amino acids may be undertaken, depending on the diagnosis that is suggested clinically. Figure 31-7 shows a simple flow diagram outlining an approach to diagnosis of metabolic acidosis in the newborn. It is important to remember that infants might not manifest an increased anion gap in the setting of lactic acidosis and thus require the direct measurement of lactate when a lactic acidosis is suspected (Lorenz et al, 1999). The morbidity and mortality of metabolic acidosis depend on the underlying pathologic process, the severity of the acidosis, and the responsiveness of the process to clinical management. By far, the most important intervention for an infant with a metabolic acidosis is to identify the pathologic process contributing to the acidosis and to take measures to correct it. The administration of base, such as sodium bicarbonate, as supportive therapy for metabolic acidosis is unproven in its efficacy (Aschner and Poland, 2008). CHAPTER 31 Acid-Base, Fluid, and Electrolyte Management ANION GAP Increased (16) Urinary ketones Moderate-large Absent-small Organic acidemias Lactic acidosis · Congenital · Hypoxia · Poor tissue Normal (16) Blood ammonia Normal RTA · Transient · Permanent Increased Urea cycle defects perfusion FIGURE 31-7 Diagnostic approach of increased anion gap and normal anion gap metabolic acidosis in the newborn. Infants with lactic acidosis may not have an increased anion gap and lactate should be measured directly if suspected by the history and physical examination. RTA, Renal tubular acidosis. (Adapted from Lorenz JM, Kleinman LI, Markarian K, et al: Serum anion gap in the differential diagnosis of metabolic acidosis in critically ill newborns, J Pediatr 135:751-755, 1999.) The use of base administration is supported by findings on the beneficial cardiovascular effects of sodium bicarbonate in preterm newborns with an arterial pH of less than 7.25 and term newborns with an arterial pH of less than 7.30 (Fanconi et al, 1993). The use of sodium bicarbonate in this study was associated with an increase in myocardial contractility and a reduction in afterload, albeit transient. However, there is concern for harm associated with the administration of base, including increased mortality and intraventricular hemorrhage (Papile et al, 1978; Simmons et al, 1974; Usher, 1967), increased cerebral blood volume regardless of rate of administration (van Alfen-van der Velden et al, 2006), and decreased intracellular pH with cellular injury (Lipshultz et al, 2003). If therapy with base is warranted, the clinician has three options: sodium bicarbonate, sodium (or potassium) acetate, and tromethamine. Sodium bicarbonate is the most widely used buffer in the treatment of metabolic acidosis in the neonatal period. Bicarbonate should not be given if ventilation is inadequate, because its administration results in an increase in Paco2 with no improvement in pH and an increase in intracellular acidosis. Therefore sodium bicarbonate should be administered slowly and in diluted form only to newborns with documented metabolic acidosis and adequate alveolar ventilation. Once a blood gas measurement has been obtained, the dose of sodium bicarbonate required to fully correct the pH can be estimated with the use of the following formula: Dose of NaHCO3 (mEq) = Base deficit (mEq/L) × Body weight (kg) × 0.3 Sodium bicarbonate is thought to be confined mostly to the ECF compartment. Although there are controversies regarding the actual bicarbonate space in humans, the 30% of total body weight in the formula represents its estimated volume of distribution in the neonate. Most clinicians would use half of the calculated total correction dose for initial therapy to avoid overcorrection of 387 metabolic acidosis. Subsequent doses of sodium bicarbonate are then based on the results of additional blood gas measurements. When clinicians are faced with a chronic metabolic acidosis caused by a prematurity-related proximal renal tubular acidosis with bicarbonate wasting, many choose to replace these losses over time. In this instance, either sodium or potassium acetate can be used as an alternative to sodium bicarbonate. Sodium acetate is a conjugate base of a weak acid (acetic acid) with a pKb of 9.25. pKb is a measure of the strength of a base, which depends on its base dissociation constant. It has been shown in one study to be an effective alternative to sodium bicarbonate in correcting this type of acid-base abnormality when added to parenteral nutrition (Peters et al, 1997). The median doses of acetate used in this randomized controlled trial were 2.6 mmol/kg/day on day 4 of life and 4.1 mmol/kg/day on day 8 of life. Infants randomized to acetate had an increased base excess and pH and increased Pco2, and they received less bicarbonate boluses compared with control infants. In certain clinical situations, tromethamine can be used as an alternative buffer to sodium bicarbonate. The theoretical advantages of tromethamine over sodium bicarbonate in the treatment of metabolic acidosis of the newborn include its more rapid intracellular buffering capability, its ability to lower Paco2 levels directly, and the lack of an increase in the sodium load (Schneiderman et al, 1993). Tromethamine lowers Paco2 by covalently binding H+ and thus shifting the equilibrium of the reaction H+ + HCO3− ↔ H2 CO3 ↔ H2 O + CO2 to the left, resulting in a decrease in CO2 and an increase in HCO3−. Because the end-product (chelated tromethamine) is a cation that is excreted by the kidneys, oliguria is a relative contraindication to the repeated use of this buffer. Tromethamine administration also has been associated with the development of acute respiratory depression, most likely secondary to an abrupt decrease in Paco2 levels as well as from rapid intracellular correction of acidosis in the cells of the respiratory center (Robertson, 1970). Furthermore, because hypocapnia is associated with decreases in brain blood flow and a higher incidence of white matter damage, especially in the immature preterm neonate, close monitoring of Paco2 is of paramount importance when tromethamine is being used. Finally, when large doses of tromethamine are administered, hyponatremia (Seri et al, 1998b), hypoglycemia, hyperkalemia, an increase in hemoglobin oxygen affinity, and diuresis followed by oliguria can occur. Because the tromethamine solution is hyperosmolar, and because rapid infusion of tromethamine can also lower blood pressure and intracranial pressure (Duthie et al, 1994), slow infusion is recommended. The suggested initial dose is 1 to 2 mEq/kg or 3.5 to 6 mL/kg given intravenously using the 0.3 M solution, with the rate of administration not exceeding 1 mL/kg/min. Once a blood gas measurement has been obtained, the dose of tromethamine required to raise the pH can be estimated using the formula Dose of tromethamine (mL) = Base deficit (mEq/L) × Body weight (kg) 388 PART VIII Care of the High-Risk Infant Finally, during the correction of metabolic acidosis, particular attention should be paid to ensuring an appropriate potassium balance. Because potassium moves from the intracellular to the extracellular space in exchange for H+ when acidosis occurs, the presence of a total body potassium deficit might not be appreciated during metabolic acidosis. Hypokalemia may become evident only as the pH increases and potassium returns to the intracellular space. Furthermore, intracellular acidosis cannot be completely corrected until the potassium stores are restored. Therefore close monitoring of serum electrolytes and potassium supplementation are important during the correction of metabolic acidosis in the sick newborn. Respiratory Acidosis Respiratory acidosis occurs when a primary increase in Paco2 develops secondary to impairments in alveolar ventilation that result in an arterial pH of less than 7.35. Primary respiratory acidosis is a common problem in newborns, and causes include hyaline membrane disease, pneumonia owing to infection or aspiration, patent ductus arteriosus with pulmonary edema, chronic lung disease, pleural effusion, pneumothorax, and pulmonary hypoplasia. The initial increase in Paco2 is buffered by the non-HCO3− intracellular buffers without noticeable renal compensation for at least 12 to 24 hours (see Table 31-5). Renal metabolic compensation reaches its maximum levels within 3 to 5 days, and its effectiveness in the newborn is influenced mainly by the functional maturity of proximal tubular HCO3− transport. Management of respiratory acidosis is directed toward improving alveolar ventilation and treating the underlying disorder. For sick newborns, adequate ventilation must often be provided by mechanical ventilation. In severe respiratory acidosis, tromethamine can be used to raise pH, because it lowers CO2 levels. Tromethamine, however, produces only a transient decrease in Paco2, and toxic doses would quickly be reached if it were used to buffer all the CO2 produced by metabolism over a sustained period. Therefore tromethamine should be used only as a temporizing measure in severe respiratory acidosis until alveolar ventilation can be improved. Metabolic Alkalosis Metabolic alkalosis is characterized by a primary increase in the extracellular HCO3− concentration sufficient to raise the arterial pH above 7.45. In the newborn, metabolic alkalosis occurs when there is a loss of H+, a gain of HCO3−, or a depletion of the extracellular volume with the loss of more chloride than HCO3−. It is important to understand that metabolic alkalosis generated by any of these mechanisms can be maintained only when factors limiting the renal excretion of HCO3− are also present. Metabolic alkalosis can result from a loss of H+ from the body, from either the gastrointestinal tract or the kidneys, that induces an equivalent rise in the extracellular HCO3− concentration. The most common causes of this type of metabolic alkalosis in the newborn period are continuous nasogastric aspiration, persistent vomiting, and diuretic treatment. Less common causes of H+ losses are congenital chloride-wasting diarrhea, certain forms of congenital adrenal hyperplasia, hyperaldosteronism, posthypercapnia, and Bartter syndrome. Metabolic alkalosis can also result from a gain of HCO3−, such as occurs during the administration of buffer solutions to the newborn. In the past, a metabolic alkalosis was intentionally created when sodium bicarbonate or tromethamine was used to maintain an alkaline pH to decrease pulmonary vasoreactivity in infants with persistent pulmonary hypertension, a practice not recommended anymore. Currently, iatrogenically produced metabolic alkalosis is primarily unintentional and due to chronic excessive administration of HCO3−, lactate, citrate, or acetate in intravenous fluids and blood products. Because excretion of HCO3− is normally not limited in the newborn, metabolic alkalosis resulting from HCO3− gain alone should rapidly resolve after administration of HCO3− is discontinued. However, if the alkalosis is severe and urine output is limited, inhibition of the carbonic anhydrase enzyme by the administration of acetazolamide may enhance elimination of HCO3−. Metabolic alkalosis can also result from a loss of ECF containing disproportionately more chloride than HCO3−—the so-called contraction alkalosis. During the diuretic phase of normal postnatal adaptation, preterm and term newborns retain relatively more HCO3− than chloride (Ramiro-Tolentino et al, 1996). The obvious clinical benefits of allowing this physiologic extracellular volume contraction to occur, especially in the critically ill newborn, clearly outweigh the clinical importance of a mild contraction alkalosis that develops after recovery. No specific treatment is needed in such cases, because with the stabilization of the extracellular volume and renal function after recovery, acid-base balance rapidly returns to normal. Contraction alkalosis due to other causes, however, may require treatment. For metabolic alkalosis to persist, factors limiting the renal excretion of HCO3− must be present. The kidneys are usually effective in excreting excess HCO3−, but this ability can be limited under certain conditions, such as decreased glomerular filtration rate, increased aldosterone production, and the more common clinical situation of volume contraction–triggered metabolic alkalosis with potassium deficiency. In the last condition, there is a direct stimulation of Na+ reabsorption coupled with H+ loss in the proximal tubule, and an indirect stimulation of H+ loss in the distal nephron by the increased activity of the renin-angiotensin-aldosterone system. Contraction alkalosis responds to administration of saline to replace the intravascular volume in conjunction with additional potassium supplementation to account for renal potassium wasting. In the other disorders, however, the primary problem of reduced glomerular filtration rate or elevated aldosterone must be treated for the alkalosis to resolve. One of the most commonly encountered clinical scenarios of chronic metabolic alkalosis actually occurs in the form of a mixed acid-base disorder in a preterm infant with chronic lung disease on long-term diuretic treatment. Such a newborn initially has a chronic respiratory acidosis that is partially compensated for by renal HCO3− retention. Prolonged or aggressive use of diuretics CHAPTER 31 Acid-Base, Fluid, and Electrolyte Management can lead to total-body potassium depletion and contraction of the extracellular volume, thus exacerbating the metabolic alkalosis. By stimulating proximal tubular Na+ reabsorption and thus H+ loss, distal tubular H+ secretion, and renal ammonium production, the diuretic-induced hypokalemia contributes to the severity and maintenance of the metabolic alkalosis. Furthermore, metabolic alkalosis per se worsens hypokalemia, because potassium moves intracellularly to replace hydrogen as the latter shifts into the extracellular space. Although the serum potassium concentration may be decreased, the serum levels in the newborn do not accurately reflect the extent of total-body potassium deficit because potassium is primarily an intracellular ion, with approximately 98% of the total body potassium being in the intracellular compartment. In addition, the condition is often accompanied by marked hypochloremia and hyponatremia. Hyponatremia occurs in part because sodium shifts into the intracellular space to compensate for the depleted intracellular potassium. If the alkalosis is severe, alkalemia (pH >7.45) can supervene and result in hypoventilation. In this situation, potassium chloride, and not sodium chloride supplementation, reverses hyponatremia and hypochloremia, corrects hypokalemia and metabolic alkalosis, and increases the effectiveness of diuretic therapy. Because chloride deficiency is the predominant cause of the increased pH, ammonium chloride or arginine chloride also corrects the alkalosis. These agents do not affect the other electrolyte imbalances such as the hypokalemia, so they should not be the only therapy given. It is important to keep ahead of the potassium losses in infants receiving long-term diuretic therapy, rather than to attempt to replace potassium after intracellular depletion has occurred. Because the rate of potassium repletion is limited by the rate at which potassium moves intracellularly, correction of total body potassium deficits can require days to weeks. In addition, there is also a risk of acute hyperkalemia if serum potassium levels are driven too high during repletion, particularly in newborns in whom an acute respiratory deterioration may occur, with worsened respiratory acidosis and the subsequent movement of potassium from the intracellular to the extracellular space. The routine use of potassium chloride supplementation and close monitoring of serum sodium, chloride, and potassium levels are therefore recommended during long-term diuretic therapy to prevent these common iatrogenic problems. 389 Respiratory Alkalosis When a primary decrease in Paco2 results in an increase in the arterial pH beyond 7.45, respiratory alkalosis develops. The initial hypocapnia is acutely titrated by the intracellular buffers, and metabolic compensation by the kidneys returns pH toward normal within 1 to 2 days (see Table 31-5). Interestingly, respiratory alkalosis is the only simple acid-base disorder in which, at least in adults, the pH can completely be normalized by the compensatory mechanisms (Brewer, 1990). The cause of respiratory alkalosis is hyperventilation, which in the spontaneously breathing newborn is most often caused by fever, sepsis, retained fetal lung fluid, mild aspiration pneumonia, or central nervous system disorders. In the neonatal intensive care unit, the most common cause of respiratory alkalosis is iatrogenic secondary to hyperventilation of the intubated newborn. Because findings suggest an association between hypocapnia and the development of periventricular leukomalacia (Okumura et al, 2001; Wiswell et al, 1996) and chronic lung disease (Garland et al, 1995) in ventilated preterm infants, avoidance of hyperventilation during resuscitation and mechanical ventilation is of utmost importance in the management of sick preterm newborns. The treatment of neonatal respiratory alkalosis consists of the specific management of the underlying process causing hyperventilation. SUGGESTED READINGS Aschner JL, Poland RL: Sodium bicarbonate: basically useless therapy, Pediatrics 122:831, 2008. Avner ED: Clinical disorders of water metabolism: hyponatremia and hypernatremia, Pediatr Ann 24:23, 1995. Brewer ED: Disorders of acid-base balance, Pediatr Clin North Am 37:430, 1990. El-Dahr SS, Chevalier RL: Special needs of the newborn infant in fluid therapy, Pediatr Clin North Am 37:323, 1990. Hartnoll G: Basic principles and practical steps in the management of fluid balance in the newborn, Semin Neonatol 8:307, 2003. Lorenz J: Fluid and electrolyte therapy in the very low-birthweight neonate, Neoreviews 9:e02, 2008. Modi N: Management of fluid balance in the very immature neonate, Arch Dus Child Fetal Neonatal Ed 89:F108, 2004. Oh W, Poindexter BB, Perritt R, et al: Association between fluid intake and weight loss during the first ten days of life and the risk of bronchopulmonary dysplasia in extremely low birthweight infants, J Pediatr 147:786, 2005. Sedin G: Fluid management in the extremely preterm infant, In Hansen TN, McIntosh N, editors: Current topics in neonatology, London, 1995, WB Saunders, pp 50-66. Complete references used in this text can be found online at www.expertconsult.com C H A P T E R 32 Care of the Extremely Low-Birthweight Infant Eric C. Eichenwald In the past two decades, the field of neonatology has experienced significant progress in medical care and improvement in overall patient survival. Advancement in technology, greater use of prenatal glucocorticoids and neonatal surfactant replacement therapy, better regionalization of perinatal and high-risk neonatal care, and a more comprehensive understanding of the physiology of the immature infant have all contributed to dramatic increases in survival of very preterm infants. Care of premature infants with birthweights between 1000 and 1500 g has become almost routine in most neonatal intensive care units (NICUs) in the United States. The most recent challenge in neonatology is the care of extremely low-birthweight (ELBW) infants (birthweight <1000 g), sometimes referred to colloquially as micropremies. These infants present one of the greatest medical and ethical challenges to the field. Although they represent a small percentage of overall births and NICU admissions, ELBW infants are often the most critically ill and at the highest risk for mortality and long-term morbidity of any NICU patient. They also contribute disproportionately to overall hospital days and consume a large percentage of NICU personnel time, effort, and costs of care. Care of these infants is in constant evolution, as a result of new discoveries in both basic and clinical research as well as to growing clinical experience. This chapter will review some of the special challenges in and practical aspects of the management of the ELBW infant. The reader is referred to specific chapters throughout the text for a more comprehensive review of specific problems and conditions. Several studies have shown increased survival in the smallest and most premature infants over the past decade (Figures 32-1 and 32-2) (Fanaroff et al, 2007; Horbar et al, 2002; Lemons et al, 2001). See Chapter 1 for a complete discussion of these changes. Improved survival has not been accompanied by a change in the incidence of several major morbidities among survivors (Tables 32-1 and 32-2). Little to no change in the frequency of severe intracranial hemorrhage, periventricular white matter injury, necrotizing enterocolitis, or bronchopulmonary dysplasia was observed in the Eunice Kennedy Shriver National Institute of Child Health and Human Development cohort of infants born in institutions participating in the Neonatal Research Network in 19951996 compared with and 1997-2002 (Fanaroff et al, 2007). Although the Vermont–Oxford NICUs reported a reduction in the incidence of severe intracranial hemorrhage in the early part of the 1990s, the incidence has remained static into the late 1990s (Horbar et al, 2002). With the current trends in survival and in hospital morbidity, the absolute number of extremely premature infants who survive to NICU discharge and are diagnosed with a major morbidity in the neonatal period is increasing. A significant percentage of these infants continues to suffer from neurodevelopmental and neurosensory disability into childhood (Hack et al, 2005; Marlow et al, 2005; Wood et al, 2000). 100 EPIDEMIOLOGY The percentage of babies born preterm in the United States has risen slowly over the past two decades (Behrman and Stith Butler, 2007). In the year 2007, preterm births (<37 weeks’ gestation) accounted for 12.7% of all births, with approximately 1% of births occurring before 28 weeks’ gestation. There is a significant racial disparity in the incidence of extreme preterm birth, with the African American ELBW birth rate being nearly double that of the Hispanic and non-Hispanic white populations. Associated with this increase in frequency of preterm births is the greater availability of assisted reproductive technologies. These technologies result in a higher incidence of LBW infants, because of the higher frequency of multiple gestations (Schieve et al, 2002). Twins, triplets, and higher-order multiple gestations currently represent almost a quarter all LBW deliveries in the United States (Martin et al, 2002) and contribute to the ELBW population. Multiple gestations add to the potential morbidity of extremely premature birth because of a higher frequency of intrauterine growth restriction and other medical complications of pregnancy. 390 Mortality % 80 1991 1996 1997-2002 2008 60 40 20 0 501-750 751-1000 1001-1250 1251-1500 Birthweight (grams) FIGURE 32-1 Mortality in very low birthweight infants receiving care in the National Institute of Child Health and Human Development Eunice Kennedy Shriver Neonatal Research Network Centers from 1991 to 2002 and the Vermont-Oxford Network in 2008. (From Eichenwald EC, Stark AR: Management and outcomes of very low birth weight, N Engl J Med 358:1700-1711, 2008. Data from Lemons J, Bauer C, Oh W, et al: Very low birth weight outcomes of the National Institute of Child Health and Human Development Neonatal Research Network, January 1995 through December 1996, Pediatrics 107:1-8, 211, 2001; Fanaroff AA, Stoll BJ, Wright LL, et al: Trends in neonatal morbidity and mortality for very low birthweight infants, Am J Obstet Gynecol 196:147.e1-147.e8, 2007; and Horbar JD, Carpenter J, Kenny M, et al, editors: Vermont Network 2008 very low birth weight database summary, Burlington, Vt, 2009, Vermont Oxford Network.) 391 CHAPTER 32 Care of the Extremely Low-Birthweight Infant PERINATAL MANAGEMENT Extremely premature infants born in perinatal centers for high-risk infants, especially those with a high volume of such infants, have better short-term outcomes than infants transferred to such centers after birth (Arad et al, 1999; Bartels et al, 2006; Chien et al, 2001; Cifuentes et al, 2002; Phibbs et al, 2007; Shah et al, 2005; Towers et al, 2000). Therefore, if clinically feasible, the pregnant woman who seems likely to deliver an extremely premature infant should be transferred to a high-risk perinatal center for the expertise in both obstetric and neonatal management. Upon arrival, the expectant mother should be evaluated for factors that may have predisposed to preterm labor and assessed for the status of the fetal membranes and the presence or absence of chorioamnionitis. In addition, best obstetric estimate of gestational age (by date of last 100% menstrual period and early ultrasonographic dating, if available), ultrasonographic assessment of fetal size and position, and the presence of other medical or obstetric complications (preeclampsia, placenta previa, abruptio placentae) should be documented. Specimens for rectovaginal cultures to detect the presence of group B streptococci should also be obtained on admission (Schrag et al, 2002), and treatment with penicillin or ampicillin (or vancomycin for the patient with a severe penicillin allergy) should be initiated until culture results are available. There is widespread agreement that prenatal glucocorticoids should be offered to any woman in whom delivery at 24 to 34 weeks’ gestation threatens; treatment at earlier gestational ages is controversial and of unclear benefit. Although unlikely to arrest labor for an extended period, tocolytic agents should be considered for women with preterm uterine contractions without evidence of chorioamnionitis. Premature rupture of the fetal membranes (PROM) occurs in 30% to 40% of women who deliver prematurely (Mazor et al, 1998). If PROM is diagnosed without evidence of chorioamnionitis, consideration should be given Mortality % 80% TABLE 32-1 Survival and Selected Morbidities 60% for Very Low Birthweight (501-1500 g) Infants Born  in NICHD Neonatal Research Network Sites  (1995-1996 Compared with 1997-2002) 40% 0% 1995-1996* n = 4438% 1997-2002† n = 18,153% Survival without morbidity 84 85 Bronchopulmonary dysplasia 70 70 Home in supplemental oxygen 23 22 Necrotizing enterocolitis 15 11 7 7 12 12 5 3 Survival 20% 501-750g 751-1000g 1001-1250g 1251-1500g 501-1500g Intact Survival Survived with Morbidity Died FIGURE 32-2 Proportion of very low birthweight infants who died, and the proportion who survived with short-term complications (bronchopulmonary dysplasia, severe intraventricular hemorrhage, necrotizing enterocolitis, or a combination of these disorders, or with no complications from 1997 to 2002 at the Eunice Kennedy Shriver National Institute of Child Health and Human Development Neonatal Research Network sites. (From Eichenwald EC, Stark AR: Management and outcomes of very low birth weight, N Engl J Med 358:1700-1711, 2008. Data from Fanaroff AA, Stoll BJ, Wright LL, et al: Trends in neonatal morbidity and mortality for very low birth weight infants, Am J Obstet Gynecol 196:147.e1-147.e8, 2007.) Severe intraventricular hemorrhage Periventricular white matter injury Late-onset sepsis NICHHD, National Institute of Child Health and Human Development. *Lemons J, Bauer C, Oh W, et al: Very low birth weight outcomes from the National Institute of Child Health and Development Neonatal Research Network, January 1995 through December 1996, Pediatrics 107:1-18, 2001. †Fanaroff AA, Stoll BJ, Wright LL, et al: Trends in neonatal morbidity and mortality for very low birth weight infants, Am J Obstet Gynecol 196:147, 2007. TABLE 32-2 Survival With Selected Morbidities for Very Low Birthweight Infants: NICHD Neonatal Research Network, 1997-2002* Birthweight 501-750 g n = 4046% 751-1000 g n = 4266% 1021-1250 g n = 4557% 1251-1500 g n = 5284% Overall survival 55 88 94 96 Overall 65 43 22 11 BPD alone 42 25 11 4 Severe IVH 5 6 5 4 NEC alone 3 3 3 2 10 4 2 <1 Survived With Morbidity BPD and severe IVH BPD, Bronchopulmonary dysplasia; IVH, intraventricular hemorrhage. *Data from Fanaroff AA, Stoll BJ, Wright LL, et al: Trends in neonatal morbidity and mortality for very low birth weight infants, Am J Obstet Gynecol 196:147, 2007. 392 PART VIII Care of the High-Risk Infant to the use of prophylactic antibiotic therapy (ampicillin or erythromycin) for the mother. Several studies have shown that such therapy prolongs the latency period, reduces the incidence of chorioamnionitis and endometritis, and improves neonatal outcome (Gibbs and Eschenbach, 1997; Mazor et al, 1998; Mercer et al, 1997). Tocolytic and antibiotic therapy in the setting of PROM may prolong latency by 48 to 72 hours in many extremely preterm pregnancies in which delivery threatens, allowing the administration of a complete course of glucocorticoids to the mother. PRENATAL CONSULTATION If possible, all parents who are at risk for delivery of an extremely premature infant should meet in consultation with a neonatologist before the infant’s birth, preferably jointly with a perinatologist caring for the mother (Finer and Barrington, 1998). There are several goals of this consultation. First, the neonatologist and perinatologist should inform the parents about the proposed management of the pregnancy and delivery of the infant, including a discussion of the advantages of administration of glucocorticoids to the mother, the possible need for cesarean delivery, and delivery room care and resuscitation of the baby. Second, the parents should be informed about the potential risks of both the extent of prematurity and the proposed therapeutic interventions (Tables 32-3 and 32-4). Third, the neonatologist should investigate the parents’ beliefs and attitudes about delivery of an extremely premature infant and the potential for long-term morbidity. This outline for prenatal consultation is especially important if delivery is expected at the borderline of viability (23 to 24 weeks’ gestation). Data are sparse about the process and results of prenatal consultation with parents of an extremely preterm infant. Some studies suggest that significant incongruity exists among the attitudes of obstetricians, nurses, neonatologists, and parents about active delivery room resuscitation and treatment of extremely premature infants (Streiner et al, 2001; Zupancic et al, 2002). Reuss and Gordon (1995) have shown that the obstetric judgment of fetal viability was associated with an 18-fold increase in survival for infants with a birthweight less than 750 g, indicating that obstetric attitudes and decisions influence outcome. Obstetricians and nurses tend to overestimate, and parents tend to underestimate, mortality and morbidity in premature infants TABLE 32-4 Incidence of Specific Interventions and Outcomes of Extremely Low Birthweight Infants (<1000 g)  in the Vermont–Oxford Network, 2008 TABLE 32-3 Major Problems of Extremely Low Birthweight Birthweight Category Infants Short-Term Problems Long-Term Problems Respiratory distress syndrome Air leaks Bronchopulmonary dysplasia Apnea of prematurity Chronic lung disease Reactive airway disease Feeding intolerance Necrotizing enterocolitis Growth failure Growth failure Failure to thrive Inguinal hernias Immunologic, infection Immune deficiency Perinatal infection Nosocomial infection Respiratory syncytial virus Central nervous system Intraventricular hemorrhage Periventricular white matter disease Cerebral palsy Neurodevelopmental delay Hearing loss Ophthalmologic Retinopathy of prematurity Blindness, retinal detachment Myopia Strabismus Cardiovascular Hypotension Patent ductus arteriosus ? Hypertension in adulthood Pulmonary hypertension System Respiratory Gastrointestinal, nutritional Renal Hematologic Endocrine Water, electrolyte imbalance Acid-base disturbances Iatrogenic anemia Frequent transfusions Anemia of prematurity Transient hypothyroxinemia ?Cortisol deficiency ? Impaired glucose regulation ? Increased insulin resistance Outcome Survival (%) 501 to 750 g n = 9963 751 to 1000 g n = 12,341 58 85 Chronic lung disease at 36 weeks (%) 66 39 Discharge home in oxygen (%) 38 20 Eye examination performed (%) 66 82 None (%) 25 51 Stages 1-2 disease (%) 47 40 ≥Stage 3 disease (%) 27 8 Laser, cryotherapy (% treated) 15 5 Cranial ultrasound examination 91 95 No intraventricular hemorrhage (%) 54 69 Grade 1-2 (%) 23 19 ≥Grade 3 (%) 23 12 Patent ductus arteriosus (%) 63 53 Indomethacin, ibuprofen treatment (%) 61 52 Surgical ligation (%) 23 12 Nosocomial infection (%) 39 26 Necrotizing enterocolitis (%) 13 10 Survival without major morbidity (%) 12 38 Respiratory Morbidity Retinopathy of Prematurity Intraventricular Hemorrhage Data from Horbar JD, Carpenter J, Kenny M, Michelle J, editors: Vermont Oxford ­Network 2008 very low birth weight database summary, Burlington, Vt, 2009, Vermont Oxford Network. CHAPTER 32 Care of the Extremely Low-Birthweight Infant (Doron et al, 1998; Haywood et al, 1994). Parents report being more in favor of intervening regardless of gestational age or condition at birth compared with health care professionals (Streiner et al, 2001). In a survey by Ballard et al (2002), the majority of neonatologists responded that they would respect parents’ wishes about resuscitation of a borderline viable infant. In practice, when parental preferences about active resuscitation are known before delivery of an extremely preterm infant, neonatologists report that they would alter their management of the infant in the delivery room accordingly after assessment of the baby after birth (Doron et al, 1998), indicating the important role for prenatal consultation. However, parental attitudes likely are strongly influenced by physician recommendations about aggressive resuscitation conveyed during the prenatal consultation. This consultation is further complicated by significant variability in physician’s opinions about the benefits and burdens of providing intensive care at the borders of viability (De Leeuw et al, 2000). Physician attitudes may be influenced by fear of litigation for not actively resuscitating an extremely premature infant even if parental choices are made clear, as well as local, regional, and national norms (Partridge et al, 2009). The earliest gestational age at which resuscitation should be initiated remains controversial among neonatologists (Finer et al, 1999), making firm recommendations for the actual content of prenatal consultation difficult (Kaempf et al, 2006). Most studies that describe the outcome of premature infants are based on birthweight rather than gestational age and thus confound the effects of extremely preterm birth with those of intrauterine growth restriction, but both parents and physicians are usually faced with making decisions based on the anticipated gestational age at delivery. However, the likelihood of survival without serious sequelae may be influenced by factors other than gestational age. Using prospectively collected data from the Eunice Kennedy Shriver National Institute of Child Health and Human Development Neonatal Research Network on infants born at 22 to 25 weeks’ gestation between 1998 and 2003, Tyson et al (2008) presented an analysis of some of these factors. In a multivariate analysis, exposure to antenatal glucocorticoids, female sex, singleton gestation, and higher birthweight (in 100-g increments) were each associated with a decrease in the risk of death or of survival with neurodevelopmental impairment. The reduced risk was similar to the risk for infants with an additional week of gestational age. Other data from several studies suggest that in terms of survival and potential for long-term disability, encouragement of active management is appropriate for pregnancies expected to deliver at 25 weeks’ gestation or later (Piecuch et al, 1997; Wood et al, 2000). The risk of death and severe morbidity is significantly higher before 25 weeks’ gestation (El-Metwally et al, 2002). Mortality approaches 80% for infants delivered at 23 weeks’ gestation, with the majority of survivors suffering from neurodevelopmental sequelae (Vohr et al, 2000; Wood et al, 2000). These data should be shared with parents during prenatal consultation, and appropriate guidance should be given for decision making about the pregnancy, delivery, and resuscitation of the infant. Some have argued that, because of the risk of long-term disability associated with extreme preterm birth and the difficulty of making an informed 393 decision immediately before delivery, parents should be given the opportunity to have these issues discussed during routine prenatal care (Harrison, 2008). GENERAL PRINCIPLES OF CARE SPECIFIC TO EXTREMELY LOW-BIRTHWEIGHT INFANTS FIRST HOURS The first few hours after admission to the NICU are critical for the ELBW infant, although there are significant variations in practice among hospitals and practitioners. Careful adherence to details in the delivery room and during the first few hours after birth is essential to help avoid some of the immediate and long-term complications of the ELBW infant. All NICUs should have a consistent approach to the initial care of these fragile infants in the delivery room and upon admission to the NICU. A suggested treatment guideline for the first few hours after birth is presented in Table 32-5, and screening guidelines for common complications are shown in Table 32-6. DELIVERY ROOM At delivery, strict attention to maintenance of body temperature by means of rapid, gentle drying of the infant and the use of adequate heat sources is paramount to avoid cold stress. Use of a polyethylene occlusive skin wrap or bag immediately after delivery may also help to prevent initial evaporative heat losses (Vohra et al, 1999). Many ELBW infants need some form of assisted ventilation immediately after birth. If positive-pressure ventilation is required, it should be provided with low inspiratory pressure to prevent overdistention of the lungs, which can result in air leak and other lung injury, and adequate positive endexpiratory pressure (PEEP) to maintain lung volume by use of a flow-inflating bag, or optimally, a T-piece resuscitator that allows for more consistent delivery of inspiratory pressure and PEEP. Hyperventilation and hyperoxia also should be avoided; many units use continuous oxygen saturation monitoring in the delivery room, with supplemental oxygen titrated to maintain oxygen saturations in the range of 85% to 93%. The use of room air for resuscitation of these infants has been proposed to protect them from hyperoxia and damage to the lungs by oxygen free radicals. This issue is under active investigation (Lefkowitz, 2002), but most centers use blended oxygen for initial resuscitation, with an initial starting Fio2 between 0.4 and 0.6 (see also Chapter 28.) Considerable practice variation exists in the use and type of ventilatory support and timing of surfactant administration for the ELBW infant after birth. Some centers routinely intubate all ELBW infants in the delivery room for respiratory support and prophylactic surfactant administration (Egberts et al, 1997; Kendig et al, 1998). Soll (2009) has reported on a metaanalysis of eight randomized trials of the use of natural surfactant as prophylaxis versus as a rescue strategy for established respiratory distress syndrome (RDS) in the ELBW infant. Administration of exogenous surfactant before 15 minutes of age resulted in a reduction in the rates of neonatal mortality, air leak, and 394 PART VIII Care of the High-Risk Infant TABLE 32-5 Treatment Guidelines for Initial Management of Extremely Low-Birthweight Infants Time after Birth Guideline Time after Birth Delivery room Ensure good thermoregulation “Gentle” ventilation as required Avoid hyperventilation and hyperoxia Administer surfactant (if prophylaxis approach) Initiate NCPAP (if early CPAP approach) First 24 to 48 Hours NICU admission Obtain weight measurement Administer surfactant within first hour (if rescue approach) Establish vascular access: Peripheral intravenous catheter Umbilical arterial catheter Umbilical venous catheter (central, double-lumen) Start intravenous fluids as soon as possible with dextrose and amino acid solution Limit evaporative water losses (humidified incubator) Minimize stimulation Avoid hyperventilation and hyperoxia Maintain target oxygen saturations between 88% and 93% Obtain specimens for complete blood count with differential, blood culture, blood glucose measurement Give antibiotics as indicated Give parents information about their child Guideline Cardiovascular Monitor blood pressure, give vasopressors as required Maintain vigilance for presence of patent ductus arteriosus Obtain echocardiogram as indicated Respiratory Give additional surfactant doses as indicated Maintain low tidal volume ventilation Avoid hyperventilation and hyperoxia Extubate infant and start on continuous positive airway pressure when possible Fluid management Obtain weight every 12 to 24 hours Monitor serum electrolyte, blood glucose, and calcium concentrations every 4 to 8 hours Limit evaporative water losses Administer skin care Hematologic Obtain second blood count Administer transfusion support as indicated Monitor bilirubin, give phototherapy as indicated Infection Consider discontinuing antibiotics if blood culture results are negative at 48 hours Nutrition Start amino acid solution, parenteral nutrition Neurologic Minimize stimulation Perform screening head ultrasonography Social Arrange to meet with family CPAP, Continuous positive airway pressure; NCPAP, nasal continuous positive airway pressure; NICU, neonatal intensive care unit. the combined outcome of bronchopulmonary dysplasia (BPD) or death compared with a selective rescue approach. However, no trials comparing prophylactic administration with early rescue administration of exogenous surfactant (within the first 30 to 60 minutes) have been performed (see Chapter 46). In some centers, ELBW infants are observed briefly after birth, and if respiratory distress develops, they are intubated and ventilated for transfer to the NICU. Exogenous surfactant is given within 1 hour of birth, after endotracheal tube position and the presence of RDS are confirmed on chest radiograph. This strategy assures uniform distribution of surfactant to both lungs and allows initial stabilization and more intensive monitoring of the infant during its administration. Recent data suggest that many spontaneously breathing, extremely premature infants can be managed successfully with continuous positive airway pressure (CPAP) started in the delivery room (Aly et al, 2005b; Lindner et al, 1999; Morley et al, 2008; SUPPORT Study Group, 2010). Routine use of CPAP immediately after delivery may obviate the need for intubation in infants with a gestational age of 24 weeks or more, and increasing experience with this approach has been shown to improve its success (Aly et al, 2005b; Finer, 2006). According to one report, mechanical ventilation was avoided in approximately one third of infants with a gestational age of 25 weeks or less, and in nearly 80% of infants with a gestational age of 28 weeks or more using this approach (Aly et al, 2005b). Whereas aggressive use of early CPAP in the delivery room avoids intubation in some ELBW infants, it remains unclear whether it improves longer term respiratory outcomes (Morley et al, 2008; SUPPORT Study Group, 2010). If intubation is required, both oral and nasal routes for endotracheal tube placement are equally effective and have similar complication rates (Spence and Barr, 2002). Preference is usually institution specific, although successful placement of a nasal endotracheal tube requires more time. ADMISSION TO THE NEONATAL INTENSIVE CARE UNIT All infants should be weighed upon admission; frequent determination of subsequent weights is a valuable tool in managing fluid and electrolyte balance. In many centers, ELBW infants are initially placed under a radiant warmer for easier access (i.e., for surfactant administration and catheter placement). Because of the high transepidermal fluid losses in these infants, intravenous solutions containing 5% to 10% dextrose should be started as quickly as possible after admission, and efforts should be made to ameliorate evaporative water losses by increasing the relative humidity surrounding the infant. A plastic heat shield used in concert with the radiant warmer may decrease transepidermal water losses; alternatively, a polyethylene tent with an infusion of warmed humidified air may be used. Several studies suggest that fluid management is improved by using a humidified incubator instead of a radiant warmer, because of lower water losses (Gaylord et al, 2001; CHAPTER 32 Care of the Extremely Low-Birthweight Infant TABLE 32-6 Recommended Screening for Common Complications of Extremely Low-Birthweight Infants Complication Screening IVH HUS on days 1-3; repeat on days 7-10 Germinal matrix hemorrhage Repeat HUS weekly until findings normal Intraventricular hemorrhage Repeat HUS every 3 to 7 days until stable or resolved IVH with ventricular dilation or intraparenchymal bleeding Repeat HUS every 3 to 7 days until stable or resolved Periventricular white matter disease HUS at day 30; repeat at 36 weeks of postmenstrual age or at discharge Consider magnetic resonance imaging if HUS findings are equivocal ROP Perform OE examination at 4 to 6 weeks of postnatal age Repeat every 2 weeks if no ROP Repeat weekly if ROP present Repeat twice weekly for prethreshold disease or rapidly progressive ROP Audiology screening Hearing screen no earlier than 34 weeks of postmenstrual age, but before discharge home Consider measurement of resistive indices for progressive ventricular dilatation HUS, Head ultrasonography; IVH, intraventricular hemorrhage; OE, ophthalmologic examination; ROP, retinopathy of prematurity. Meyer et al, 2001). Exposure to high humidity may raise the rate of skin colonization with gram-negative organisms, although no increase in the rate of nosocomial infections was observed in several randomized trials comparing incubators with radiant warmers (Flenady and Woodgate, 2002). However, temperature regulation for the smallest infants in the first few days may be more difficult in an incubator than in a servo-controlled radiant warmer, because of rapid drops in air temperature as the incubator doors are opened to care for the infant (Meyer et al, 2001; see also Chapter 30). VASCULAR ACCESS Close monitoring of blood pressure, arterial blood gases, and serum chemistries during the first few days after birth is required in most sick ELBW infants; therefore it is advantageous to insert an umbilical arterial catheter for reliable access in infants who require assisted ventilation. Infusion of half-normal saline with 0.5 unit of heparin per milliliter at a low rate (0.5 to 1 mL/h) is usually enough to maintain catheter patency. Although using a saline solution rather than a dextrose-water solution in the umbilical arterial line may complicate fluid and electrolyte management, this disadvantage is offset by the advantage of reliable measurement of blood glucose levels, which are frequently required, without disturbing the infant. Placement of a central umbilical venous catheter (tip at the inferior vena cava–right atrial junction) at the same time the umbilical arterial catheter is placed also provides the clinician with reliable venous access for infusion of fluids, medications, 395 and blood products. Use of a double-lumen umbilical venous catheter often obviates insertion of a peripheral intravenous line over the first few days after birth and helps to preserve intravenous sites and skin integrity. The length of time that umbilical catheters are left in place varies by hospital. In most centers, umbilical lines are generally discontinued after 7 to 10 days because of the potential for catheter-related infection and vascular complications, although it is unusual for arterial access to be needed beyond a few days after birth. Before the umbilical venous catheter is removed, it is advisable to insert a percutaneous central venous catheter with its tip at the junction of the superior or inferior vena cava and the right atrium, dedicated to infusion of parenteral nutrition (see Nutritional Management, later). This catheter helps to maintain intravenous access for nutritional purposes without raising the risk of catheter-related infections and reduces the need to establish and maintain peripheral intravenous lines (Janes et al, 2000; Parellada et al, 1999). The incidence of complications of percutaneous central catheters, including catheter-related infections, is lower if a limited number of NICU personnel insert and maintain the lines. SKIN CARE The skin of an infant born at 23 to 26 weeks’ gestation is extremely immature and is ineffective as an epidermal barrier. Poor epidermal barrier function in the extremely preterm infant leads to disturbances in temperature regulation and water balance as well as breakdown in skin integrity, which can increase the risk of infection. The stratum corneum, which is responsible for epidermal barrier function, does not become functionally mature in the fetus until approximately 32 weeks’ gestation (Rutter, 2000). However, acceleration of the maturation process occurs after birth, so most extremely premature infants have a mature epidermal barrier by approximately 2 weeks of postnatal age. Until that time, full-thickness skin injury can occur in the ELBW infant from seemingly innocuous causes, such as local pressure from body positioning, removal of adhesives, and prolonged exposure to products containing alcohol or iodine. Such injury can lead to larger transepidermal water losses, an even greater risk of nosocomial infection, and significant scarring. Because of these risks, preservation of skin integrity should be incorporated into the care of the extremely preterm infant (Table 32-7). Limited use of adhesives and extreme care upon their removal, frequent repositioning of the infant to avoid pressure points on the skin, and use of soft bedding or a water mattress are the minimum requirements. Hydrocolloids (e.g., DuoDerm, ConvaTech, U.S.) applied to the baby in areas where adhesive tape may come in prolonged contact with the skin (i.e., umbilical catheter or endotracheal tube fixation) to prevent the direct application of tape to the baby may be useful, because it is more easily removed than standard adhesive tape. Polyurethane adhesive dressings (Tegaderm 3m, St. Paul, Minnesota) or hydrogels (Vigilon, Bard Medical, Covington, Georgia) may also be used to protect areas of skin friction and superficial wounds. Prophylactic application of preservative-free emollient ointments (Aquaphor, Beiersdorf AG, Hamburg, Germany) to protect the skin of the ELBW infant has been studied 396 PART VIII Care of the High-Risk Infant TABLE 32-7 Practical Guidelines for Skin Care of Extremely Low-Birthweight Infants Interventions Guidelines Adhesive application Increase adhesive tack by applying to dry, clean skin surface Avoid alcohol for skin cleansing Use smallest amount of tape possible Use a hydrocolloid or pectin-based layer on the skin, before application of heavy adhesive Avoid using adhesive over areas of skin breakdown Avoid adhesive bonding agents (e.g., benzoin) Use hydrophilic gel or pectin-based adhesives preferentially Adhesive removal Avoid adhesive removers and solvents Use a warm, wet cotton ball to periodically saturate hydrogel adhesives (avoid overdrying and oversaturation) Facilitate removal of adhesive with mineral oil, petrolatum, and emollients if reapplication is not necessary Emollient application Infants born at <27 weeks’ gestation may benefit from emollient use Avoid multidose containers (e.g., large jars) Use nonperfumed, nonirritating hydrophilic emollients Recognize potential for emollients to interfere with adhesive and conductive properties of monitoring devices Emollient removal Wipe off gently with a soft cloth or gauze if site is contaminated Avoid repeated attempts to thoroughly cleanse the skin (undesired friction effect) Remove emollients before attaching thermistors or other monitoring devices Data from Hoath S, Narendran V: Adhesives and emollients in the preterm infant, Semin Neonataol 5:289-296, 2000. (Edwards et al, 2004; Lane and Drost, 1993; Nopper et al, 1996). Several small studies demonstrated smaller transepidermal water losses, improved skin condition, and lower risk of suspected or proven nosocomial infection with prophylactic application of emollient ointments. However, one report documented an increase in the rate of systemic yeast infections in a single NICU coincident with a change to use of prophylactic emollients in ELBW infants, who returned to baseline after such use was discontinued (Campbell et al, 2000). In the largest study to date, infants randomized to prophylactic emollient had better skin integrity during the first month after birth, but had a higher rate of nosocomial bacterial infection compared with a control group with no emollient use (Edwards et al, 2004). Given these concerns, routine use of prophylactic emollients is not recommended; however, selective use in ELBW infants at risk for significant skin breakdown may be effective as an adjunct to other types of local skin care already described. MECHANICAL VENTILATION AND NASAL CONTINUOUS AIRWAY PRESSURE A high percentage of ELBW infants require some level of assisted ventilation to survive. Data from animal models of RDS indicate that positive-pressure ventilation with large tidal volumes damages pulmonary capillary endothelium, alveolar and airway epithelium, and basement membranes. This mechanical damage results in leakage of fluid, protein, and blood into the airways, alveoli, and interstitial spaces, leading to inhibition of surfactant activity and further damage to the lungs. These data suggest that a ventilator strategy that avoids large changes in tidal volume may reduce ventilator-induced lung injury in ELBW infants, an important management goal. The optimal mode, timing, and application of ventilatory support used in the initial management of the ELBW infant to meet this goal remain controversial. However, the objectives of all strategies of assisted ventilation in the ELBW infant should be similar—(1) to provide the lowest level of ventilatory support possible that will both support adequate oxygenation and ventilation and prevent atelectasis and (2) to try to reduce acute and chronic lung injury secondary to barotrauma, volutrauma, and oxygen toxicity (Clark et al, 2000). Data also suggest that targeting oxygen saturations (88% to 93%) lower than those used historically in ELBW infants who are receiving supplemental oxygen may protect the lung from oxidative injury and result in better outcomes (Askie et al, 2003). CONTINUOUS POSITIVE AIRWAY PRESSURE In a classic study comparing the incidence of BPD between NICUs, Avery et al (1987) reported that the NICU with the lowest incidence used CPAP more frequently and more aggressively than the other units. Van Marter et al (2000) confirmed these observations. Theoretically, early CPAP protects the immature lung from injury caused by positive-pressure tidal breaths, by preserving surfactant function, increasing alveolar volume and functional residual capacity, enhancing alveolar stability, and improving ventilation-perfusion matching. It also may stimulate the growth of the immature lung. Some NICUs now use CPAP (delivered via nasal prongs or nasopharyngeal tube at 5 to 7 cm H2O) as the initial mode of assisted ventilation for ELBW infants, starting almost immediately in the delivery room (Lindner et al, 1999). CPAP used this way has been reported in retrospective studies to decrease the need for mechanical ventilation (Gittermann et al, 1997; Poets and Sens, 1996), the need for surfactant treatment, and the incidence of BPD (Aly, 2001; Aly et al, 2005b; de Klerk and de Klerk, 2001; Lindner et al, 1999). In one prospective study (Morley et al, 2008), infants with a gestational age of 25 to 28 weeks who were breathing spontaneously, but required ventilatory assistance at 5 minutes after birth, were randomly assigned to treatment with nasal CPAP or intubation and mechanical ventilation. In the infants assigned to CPAP, 56% did not require intubation, and surfactant use was halved. Although respiratory outcomes at 36 weeks postmenstrual age were equivalent in the two study groups, a greater number of infants who were assigned to initial treatment with CPAP had pneumothorax (9% versus 3%). Alternatively, CPAP may be used after prophylactic or rescue surfactant therapy is given during a brief period of intubation and positive-pressure ventilation (Booth et al, 2006; Verder et al, 1999). Observational studies of this approach show that approximately one fourth of infants born before CHAPTER 32 Care of the Extremely Low-Birthweight Infant 27 weeks’ gestation do not require a subsequent course of mechanical ventilation and are less likely to develop BPD (Booth et al, 2006; Dani et al, 2004). CPAP can be delivered by a conventional mechanical ventilator with continuous flow, a variable flow device that adjusts flow through the respiratory cycle, or via “bubble” CPAP, in which a tube is immersed in water to the desired depth to generate CPAP with continuous gas flow bubbling through the immersed tube. Animal data suggest that bubble CPAP improves lung volume and gas exchange compared with conventional CPAP (Pillow et al, 2007), perhaps secondary to effects of the higher intranasal pressures generated by the bubbling water (Kahn et al, 2008). Other data suggest that work of breathing and thoracoabdominal asynchrony may be lessened with variable flow devices (Liptsen et al, 2005). Whether these differences are clinically relevant remains to be elucidated. CONVENTIONAL MECHANICAL VENTILATION Most NICUs continue to use pressure-limited, timecycled conventional ventilators for the initial respiratory management of the ELBW infant requiring mechanical ventilation. Synchronized intermittent mandatory ventilation (SIMV) remains the preferred mode of conventional ventilation for the ELBW infant. In the premature infant, SIMV, in which the inspiratory cycle is synchronized with the patient’s own effort, is better than conventional IMV in terms of oxygenation, ventilation, work of breathing, and blood pressure variability (Cleary et al, 1995; Hummler et al, 1996; Jarreau et al, 1996). Technologic limitations in the ability of some ventilators to synchronize breaths with very weak inspiratory efforts may prevent use of SIMV in the smallest infants. With further advancement in ventilator design, other modes of patient-triggered ventilation, including volume guarantee, assist control, and pressure support, have become available to clinicians. These designs generally allow ventilation at lower pressures than conventional SIMV. These modes of ventilation are under investigation but have not been proved to produce better pulmonary outcomes than SIMV in ELBW infants (Donn and Sinha, 1998; Herrera et al, 2002), although one report suggested a shorter time to extubation in infants babies managed with pressure-support ventilation (Reyes et al, 2006). Currently accepted conventional ventilatory strategies in the ELBW infant stress the avoidance of excessive tidal volumes by limiting peak inspiratory pressures and provision of adequate PEEP to maintain lung volume (Thome et al, 1998). This strategy helps to prevent repeated cycles of atelectasis and lung overdistention, a risk factor for ventilator-induced lung injury (Dreyfuss and Saumon, 1998). Hyperventilation (Paco2 <35 mm Hg) has been associated with a higher risk for the development of BPD (Garland et al, 1995; Van Marter et al, 2000) and neurodevelopmental sequelae in ELBW infants (Wiswell et al, 1996). In response to this suggestion of worse pulmonary outcome in hyperventilated infants, a strategy of minimal ventilation, or permissive hypercapnia, has been proposed for conventional ventilation in ELBW infants. In 397 the largest study to date of this ventilatory strategy, Carlo et al (2002) randomly assigned 220 infants with birthweights between 501 and 1000 g to receive either minimal ventilation (target Paco2 >52 mm Hg) or routine ventilation (target Paco2 <48 mm Hg). No difference in the rate of BPD, death, or other major short-term morbidities was seen between the two groups. Potential risks of a high Paco2 include increases in both cerebral perfusion and pulmonary vascular resistance and lower pH. Without clear data to support the benefit and safety of higher levels of hypercapnia, most centers target their conventional ventilatory strategy to maintain Paco2 between 45 and 55 mm Hg in the first several days of mechanical ventilation in the ELBW infant. In infants whose lung disease prevents extubation in the first several days after birth, changes in lung dynamics over the first 1 to 2 weeks often necessitate a change in ventilatory strategy. In the early stages of chronic lung disease, increased airway resistance and decreased lung compliance may require higher values for mean airway pressure, peak inspiratory pressure, PEEP, and inspiratory time than are usually used in initial ventilatory management. Once the need for more prolonged mechanical ventilation is established, many centers tolerate higher target Paco2 values in an attempt to limit further ventilator-induced lung injury. HIGH-FREQUENCY VENTILATION Considerable interest has been generated over the past 15 years in the application of high-frequency ventilation (HFV) in newborns who have respiratory failure, because this technique allows ventilation with small tidal volumes. Results of studies using HFV in animal models of RDS have been promising in the prevention of lung injury, but results of clinical studies of this ventilatory technique have not. Despite many clinical trials, controversy continues to surround the indications for HFV in ELBW infants, whether HFV is more effective than other modes of ventilation for RDS, whether HFV reduces adverse outcomes (specifically BPD), and whether HFV is more likely to have significant long-term complications than conventional mechanical ventilation. Early trials of HFV before surfactant replacement therapy demonstrated no pulmonary advantage of HFV over conventional mechanical ventilation and suggested an increase in rates of air leak and intracranial abnormalities in the HFV-treated infants (HIFI Study Group, 1989). Later trials in ELBW infants who were treated with surfactant also failed to demonstrate any reduction in the incidence of BPD in the HFV-treated infants (Johnson et al, 2002; Rettitz-Volk et al, 1998; Thome et al, 1999). A rigorously controlled trial of HFV as the primary mode of assisted ventilation compared with conventional mechanical ventilation was the first to suggest a small advantage to the early use of HFV in reduction of BPD or death without an increase in the incidence of short-term complications in ELBW infants. Courtney et al (2002) compared high-frequency oscillatory ventilation (HFOV) with synchronized IMV in a randomized trial that enrolled 500 ELBW infants. These investigators found a small but 398 PART VIII Care of the High-Risk Infant significant decrease in the incidence of BPD in survivors, requirement of fewer doses of exogenous surfactant, and shorter time to successful extubation in the HFOV-treated group. No differences in other complications of prematurity were observed between the two groups. These results suggest that, when used in experienced hands according to strict protocol, HFV confers some protection from lung injury in ELBW infants. It remains unclear whether this same advantage is gained when HFV is used in usual clinical circumstances in less experienced centers and whether the incidence of other complications may be affected by mode of ventilation. Some NICUs with the most experience with HFV use it routinely as the initial mode of ventilation for ELBW infants. Most centers continue to use conventional ventilation with low tidal volumes and reasonable ventilation goals as the initial mode of ventilation for ELBW infants, reserving HFV for infants in whom conventional ventilation and surfactant fail. This latter practice seems advisable, given the potential risks of HFV, including inadvertent lung overdistention, impaired cardiac output, and increased central venous pressure that may lead to intracranial hemorrhage. POST-EXTUBATION CPAP FOR RESPIRATORY DISTRESS SYNDROME AND APNEA CPAP is commonly used in ELBW infants after extubation to stabilize functional residual capacity and reduce the frequency of apneic spells after the lung disease has improved. All ELBW infants should be extubated as soon as they have recovered from acute RDS and should be given a trial of CPAP to protect them from further ventilator induced lung injury. When used in combination with methylxanthine therapy, CPAP decreases the need for reintubation because of progressive respiratory distress or apnea (Davis et al, 2003). Methylxanthine (aminophylline or caffeine) treatment before extubation has also been shown to lower the incidence of apnea and the need for reintubation in premature infants. Many ELBW infants benefit from prolonged use of CPAP by nasal prongs after extubation, especially those with frequent or severe episodes of apnea and bradycardia resistant to methylxanthine. Some centers currently provide assisted ventilation through nasal prongs (nasal intermittent positive pressure ventilation) to avoid intubation (Khalaf et al, 2001); this method may be more successful than conventional CPAP in avoiding postextubation failure in some infants (Bhandari et al, 2009). However, the smallest and least mature infants may need prolonged mechanical ventilation because of frequent, severe apneic spells that are unresponsive to other therapies. ADJUNCTIVE THERAPIES TO PREVENT BRONCHOPULMONARY DYSPLASIA Vitamin A Supplementation ELBW infants have low stores of vitamin A. Because of the role of vitamin A in promoting lung healing, its deficiency has been linked to a higher risk for development of BPD. In a large multicenter trial, vitamin A supplementation (5000 IU given intramuscularly three times per week for 4 weeks) reduced biochemical evidence of vitamin A deficiency and decreased the incidence of BPD by 12% without adverse effects in ELBW infants who required mechanical ventilation or supplemental oxygen at 24 hours of age (Tyson et al, 1999). Given these efficacy data and apparent safety, many NICUs choose to administer vitamin A supplementation as described previously to all ELBW infants starting at 24 hours after birth. However, widespread acceptance of this therapy has been limited by concerns over the need for thrice weekly intramuscular injections and its associated pain and stress to the infant. Caffeine In a prospective, masked, randomized trial involving infants with a birthweight of 500 to 1250 grams, the initiation of treatment with caffeine citrate (20 mg/kg loading dose, followed by 5 mg/kg/day) in the first 10 days after birth decreased the rate of BPD (a secondary outcome), as compared with placebo (36% versus 47%) without adverse effects (Schmidt et al, 2006). The composite primary outcome of death, cerebral palsy, cognitive delay, deafness or blindness at 18 to 22 months of age was also reduced in the subjects randomized to caffeine (Schmidt et al, 2007). The effect of caffeine on the incidence of bronchopulmonary dysplasia may be secondary to less ventilator-induced lung injury in the caffeine-treated infants, because shorter duration of positive pressure ventilation and supplemental oxygen therapy was associated with its use (Schmidt et al, 2006). These results suggest that caffeine should be administered routinely in ELBW infants in the first 10 days after birth, even if they continue to require mechanical ventilation. Inhaled Nitric Oxide Inhaled nitric oxide (iNO) may improve the pulmonary outcome in some VLBW infants, through mechanisms that are thought to involve decreased pulmonary vascular resistance or improved ventilation-perfusion matching, bronchodilatation, antiinflammatory effects, promotion of lung remodeling in response to injury, or normalized surfactant function. iNO used as rescue therapy for ELBW infants with severe hypoxic respiratory failure does not improve the pulmonary outcome or survival, and it may be associated with increased mortality or an increased incidence of intraventricular hemorrhage (Van Meurs et al, 2005). Results of trials of iNO involving premature infants with less severe lung disease who were at risk for BPD are mixed. In the Nitric Oxide for the Prevention of Chronic Lung Disease Trial (NO-CLD) study, a multicenter trial involving ventilator-dependent infants with a birthweight of 500 to 1250 g, treatment with iNO started at 7 to 14 days and continued for an average of 23 days of increased survival without BPD as compared with placebo (Ballard et al, 2006). In contrast, in a multicenter trial of iNO, treatment started before 48 hours of age and continued for 21 days (20 ppm for 3 to 4 days, then weaned to 10 ppm, 5 ppm, and 2 ppm for 7 days each) in infants with a birthweight of 500 to 1250 g who continued to require mechanical ventilation, only iNO-treated infants CHAPTER 32 Care of the Extremely Low-Birthweight Infant with a birthweight greater than 1000 g had a pulmonary benefit (Kinsella et al, 2006). Treated infants in this latter study were also less likely to have ultrasonic evidence of brain injury than were control infants. Long-term followup from the latter two studies suggests that iNO is safe and may improve long term pulmonary outcomes (Hibbs et al, 2008; Watson et al, 2009). iNO use for the prevention of BPD is not currently approved by the United States Food and Drug Administration. Routine use of iNO for the prevention of BPD should be discouraged until additional studies are available to define dosing and duration of therapy. Systemic Corticosteroids Inflammation also plays an important role in the pathogenesis of BPD; therefore pharmacologic doses of systemic corticosteroids have been widely used for prevention and treatment of BPD in ELBW infants. Several studies have examined early (<96 hours of age) administration of corticosteroids, usually dexamethasone, to prevent the development of BPD in infants at risk (Garland et al, 1999; Rastogi et al, 1996; Stark et al, 2001). A metaanalysis of studies in which corticosteroids were used to prevent rather than treat established BPD suggested that early corticosteroid treatment in ELBW infants results in more rapid extubation and a lower incidence of BPD (Halliday et al, 2002), although no effect on mortality was observed. However, a higher incidence of short-term complications, including hyperglycemia, hypertension, poor growth, and intestinal perforations and bleeding, was observed in the corticosteroid-treated infants (Garland et al, 1999; Stark et al, 2001; Watterberg et al, 2004). More importantly, long-term follow-up data suggest that exposure to corticosteroid for prevention or treatment of BPD raises the risk of neurological sequelae in treated infants, including poor head growth, cerebral palsy, and developmental impairment (American Academy of Pediatrics, 2010). The apparently higher risk of long-term sequelae without an effect on overall mortality has tempered enthusiasm for systemic corticosteroid treatment to prevent or treat BPD in ELBW infants. The possible role of postnatal corticosteroids in the prevention of BPD in selected infants, such as those exposed to chorioamnionitis, was suggested by a trial of early treatment with hydrocortisone (Watterberg et al, 2004). The risk of impaired neurodevelopment may be lower with hydrocortisone than dexamethasone (Rademaker et al, 2007; Watterberg et al, 2007), and some centers are substituting hydrocortisone for dexamethasone if postnatal corticosteroids are used. However, it seems prudent to avoid routine use of systemic corticosteroids in ELBW infants for the prevention or treatment of BPD until further data are available (American Academy of Pediatrics, 2010). NUTRITIONAL MANAGEMENT Provision of adequate nutrition is central to effective care of ELBW infants (see Chapters 66 and 67). These infants are born with limited nutrient reserves, immature pathways for nutrient absorption and metabolism, and higher 399 nutrient demands. In addition, medical conditions associated with extreme prematurity both alter requirements for and complicate the adequate delivery of nutrients. The goals of nutritional management of the ELBW infant are preservation of endogenous body stores, achievement of postnatal growth similar to intrauterine weight accretion and body composition, and maintenance of normal physiologic and metabolic processes concomitant with minimizing complications and side effects. However, few ELBW infants are able to meet these goals despite the use of central parenteral nutrition and caloric supplementation of enteral feedings. As a result, significant growth failure is commonplace (Berry et al, 1997; Ehrenkranz et al, 1999; Martin et al, 2009). Enteral Nutrition Medical problems of ELBW infants sometimes preclude initiation of enteral feedings for several days to weeks. However, the structural and functional integrity of the gastrointestinal tract depends on the provision of enteral feedings. Withholding enteral feedings at birth imposes risks for all the complications of luminal starvation, including mucosal thinning, flattening of the villi, and bacterial translocation. Early initiation (within the first few days after birth) of low volumes of milk (10 to 20 mL/kg/day, preferably with expressed breast milk; trophic feedings, or “gut priming”) has been studied in several small trials in premature infants (McClure and Newell, 2000; Schanler et al, 1999a). Trophic feedings are not meant to give the infant significant nutrition, rather to promote continued functional maturation of the gastrointestinal tract. Documented benefits of trophic feedings include higher plasma concentrations of gastrointestinal hormones, a more mature gut motility pattern, lower incidence of cholestasis, increased calcium and phosphorus absorption, and improved and earlier tolerance of enteral feedings. Trophic feedings have not been associated with a higher risk of necrotizing enterocolitis or other adverse outcomes; therefore there is no clinical advantage to delaying initiation of feedings in the medically stable ELBW infant. (See also Chapter 66.) Feeding intolerance, indicated by gastric residuals that exceed 25% to 50% of the volume fed, abdominal distention, or microscopic blood in the stool, is common in ELBW infants and may be difficult to differentiate from early stages of necrotizing enterocolitis. Feeding intolerance may preclude the advance of enteral nutrition for days to weeks, complicating nutritional management and prolonging the need for parenteral nutrition. Numerous feeding strategies to avoid episodes of feeding intolerance have been used in ELBW infants, including slow increase in enteral volume (<10 mL/kg/day), use of dilute rather than full-strength milk, continuous versus bolus tube-feeding, and use of prokinetic agents. None of these feeding strategies has been found to be clearly superior, although bolus feedings may decrease episodes of gastric residuals compared with continuous tube feedings and may allow a more rapid advance to full enteral volumes as well as promote better growth. Episodes of feeding intolerance also are reduced in ELBW infants fed human milk rather than specialized 400 PART VIII Care of the High-Risk Infant formulas for premature infants (Schanler et al, 1999b). Other benefits of giving human milk in ELBW infants are a more rapid advance to full enteral volumes and its positive immunologic effects, with an associated reduction in the risk of necrotizing enterocolitis and late-onset sepsis. Data also suggest that neurodevelopmental outcome may be improved in ELBW infants fed expressed breast milk (Vohr et al, 2006). Human milk must be fortified with calcium, phosphorus, sodium, protein, and other minerals to provide adequate nutrition in the ELBW infant. In addition to commercially available human milk fortifiers, which increase the caloric density to approximately 24 calories per ounce, human milk can be fortified further to higher caloric densities with medium chain triglycerides, glucose polymers, and added protein. Premature infants who are fed fortified human milk may grow more slowly than infants fed premature formulas (Schanler et al, 1999b), perhaps because of the variability in fat and caloric content of pumped breast milk or changes in the nutrient composition of human milk with fortification that affect fat absorption. Despite the potential for slower growth in infants fed human milk, its use should be strongly encouraged in ELBW infants because of the immunologic and other nutritional benefits. Further research on how best to fortify human milk is necessary to promote the best rate of growth in ELBW infants. (See also Chapters 65 and 66.) In addition, even after recommended enteral dietary intakes are reached, many ELBW infants continue to have a cumulative energy and protein deficit, which in part explains their later growth failure (Ehrenkranz et al, 1999; Embleton et al, 2001; Martin et al, 2009). Early Parenteral Nutrition Protein losses in ELBW infants receiving a glucose infusion alone begin immediately after birth and can approach 1.5 g/kg/day in the first 24 to 72 hours. Fortunately, these losses can be offset by early administration of an amino acid solution, even at low caloric intakes. Several studies have demonstrated the safety of early administration (within 24 hours after birth) of an amino acid solution, with no abnormal elevations of ammonia or blood urea nitrogen even in the most immature infants (Rivera et al, 1993; Van Goudoever et al, 1995). To prevent early protein deficit, ELBW infants should be given a source of parenteral protein as soon as possible after birth (see also Chapter 67). In one study, ELBW infants who received 3 g or more of protein per day at 5 days of age or less were less likely to have a weight below the 10th percentile at 36 weeks postmenstrual age and suboptimal head growth at 18 months of age when compared with ELBW infants who received less protein supplementation after birth (Poin­ dexter et al, 2006). In most centers, parenteral nutrition is used exclusively during the first few days after birth and then gradually reduced as enteral feedings are introduced. Longer duration of parenteral nutrition is associated with the development of a number of complications, including cholestasis, osteopenia, and sepsis. For example, the risk of an episode of late-onset sepsis in premature infants is 22-fold higher if parenteral nutrition is continued for more than 3 weeks compared with 1 week or less (Stoll et al, 2002b). MANAGEMENT AND PREVENTION OF INFECTION Bacterial and fungal infections are an important cause of illness and death among ELBW infants. In addition to the immediate morbidity and mortality, local and systemic inflammation caused by infections may increase the risk for development of other complications of prematurity, including BPD and brain injury. ELBW infants are frequently exposed to perinatal and delivery complications that raise their risk of early-onset (<72 hours) infections. The need for prolonged intravenous access, exposure to parenteral nutrition, and mechanical ventilation also subject the ELBW infant to a high risk of late-onset (>72 hours) nosocomial infections. The frequent infections seen in the ELBW population are related to immaturity of both humoral and cellular immunity (see Chapter 36). In addition to the judicious use of antimicrobial therapy, environmental controls, nursery surveillance, and modulation of the immature immune response have been proposed as possible interventions to prevent infections in extremely premature infants. Early-Onset Infections The incidence of early-onset bacterial infections in very LBW (VLBW) infants is approximately 1% to 2%, with a mortality of approximately 40% to 50% (Stoll et al, 2002a; see also Chapter 39). The major risk factor for the development of perinatally acquired bacterial infections is PROM with chorioamnionitis, which frequently complicates premature deliveries. However, a significant percentage of extremely premature births may be associated with intrauterine infection before membrane rupture (Goldenberg et al, 2000). In one study, 41% of premature infants with early-onset sepsis were born less than 6 hours after membrane rupture (Stoll et al, 2002a). Because the clinical signs of perinatally acquired infection are nonspecific, the index of suspicion and the concern about the possibility of intrauterine infection should always be high in the presence of premature birth. All ELBW infants, except for those delivered for maternal indications with no labor, should be evaluated for infection at birth by means of a complete blood count with differential and blood culture, and empiric antibiotic therapy with ampicillin and an aminoglycoside should be initiated. A white blood cell count less than 5000 cells/μL, a ratio of immature to total neutrophils ratio greater than 0.2 to 0.3, and neutropenia (absolute neutrophil count less than 1000 cells/μL) are all suggestive of infection, but may also be seen in infants with other conditions, including maternal preeclampsia and hypertension. The duration of initial antibiotic therapy depends on the results of the blood culture, blood counts, the clinical course, and the perinatal history. If the blood culture is negative for bacterial growth at 48 hours and the infant has improved clinically, consideration should be given to discontinuing antibiotics. Prolonged exposure to antibiotic therapy increases the likelihood of colonization with multiple antibiotic-resistant 401 CHAPTER 32 Care of the Extremely Low-Birthweight Infant organisms, the development of fungemia, and necrotizing enterocolitis (Cotten et al, 2009); therefore it should be reserved for infants with documented infection or a very high index of suspicion of infection based on clinical or historical factors. The distribution of pathogens causing early-onset sepsis in VLBW infants has changed, likely because of increased use of intrapartum antibiotics for prevention of group B streptococcal infections and treatment of preterm PROM (Puopolo and Eichenwald, 2010). As the rate of infections from group B streptococci has diminished with intrapartum antibiotic prophylaxis, the proportion of documented infections from gram-negative organisms has risen (Table 32-8; Stoll et al, 2002a). In an additional change of the epidemiology of early-onset infections in premature infants, possibly related to increased use of intrapartum antibiotics, the frequency of infections owing to ampicillin-resistant Escherichia coli strains has increased in some centers (Bizzarro et al, 2008; Joseph et al, 1998; Stoll et al, 2002a). Current recommendations for empiric antibiotic therapy for ELBW infants at risk of early-onset sepsis have not changed, but continued surveillance of the epidemiology and antibiotic resistance patterns of isolates within individual units is warranted. In infants with severe illness that may be caused by sepsis, broadening initial antibiotic coverage to include a third-generation cephalosporin should be considered. LATE-ONSET INFECTIONS Nosocomial infection is a common though preventable complication of intensive care of the ELBW infant. The incidence of late-onset sepsis in ELBW infants who survive beyond 3 days of age is 25% to 50%, depending on gestational age and birthweight, with the median age at onset of the first episode approximately 2 weeks (Fanaroff et al, 1998; Stoll et al, 2002b). The overall mortality rate is approximately 20% but may be as high as 80%, depending on the organism causing sepsis. The risk of neurodevelopmental impairment is increased by approximately 1.5-fold in ELBW survivors of nosocomial bloodstream infections (Stoll et al, 2004). Risk factors for the development of late-onset sepsis include prolonged hyperalimentation and lipid use, the presence of a central venous catheter, longer duration of mechanical ventilation, and delay in initiation of enteral feedings (Stoll et al, 2002b). These practices and procedures are common events that may be unavoidable in the care of ELBW infants. However, large variations have been observed among NICUs in the rate of late-onset infections in premature infants (Brodie et al, 2000; Stoll et al, 2002b), suggesting that individual NICU practices may affect the incidence of nosocomial infections. The most common cause of late-onset infections in ELBW infants is coagulase-negative Staphylococcus (CoNS). The most significant risk factor for the development of CoNS infection is the use of a fat emulsion (e.g., Intralipid) infusion (Freeman et al, 1990). Infection with CoNS is almost never fatal but is associated with significant morbidity, such as prolonged ventilator use and hospital stay (Gray et al, 1995). Other organisms that cause late-onset infections in ELBW infants are associated with a much higher morbidity and mortality. The distribution TABLE 32-8 Distribution of Pathogens among 84 Cases of Early-Onset Sepsis* Organism No. % Gram-negative: 51 60.7 Escherichia coli 37 44.0 Haemophilus influenzae 7 8.3 Citrobacter 2 2.4 Other 5 6.0 31 36.9 Group B streptococci 9 10.7 Viridans streptococci 3 3.6 Other streptococci 4 4.8 Listeria monocytogenes 2 2.4 Coagulase-negative staphylococci 9 10.7 Other 4 4.8 Gram-positive: Fungi: Candida albicans 2 84 Total 2.4 100 Data from Stoll B, Hansen N, Fanaroff A, et al: Changes in pathogens causing early onset sepsis in very low birth weight infants, N Engl J Med 347:240-247, 2002. *Occurring in 5447 infants born between September 1, 1998, and August 31, 2000. TABLE 32-9 Distribution of Pathogens Associated With the First Episode of Late-Onset Sepsis* Organism No. % Gram-positive: 922 70.2 Coagulase-negative staphylococci 629 47.9 Staphylococcus aureus 103 7.8 43 3.3 30 2.3 Enterococcus spp. Group B streptococci Other Gram-negative: 117 8.9 231 17.6 Escherichia coli 64 4.9 Klebsiella spp. 52 4.0 Pseudomonas spp. 35 Enterobacter spp. 33 2.5 Serratia spp. 29 2.2 Other Fungi: Candida albicans 27 18 1.4 160 12.2 76 5.8 Candida parapsilosis 54 4.1 Other 30 2.3 Total 1313 100 Data from Stoll B, Hansen N, Fanaroff A, et al: Late-onset sepsis in very low birth weight neonates: experience of the NICDH Neonatal Research Network, Pediatrics 110:285-291, 2002. *In National Institute of Child Health and Development Neonatal Research Network institutions, September 1, 1998, through August 31, 2000. of pathogens associated with the first episode of late-onset sepsis among 1313 infants in a cohort of VLBW babies over a 2-year period is shown in Table 32-9. Presenting features of late-onset sepsis include increased apnea, feeding intolerance, abdominal distention, guaiacpositive stools, increased respiratory support, and lethargy and hypotonia (Fanaroff et al, 1998). Because these 402 PART VIII Care of the High-Risk Infant symptoms are nonspecific, ELBW infants are frequently evaluated for infection and treated with empiric antibiotic therapy. In one study, use of both vancomycin and antifungal therapy was inversely related to birthweight; approximately three fourth of infants with a birthweight less than 750 g was treated with vancomycin during their hospital stay, and approximately one third was treated with antifungals (Stoll et al, 2002b). Central catheters should be removed immediately to ensure adequate treatment of infants in whom sepsis is diagnosed, except for that caused by CoNS (Benjamin et al, 2001). Endotracheal tube colonization with multiple organisms is common in infants who require prolonged mechanical ventilation. In general, such colonization should not be treated with antibiotics unless there is evidence of pneumonia or significant inflammation indicative of tracheitis. Prospective surveillance of common isolates and antimicrobial resistance patterns within individual NICUs can help to guide empiric antibiotic therapy in ELBW infants being evaluated and treated for presumed sepsis. However, indiscriminate use of broad-spectrum antibiotics in the absence of true infection can alter antimicrobial resistance patterns (Goldmann et al, 1996), raising the risk of lateonset infections and complicating therapy. Prevention of Nosocomial Infection Because of the frequency and potential severity of lateonset sepsis in ELBW infants, several strategies to prevent infection have been proposed. Using these practices as a guideline, Horbar et al (2001) observed a decrease in the CoNS infection rate from 22% to 16.6% over a 2-year period in VLBW infants in six study NICUs. Other investigators have confirmed that changes in practice, primarily surrounding the use and care of central venous catheters, can reduce the overall burden of late onset infections in individual units (Aly et al, 2005a; Kilbride et al, 2003). Some NICUs routinely screen ELBW infants by stool or respiratory secretion cultures for the presence of multiple antibiotic-resistant organisms, which would necessitate isolation (Gregory et al, 2009). Clusters of infections with unusual organisms should prompt surveillance cultures of infants, potential NICU environmental sources, and NICU staff (Foca et al, 2000). Restriction of broadspectrum antibiotic use by hospital policy or treatment guidelines may limit the local spread of resistant organisms (Goldmann et al, 1996). Strict adherence to hand hygiene before and after every patient contact and avoidance of overcrowding within NICUs also help to decrease the incidence of infection. The use of alcohol-based hand gels at the bedside may improve compliance with hand hygiene (Harbarth et al, 2002). In addition to practice and environmental controls, prophylactic use of antibiotics and modulation of the immune response of ELBW infants have been studied as methods to reduce the incidence of late-onset sepsis. Low-dose vancomycin given continuously via hyperalimentation solutions (Baier et al, 1998; Spafford et al, 1994), intermittently via peripheral vein (Cooke et al, 1997), or vancomycin lock of the central venous catheter (Garland et al, 2005) has been shown to reduce the incidence of CoNS in premature infants at risk. Concern about the emergence of vancomycin-resistant organisms and the low mortality associated with CoNS infections has prevented widespread use of this approach. Prophylactic fluconazole given for 6 weeks lowered the incidence of fungal colonization and invasive disease in ELBW infants without associated complications or the emergence of resistant organisms (Kaufman et al, 2001, 2005; Manzoni et al, 2007). Such an approach might be advisable for NICUs with a high incidence of fungal infections in their ELBW population, but more study is needed to define any potential short- and long-term risks. Approaches that have been used with success in immunocompromised adult patients, such as antiseptic-impregnated central catheters, are promising but have not yet been studied adequately in premature infants. Prophylactic intravenous administration of polyclonal immunoglobulin (IVIG) to prevent late-onset sepsis has been studied extensively in premature infants. Several trials have shown a decrease in the incidence of documented sepsis by a small but significant amount in premature infants treated with prophylactic IVIG and no effect on mortality or other complications of prematurity (Lacy and Ohlsson, 1995). The costs associated with this therapy to achieve the small decrease in infection rates, as well as the increased exposure to blood products, have limited its use; it is unclear whether selective prophylactic IVIG treatment of ELBW infants at highest risk for sepsis is warranted. However, IVIG therapy in addition to antibiotic therapy may be of benefit in reducing mortality in infants with established sepsis (Jenson and Pollock, 1997). Development of more targeted polyclonal or monoclonal γ-globulin preparations for specific organisms that cause sepsis in premature newborns may alter the use of IVIG in the future (Lamari et al, 2000; Weisman et al, 2009). Another promising strategy under investigation for modulation of the immature immune response to help prevent infections in premature infants is treatment with hemopoietic colony-stimulating factors, including granulocyte colony-stimulating factor and granulocyte-macrophage colony-stimulating factor (Modi and Carr, 2000). Most studies of these factors have been conducted in neutropenic, small-for-gestational-age infants or infants delivered to women with preeclampsia. Treatment with granulocyte colony-stimulating factor in neutropenic infants resulted in an increase in neutrophil counts and reduced the incidence of sepsis (Kocherlakota and La Gamma, 1998). Prophylactic treatment with granulocyte-macrophage colony-stimulating factor in premature infants with normal neutrophil counts prevented the development of neutropenia in episodes of sepsis, but it remains unclear whether this type of therapy to address cellular immune deficiency in ELBW infants will reduce the incidence of infection without additional complications (Miura et al, 2001; Modi and Carr, 2000). NEUROSENSORY COMPLICATIONS The major neurosensory complications associated with extreme premature birth are intraventricular hemorrhage, periventricular white matter injury, and retinopathy of prematurity. Although the incidence of severe intraventricular hemorrhage has fallen with improvements in management and increased antenatal steroid use, it remains a major cause of brain injury with consequent abnormal CHAPTER 32 Care of the Extremely Low-Birthweight Infant neurodevelopment. Pharmacologic approaches to its prevention after birth have been generally unsuccessful. Prophylactic indomethacin reduces the incidence of severe intraventricular hemorrhage, but does not improve longterm neurodevelopment (Schmidt et al, 2001). Periventricular white matter injury is the predominant form of brain injury in extremely preterm infants and correlates strongly with the development of cerebral palsy. Its pathogenesis is poorly understood, and no specific neuroprotective strategy is known. In some infants, cerebral blood flow and oxygen delivery measured with near infrared spectroscopy varies during variations of blood pressure considered to be in the normal range, and this lack of autoregulation of cerebral blood flow may lead to ischemic white matter injury (Evans, 2006). Whether aggressive treatment of hypotension in ELBW infants prevents or may lead to subsequent brain injury is uncertain, probably because blood pressure, which is easily measured, does not correlate well with systemic or cerebral blood flow (Fanaroff et al, 2006; Limperopoulos et al, 2007). A higher rate of white matter injury occurs in the setting of maternal or neonatal infection, or with elevated proinflammatory cytokines in amniotic fluid or cord blood, suggesting that inflammation has a role in the pathogenesis (Viscardi et al, 2004). Advanced magnetic resonance imaging techniques in infants with white matter injury show disturbances in cerebral growth, with reduced volume of both gray and white matter (Inder et al, 2005). These observations might serve to explain the motor and cognitive dysfunction often seen in infants with white matter injury. Retinopathy of prematurity, a vascular proliferative disorder that affects the incompletely vascularized retina of preterm infants, is a major cause of blindness in these children. Severe retinopathy is 18-fold more likely to develop in infants delivered at less than 25 weeks’ gestation compared with 28 weeks’ gestation (Fanaroff et al, 2007). Periods of hyperoxia owing to exposure to excessive inspired oxygen concentration contribute to the development of retinopathy (Saugstad, 2006); however, the optimal target range of oxygen saturation is not known. Because fetal hemoglobin shifts the hemoglobin oxygen saturation curve to the left, oxygen saturations greater than 95% may be associated with arterial oxygen tension greater than 80 mm Hg, possibly excessive for the ELBW infant. Conversely, oxygen saturation that is too low can increase the risk of injury to the brain or other end organs (Deulofuet et al, 2006). Adjusting inspired oxygen concentration to target lower oxygen saturations in extremely preterm infants may decrease the rate of severe retinopathy. In a prospective observational study, ELBW infants treated in centers with a restrictive approach to oxygen delivery (i.e., saturation alarm limits of 70% to 90%) had less retinopathy requiring cryotherapy (6.3 versus 27.7 %) than did those in units with a liberal approach (i.e., alarm limits of 88% to 98%), and neurodevelopmental outcome at 1 year of age was similar (Tin et al, 2001). In two other studies with historical controls, the incidence of severe retinopathy decreased after oxygen saturation alarm limits were lowered from 87%-97% to 85%-93% for infants born at 28 weeks’ gestation and less until they reached 32 weeks postmenstrual 403 age (Chow et al, 2003; Vanderveen et al, 2006). In a masked randomized trial comparing two different target oxygen saturation ranges (85% to 89% compared with 91% to 95%) in infants born between 24 and 27 weeks, ROP occurred less frequently in survivors assigned to the lower oxygen saturation group. However, death before discharge occurred more frequently in the lower oxygen saturation group, suggesting that the best target range for oxygen saturation remains unclear (SUPPORT Study Group, 2010). DEVELOPMENTAL AND PARENTAL CARE ELBW infants are particularly vulnerable to the potentially noxious stimuli of the NICU environment, including light, noise, frequent disturbances, and painful procedures. The ELBW infant reacts to the noisy and well-lit environment of many NICUs with greater variability of blood pressure, ventilatory requirements, and oxygen saturation as well as behavioral disorganization, which may have both short- and long-term effects on outcome (Jacobs et al, 2002). Modification of the NICU environment to limit exposure of ELBW infants to such stresses—by lowering ambient light and reducing noise, clustering caregiving periods and procedures to allow periods of uninterrupted sleep, and using positioning aids to promote containment—is an intuitive part of their care. Newer NICU designs, transitioning from open common rooms to private room settings, may also facilitate a better environment for vulnerable infants and enhance parental involvement. These environmental and developmental interventions in the NICU can improve physiologic stability and some short-term outcomes in preterm infants, including decreased severity of BPD and shorter length of hospital stay. It remains unclear whether individualized, developmentally supportive care or other developmental interventions started in the NICU improve long-term outcome in ELBW infants (Symington et al, 2009). In addition to environmental modifications, NICUs should promote parental involvement with infants, even when they are critically ill. Open family presence guidelines and encouragement of parental caregiving when appropriate may help parents to bond with their baby. Many NICUs have embraced a philosophy of familycentered care, in which a stronger bond is forged with the families of NICU patients by encouraging collaboration among family members and care providers in policy and program development, professional education, and aspects of the delivery of care (Dunn et al, 2006). Skin-to-skin (kangaroo) care, in which the infant is placed unclothed on the mother or father’s bare chest, was originally developed in nonindustrialized countries to maintain temperature regulation in premature infants. It is now used in many NICUs to promote parental attachment. Skin-to-skin care may have a positive effect on infant state organization and respiratory patterns, increase the rate of infant weight gain, improve maternal milk production, and have long-term benefits in infant development and parents’ perceptions of their babies (Feldman et al, 2002). Skin-to-skin care can be initiated in ELBW infants within the first 2 weeks after birth, when they are more medically stable. 404 PART VIII Care of the High-Risk Infant FUTURE DIRECTIONS This chapter presents some of the special needs of the ELBW infant. The medical care of the ELBW infant is a complex combination of knowledge of developmental physiology, evidenced-based interventions, and clinical experience. Wide variability in approaches to care of these infants exists among practitioners and NICUs, as does variability in outcomes. Nevertheless, NICUs involved in treating ELBW infants should develop a coherent approach to the medical and ethical aspects of their care. Future research should focus on identifying best practices to narrow the variability in approach to care and with the goal to prevent long-term disability. As more of these tiny infants survive, it is the responsibility of neonatologists to stay abreast of clinical improvements and the short- and long-term consequences of established and newly proposed medical interventions to provide the best care for these vulnerable infants and to keep parents informed and involved. Complete references used in this text can be found online at www.expertconsult.com C H A P T E R 33 Care of the Late Preterm Infant Sowmya S. Mohan and Lucky Jain Had he been alive today, Patrick Bouvier Kennedy would have been hailed as a triumph of neonatal care—after all, he was the son of the former United States President, John F. Kennedy and former First Lady Jacqueline B. Kennedy. Born prematurely at 34 weeks gestation, he would have been aptly labeled as a late preterm neonate; however, based on his birth weight (2.1 kg) and gestational age, few would have predicted the outcome he had then, were he born in 2009. But then, in 1963, little was available to the clinician for the management of hyaline membrane disease—no routine use of neonatal ventilators, no device to provide airway positive pressure, no surfactant, and no antenatal steroids. He died two days after his birth; the New York Times obituary said that “the battle for the Kennedy baby was lost because medical science has not advanced far enough.” (Jain and Carlton, 2009) With more than 4 million live births per year (Martin et al, 2007), the United States has one of the highest birth rates among industrialized countries; it also has the stigma of having a disproportionately high prematurity rate. Decades of efforts to reduce preterm births have not affected this formidable problem. In recent years the problem has been highlighted by the rise in births between 34 and 366⁄7 weeks’ gestation, a group referred to as late preterm infants (Figure 33-1). Late preterm infants have a checkered history, having been passed off as nothing more than “near term” infants, yet being feared as the “quick to spiral down group” when they develop respiratory distress syndrome or other complications. As the number of late preterm infants has grown, so has the awareness of their unique set of problems, such as delayed neonatal transition, wet lung syndrome, hypothermia, hypoglycemia, and hyperbilirubinemia (Figure 33-2). Although not unique to this population, these complications have sufficient differences in their manifestations and management, prompting the editors to add an entirely new chapter to this textbook devoted to the health issues of late preterm infants. In Chapter 14, the obstetric issues and epidemiology related to prematurity are addressed. This chapter focuses on the special considerations applicable to the clinical course and management of late preterm infants. There has been a shift in the distribution of births away from term and post term and toward earlier gestational ages (Davidoff et al, 2006). This shift has resulted in a disproportionately high rate of premature births with estimates of up to 12.7% of live births being premature (Martin et al, 2007)—defined as <37 completed weeks’ gestation or <260 days, counting from the first day of the last menstrual period (Raju, 2006). Within this group of premature babies, up to 75% are classified as late preterm infants (Adamkin, 2009). Although the reasons for such a high number of late preterm births are multifactorial, higher rates of induced deliveries, cesarean births, and efforts to reduce stillbirths may have contributed to the increase. Late preterm babies currently account for up to one third of all neonatal intensive care unit (NICU) admissions in the United States (Angus et al, 2001), adding strain to the overburdened system of health care delivery, particularly in community hospitals and rural areas. These admissions range from short stays, for problems such as transient tachypnea of the newborn (TTNB), to more complicated or extended NICU stays for problems such as persistent pulmonary hypertension of the newborn (PPHN). With the average NICU stay costing up to $3500 per day, the economic impact of caring for the late preterm baby can be significant. For example, in 1996 the State of California alone could have saved $49.9 million in health care costs by preventing non–medically indicated deliveries between 34 and 37 weeks’ gestation (Gilbert et al, 2003). In addition to the expense of the initial hospitalization, the cost of caring for a late preterm baby can also be compounded by the increased incidence of hospital readmissions and the long-term care issues related to persistent problems. The effects of the increasing number of late preterm births create a societal burden in lost productivity, as parents take extended leave from work to be with their fragile newborns. More importantly, there may be lasting effects with neurodevelopmental delays extending into early school age. Because a significant proportion of brain growth occurs during the last 6 weeks of gestation (Adams-Chapman, 2006), late preterm infants are vulnerable to neuronal injury and disruption of normal brain development. Whereas more longitudinal studies are needed, preliminary studies show that late preterm infants are more likely to have a diagnosis of developmental delay within the first 3 years of life, require special needs preschool resources, and have more problems with school readiness (Morse et al, 2009). Given their large numbers, the overall socioeconomic effects of the late preterm births can be significant. Strategies are required that can reduce the preventable fraction of late preterm births and work toward reducing the morbidity in others, when continuation of the pregnancy is deemed harmful to the fetus or the mother. This chapter explores the pathophysiology of the major morbidities that affect late preterm infants and discusses the unique challenges faced by clinicians in the management of these conditions. DEFINITION Late preterm birth is an accepted term used for infants born between 34 and 366⁄7 weeks’ gestation (see Figure 33-1) (Raju et al, 2006). This group of infants was initially referred to as near term, but the misleading implication of 405 406 PART VIII Care of the High-Risk Infant Late preterm First day of last menstrual period Day # 1 239 Early term 259 260 274 BOX 33-1 C  haracteristics of Late Preterm Infants 294 ll ll Week # 0/7 34 0/7 36 6/7 37 38 0/7 6/7 41 6/7 Term Preterm Post term FIGURE 33-1 Definitions of late preterm and early term. (Adapted from Engle WA, Kominiarek MA: Late preterm infants, early term infants, and timing of elective deliveries, Clin Perinatol 35:325-341, 2008.) CLINICAL OUTCOMES Temperature instability Full term Near term Hypoglycemia Intravenous infusion Respiratory distress ll Late preterm infants—defined as born at 34 to 366⁄7 weeks’ gestation Physiologically immature with limited compensatory responses to extrauterine environment compared to term infants Greater risk than term infants for mortality and morbidities such as: ll Temperature instability ll Hypoglycemia ll Respiratory distress ll Apnea ll Jaundice ll Feeding difficulties ll Dehydration ll Suspected sepsis Adapted from Engle WA, Tomashek KM, Wallman C: “Late preterm” infants: a population at risk, Pediatrics 120:1390-1401, 2007. physiological jaundice, late neonatal sepsis (Raju, 2006), thermoregulation issues, hypoglycemia, feeding difficulties (Dudell and Jain, 2006; Escobar et al, 2006; Fuchs and Wapner, 2006), and risk of injury to the developing brain, which can lead to neurodevelopmental problems. These problems account for a substantially higher number of NICU admissions (see Figures 33-2 and 33-3). PATHOPHYSIOLOGY AND CLINICAL COURSE Clinical jaundice 0 10 20 30 40 50 60 (%) FIGURE 33-2 Graph of clinical outcomes in near-term (35 to 366⁄7 weeks) and full-term infants as percentage of patients studied. (Adapted from Wang ML, Dorer DJ, Fleming MP, et al: Clinical outcomes of nearterm infants, Pediatrics 114:372-376, 2004.) maturity has prompted the name change to late preterm (Box 33-1). This notion is further validated by recent studies showing that term infants born at 37 to 38 weeks’ gestation have higher morbidity and mortality than those born at 39 weeks’ gestation (Hansen et al, 2008; Madar et al, 1999; McIntire and Leveno, 2008; Shapiro-Mendoza et al, 2008); this has prompted the use of early term to describe births at 37 to 38 weeks’ gestation. The term late preterm and the gestational age limits were established by a panel of experts convened by the National Institutes of Health and the National Institute of Child Health and Human Development in 2005. While developing these criteria, the group considered many factors, including the obstetric guidelines that consider 34 weeks to be a maturational milestone. Beyond 34 weeks’ gestation, surfactant is generally considered to be adequate and antenatal steroids are not offered to mothers with anticipated delivery (Raju et al, 2006). Unlike the smaller, more typical premature infant, late preterm infants appear mature because of their larger size, but have a higher incidence of transient tachypnea of the newborn (McIntire and Leveno, 2008; Wang et al, 2004), respiratory distress syndrome (RDS) (Clark, 2005; Wang et al, 2004), PPHN (Roth-Kleiner et al, 2003), respiratory failure, prolonged Although many of the diseases discussed in this section are not specific or unique to late preterm infants and are being covered in other chapters in this book, it is important to understand and recognize them as part of the unique challenge of caring for late preterm newborns. RESPIRATORY Several studies have consistently shown that late preterm infants have higher respiratory morbidity and mortality compared with full-term infants. Many late preterm infants develop respiratory distress soon after birth (sustained distress for more than 2 hours after birth accompanied by grunting, flaring, tachypnea, retractions, or supplemental oxygen requirement), which studies show occurs more often in late preterm infants than in term newborns (28.9% versus 4.2%, respectively) (Wang et al, 2004). In addition, within the early term and late preterm groups, infants born at 37 weeks’ gestation are fivefold more likely, and babies born at 35 weeks’ gestation are ninefold more likely, to have respiratory distress compared with babies born at 38 to 40 weeks’ gestation (Escobar et al, 2006). For each gestational week of age, infants delivered by elective cesarean section tend to do worse (Figure 33-4). In fact, Madar et al (1999) found that the incidence of respiratory distress was significantly increased with every week of gestation less than 39 weeks: 30 in 1000 infants born at 34 weeks’ gestation developed respiratory distress, 14 in 1000 born at 35 weeks’ gestation, and 7.1 in 1000 born at 36 weeks’ gestation. The etiology of respiratory distress is diverse and includes transient tachypnea of the newborn, RDS, persistent pulmonary hypertension, and apnea. Not surprisingly, of the affected babies, the incidence of respiratory 407 CHAPTER 33 Care of the Late Preterm Infant 30000 Number of patients 25000 20000 15000 10000 5000 0 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 Estimated gestational age (wks) 3.5 3.3 PVD ECD EmCD 45 40 35 3.0 2.5 30 Percent of infants Respiratory morbidity, % (95% CI) 50 FIGURE 33-3 Distribution of neonatal intensive care unit (NICU) admissions by gestational age, highlighting the contribution made by late preterm and early preterm infants. Data were obtained from a large consortium of NICUs under a common management. (Adapted from Clark RH: The epidemiology of respiratory failure in neonates born at an estimated gestational age of 34 weeks or more, J Perinatol 25:251-257, 2005.) 25 20 15 10 5 2.0 1.5 1.0 0.5 0 34 35 36 37 38 39 40 41 42 43 GA stratum, wk FIGURE 33-4 Respiratory morbidity in late preterm and early term infants and the impact of mode of delivery. ECD, Elective cesarean delivery; EmCD, emergency cesarean delivery; PVD, planned vaginal delivery. (Adapted from De Luca R, Boulvain M, Irion O, et al: Incidence of early neonatal mortality and morbidity after late-preterm and tern cesarean section, Pediatrics 123:e1064-e1071, 2009.) distress requiring mechanical ventilation corresponded to the degree of prematurity: 3.3% of late preterm infants born at 34 weeks’ gestation, 1.7% at 35 weeks’ gestation, and 0.8% at 36 weeks’ gestation (Figure 33-5) (McIntire and Leveno, 2008). Whereas respiratory issues often tend to be transient in a vast majority of these neonates, some develop into PPHN or severe hypoxic respiratory failure requiring additional therapies such as nitric oxide, high frequency ventilation, and extracorporeal membrane oxygenation (ECMO; Heritage and Cunningham, 1985; Keszler et al, 1992). Studies have shown that pulmonary hypertension is more likely in preterm infants (born at 34 to 37 weeks’ gestation) who develop RDS than in similar infants born at 32 weeks’ gestation. Such predisposition is attributed to a developmental increase in smooth muscle in the walls of pulmonary blood 1.7 0.8 0.3 0 34 wks 35 wks 36 wks 39 wks FIGURE 33-5 Percentage of infants born at late preterm gestations who require mechanical ventilation. (Adapted from McIntire DD, Leveno KJ: Neonatal mortality and morbidity rates in late preterm births compared with births at term, Obstet Gynecol 111:35-41, 2008.) vessels. PPHN is associated with increased pulmonary vascular resistance that eventually leads to right-to-left shunting by means of fetal pathways and ventilation-perfusion mismatching (Dudell and Jain, 2006). Management of neonates who develop significant pulmonary hypertension can be challenging, given the self propagated nature of hypoxia-induced pulmonary vasoconstriction. Treatment options include exogenous surfactant (shown to be effective if used earlier in the disease course; Dudell and Jain, 2006), inhaled nitric oxide (selectively lowers pulmonary vascular resistance and decreases extrapulmonary right-toleft shunting) (Kinsella et al, 1992, 1993), high-frequency ventilation, and ECMO. A review of the Extracorporeal Life Support Organization Neonatal Registry from 1989 to 2006 by Dudell and Jain (2006) found that 14.5% of the ECMO patients during that time period were late preterm infants and had a mean gestational age of 35.3 weeks. Interestingly, affected 408 PART VIII Care of the High-Risk Infant Alveolus AMILORIDE Na Na, K Cl ENaC NSC Paracellular Na, K, Cr H2O ENaC HSC CNGC Na, K AQP5 ENaC ENaC HSC NSC Na Na, K H2O AQP5 T1 cell Na Interstitium K Na, K-ATPase Cl CFTR Paracellular Na, K, Cl T2 cell CT CLC OUABAIN Na K Na, K-ATPase Pulmonary capillary FIGURE 33-6 Epithelial sodium (Na) absorption in the fetal lung near birth. Na enters the cell through the apical surface of both ATI and ATII cells via amiloride-sensitive epithelial Na channels (ENaC), both highly selective channels (HSC) and nonselective channels (NSC), and cyclic nucleotide gated channels (seen only in ATI cells). Electroneutrality is conserved with chloride movement through cystic fibrosis transmembrane conductance regulator (CFTR) or through chloride channels (CLC) in ATI and ATII cells, and/or paracellularly through tight junctions. The increase in cell Na stimulates Na-K-ATPase activity on the basolateral aspect of the cell membrane, which drives out three Na ions in exchange for two K ions, a process that can be blocked by the cardiac glycoside ouabain. If the net ion movement is from the apical surface to the interstitium, an osmotic gradient would be created, which would in turn direct water transport in the same direction, either through aquaporins or by diffusion. (From Jain L: Respiratory morbidity in late-preterm infants: prevention is better than cure! Am J Perinatol 25:75-78, 2008.) infants were more likely to require ECMO secondary to hypoxic respiratory failure or RDS instead of aspiration syndromes, which is the primary insult for term infants requiring ECMO (Dudell and Jain, 2006). In addition, late preterm infants were older at cannulation (likely because most of these infants are asymptomatic at birth, but gradually develop an increasing oxygen requirement with subsequent development of PPHN), had a longer duration of ECMO support, and were more likely to have intraventricular hemorrhage and other neurologic complications than term infants. Overall survival rate was significantly lower (74%) for late preterm infants compared with term infants (87%) (Dudell and Jain, 2006). Why is it that, even in situations in which amniotic fluid testing shows a mature surfactant profile, late preterm infants are at risk for developing respiratory distress? Part of the answer lies in the delay in clearing fetal lung fluid. Throughout much of gestation, fetal lungs actively secrete fluid into alveolar spaces via a chloride secretory mechanism. This process can be blocked by inhibitors of Na-K-2Cl co-transport. The fluid that accumulates in the developing lung plays a critical role by providing a structural template that prevents the collapse of the developing lung and promotes its growth. At the time of delivery, the lung epithelium becomes integral in the process of switching from placental to pulmonary gas exchange (Bland, 2001; Jain, 1999). For effective gas exchange to occur in the lungs, alveolar spaces must be cleared of excess fluid, and pulmonary blood flow must be increased to match ventilation with the perfusion that is taking place. If either the ventilation or perfusion is inadequate, the infant will have a difficult time transitioning and will develop respiratory distress. In addition, during fetal development, many abnormalities can occur and interfere with the normal production of this lung fluid. Some problems during development include pulmonary artery occlusion, diaphragmatic hernia, and uterine compression of the fetal thorax from chronic leak of amniotic fluid. All these conditions inhibit normal lung development and growth (Jain and Eaton, 2006). Although a small role in the clearance of this fluid can be attributed to Starling forces and “vaginal squeeze” (Bland, 2001; Jain, 1999), amiloride-sensitive sodium transport by lung epithelial cells through epithelial sodium channels (ENaCs) has emerged as a key event in the transepithelial movement of alveolar fluid (Figure 33-6) (Bland, 2001; Jain et al, 2001). Research has shown that these ENaCs orchestrate the clearing of fluid from the fetal lungs, and disruption of their function has been implicated in several disease processes affecting the newborn, including transient tachypnea of the newborn and hyaline membrane disease. The late preterm infant is more susceptible to these problems, in part because ENaC expression is developmentally regulated and peak expression in the alveolar epithelium is achieved only at term gestation, which leaves the preterm infant with lower expression of these channels, thus reducing their ability to clear fetal lung fluid after birth (Smith et al, 2000). High doses of glucocorticoids have been shown to stimulate transcription of ENaCs in several sodium transporting epithelia and in the lung (Tomashek et al, 2007). In the alveolar epithelia, glucocorticoids were found to induce lung sodium reabsorption in the late gestation fetal lung (Tomashek et al, 2007). In addition to increasing transcription of sodium channel subunits, steroids increase the number of available channels, by decreasing the rate at which membrane-associated channels are degraded, and increase the activity of existing channels. Glucocorticoids have also been shown to enhance the responsiveness of lungs to β-adrenergic agents and thyroid hormones (Venkatesh and Katzberg, 1997). In addition to problems with lung fluid clearance, several other factors may contribute to the overall burden of respiratory morbidity (Hansen et al, 2008; Kolas et al, 2006; Levine et al, 2001; Morrison et al, 1995; Roth-Kleiner et al, 2003; Villar et al, 2007). Given the shortcomings clinicians face in accurate estimation of gestational age, elective induction and cesarean section may have increased the burden of iatrogenic CHAPTER 33 Care of the Late Preterm Infant prematurity. In an attempt to minimize the occurrence of iatrogenic RDS in light of the increasing frequency of elective cesarean sections—commonly performed between 37 and 40 weeks’ gestation (Hales et al, 1993)— fetal lung maturity testing was recommended before elective cesarean sections. Because of the risks and complications associated with amniocentesis, this testing is done infrequently (Dudell and Jain, 2006), especially in light of recent studies showing that even late preterm infants and some early term infants born by cesarean section before the onset of labor have respiratory distress despite having mature surfactant profiles. This finding prompted the American College of Obstetrics and Gynecologists (2002) to recommend scheduling elective cesarean section at 39 weeks or later or waiting for the onset of spontaneous labor, but unfortunately factors related to the convenience of scheduled elective cesarean section deliveries for both families and providers will continue to influence the timing of elective cesarean section (Dudell and Jain, 2006). GASTROINTESTINAL Nutrition Feeding problems are one of the primary reasons for delay in the discharge of late preterm infants (Adamkin, 2006). Late preterm infants often have poor coordination of sucking and swallowing because of neuronal immaturity, decreased oromotor tone, and inability to generate adequate intraoral pressures during sucking (Engle et al, 2007; Kinney, 2006; Polin et al, 2003; Raju et al, 2006). Breastfeeding has also been shown to be more difficult for early term or late preterm infants compared with term infants (Raju, 2006). These problems can lead to poor caloric intake and dehydration. These problems are compounded by the variations in practice and nutritional management of these infants, given the paucity of published studies in this regard. Recent studies have shown that issues such as hypoglycemia and poor feeding contributed to 27% of all late preterm babies requiring intravenous fluids, compared with only 5% of their term counterparts (Wang et al, 2004). In the face of poor enteral intake, TPN is indicated and can become an important therapy in the care of the late preterm infant, but is often delayed in anticipation of a quick recovery (Adamkin, 2006). The challenge then becomes providing adequate nutrition to support growth and equate the energy expenditure that can occur when the infant faces issues such as hypothermia, sepsis, and respiratory distress, which are often seen in late preterm infants. Studies show that the energy expenditure of nongrowing low-birthweight infants (birthweight less than 2500 g) is 45 to 55 cal/ kg/day (Adamkin, 2006). These calories come from several sources in the TPN including amino acids and lipids. Late preterm infants are more adept at handling amino acids, allowing the protein content in TPN to be started at 2 g/kg/day. With a protein intake of 2.5 to 3 g/kg/day (with adequate caloric intake), a late preterm infant can achieve weight gain similar to a term infant fed human milk (Adamkin, 2006). More controversial is the use of intravenous lipids in late preterm infants. Of the late 409 preterm infants with respiratory distress or disease, there are two subgroups: infants with parenchymal lung disease without increased pulmonary vascular resistance (PVR) and those with signs of PPHN or increased PVR (Adamkin, 2006). The concern over the use of lipids in the late preterm infant with lung disease stems from adult studies showing that failure to clear infused lipids has an adverse effect on gas exchange in the lungs (Greene et al, 1976). Contrary to those findings, preterm neonates randomized to different lipid infusion rates did not demonstrate any effect on alveolar-arterial oxygen gradient, arterial blood pH, or oxygenation when randomly assigned to modest doses of lipids (0.6 to 1.4 g/kg/day) over the first week of life (Adamkin, 2006). The other argument for restricted use of lipids in late preterm infants specifically addresses the infants with increased PVR and respiratory disease. The concern is that the high polyunsaturated fatty acid content of lipid emulsions (with excess omega 6-linoleic acid) feeds into the arachidonic acid pathways, leading to synthesis of prostaglandins and leukotrienes, which can increase vasomotor tone and result in hypoxemia (Adamkin, 2006). Despite the lack of firm evidence for the effects of lipid emulsions in infants with severe respiratory failure with or without pulmonary hypertension, the recommendation is that infants with respiratory disease, but not increased PVR, should receive adequate lipids to prevent essential fatty acid deficiency; in infants with elements of PPHN, lipids should be avoided during the critical stages of their illness (Adamkin, 2006). Because of these issues and other concerns about parenteral nutrition (e.g., difficult in optimizing nutrition, the need for intravenous access and the potential for infiltrates or infection, risk of cholestatic jaundice with prolonged use of parenteral nutrition) enteral feeds should be started as soon as clinically possible while the infant is slowly weaned from the parenteral nutrition. In general, nutritional experts recommend that 34- and 35-week late preterm infants receive nutrient-enriched (22 kcal/oz) milk, whereas older 36- and 37-week late preterm infants with an uncomplicated neonatal course be fed unfortified milk after discharge (Adamkin, 2006). These nutrient-enriched formulas have a higher protein content (1.9 versus 1.4 g/dL), increased energy (22 versus 20 kcal/oz), additional calcium, phosphorous, zinc, trace elements, and vitamins compared with standard formulas (Adamkin, 2006). This enrichment becomes essential to the late preterm infant who was born at 34 to 35 weeks’ gestation or the older group of late preterm infants who had a difficult NICU course, where the goal is to compensate for earlier deprivation of adequate nutrition and allow for somatic and brain growth during the first year of life. These issues notwithstanding, these nutritional guidelines are not always followed, leading to variability in nutritional practices by providers. One study showed that, although nearly 46% of late preterm infants were discharged home with recommendations to use formula that contained more than 20 kcal/oz, this practice recommendation had a broad range of followers (4% to 72%) (Adamkin, 2006). The issue becomes more pressing in the late preterm infant with chronic conditions, such as bronchopulmonary dysplasia, that are often associated with growth failure caused by inadequate nutrient intake. 410 PART VIII Care of the High-Risk Infant BOX 33-2 Causes of Hypoglycemia in the Late Preterm Infant TRANSIENT HYPOGLYCEMIA IN THE LATE PRETERM INFANT Maternal Conditions ll Glucose infusion in the mother ll Preeclampsia ll Drugs: tocolytic therapy, sympathomimetics ll Infant of diabetic mother Neonatal Conditions ll Prematurity ll Respiratory distress syndrome ll Twin gestation ll Neonatal sepsis ll Perinatal hypoxia-ischemia ll Temperature instability: hypothermia ll Polycythemia ll Specific glucose transporter deficiency ll Isoimmune thrombocytopenia, Rh incompatibility PERSISTENT HYPOGLYCEMIA IN THE LATE PRETERM INFANT Endocrine Disorders ll Pituitary insufficiency ll Cortisol deficiency ll Congenital glucagon deficiency Inborn Errors of Metabolism Carbohydrate metabolism: glycogen storage disease, galactosemia, fructose1,6-diphoshatase deficiency ll Amino acid metabolism: maple syrup urine disease, propionic academia, methylmalonic academia hereditary tyrosinemia ll Fatty acid metabolism: acyl-coenzyme dehydrogenase defect, defects in carnitine metabolism, beta-oxidation defects ll Defective glucose transport ll From Garg M, Devaskar SU: Glucose metabolism in the late preterm infant, Clin Perinatol 33:853-870, 2006. Reprinted with permission. For mothers who choose to breastfeed their late preterm infant, it can often be more challenging compared with nursing a full-term infant. The challenge often lies in initiating and establishing breastfeeding because these infants are sleepier; have less stamina; have more difficulty maintaining body temperature; have problems with latching, sucking, and swallowing; and have more respiratory instability than full-term infants (Adamkin, 2006). Despite these obstacles, mothers should still be encouraged to provide breast milk given the numerous proven benefits of breast milk. In fact, recent studies have shown that the advantages of breast milk feeding for premature infants might be even greater than those for term infants (Adamkin, 2006). Hypoglycemia Hypoglycemia is defined as low circulating glucose concentrations, but the actual neonatal threshold value is still debated. A physiologic definition was established, based on abnormal electroencephalograms at glucose levels that were lower, with a glucose level less than 45 mg/dL being considered hypoglycemia (Koh et al, 1988). Hypoglycemia is often missed in late preterm infants, mainly because of the early transition of these infants to the well baby nursery or the mother’s room in an effort to triage the limited number of acute care beds in the NICU and to allow the mother to bond with her new baby (Garg and Devaskar, 2006). However, developmental immaturity is associated with multiple problems, including decreased glycogen stores and feeding difficulties, both of which can lead to hypoglycemia (Box 33-2). Not surprisingly, the incidence of hypoglycemia in preterm infants is threefold greater than in full-term infants (Wang et al, 2004). In addition, severe hypoglycemia is a well-known risk factor for neuronal cell death and adverse neurodevelopmental outcomes (Garg and Devaskar, 2006). Therefore, if the hypoglycemia is not recognized and treated in a timely manner with intravenous fluids or feedings, the infant can develop neurodevelopmental abnormalities because the compensatory mechanisms for protecting the brain from hypoglycemia are not fully developed (Cornblath and Ichord, 2000; Cornblath et al, 2000; Rozance and Hay, 2006; Vannucci and Vannucci, 2001). Therefore early recognition, diagnosis, and treatment of hypoglycemia are crucial to the late preterm infant’s long-term outcome. It is recommended that institutions develop protocols for routine testing of blood sugars in late preterm infants. One can use existing serum glucose screening protocols for infants at high risk for hypoglycemia (i.e., small for gestational age, large for gestational age, infant of a diabetic mother). If none are available, the following is recommended: glucose checks between 1 and 2 hours after birth, followed by testing before the next three consecutive feeds and then before alternate feedings for the remainder of the first 24 hours. If the blood sugar is less than 40 to 45 mg/dL, the hypoglycemia protocol should be followed for management. Hypoglycemia is not a problem in utero, because the fetus receives a steady supply of glucose primarily by maternal transfer through the placenta. Once the baby is delivered, this constant supply of glucose is abruptly stopped and the infant has to rely on glucose production primarily via hepatic glycogenolysis and gluconeogenesis (Halamek et al, 1997). After birth, the baby experiences a surge in catecholamines, glucagon, and corticosteroids, which play a key role in maintaining a euglycemic state. The increase in catecholamines leads to a surge in glucagon concentration and a decline in circulating insulin concentrations, which both contribute to maintaining a normal serum glucose level. Glucose levels are also affected by the unregulated insulin production by the immature pancreatic β cells (Garg and Devaskar, 2006). As a result, the late preterm newborn can experience significant hypoglycemia secondary to developmentally immature hepatic enzyme systems for gluconeogenesis, glycogenolysis, and hormonal dysregulation (Engle et al, 2007; Garg and Devaskar, 2006; Raju et al, 2006). The neonatal glucose requirement is 6 to 8 mg/kg/min, which is a higher value than that observed in adults (3 mg/ kg/min) (Bier et al, 1977). This demand for glucose increases if the late preterm infant has coexisting conditions such as sepsis, birth asphyxia, or cold stress (Greisen and Pryds, 1989; Halamek et al, 1997; Halamek and CHAPTER 33 Care of the Late Preterm Infant Stevenson, 1998). Treatment options for hypoglycemia in the late preterm infant include establishing early feeds (supplementing with formula if quantity of breast milk is insufficient), glucose infusion through intravenous fluids, hydrocortisone, glucagon, epinephrine, diazoxide, and octreotide (Garg and Devaskar, 2006). The treatment choice for the late preterm infant is based on the underlying cause of the hypoglycemia. Regardless of the etiology of the hypoglycemia, it is important to have constant monitoring of the glucose levels until they stabilize and the baby is tolerating adequate nutrition. Hyperbilirubinemia Hyperbilirubinemia is the most common clinical condition requiring evaluation and treatment in the late preterm newborn and the most common cause for readmission during the first postnatal week of life (Bhutani et al, 2004; Brown et al, 1999; Escobar et al, 2005; Maisels and Kring, 1998). Studies show that late preterm infants are more likely than term infants to be rehospitalized for jaundice (4.5% versus 1.2% in term infants) (Escobar et al, 2005). In general, neonatal hyperbilirubinemia in late preterm infants is more prevalent, more pronounced, and more protracted than in their term counterparts. Important risk factors for severe jaundice are summarized in Figure 33-7, A. A study by Newman et al (1999) showed that infants born at 36 weeks’ gestation have an eightfold increase in the risk of developing a total serum bilirubin concentration greater than 20 mg/dL (343 μmol/L) when compared with those born at 41 weeks’ gestation or later. Part of the reason for this increased risk is the immature hepatic metabolic pathways for bilirubin and the overall immaturity of gastrointestinal function and motility. The decreased ability for hepatic uptake and conjugation puts the late preterm infant at increased risk of elevated serum bilirubin levels, and the jaundice then becomes more prolonged, prevalent, and severe (Bhutani and Johnson, 2006). In addition, late preterm infants are at increased risk of kernicterus at bilirubin levels equal to or lower than that of term infants (Bhutani and Johnson, 2006). Kernicterus is a devastating, chronic, and disabling condition characterized by the tetrad of choreoathetoid cerebral palsy, neural hearing loss, palsy of vertical gaze, and dental enamel hypoplasia (Watchko, 2006). Whereas the need for universal predischarge bilirubin testing in neonates is debated, it is generally accepted that late preterm infants are at higher risk and should not be included with term infants. The current recommendation from the American Academy of Pediatrics from 2004 for the management of hyperbilirubinemia recommends that all newborns be assessed for their risk of developing hyperbilirubinemia by using predischarge total serum bilirubin or transcutaneous bilirubin measurements (Kuzniewicz et al, 2009). The effectiveness of this policy was studied by Kuzniewicz et al (2009), and they concluded that universal bilirubin screening, whether using transcutaneous bilirubin or total serum bilirubin for measurements, was associated with increased identification of newborns needing phototherapy and a significantly lower incidence of severe hyperbilirubinemia (see Figure 33-7, B and C). 411 This finding underscores the need for close monitoring of late preterm infants, particularly breastfed infants whose mothers may not have a proper milk supply before being discharge home. Early discharges should be avoided until proper feeding has been established, and early follow-up should be arranged. INFECTIOUS DISEASES Late preterm infants are also more susceptible to infections because of their immunologic immaturity. The timing of the infection categorizes them as congenital (acquired before delivery), early onset (usually acquired during delivery and presenting within the first 72 hours), or late onset (often acquired in the hospital and presents after 72 hours of life). Congenital infections are commonly attributed to rubella, cytomegalovirus, herpes simplex virus, and HIV (Benjamin and Stoll, 2006). The severity of the infection and its effect on the late preterm infant depends on the stage of pregnancy at which the maternal infection occurred. With both herpes simplex virus and HIV, maternal-infant transmission is more frequent if a mother has a primary infection at the time of delivery, whereas maternal viral load is the main risk factor for HIV transmission from mother to newborn (Benjamin and Stoll, 2006). Early-onset sepsis is almost always caused by perinatally acquired infections. In most cases, the late preterm infant is initially colonized by exposure to various organisms in the maternal genital tract including group B Streptococcus (GBS), Escherichia coli, and Candida spp. Additional risk factors for developing sepsis are prolonged rupture of membranes (greater than 18 hours), maternal fever, and chorioamnionitis (Centers for Disease Control and Prevention, 2002). The third group—the late-onset sepsis—may be caused by perinatally or postnatally acquired organisms, but usually is a consequence of nosocomial transmission. The most common organisms are gram-negative rods, but sepsis could be caused by Staphylococcus aureus, Enterobacter spp., or Candida spp. Although the mortality rate is low for late preterm infants, infections increase the risk of complications and often involve longer hospital stays (Benjamin and Stoll, 2006). In addition, research shows that late preterm infants undergo testing for sepsis more often than term infants (36.7% versus 12.6%; odds ratio, 3.97; 95% confidence interval, 1.82 to 9.21; p = 0.00015) and receive antibiotics more often and for a longer duration (7-day course 30% versus 17% in term infants) (Wang et al, 2004). Other studies show that the need for a sepsis evaluation increases with decreasing gestational age; 33% were evaluated for possible sepsis at 34 weeks’ gestation compared with 12% at 39 weeks’ gestation (p <0.01), of which only 0.4% of infants had culture-proven sepsis (McIntire and Leveno, 2008). This higher frequency of screening late preterm newborns for sepsis compared with term infants may be multifactorial. First, records show that one third of all preterm deliveries occur after prolonged premature rupture of membranes, which can put the newborn at a significantly higher risk of infection. In addition, this higher rate of sepsis workups in the late preterm baby may be 412 PART VIII Care of the High-Risk Infant A IMPORTANT RISK FACTORS FOR SEVERE HYPERBILIRUBINEMIA • Predischarge total serum bilirubin (TSB) or transcutaneous bilirubin (TcB) measurement in the high-risk or high-intermediate-risk zone • Lower gestational age • Exclusive breastfeeding, particularly if nursing is not going well and weight loss is excessive • Jaundice observed in the first 24 h • Isoimmune or other hemolytic disease (e.g., G6PD deficiency) • Previous sibling with jaundice • Cephalohematoma or significant bruising • East Asian race HOURLY PROGRESSION OF TOTAL SERUM BILIRUBIN (TSB) LEVELS: RISK CATEGORIES 25 428 20 342 15 257 10 171 5 Infants at lower risk (38 wk and well) Infants at medium risk (38 wk  risk factors or 35–37 6/7 wk and well) Infants at higher risk (35–37 6/7 wk  risk factors) 0 mol/L TSB, mg/dL B 85 0 Birth 24 h 48 h 72 h 96 h 5d 6d 7d Age C GUIDELINES FOR PHOTOTHERAPY IN HOSPITALIZED INFANTS 35 WEEKS’ GESTATION Note that these guidelines are based on limited evidence and that the levels shown are approximations. The guidelines refer to the use of intensive phototherapy, which should be used when the TSB level exceeds the line indicated for each category. • Use total bilirubin. Do not subtract direct-reacting or conjugated bilirubin. • Risk factors are isoimmune hemolytic disease, G6PD deficiency, asphyxia, significant lethargy, temperature instability, sepsis, acidosis, or an albumin level of 3.0 g/dL (if measured). • For well infants at 35 to 37 6/7 weeks’ gestation, on can adjust TSB levels for intervention around the medium-risk line. It is an option to intervene at lower TSB levels for infants closer to 35 weeks’ gestation and at higher TSB levels for those closer to 37 6/7 weeks’ gestation. • It is an option to provide conventional phototherapy in the hospital or at home at TSB levels of 2 to 3 mg/dL (35–50 mol/L) below those shown, but home phototherapy should not be used in any infant with risk factors. FIGURE 33-7 Hyperbilirubinemia. A, Risk factors. B, Hourly progression of total serum bilirubin levels. C, Management guidelines. (Adapted from Maisels MJ, Bhutani VK, Bogen D et al: Hyperbilirubinemia in the newborn infant > or = 35 weeks’ gestation: an update with clarifications, Pediatrics 124:1193-1198, 2009.) a reflection of a standard protocol used for admissions to the NICU or due to their clinical presentation (e.g., respiratory distress, hypothermia, hypoglycemia), which could be a sign of sepsis or a reflection of the infants’ immaturity. THERMOREGULATION Because of their relatively smaller size, the late preterm infant is susceptible to periods of hypothermia or cold stress. Unfortunately, as for other problems discussed earlier, this may be difficult to assess if the late preterm newborn has 413 CHAPTER 33 Care of the Late Preterm Infant been sent to the mother’s room or is not closely observed. Usually cold stress will manifest as tachypnea or apnea, poor feeding, poor color caused by peripheral vasoconstriction, and metabolic acidosis. Hypothermia and its related consequences can delay the respiratory transition and exacerbate hypoglycemia; these signs and symptoms may also be misinterpreted as possible sepsis, which then leads to unnecessary interventions and workups. The reason that late preterm infants are particularly more susceptible to temperature instability is because of their physiologic immaturity of thermoregulation, which in turn is dependent on three main things: the amount of brown adipose tissue, white adipose tissue, and body surface area (Engle et al, 2007; Martin et al, 2006; Polin et al, 2003). Nonshivering thermogenesis is controlled by the hypothalamic ventromedial nucleus through the sympathetic nervous system, which releases the neurotransmitter norepinephrine. The norepinephrine then causes the brown adipose tissue to liberate free fatty acids, which are eventually oxidized and produce heat (Engle et al, 2007; Martin et al, 2006; Polin et al, 2003). Late preterm infants have decreased stores of brown adipose tissue and the hormones responsible for brown fat metabolism (i.e., prolactin, norepinephrine, triiodothyronine, and cortisol). These hormones peak at term gestation and the late preterm infant misses those last few weeks of in utero development (Engle et al, 2007; Polin et al, 2003). In addition to the decreased stores of hormones leading to thermogenesis, late preterm infants also have problems with hypothermia because of a decreased amount of white adipose tissue, which leads to less insulation, and their smaller size. The late preterm infant’s relatively smaller size, compared with term infants, leads to an increased ratio of surface area to body weight, which allows for greater heat loss to the environment (Engle et al, 2007; Martin et al, 2006; Polin et al, 2003). Appropriate monitoring and triaging of the late preterm infant who is susceptible to temperature instability can avoid unnecessary morbidity, workups, interventions, and prolonged hospitalizations. NEURODEVELOPMENTAL During pregnancy, the infant’s lungs and brain are among the last organs to mature and are therefore more prone to injury. Not surprisingly, research has shown that even healthy near-term or late preterm infants are at risk for developmental delays through the first 5 years of life (Raju, 2006). During the final few weeks of gestation, many aspects of brain maturity are still in progress. These aspects include maturing oligodendroglia, increasing neuronal arborization and connectivity, maturation of neurotransmitter systems, and continued brain growth that accounts for a 30% increase in brain size during the last few weeks of gestation (Figure 33-8) ( Jain and Raju, 2006). At 34 weeks’ gestation, the brain weighs only 65% of the weight of a 40-week term infant (Billiards et al, 2006; Kinney, 2006). Researchers have shown that the brain of a late preterm infant is still immature and continues to grow until 2 years of age, when it reaches 80% of adult brain volume. In addition, the cerebral cortex is still smooth and the gyri and sulci are not fully formed, and myelination and interneuronal connectivity is still 20 weeks 35 weeks 40 weeks Near-term FIGURE 33-8 The immaturity of the laminar position and dendritic arborization of neurons, as demonstrated by Golgi drawings, in the cerebral cortex in the late preterm infant at 35 weeks’ gestation is striking in comparison with neurons at mid-gestation (20 weeks) and at term (40 weeks). (From Kinney HC, Armstrong DD: Perinatal neuropathology. In Graham DI, Lantos PE, editors: Greenfield’s neuropathology, ed 7, London, 2002, Arnold, pp 557-559.) incomplete. Multiple insults during this critical phase of neuronal and glial maturation cause white and grey matter injury, especially in the thalamic region and the periventricular white matter (Kinney, 2006). These events can be correlated subsequently to delayed development and special education needs; therefore it is important to start early developmental follow-up, anticipatory guidance, and interventions for infants born at 32 to 36 weeks’ gestation (Chyi et al, 2008). HOSPITALIZATION OF THE LATE PRETERM INFANT Based on the need for close monitoring and management of the various medical problems identified in late preterm infants, they are more likely than a full-term infant to require admission to an intensive care unit. Despite this finding, individual hospitals and nurseries follow different criteria regarding which infants to admit to the NICU, an intermediate care unit, or an observation area. Some routinely admit all infants <35 weeks, gestation to the NICU, whereas others do so on an individual basis. The most common reasons for admission include temperature instability, jaundice, respiratory distress, dehydration, poor feeding, and hypoglycemia (Vachharajani and Dawson, 2009; Wang et al, 2004). Studies have shown that 88% of infants born at 34 weeks’ gestation, 12% born at 37 weeks’ gestation, and 2.6% born at 38 to 40 weeks’ gestation were admitted to the NICU (Engle and Kominiarek, 2008). Other studies have shown similar rates, and the overall trend was that the late preterm infant had significantly higher rates of NICU admission than the 39-week infants (McIntire and Leveno, 2008). In addition, the duration of hospitalization for the late preterm infant is inversely proportional to the baby’s gestational age, which means that late preterm infants require longer hospitalization after birth than their term counterparts. On average, infants are hospitalized for 6 to 11 days at 34 weeks’, 4 to 6 days at 35 weeks’, and 3 to 4 days at 36 weeks’ gestation (Escobar 414 Early Neonatal Mortality (1-7 days) Infant ­Mortality (1-365 days) Gestational Age (weeks) Mortality Rate Risk Ratio Mortality Rate Risk Ratio 34 7.2 25.5 12.5 10.5 35 4.5 16.1 8.7 7.2 36 2.8 9.8 6.3 5.3 37 0.8 2.7 3.4 2.8 38 0.5 1.7 2.4 2.0 39 0.2 0.8 1.2 1.2 From Young PC, Glasgow TS, Li X, et al: Mortality of late-preterm (near-term) newborns in Utah, Pediatrics 119:e659-e665, 2007. et al, 2005; Gilbert et al, 2003; Khashu et al, 2009; McIntire and Leveno, 2008; Phibbs and Schmitt, 2006; Vachharajani and Dawson, 2009). MORTALITY Preterm birth is consistently recognized as the most pressing public health problem in perinatology by both clinicians and researchers, given its overall contribution to infant mortality. It also contributes substantially to neurocognitive, pulmonary, and ophthalmologic morbidity (Kramer et al, 2000). On review of infant birth and death files from 1995 to 2002 in the United States, Tomashek et al (2007) compared the overall and cause-specific mortality rates between singleton late preterm infants and term infants. They found that, despite significant declines since 1995 in mortality rates for late preterm and term infants, the infant mortality rate in 2002 was threefold higher in late preterm infants than in term infants (7.9 versus 2.4 deaths per 1000 live births); early, late, and postneonatal mortality rates were sixfold, threefold, and twofold higher, respectively. Another study by Young et al (2007) found that in a large cohort from Utah, the relative risk of death increased for every decreasing week in gestational age less than 40 weeks (Table 33-1). In addition, a large study involving 133,022 infants born at 34 to 40 weeks’ gestation found that neonatal mortality rates were significantly higher for late preterm infants (1.1, 1.5, and 0.5 per 1000 live births at 34, 35, and 36 weeks, respectively, compared with 0.2 per 1000 live births at 39 weeks’ gestation p <0.001) (McIntire and Leveno, 2008). In 2005, the U.S. infant mortality rate for late preterm infants was 7.3 versus 2.43 per 1000 live births in term neonates. Surprisingly, the mortality rate is 30% higher even for infants born between 37 and 39 weeks’ gestation (early term) (Mathews and MacDorman, 2008). These studies and statistics again emphasize the fact that infants born just a few weeks early are at a much greater risk of morbidity and mortality than those born at term gestation. However, perinatal data collected over similar periods also reveal a remarkable decline in stillbirth rate (Figure 33-9). Perinatologists argue that this reduction in fetal demise is directly related to close monitoring of the fetus and early intervention (delivery) when needed. 10 Infant deaths 9 9 Late preterm births 8 8 7 7 Stillbirths 6 6 0 infant deaths per 1000 births at Late Preterm and Early Term Gestations 10 Stillbirths and TABLE 33-1 Mortality (rate per 1000 live births) in Infants Born Percent of late preterm births (34–36 weeks) PART VIII Care of the High-Risk Infant 0 1990 1995 2000 2005 Year FIGURE 33-9 Trends in late preterm birth, stillbirth, and infant mortality in the United States, 1990 to 2004. The left axis shows trends in stillbirth and infant mortality rates; the right axis shows trends in late preterm births (34 to 36 weeks). Late preterm birth rates are shown per 100 live births; stillbirth rates per 1000 total births, and infant death rates per 1000 live births. (Adapted from Ananth CV, Gyamfi C, Jain L: Characterizing risk profiles of infants who are delivered at late preterm gestations: does it matter? Am J Obstet Gynecol 199:329-331, 2008.) RECOMMENDATIONS ADMISSION CRITERIA With more than 80% of all deliveries occurring in community hospitals (Jain and Raju, 2006), many of which have a relatively small number of deliveries, and health care teams that might not always be equipped to assess and manage the needs of a late preterm infant, it becomes increasingly important to establish safeguards for the screening, identification, and appropriate triage of these patients. A subcommittee on the American Academy of Pediatrics Committee on the Fetus and Newborn (Engle et al, 2007) outlined recommendations for the care of the late preterm newborn. Based on these guidelines and the potential complications associated with late preterm infants, we recommend that all infants born before 35 weeks’ gestation weighing less than 2300 g should be admitted to a transitional nursery where the infant can be monitored closely until there has been adequate time to assess the baby’s vital signs, feeding abilities, and thermoregulation, among other issues, before sending the baby to the mother’s room. Sicker neonates who require intensive care obviously will need to be admitted to higher levels of care. In addition, each nursery should establish guidelines for frequency of monitoring vital signs, assessment for sepsis and use of antibiotics, and the use of supplemental oxygen. It is also important to determine a threshold (based on comfort level, staff training, and available resources) for transferring the newborn to a tertiary care center when the disease process associated with the late preterm infant continues to progress or worsen. Box 33-3 shows the recommendations for admission, management, and discharge of the late preterm infant. This list is not all-inclusive and was designed to be used as a guideline, and not as a replacement for good clinical judgment. CHAPTER 33 Care of the Late Preterm Infant BOX 33-3 A  dmission and Discharge Criteria and Management of Late Preterm Infants ADMISSION CRITERIA Admit all infants born before 35 weeks’ gestation or weighing less than 2300 g at birth ll They should not to be sent to their mother’s rooms in the first 24 hours until stable, unless arrangements can be made to provide transitional care and close monitoring in the mother’s room. ll HOSPITAL MANAGEMENT Physical examination on admission and discharge ll Determination of accurate gestation age on admission examination ll Vital signs and pulse oximeter check on admission, followed by vital signs every 3 to 4 hours in the first 24 hours, and every shift thereafter ll Caution against use of oxyhoods with high Fio2; consider transfer to NICU or tertiary care center if Fio2 exceeds 0.4 ll A feeding plan should be developed. Formal evaluation of breastfeeding and documentation in the record by care givers trained in breastfeeding at least twice daily after birth ll Serum glucose screening per existing protocols for infants at high risk of hypoglycemia ll DISCHARGE CRITERIA Discharge should not be considered before 48 hours after birth ll Vital signs should be within normal range for the 12 hours preceding discharge ll Respiratory rate less than 60 breaths/min ll Heart rate of 100 to 160 beats/min ll Axillary temperature of 36.5° to 37.4° C measured in an open crib with appropriate clothing ll Passage of one stool spontaneously ll Adequate urine output ll Twenty-four hours of successful feeding: ability to coordinate sucking, swallowing, and breathing while feeding ll If weight loss is greater than 7% in 48 hours, consider further assessment before discharge ll Risk assessment plan for jaundice for infants discharged within 72 hours of birth ll No evidence of active bleeding at circumcision site for at least 2 hours ll Initial hepatitis B vaccine has been given or an appointment scheduled for its administration ll Metabolic and genetic screening tests have been performed in accordance with local or hospital requirements ll The late preterm infant has passed a car seat safety test ll Hearing assessment is performed and results documented in the medical record; follow-up if necessary has been arranged ll Parents have been trained and demonstrate competency in caring for the infant ll Family, environmental, and social risk factors have been assessed; when risk factors are present, discharge should be delayed until a plan for future care has been generated ll Identification of a physician with a follow-up visit arranged for 24 to 48 hours after discharge with a possibility of additional visits initially until the infant can demonstrate a consistent pattern of weight gain ll Adapted from Engle WA, Tomashek KM, Wallman C: “Late Preterm” Infants: a population at risk, Pediatrics 120: 1390-1401, 2007. Fio2, Fractional concentration of oxygen in inspired gas; NICU, neonatal intensive care unit. DISCHARGE CRITERIA Because of the morbidities and risk factors associated with late preterm babies, they should not be discharged before 48 hours after birth. Before discharge, 415 while the baby is still in the NICU, the following are recommended: 1. Vital signs should be within normal range for at least 12 hours preceding discharge; this includes respiratory rate less than 60 breaths/min, heart rate of 100 to 160 beats/min, and axillary temperature 36.5 to 37.5° C in an open crib with appropriate clothing. 2. There should be documentation of passage of at least one stool spontaneously. 3. Adequate urine output should be accompanied by educating the parents about ways to assess the adequacy of output and appropriate interventions if the urine output appears to decrease, with at least 24 hours of successful feeding with adequate coordination of sucking, swallowing, and breathing during feedings. 4. Weight loss should not exceed 7% of birthweight in the first 48 hours of life. 5. Serum or transcutaneous bilirubin check—a transcutaneous bilirubin higher than 12 mg/dL should warrant a serum bilirubin check, which will then be stratified into risk category by using a bilirubin nomogram. Parents should also be educated on what to look for and what to do if their baby appears jaundiced. 6. Hearing screen, a car seat test, and metabolic and genetic screening tests should have been performed in accordance with state, local, and hospital protocols; if the baby is circumcised, there should be no bleeding at the site for at least 2 hours. 7. Hepatitis B vaccine should be given or an appointment should be made for its administration. 8. Parents should be educated about umbilical cord and skin care, identification of common signs and symptoms of illness, sleeping patterns and positions, instructions on using the thermometer and parameters for normal measurements, and instructions regarding responses to an emergency (i.e., CPR training before discharge). If these guidelines (or other criteria outlined as standard of care) are not met, we recommend considering postponing discharge until the baby has been observed for a longer period of time and the issues have been resolved. When additional risk factors are present (e.g., twin or multiple gestation, teenage mother), discharge should be delayed until an appropriate care plan has been generated. In addition, when indicated, it may be appropriate to arrange a nursing home health visit for closer monitoring, but this should not be used to replace the due diligence that must be done while the baby is in the hospital and the appropriate and timely follow-up with a pediatrician. Follow-up after Discharge After discharge from the hospital, the majority of the medical care a newborn baby receives occurs in two main settings: the primary care physician’s office and the emergency department. To avoid fragmented care by multiple emergency department visits and to allow for an early assessment of the baby, it is recommended that the late preterm baby should be brought for a checkup by their pediatrician within 24 to 48 hours after discharge from the hospital. 416 PART VIII Care of the High-Risk Infant In addition, because of their increased risk for developmental delays, these infants should be monitored closely to ensure that all milestones are achieved appropriately and that early intervention (e.g., physical therapy, occupational therapy, speech therapy) is in place if needed. Early developmental testing can also be useful in determining any cognitive delays, which can then be addressed with individualized educational programs. Readmission to the Hospital The late preterm infant is susceptible to many of the problems of smaller preterm infants. Because of the multiple factors discussed throughout this chapter, the close observation and intensity of management provided to smaller premature infants is often lacking. In addition, there are often more lenient criteria for discharge, which sets them (and their parents) up for failure and eventual readmission to the hospital. Tomashek et al (2007) looked at late preterm infants who were discharged early (<2 days after birth) from the hospital and found that 4.3% of late preterm and 2.7% of term infants were either readmitted or had an observational hospital stay. Recent data suggest that the most frequent causes of emergency department visits by this subgroup of premature infants after being discharged home are dehydration, feeding problems, respiratory distress, apnea, fever, infection, and jaundice (Jain and Cheng, 2006). Whether they are evaluated in the emergency department or a pediatrician’s office, it is important to have a lower threshold for readmitting these infants because of their vulnerability to serious complications. Their fragility is evidenced by their higher rate (4.4%) of rehospitalization than term infants (2%; Escobar et al, 2005). In addition, studies show that 34- to 36-week infants who were never admitted to the NICU (or if admitted, their NICU stay lasted less than 24 hours) had nearly a threefold and 1.3-fold higher risk of readmission compared with term infants, respectively (Escobar et al, 1999, 2005, 2006). Because late preterm infants are at a much higher risk for rehospitalization, they need close follow-up after discharge to assess breastfeeding and nutrition and to monitor for jaundice. BOX 33-4 U  nanswered Questions and Future Directions EPIDEMIOLOGY, TRENDS, ETIOLOGY, AND PREVENTION OF LATE PRETERM BIRTHS ll Is there a subset of preventable late preterm births? ll What is the effect of continuing such pregnancies (preterm rupture of membranes, preterm labor, medically indicated birth) on perinatal outcomes? ll Is there a role for antenatal steroids in late preterm gestation pregnancies threatened with preterm delivery? ll Should singleton and multiple gestation late preterm deliveries be treated differently? CLINICAL MANAGEMENT AND OUTCOMES OF LATE PRETERM INFANTS ll Can standardized “care paths” for late preterm infants improve outcomes? ll Optimal discharge strategies that minimize readmissions and other problems after discharge ll Strategies for management of rapidly progressive and/or severe morbidities in late preterm infants ll Standardized approach to central nervous system imaging and postdischarge follow-up. LONGTERM OUTCOMES Prospective studies of long-term outcomes of symptomatic and asymptomatic late preterm births to determine which infants should have long-term follow-up ll the understanding of the spectrum of morbidities in late preterm infants; they have also formed the basis for current strategies to standardize management and optimize outcomes. SUGGESTED READINGS Definition, Epidemiology, and Background Davidoff MJ, Dias T, Damus K, et al: Changes in the gestational age distribution among U.S. singleton births: impact on rates of late preterm birth, 1992 to 2002, Semin Perinatol 30:8-15, 2006. Engle WA, Tomashek KM, Wallman C: Late-preterm” infants: a population at risk, Pediatrics 120:1390-1401, 2007. Raju TN, Higgins RD, Stark AR, et al: Optimizing care and outcome for latepreterm (near-term) infants: a summary of the workshop sponsored by the National Institute of Child Health and Human Development, Pediatrics 118:1207-1214, 2006. Pathophysiology and Clinical Course FUTURE RESEARCH Since 1981, the number of premature births in the United States has increased by 30%. The majority of this increase is attributable to the increased number of late preterm infants (Pulver et al, 2009)—a unique and high-risk subgroup of premature infants. The identification and care of this group and their unique challenges have only recently been recognized as a high-priority area of research in the field of neonatology. There are many questions still to be answered regarding the care and outcomes associated with late preterm infants (Box 33-4). Topping this list of priorities is the need to know which babies are better off being delivered early, and for the rest, what can be done to take pregnancy as close to term as possible. Because answers to many issues related to preterm birth still remain elusive, clinicians have, rightly so, focused on ways to improve the postnatal management of these premature babies. New studies, albeit mostly retrospective, have added to Jain L: Alveolar fluid clearance in developing lungs and its role in neonatal transition, Clin Perinatol 26:585-599, 1999. McIntire DD, Leveno KJ: Neonatal mortality and morbidity rates in late preterm births compared with births at term, Obstet Gynecol 111:35-41, 2008. Wang ML, Dorer DJ, Fleming MP, et al: Clinical outcomes of near-term infants, Pediatrics 114:372-376, 2004. Mortality Kramer MS, Demissie K, Yang H, et al: The contribution of mild and moderate preterm birth to infant mortality. Fetal and Infant Health Study Group of the Canadian Perinatal Surveillance System, J Am Med Assoc 284:843-849, 2000. Tomashek KM, Shapiro-Mendoza CK, Davidoff MJ, et al: Differences in mortality between late-preterm and term singleton infants in the United States, 19952002, J Pediatr 151:450-456, 2007. Longterm Outcome Chyi LJ, Lee HC, Hintz SR, et al: School outcomes of late preterm infants: special needs and challenges for infants born at 32 to 36 weeks gestation, J Pediatr 153:25-31, 2008. Kinney HC: The near-term (late preterm) human brain and risk for periventricular leukomalacia: a review, Semin Perinatol 30:81-88, 2006. Complete references and supplemental color images used in this text can be found online at www.expertconsult.com C H A P T E R 34 Pharmacokinetics, Pharmacodynamics, and Pharmacogenetics Robert M. Ward, Steven E. Kern, and Ralph A. Lugo Dynamic changes in growth and physiologic maturation in newborns create unique complexities in drug therapy that affect absorption, distribution, metabolism, and elimination. Pharmacologic studies during this period of rapid growth and physiologic maturation reveal different patterns of change in kinetics, some related to postnatal age and others to postmenstrual age. These differences also provide information about varied patterns of maturation in drug metabolizing enzymes and pathways of elimination during early infancy. Therapeutic drug monitoring is used for a few drugs in neonatology, such as phenobarbital and gentamicin, when concentrations correlate with desired effects as well as adverse effects. Dosage adjustments to reach desired concentrations can be estimated at the bedside using simple calculations. Pharmacogenetics and pharmacogenomics of enzymes and receptors explain many variations among individuals in their responses to drugs. The immature, preterm newborn adds another level of complexity to drug metabolism, because of variability in the timing of the expression of these enzymes. PRINCIPLES OF NEONATAL THERAPEUTICS A thorough understanding of factors that affect drug concentrations helps to provide accurate, effective drug therapy and to identify the causes of therapeutic failure in neonates. Many important factors are not chosen consciously in a therapeutic plan, but have tremendous impact on its effectiveness. Because pharmacokinetics and pharmacodynamics in newborns follow the same general principles that govern drug actions in patients of any age, the diagnosis, drug selection, and administration needed to achieve a therapeutic goal must consider the effects of absorption, distribution, metabolism, and excretion on the dose-exposure relationship. When applied to the newborn, these principles should adjust for several unique physiologic and pharmacologic features of these immature patients, as outlined in Box 34-1 and discussed in detail here. DIAGNOSIS Effective treatment begins with an accurate diagnosis and assessment of clinical signs and symptoms. Although this principle applies to all areas of therapeutics, treatment in newborns presents special diagnostic challenges because the small size and fragility of such nonverbal patients may preclude useful, but inordinately invasive, diagnostic procedures. For example, many small immature newborns with chronic lung disease are treated for bronchospasm based on the findings of decreased air entry associated with desaturation and abnormal breath sounds. Relief of these symptoms with aerosolized bronchodilators may be interpreted as confirmation of the diagnosis. Although this interpretation may be correct, increased humidity or movement of the endotracheal tube bevel away from a pliable tracheal wall during an aerosol treatment may also produce the improvement. Therefore evaluation of any ineffective therapy should include reconsideration of the original diagnosis, just as conclusions about why a therapy succeeded should be made with a degree of skepticism. ABSORPTION Although most drug therapy for acute problems in the intensive care setting involves intravenous administration to ensure drug delivery to the site of action, this route is not always reliable in newborns (Gould and Roberts, 1979; Roberts, 1984). In a critical care setting, drugs should be infused toward the patient as close as possible to the site of vascular access. If a drug is injected away from the infant into a catheter with a slow infusion rate, the drug may reach the patient too slowly to achieve effective concentrations. Infusion solution filters may also prevent effective drug delivery by blocking large molecules, such as amphotericin, by direct adsorption of the drug to the filter, or by allowing a heavier drug to settle in the filtration chamber and mix slowly with the infusion solution. For drug therapy in which the driving force for tissue penetration is a concentration gradient between the circulation and the tissue (e.g., in meningitis), sustained low drug concentrations might be therapeutically suboptimal. Intramuscular administration of drugs to newborns is generally suboptimal, but might be used when there is difficulty establishing intravenous access. Absorption of drugs from an intramuscular injection site is directly related to muscle blood flow, which is usually reduced in patients experiencing hypothermia or shock, when intravenous access can be difficult, but intramuscular doses are unlikely to be absorbed effectively. Furthermore, intramuscular administration of drugs can sclerose tissue and cause sterile abscesses or create large intramuscular collections of the drugs, which are then absorbed slowly, producing a “depot effect” in which serum concentrations rise slowly over a prolonged period. Intramuscular administration of drugs, especially for multiple doses, should be avoided in newborns. Although oral administration of drugs is preferred for treatment of chronic illnesses in newborns, this route is not well studied, particularly in acutely ill premature infants. Studies in adults show that less drug is usually absorbed from the stomach than from the intestinal tract because of the smaller surface area and differences in pH. Many newborns experience gastroesophageal reflux associated with delayed gastric emptying, which might also alter drug bioavailability. Delayed gastric emptying postpones 417 418 PART VIII Care of the High-Risk Infant BOX 34-1 P  harmacologic Principles and Pitfalls in Management of the Very Low-Birthweight Infant I. Diagnosis A. Limited diagnostic procedures II. Absorption A. Intravenous 1. Drug injection away from patient 2. Uneven mixing of drugs and intravenous fluids 3. Delayed administration because of low flow 4. Part of the dose discarded with tubing changes B. Intramuscular 1. Poor perfusion limits absorption 2. Danger of sclerosis or abscess formation 3. Depot effect C. Oral 1. Poorly studied 2. Affected by delayed gastric emptying 3. Potentially affected by reflux 4. Passive venous congestion may occur with chronic lung disease, decreasing absorption III. Distribution, affected by A. Higher (85%) total body water (versus 65% in adults) B. Lower body fat—that is, approximately 1% body weight (versus 15% in term infants) C. Low protein concentration D. Decreased protein affinity for drugs IV. Metabolism A. Half-life prolonged and unpredictable B. Total body clearance decreased C. Affected by nutrition, illness, and drug interaction D. Affected by maturational changes V. Excretion: decreased renal function, both glomerular filtration rate and tubular secretion reaching peak serum drug concentrations and prolongs the absorption phase, which reduces the peak concentration. If the total absorbed dose is reduced, the area under the concentration time curve (AUC) will decrease along with drug exposure. Passive venous congestion of the intestinal tract from elevated right atrial pressures decreases drug absorption in adults and may do so in premature infants with severe bronchopulmonary dysplasia complicated by cor pulmonale (Peterson et al, 1980). The administration of medications to newborns in small volumes of formula or during continuous gastric feedings may also alter drug absorption by binding to proteins, lipids, carbohydrates, or minerals in the feeding. When enteral drug therapy fails, possible effects of feeding patterns on drug absorption and action should be considered. DISTRIBUTION In pharmacokinetics, distribution is the partitioning of drugs among various body fluids, organs, and tissues. The distribution of a drug within the body is determined by several factors, including organ blood flow, pH and composition of body fluids and tissues, physical and chemical properties of the drug (e.g., lipid solubility, molecular weight, ionization constant), and the extent of drug binding to plasma proteins and other macromolecules (Plonait and Nau, 2004; Ward and Lugo, 2005). Important differences among premature infants, children, and adults affect the distribution of drugs. Total body water varies from 85% in premature newborns to 75% in term newborns to 65% in adults (Friis-Hansen, 1961, 1971). Conversely, body fat content varies from 0.7% or less in extremely premature newborns to approximately 12% in term newborns (Friis-Hansen, 1971; Ziegler et al, 1976). These differences change the distribution of many drugs, especially polar, water-soluble drugs such as the aminoglycosides. Protein binding of drugs in the circulation is decreased in the premature newborn because of a smaller total amount of circulating protein and lower binding affinity of the protein itself (Aranda et al, 1976). With rare exceptions, only the free (not bound to protein) drug molecules cross membranes, exert pharmacologic actions, and undergo metabolism and excretion. Clinical measurements of serum or plasma drug concentrations usually reflect total circulating drug concentrations, which consist of both free and protein-bound drug. Thus, even when total circulating drug concentrations in the newborn may be low by adult standards, the free drug concentrations may be equivalent or even higher than those in the adult because of decreased protein binding in the newborn. METABOLISM Many drugs require metabolic conversion before elimination from the body. Biotransformation of a drug usually produces a more polar, less lipid-soluble molecule that can then be eliminated rapidly by renal, biliary, or other routes of excretion. Drug biotransformation is classified into two broad categories: nonsynthetic (phase I) reactions, which include oxidation, reduction, and hydrolysis; and synthetic or conjugation (phase II) reactions, which include glucuronidation, sulfation, and acetylation. Multiple forms of phase I and phase II enzymes exist to perform these metabolic functions. Although the liver is considered the major organ responsible for drug biotransformation, many other organs also contribute to drug metabolism. For many drugs in the newborn, the half-life is prolonged and total body clearance is decreased compared with older children and adults. Important variations occur, however, among drug classes and among individuals that prevent broad generalizations. Glucuronide conjugation of bilirubin by uridine 5′-diphospho-glucuronosyltransferase (UGT) 1A1 is usually low at birth unless this enzyme has been induced in utero through maternal exposure to drugs, cigarette smoke, or other inducing agents (Maurer et al, 1968). In contrast, conjugation through sulfation is usually active at birth. Various factors after birth, such as nutrition, illness, and drug interactions, may hasten or retard the maturation of enzymes and organs responsible for drug metabolism in the newborn. Maturational changes in hepatic blood flow, drug transport into hepatocytes, synthesis of serum proteins, protein binding of drugs, and biliary secretion—alone and in combination— confound accurate predictions about drug metabolism after birth, leading to empiric dose adjustments (Morselli et al, 1980). Because smaller and more immature newborns are now surviving, many of these factors that were studied in larger, more mature neonates must be reassessed in this less mature population. CHAPTER 34 Pharmacokinetics, Pharmacodynamics, and Pharmacogenetics The primary types of enzymes involved in phase I reactions are cytochromes P450 (CYPs). CYPs are microsomal, mixed-function oxidases that catalyze chemical changes within a molecule, such as hydroxylation, methylation, demethylation, addition of oxygen, or removal of hydrogen (Leeder and Kearns, 1997). There are thousands of different CYP450 enzymes found in plants and animals on land and in water. CYPs with more than 40% of the same amino acid sequence are grouped as a numbered family (e.g., CYP3). Those with greater than 67% polypeptide homology belong to the same subfamily (e.g., CYP3A), and specific genes are denoted with a number (e.g., CYP3A5). The amino acid sequence of these enzymes determines the tertiary structure that creates a hydrophobic pocket with selective binding for chemicals and drugs. This substrate specificity creates groups of drugs, often with similar structure and function that are metabolized by the same CYP. In the preterm newborn, CYPs develop at different rates and in different patterns. Maturation correlates best with postmenstrual age for some CYPs and with postnatal age for others. Some CYPs do not reach adult activity until several years of age, whereas others develop activities twice that of the adult during childhood; therefore generalizations about patterns of maturation for CYPs are seldom valid. In addition, single nucleotide polymorphisms (SNPs) or substitutions in the DNA sequence for a CYP may reduce its metabolizing activity or completely eliminate it if the polypeptide cannot be formed. The phase II conjugation enzymes, such as the UGTs, have several forms with different substrate specificity (e.g., UGT1A1 for bilirubin, UGT2B7 for morphine), although they are not as absolutely selective as the CYPs (Leeder and Kearns, 1997). Some phase II enzymes complement each other at birth for conjugation of specific drugs, such as acetaminophen that is conjugated primarily by sulfotransferases at birth and less so by UGT. Moreover, by adolescence, UGT has become the predominant conjugation pathway for acetaminophen. EXCRETION Drug elimination from the body can occur through several mechanisms, including renal excretion, biliary excretion, transcutaneous loss, gastrointestinal loss, and pulmonary exhalation. Renal excretion is one of the most important pathways for elimination of metabolized and unchanged drug. Neonatal renal function is diminished in absolute terms and when normalized to body weight or surface area. The neonatal glomerular filtration rate averages 30% of the adult rate per unit surface area. Glomerular function rises steadily after birth, whereas tubular function matures more slowly, causing a glomerular and tubular imbalance (Aperia et al, 1981). The postnatal increase in glomerular function reflects greater cardiac output, reduced renal vascular resistance, redistribution of intrarenal blood flow, and changes in intrinsic glomerular basement membrane permeability (Morselli et al, 1980). The dynamics of neonatal renal function markedly influence drug excretion. The rate of change of renal function and its susceptibility to hypoxemia, nephrotoxic drugs, and underperfusion prevent accurate predictions of drug elimination rates in newborns, which often must be measured empirically. 419 Some drugs, such as nafcillin and spironolactone, have metabolites that must be eliminated through biliary excretion. Drugs that are conjugated within the liver may also be excreted through bile, enter the intestinal tract, and undergo deconjugation and enterohepatic recirculation, similar to bilirubin. Although biliary excretion is not well studied in newborns, clinical conditions such as parenteral nutrition– associated cholestasis suggest that it may be highly variable among specific patients and conditions. Transporters play important roles in removing drugs and preventing drug absorption. Organic anion transporter polypeptides provide facilitated transport of anions in many tissues, including the kidney and liver. Permeability glycoprotein is an efflux transporter that belongs to the adenosine triphosphate–binding cassette–multiple drug resistance family of transporters. Permeability glycoprotein prevents the absorption of many compounds across the intestinal wall or into the brain, where it functions as a significant portion of the blood brain barrier. PHARMACOGENETICS AND PHARMACOGENOMICS The human genome project described the protein structures of many enzymes involved in drug metabolism and identified genetic variants in the proteins that alter activity, many of which are produced by a change in a single nucleotide to create SNPs. As discussed in the previous sections, genetic variation in drug metabolizing enzymes and in drug transporters can have a significant influence on the relative activity of these systems within a particular individual. Knowledge of pharmacogenetic and pharmacogenomic factors that can effect pharmacokinetics, particularly with a relationship to drug metabolism and elimination, has been important for understanding ways to avoid unanticipated drug effects in adults and to some extent older children. In neonates, however, this knowledge is only beginning to develop. Moreover, because many of these drug-metabolizing systems that are subject to pharmacogenetic variation continue to mature during the first few years of life, it can be difficult to determine whether genetic variability or physiologic maturation are contributing to differences in drug metabolism by neonates. What should be appreciated from this limited knowledge, however, is that drug metabolism is often reduced in neonates, and scaling of drug dosage by simple body weight or allometrically with an exponent will not fully compensate for differences in clearance that exist in this newborn population. Clearance often varies several-fold among adults, and the same degree of variation is emerging among neonates whether caused by maturation of the expression of these enzymes by gestational age or induction of protein synthesis after birth. One of the most prominent phase 1 metabolizing enzymes, CYP3A4, illustrates several important aspects of CYPs. CYP3A4 exists as a fetal form at birth (CYP3A7) that decreases over the first months of life as it is replaced by the adult form, CYP3A4, by 1 year of age (Lacroix et al, 1997). However, these changes are not linear, which confounds making accurate dosing adjustments by age. The maturation of metabolic activity of phase I enzymes to adult activity varies among the specific enzymes, requiring 420 PART VIII Care of the High-Risk Infant months for some and years for others (Kearns et al, 2003). During the fetal–neonatal period some CYPs increase in activity, but the most rapid increase in activity usually occurs after birth, regardless of the gestational age at birth (Koukouritaki et al, 2004). Drug-drug interactions influence activity of CYPs as well. Macrolide antibiotics, such as erythromycin and clarithromycin, and the azole antifungal drugs, such as fluconazole, inhibit the activity of CYP3A4 (Leeder and Kearns, 1997). These drugs can reduce clearance for drugs that are a substrate for CYP3A4, such as fentanyl or midazolam, and lead to toxicity without dosage adjustments. Conversely, inducers of CYP3A4 activity, such as rifampin and phenobarbital, can increase clearance of drugs that are substrates and reduce the concentration and effectiveness. Large interindividual variation occurs for most of these enzymes and is complicated by inherited differences in activity. An SNP in the DNA for one of these enzymes is designated with an asterisk, such as CYP2C9*1, and can alter the protein structure enough to reduce or even inactivate its enzymatic activity. SNPs have been identified for most CYPs, especially CYP2D6, one of the first recognized inherited variations in activity for debrisoquine (Leeder and Kearns, 1997). Ethnic variations in these SNPs help to predict when activity is likely to be reduced or increased. For example, several inactivating SNPs of CYP2D6 have been recognized, indicating that 50% or more of Asians lack enough active CYP2D6 to metabolize codeine to morphine for a more potent analgesic effect (Ingelman-Sundberg, 2005). Study of the maturational changes in the pharmacokinetics of drugs metabolized by specific CYPs expands current knowledge of how these enzymes mature. A better understanding of the rates and patterns of maturation of different CYPs will help to guide appropriate dosage adjustments to improve drug therapy for newborns. PHARMACOKINETIC PRINCIPLES Pharmacokinetics describes the time course of changes in drug concentrations within the body. Although rates of change are often described with differential equations, useful concepts at the bedside are emphasized in this section. More detailed mathematical discussions of pharmacokinetics can be found elsewhere (Buxton, 2006; Rowland and Tozer, 2010). COMPARTMENT In pharmacokinetics, compartment refers to fluid and tissue spaces into which drugs penetrate. These compartments may or may not be equivalent to anatomic or physiologic fluid volumes. In the simplest case, the compartment may correspond to the vascular space and equal the volume of a real body fluid (i.e., blood). Large or highly polar molecules can be confined to this central compartment until they are eliminated by excretion or metabolism. Many drugs, however, diffuse reversibly out of the central compartment into tissues or other fluid spaces, generically referred to as peripheral or tissue compartments. Such compartments are seldom sampled directly, but their involvement in kinetic processes may be recognized from the graphic or mathematical description of the kinetics of a drug. The number of exponential terms necessary to adequately describe the kinetic profile of a drug designates the number of compartments involved, recognizing that many more compartments may exist. For appropriate clinical application, rarely are more than three compartments required to describe a drug’s pharmacokinetics. APPARENT VOLUME OF DISTRIBUTION The apparent volume of distribution might be better termed volume of dilution, because it is a mathematical description of the volume (L or L/kg) that dilutes a dose (mg or mg/kg) to produce the observed circulating drug concentration (mg/L or μg/mL). (To simplify cancellation of units, concentrations are expressed as mg/L, which is the same as μg/mL, the more conventional unit for drug concentrations.) Concentration (mg/L) = Dose (mg/kg) Apparent volume of distribution (L/kg) For many drugs, the volume of distribution does not correspond to a specific physiologic body fluid or tissue— hence the term apparent. In fact, the volume of distribution for drugs that are bound extensively in tissues may exceed 1.0 L/kg, a physiologic impossibility that emphasizes the arithmetic, nonphysiologic nature of the apparent volume of distribution. The determination of distribution volume is described later. FIRST-ORDER KINETICS Removal of most drugs from the body can be described by first-order (exponential or proportional) kinetics, in which a constant proportion or percentage of a drug is removed per unit of time (e.g., 50% in one half-life interval), rather than a constant amount per unit of time. For drugs exhibiting first-order kinetics, the higher the concentration, the greater the amount removed during an interval of time. The following equations describe the concentration (C) of a drug whose first-order kinetics have a constant rate k (hour−1), at time t, and an initial concentration of C0 achieved after administration of a dose. In differential equation form, the change in C with time is dC = − kC dt The solution to this differential equation gives the exponential form, which describes C at time t: Ct = C0 e − kt If this equation is transformed using the natural logarithm (ln), it results in: ln Ct = ln C0 − kt The last equation fits the equation of a straight line, so that a graph that plots ln Ct versus t has an intercept of ln C0 at t = 0 and a slope of –k, the rate constant for the change in concentration; this can be used to calculate the half-life and to estimate appropriate dosages. Multiple rate constants in CHAPTER 34 Pharmacokinetics, Pharmacodynamics, and Pharmacogenetics 1000 10 DISTRIBUTION PHASE (ALPHA) (Distribution + Elimination) Log serum concentration Plasma Concentration 800 400 200 T1/2 421 ELIMINATION PHASE (BETA) (Elimination) 1 Slope = β/2.303 .1 Slope = α/2.303 100 0 30 60 90 120 150 180 Minutes after Dose FIGURE 34-1 Apparent single-compartment, first-order plasma drug disappearance curve illustrating graphic determination of half-life from best-fit line of serial plasma concentrations. more complex equations are distinguished with the letter k and numbered subscripts or with Greek letters. HALF-LIFE The drug half-life (t1/2) is the time required for a drug concentration to decrease by 50%. Half-life is a first-order kinetic process because the same proportion, 50%, of the drug is removed during equal periods. Half-life can be determined mathematically from the elimination rate constant k as: Natural logarithm 2 0.693 t1/2 = = k k Figure 34-1 illustrates a graphical method for determination of half-life. Drug concentrations measured serially are graphed on semilogarithmic axes, and the best-fit line is determined either visually or by linear regression analysis. In this illustration of first-order kinetics, the concentration decreases 50% (from 800 to 400) during the first hour and decreases another 50% (from 400 to 200) during the second hour. Thus the half-life is 1 hour. More drug is removed during one half-life at higher concentrations, although the proportion removed remains constant. The exponential equation for this graph is: C = 800e − 0.693t where k = 0.693/h and C0 = 800, allowing a mathematical calculation of half-life using the equation described previously: 0.693 0.693 t1/2 = = 1 hour = -1 0.693/h k (hour ) MULTICOMPARTMENT, FIRST-ORDER KINETICS The rate of removal of many drugs from the circulation is often biphasic. An initial rapid decrease in concentration is the distribution (α) phase, often lasting 15 to 45 minutes, which is followed by a sustained slower rate of removal, the elimination (β) phase. Such biphasic processes are best visualized from semilogarithmic graphs of concentration versus time. When such semilogarithmic graphs show .01 0 120 240 Minutes after dose FIGURE 34-2 Multicompartment serum drug disappearance curve. kinetics that best fit two straight lines, the kinetics are described as biexponential, or reflective of a drug that shows two-compartment first order pharmacokinetics (Figure 34-2). Two exponential terms are needed to describe the change in concentration over time, as: C = Ae − αt + Be − βt In this equation, the rate constant for distribution is designated α to discriminate it from the rate constant for terminal elimination (β), where A and B are the time = 0 intercepts for the lines describing distribution and elimination, respectively. Division by 2.303 converts logarithms to natural logarithms. After an intravenous dose, drug loss from the vascular space during the distribution phase occurs through both distribution and elimination (see Figure 34-2). The rate constant of distribution (α) can be determined by plotting the difference between the total amount of drug lost initially and the amount of drug lost through elimination (Greenblatt and Koch-Weser, 1975). This determination produces the line with the steeper slope (equal to α/2.303) below the serum concentration graph in Figure 34-2. The single slope of the distribution phase and of the terminal elimination phase does not imply that distribution or elimination occurs through a single process. The observed rates usually represent the summation of several simultaneous processes, each with differing rates, occurring in various tissues. When the time course of drug elimination is observed for prolonged periods, a third rate of elimination, or γ phase, may also be observed and is usually attributed to elimination of drug that has reequilibrated from deep tissue compartments back into the plasma. Such kinetics are designated three-compartment first-order pharmacokinetics. The kinetics of a drug are expressed with the smallest number of compartments that accurately describe its concentration changes over time. APPARENT SINGLE-COMPARTMENT, FIRST-ORDER KINETICS When a semilogarithmic graph of concentration versus time reveals a single slope with no distribution phase, the kinetics are characterized as apparent one-compartment, 422 PART VIII Care of the High-Risk Infant BOX 34-2 D  rugs That ­Demonstrate ­Saturation ­Kinetics With Therapeutic Doses in ­Newborns ll ll ll ll ll ll FIGURE 34-3 Representation of saturation, or zero-order (serum concentration–dependent), and first-order (serum concentration– independent) pharmacokinetics. first-order (see Figure 34-1). Such kinetics can occur when a drug remains entirely within the vascular space or central compartment or when a drug passes rapidly back and forth between the circulation and peripheral sites until it is metabolized or excreted by first-order kinetics. The adjective apparent is used because careful study often shows that distribution occurs even though the kinetic curve has only a single slope. Single-compartment kinetics implies that the drug rapidly and completely distributes homogeneously throughout the body, which rarely occurs clinically. In many pharmacokinetic studies in newborns, blood samples are not obtained early enough to allow calculation of the distribution phase, and the kinetics are described as single-compartment. If sampling begins after the distribution phase, the concentration time points may fit a single-compartment, first-order model, which determines the elimination rate constant (β). The kinetics cannot be assumed, however, to fit a single-compartment model from such a limited study. The more clinically limited but accurate approach to kinetic analysis, noncompartmental analysis, makes no assumptions about the number of compartments (Rowland and Tozer, 2010). ZERO-ORDER KINETICS Some drugs demonstrate zero-order kinetics, in which a constant amount of drug, rather than a constant proportion or percentage, is removed per unit of time. This relationship can be expressed as: dC = −k dt It is important to understand when zero-order kinetics occurs, how to recognize it, and how it affects drug concentrations. Zero-order kinetics is sometimes referred to as saturation kinetics, because it can occur when excess amounts of drug completely saturate enzymes or transport systems so that they metabolize or transport only a constant amount of drug over time. Zero-order processes produce a curvilinear shape in a semilogarithmic graph of concentrations versus time (Figure 34-3). When drug concentrations are high from a drug overdose or the pathway for elimination is impaired as in renal dysfunction, kinetics Caffeine Heparin Theophylline Furosemide Indomethacin Phenytoin may become zero-order initially and be followed by firstorder kinetics at lower concentrations. For drugs exhibiting zero-order kinetics, small increments in dose can cause disproportionately large increments in serum concentration. Certain drugs administered to newborns exhibit zero-order kinetics at therapeutic doses, and concentrations and must be recognized for their potential accumulation (Box 34-2). NONCOMPARTMENTAL ANALYSIS Noncompartmental analysis is based on describing drug exposure measured by the AUC without any assumptions about the pattern of elimination or number of compartments. Central to this analysis is the determination of drug clearance from the dose and the AUC: Cl = Dose / AUC If the dose is administered intravenously, then noncompartmental analysis allows the direct determination of drug clearance using this relationship. Estimation of the elimination half-life is generally done using the slope of the log transformed concentration measurements made during the end of a pharmacokinetic study. With an estimation of clearance and elimination rate, the apparent distribution volume can also be estimated from: Cl = V × K Thus noncompartmental pharmacokinetics provides a simple means to assess fundamental pharmacokinetic parameters, which may be useful for dosing patients when detailed knowledge of the complete pharmacokinetic profile is not needed. POPULATION PHARMACOKINETICS Most pharmacokinetic studies in newborns are limited by the amount of blood that can be removed safely for sampling. To determine the kinetics safely, population pharmacokinetic approaches are valuable, particularly because they can accommodate unbalanced study designs for drug sampling. The population approach describes the ­concentration-versus-time profile for all the patients enrolled in a given study simultaneously, estimating population parameters that describe the general pharmacokinetic profile of the complete study group and patient specific parameters that define the individual patients in the study. The population approach can use fewer samples taken from each patient, if the samples are taken at different specified times over the complete time profile of 423 CHAPTER 34 Pharmacokinetics, Pharmacodynamics, and Pharmacogenetics interest for the study analysis. For example, one group of 28- to 32-week premature infants might have samples drawn at 1, 4, and 12 hours, whereas another group of 28to 32-week premature infants might be sampled at 0.5, 2, and 8 hours. The concentrations from these two groups of similar patients are then analyzed in aggregate to provide information during both distribution and elimination phases, thus describing the kinetics with a limited volume of blood sampled from each patient. Furthermore, the population approach allows for the investigation of patient covariates of interest that might explain differences seen within the population of patients enrolled in the trial. Typical covariates such as gestational age, gender, and disease conditions can also be assessed for their contribution to differences seen between subjects in a clinical study. These covariates can be helpful for gaining a better understanding of factors that may alter the pharmacokinetics of infants that might otherwise be considered similar. TARGET DRUG CONCENTRATION STRATEGY Drug treatment of newborns commonly uses the target drug concentration strategy (Box 34-3), in which drug therapy corrects a specific problem by producing an effective concentration of free drug at a specific site of action (Rowland and Tozer, 2010). The target site of drug action is usually inaccessible for monitoring concentrations. A specific concentration or range of circulating concentrations is correlated with the effective concentration at the site of action, which provides a “therapeutic” concentration range. The requirements for effective and accurate application of the target drug concentration treatment strategy in adults have been discussed by Spector et al (1988). When applied to newborns, these requirements highlight the special problems of drug therapy in these patients and the special circumstances in which clinical drug concentration monitoring is appropriate. Some of these requirements are as follows: ll An available analytic procedure for accurate measurement of drug concentrations in small volumes of blood ll A wide variation in pharmacokinetics among individuals with the knowledge that population-based kinetics do not accurately predict individual kinetics ll Drug effects are proportional to plasma drug concentrations ll A narrow concentration range between efficacy and toxicity (narrow therapeutic index) ll Constant pharmacologic effect over time, in which tolerance does not develop ll Clinical studies that have determined the therapeutic and toxic drug concentration ranges THERAPEUTIC DRUG MONITORING Box 34-3 illustrates the basic assumptions of therapeutic drug monitoring—that total plasma drug concentrations correlate with dose, circulating unbound drug concentrations, unbound drug concentration at the site of action. Clinical measurements of drug concentrations usually include both bound and unbound drug, and the active portion is the portion that is unbound (see Distribution, earlier). The two broad indications for monitoring drug BOX 34-3 T  arget Drug Concentration ­Strategy Drug dose ↔ Plasma total drug concentration ↔ Plasma unbound drug concentration ↔ Target site unbound drug concentration ↔ Desired pharmacologic effect Data from Shenier LB, Tozer TN: Clinical pharmacokinetics: the use of plasma concentrations of drugs. In Melmon KL, Morrelli HF, editors: Clinical pharmacology: basic principles in therapeutics, ed 2, New York, 1978, Macmillan, p 71. concentrations are attainment of effective concentrations and avoidance of toxic concentrations. As Kauffman (1981) has indicated, drug concentration ranges are not absolute reflections of effective therapy. Patient response, not a specific drug concentration range, is the endpoint of therapy. Although concentrations of aminoglycoside antibiotics, such as gentamicin, are monitored frequently in newborns, toxicity is rare (McCracken, 1986). Because of the limited evidence of toxicity in newborns, it is more important to measure aminoglycoside concentrations to achieve effective concentrations for treatment of culture-proven infections than to avoid toxicity. In newborns with serious therapeutic problems, measurement of serum drug concentrations should be used to achieve effective concentrations and to avoid toxicity. When the desired concentration range and kinetic parameters are known, doses may be estimated to reach that concentration with single bolus doses or bolus doses followed by continuous infusions. PHARMACOKINETIC-BASED DOSING The following equations can be used both to guide dosing and to derive kinetic parameters for individual patients. Dose = ΔC × Vd = (Cdesired − Cinitial ) × Vd mg/kg = (mg/L) × (L/kg) = (mg/L) × (L/kg) where C = concentration, and Vd = volume of distribution. This equation may be used to estimate dosage changes needed to increase or decrease concentration. For the first dose, the starting concentration is zero; afterward, the calculation of distribution volume should use the change (Δ) in concentration from the preceding trough to the peak associated with that dose. To reach a desired concentration rapidly, a loading dose can be administered followed by a sustaining infusion. The equation for calculation of infusion doses to maintain a constant ­concentration is shown: Infusion rate = k × Vd × C (mg/kg) (min−1 ) = min−1 × (L/kg) × (mg/L) where C = concentration, Vd = volume of distribution, and k = rate constant of elimination. Steady state is reached when tissue concentrations are in equilibrium and the amount of drug removed equals the amount of drug infused. The time needed to reach a steady state depends on the elimination half-life. Whereas the time is not shortened by the administration of a loading dose, a loading dose can allow the patient to achieve 424 PART VIII Care of the High-Risk Infant and trough concentrations after the first dose); therefore the shorter the dosing interval–to–half-life ratio, the higher the drug accumulation. As noted during infusions, the length of time required to reach steady-state concentrations depends primarily on the elimination half-life, not the dosing interval. CLEARANCE FIGURE 34-4 Representation of multiple dosing with accumulation of serum drug levels to steady-state concentration. therapeutic concentrations rapidly that can be maintained with an infusion during steady state. It is important to consider that if drug clearance decreases, the steady state concentration during an infusion will increase proportionally. In addition, the half-life increases and the rate constant decreases. Because concentrations are not measured for most of the drugs administered by continuous infusion in the neonatal intensive care unit (NICU), it is important to adjust dosages for factors that reduce clearance, such as kidney or liver dysfunction or reduced kidney or liver blood flow, to avoid high and toxic concentrations. REPETITIVE DOSING AND THE “PLATEAU PRINCIPLE” During the typical course of drug therapy, drug doses are administered before complete elimination of previous doses, and the drug accumulates in the body (Rowland and Tozer, 2010). During repeated administration, the peak and trough levels after each dose increase for a time. Steady-state, or plateau, concentrations are reached when the amount of drug eliminated equals the amount of drug administered during each dosing interval. During repetitive dosing, the steady-state concentrations achieved are related to the half-life, dose, and dosing interval relative to the half-life (Buxton, 2006; Rowland and Tozer, 2010). Figure 34-4 illustrates a hypothetical concentration-time curve for a drug with a half-life of 4 hours administered orally every 4 hours, so that the dosing interval corresponds to one half-life. Several important principles of pharmacokinetics are illustrated in this figure; the mathematics are described in detail elsewhere (Buxton, 2006). Drug concentrations rise and fall with drug administration (absorption) and elimination. For dosing intervals of one half-life, accumulation is 88% complete after the third dose, 94% complete after the fourth dose, and 97% complete after the fifth dose. At steady state, the peak and trough concentrations between doses are the same after each dose. If a drug is administered with a dosing interval equal to one half-life, the steady-state peak and trough concentrations are twofold those reached after the first dose. If the dosing interval is shortened to half of a half-life, the concentration decreases less before the next dose, more total drug is administered per day, and the steady-state peak and trough concentrations are considerably higher (3.4-fold the peak Clearance of drugs, as for creatinine, describes the volume of blood from which all the drug is removed per unit of time. Clearance is proportional to organ blood flow and the intrinsic capacity of organs to metabolize or remove drug from the circulation. In its simplest form, clearance is proportional to the flow to a single organ (Q) and to the arterial-venous difference in drug concentrations compared with the amount of drug in the arterial circulation, expressed as C − Cvenous CL = Q × arterial Carterial Total body clearance usually reflects the combined clearance of multiple organs with different enzyme activities and different rates of blood flow. Clearance can be measured by the rate of appearance of drug outside the body (similar to urinary creatinine clearance) or by the rate of disappearance of drug from the circulation compared with the circulating concentration. For calculations, clearance (CL) is defined as the dose divided by the area under the ­plasmaconcentration versus AUC and by the rate of drug input per Css average, where rate of input is the dosing interval (τ) and Css is the steady-state concentration. For a drug administered by continuous infusion, this is simply the infusion rate (mg/kg)hr−1 divided by Css as follows: Dose Dose/τ Infusion rate CL = = = AUC Css Css Once clearance is known, this equation can be rearranged to solve for the dose necessary to achieve any desired steady state concentration: Infusion rate = Css × CL −1 −1 (mg/kg) h = (mg/L) × (mL/kg) h Clearance changes significantly for some drugs during fetal and infant development because the activity of metabolic enzymes increases with advancing gestational and postnatal age. Values for clearance and volume of distribution at different stages of preterm development are available for a few drugs and can be used to estimate the doses needed to achieve and maintain therapeutic concentrations associated with desired clinical responses. Studies of the analgesic fentanyl illustrate the developmental changes in its kinetics and how they can be used to calculate dosages to reach and maintain concentrations associated with effective analgesia. Analgesia has been associated with a serum fentanyl concentration of 1 to 2 ng/mL (Santeiro et al, 1997). If analgesic treatment is initiated with a continuous infusion of fentanyl, five half-lives are needed to reach a steady state. The fentanyl half-life ranges from 3 hours in term newborns to 12.7 hours in premature newborns (Koehntop et al, 1986; Santeiro et al, 1997). Because of this prolonged half-life, the patient may be inadequately CHAPTER 34 Pharmacokinetics, Pharmacodynamics, and Pharmacogenetics TABLE 34-1 Fentanyl Development Kinetics Gestational Age (wk) Clearance at 0-47 Hours after Birth (mL/[min·kg]) 29 9.6 33 11.4 37 13.2 41 15.0 Data from Saarenmaa E, Neuvonen PJ, Fellman V: Gestational age and birth weight defects of plasma clearance of fentanyl in newborn infants, J Pediatr 136:767-770, 2000. treated for a long time, unless a loading dose is administered to reach an effective concentration more rapidly. In general, the initiation of analgesic treatment and increases in infusion doses of analgesics should begin with a loading dose based on the estimated volume of distribution in the central compartment (circulation) and desired concentration. The use of a loading dose shortens the time to reach higher effective analgesic concentrations, but also increases the likelihood of toxicity, as has been reported with digoxin. Limited data are available regarding the gestational age– related changes in fentanyl clearance, but two studies show that it increases with advancing gestational age (Koehntop et al, 1986; Santeiro et al, 1997) and with increasing age after birth (Gauntlett et al, 1988; Santeiro et al, 1997). The linear graph of clearance versus gestational age from 38 neonates who began treatment within 47 hours after birth was used to derive mean rates of clearance at different gestational ages, as shown in Table 34-1. Other investigators studied single-dose fentanyl kinetics during anesthesia and found an apparent central volume of distribution of fentanyl in neonates of 1.45 L/kg (Koehntop et al, 1986). Note that this distribution volume is smaller than the steady-state volume of distribution of 5.1 L/kg, also calculated after a single dose of fentanyl (Koehntop et al, 1986). In turn, the apparent steady-state volume of distribution after a single bolus dose of a lipophilic drug is usually smaller than that associated with continuous drug infusions, during which tissues throughout the body become saturated with drug. The steady-state distribution volume for fentanyl during continuous infusions was calculated as 17 L/kg (Santeiro et al, 1997). It should be noted that because fentanyl is highly lipid soluble, it distributes ­rapidly from the central compartment into the peripheral tissue compartment. This large distribution volume likely reflects the period during the infusion when the drug is leaving the circulation to penetrate peripheral tissues, such as fat. Because it may take 15 to 60 hours to achieve a steady-state concentration (five half-lives) after a fentanyl infusion is begun or the infusion rate is increased, a patient may need repeated bolus doses to maintain effective plasma concentrations in the central compartment. The best approach is to repeat the calculated loading dose until the desired clinical effect is achieved. This also illustrates why, for sedation specifically, dosing should be adjusted to achieve the desired clinical effect. Clearance calculations, however, can guide the starting doses to achieve effective sedation, as illustrated later. The kinetic parameters for fentanyl in premature infants reported by Koehntop et al (1986) can be used to calculate a loading and infusion dose to reach a fentanyl concentration 425 of 2 ng/mL, which is considered an analgesic concentration (Saarenmaa et al, 2000). Fentanyl analgesic concentration is estimated for a premature newborn at a gestational age of 33 weeks (note that ng/mL is equivalent to μg/L) as follows: C (μg/L) = Loading dose (μg/kg) Vdcentral (L/kg) Loading dose = 2 µg/L × 1.45 L/kg = 2.9 µg/kg [1] [2] Infusion rate (µg/kg × h) = C (µg/L) × CL (mL/kg × h) = 2 µg/L × 11.4 mL/kg/min × 1 L/1000 mL × 60 min/h = 1.4 µg/kg × h [3] Two studies have observed rises in fentanyl clearance with increasing postnatal age (Gauntlett et al, 1988; Santeiro et al, 1997). This postnatal rise in clearance of fentanyl likely relates either to maturation of cytochrome P450 3A4 (the enzyme responsible for fentanyl metabolism) activity or to increased hepatic blood flow after birth, because fentanyl has a high hepatic extraction rate. For drugs like fentanyl with a high hepatic extraction ratio, the rate-limiting factor in clearance is the flow of blood to the liver (Saarenmaa et al, 2000). Some researchers have observed that increased intraabdominal pressure reduces fentanyl clearance, which is likely caused by reduced hepatic blood flow (Gauntlett et al, 1988; Koehntop et al, 1986). Clinical changes known to increase or decrease fentanyl clearance should be used to adjust starting dosages, but dosing should be adjusted primarily for the desired clinical effect. MODELING AND SIMULATIONS Pharmacokinetic modeling and simulations can be used to evaluate the effects of important developmental changes and disease processes upon pharmacokinetic parameters for distribution volume, elimination rate, and clearance. These models can identify clinical situations and conditions when doses are likely to require modification. Mathematical simulations create theoretical pharmacokinetic profiles for patients after a dose using the range of pharmacokinetic parameters determined from a patient population. These can then be calculated for 100 to 1000 hypothetical patients to define the expected range of concentrations that are likely after a dose. For drugs, such as antiinfectives with which serum concentrations have been correlated with effectiveness, this provides estimates of how large a dose is needed to reach effective concentrations. A recent study of fluconazole kinetics in newborns illustrates the application of this process (Wade et al, 2008). PHARMACOKINETIC PRACTICAL EXAMPLES ESTIMATED DOSE ADJUSTMENTS Gentamicin and phenobarbital can be used to illustrate the practical application of the principles of pharmacokinetics and therapeutic drug monitoring already discussed. 426 PART VIII Care of the High-Risk Infant The calculations can be performed with standard arithmetic calculators and provide close enough estimates of the kinetics for drugs with a long half-life to adjust dosages at the bedside. GENTAMICIN Assume that optimal gentamicin concentrations are Peak = 6 to 10 µg/mL Trough = 0.5 to 2 µg/mL After the fourth dose (2.5 mg/kg) of gentamicin to an edematous premature newborn, the peak concentration was 5.0 μg/mL; 18 hours later, the trough was 2.5 μg/mL. It appears that the distribution volume is greater than anticipated, because the peak concentration is lower than expected, and the half-life is longer than anticipated because the trough is higher than expected. The time of drug administration and blood sampling were confirmed (an important step), so the half-life is 18 hours, because the concentration decreases 50% from 5.0 to 2.5 μg/mL in 18 hours, assuming that the kinetics are linear and first-order. Vd (mL/kg) = = Dose (mg/kg) Δ C (μg/mL) × 1 mg/1000 μg 2.5 mg/kg × 1000 μg/mg (5.0 − 2.5) (μg/mL) = 1000 mL/kg To ensure a trough concentration of 2.0 μg/mL or less, doses are administered every two half-lives or every 36 hours. When two half-lives have passed after the fourth dose, the gentamicin concentration should be approximately 1.25 μg/mL (50% of 2.5 μg/mL). Increasing the concentration from the 1.25 μg/mL (trough) to greater than 6 μg/mL (peak) requires a concentration difference of 4.75 μg/mL or greater. With a distribution volume of 1000 mL/kg, a dose of 4.75 mg/kg should raise the concentration from a trough of 1.25 μg/mL to a peak of 6.00 μg/mL. In one half-life, this concentration will decrease to 3.0 μg/mL, and in two half-lives or 36 hours to 1.5 μg/mL. An additional 4.75-mg/kg dose will raise the peak concentration to 6.25 μg/mL, which will fall to 3.12 μg/mL in one half-life and to 1.6 μg/mL in two half-lives. The variation between the peak and trough concentrations after the last dose is within the measurement error for gentamicin and should achieve the optimum concentrations defined previously. the phenobarbital concentration was checked daily. The results were as follows: ll 24 hours: 40 μg/mL ll 48 hours: 31 μg/mL ll 72 hours: 25 μg/mL ll 96 hours: 21 μg/mL The maintenance dose (7 mg/kg) was resumed immediately after the 21 μg/mL concentration was measured and produced a peak concentration of 30 μg/mL after administration of the dose. These concentrations and doses can be used to calculate the volume of distribution and a dose to maintain the phenobarbital concentration between 20 and 30 μg/mL as follows: Dose (mg/kg) Vd (L/kg) = Δ C (μg/mL = mg/L) = = 7.0 (mg/kg) (30 − 21) (mg/L) 7 mg/kg 9 mg/L = 0.78 L/kg Half-life can be determined from inspection, because the concentration decreased from 50 to 25 μg/mL in 72 hours; therefore it should take 72 hours for the concentration to decrease by one half-life from 30 to 15 μg/mL. The concentration will decrease approximately 5 μg/mL every 24 hours, or one third of a half-life. Dividing the halflife into fractions is an approximation because it estimates the change in concentration as linear rather than exponential. To be more accurate, the concentration decreases 59% in half of one half-life. Although this approximation violates certain principles of pharmacokinetics, it allows estimation of the change in concentration for each one third of a half-life as one third of the change during one half-life. Therefore the concentration decreases approximately 5 μg/mL in 24 hours. The following approach can be used to estimate the daily phenobarbital dose needed to return the concentration to 30 μg/mL, a change in concentration of 5 μg/mL: Dose (mg/kg) Δ C (mg/L) = Vd (L/kg) 5 mg/L = Dose (mg/kg) 0.78 L/kg 3.6 mg / kg = Dose (mg/kg) PHENOBARBITAL DRUG-INDUCED ILLNESS Seizures that were difficult to control developed in a 3.6-kg asphyxiated newborn. Seizures continued after two 20-mg/kg phenobarbital doses until an additional 10-mg/kg dose was administered. A maintenance dose of 7 mg/kg per day was started 24 hours after the loading doses were administered. At 10 days, the child was increasingly somnolent. The phenobarbital level measured in a blood specimen taken 2 hours after administration of the oral maintenance dose was 50 μg/mL. Additional doses were withheld, and The extensive exposure of newborns to drugs in the NICU is not benign. Neonates in the NICU have higher rates of medication errors and potential adverse error rates (91 in 100 admissions) than do neonates in other areas of the hospital, and physician orders are responsible for 74% of errors and 79% of potential errors (Kaushal et al, 2001). Physician reviewers judged more than 90% of these errors to be preventable, and 16% were potentially lifethreatening or fatal (e.g., heparin and digoxin overdoses). CHAPTER 34 Pharmacokinetics, Pharmacodynamics, and Pharmacogenetics Similar problems have been observed worldwide (Kunac et al, 2009). The causes of this drug-related morbidity and mortality are complex. Pharmacologic studies in pediatric patients are difficult because of a variety of problems ranging from ethics to study design (Ward and Green, 1988). The difficulty of studying therapeutics in the newborn has created a situation in which a plethora of drugs is administered with a paucity of pharmacologic data. For the smaller and more immature newborns who are now surviving, gestational age–appropriate pharmacologic data for efficacy, dose-response, and kinetics for most drugs remain limited and need further study (Ward and Kern, 2009). Furthermore, drug-induced illness is seldom considered in newborns. Failure to recognize drug-induced illness in the newborn often leads to further pharmacologic treatment as the first approach to correct unrecognized druginduced problems. This fact may reflect an expectation that drug therapy is usually effective and safe. Prudent management of newborns must recognize and weigh the potential benefits of unstudied drug therapy against potential druginduced adverse effects, morbidity, and mortality. Some examples from the history of drug-induced mortality and morbidity in newborns should serve as a reminder of how more harm than good may accrue from uncontrolled or unstudied drug therapy in the NICU. LESSONS FROM CHLORAMPHENICOL Chloramphenicol was released for use in the 1940s, and reports of its efficacy for treatment of Salmonella spp. infections included pediatric patients. The manufacturer recommended doses of 50 to 100 mg/kg per day for patients weighing 15 kg or less. When Sutherland (1959) reported three cases of sudden death in newborns treated with high doses of chloramphenicol (up to 230 mg/kg per day), the drug was considered “well tolerated and nontoxic.” Later the same year, Burns et al (1959) reported the disturbing results of a controlled trial of the following four prophylactic treatment regimens for newborn sepsis: (1) no treatment, (2) chloramphenicol alone, (3) penicillin and streptomycin, and (4) penicillin, streptomycin, and chloramphenicol. The groups that received chloramphenicol (100 to 165 mg/kg per day), in regimens 2 and 4, had overall mortality rates of 60% and 68%, respectively, whereas groups receiving regimens 1 and 3 had mortality rates of 19% and 18%, respectively. The deaths of these newborns demonstrated the stereotyped sequence of symptoms and signs caused by chloramphenicol, designated the gray syndrome, which consisted of abdominal distention with or without emesis, poor peripheral perfusion and cyanosis, vasomotor collapse, irregular respirations, and death within hours of the onset of these symptoms. Weiss et al (1960) attributed the gray syndrome in newborns to high concentrations of chloramphenicol secondary to its prolonged half-life in newborns who received dosages of more than 100 mg/kg per day, which are usually used in older children. They recommended maximum doses of 50 mg/kg per day in term infants younger than 1 month, half that dose for premature infants, and careful monitoring of chloramphenicol blood concentrations. The discovery and explanation of chloramphenicol toxicity in newborns illustrate several important aspects 427 of neonatal pharmacology. Because chloramphenicol was considered well tolerated in older children and adults, it was regarded as nontoxic for newborns. Chloramphenicol was so effective in newborns that higher doses were used without pharmacokinetic study. Higher doses were administered to newborns despite recognition that its clearance required glucuronide conjugation, which was known to be immature in newborns. The unexpected finding that chloramphenicol in doses of 100 to 165 mg/kg per day could be lethal to newborns was demonstrated because the study conducted by Burns et al (1959) included appropriate control groups. Because the mortality rate from the most effective antibiotic treatment regimen was equivalent to that of no antibiotic treatment, these investigators discontinued prophylactic use of antibiotics in the nursery. Similar pharmacologic comparisons are needed for other drugs used in newborns. Additional thoughtful consideration should be given to clinicians‘ response to therapeutic failure. Fewer drugs and lower doses may be safer and more effective than additional drugs in higher dosages. REDUCTION AND PREVENTION OF MEDICATION ERRORS IN NEWBORN CARE Drug treatment is one of the most common approaches used in the care of sick newborns. At Primary Children’s Medical Center 35-bed NICU in Salt Lake City, Utah, with an average of 785 patient-days per month, patients receive an average of 8700 (range 6990 to 11,290) doses of medications, pharmacy-formulated intravenous solutions, and aerosols each month (unpublished observations, 1994). These doses are usually prepared by pharmacists and administered by nurses, respiratory therapists, and (rarely) physicians. In such a large and complex system that produces so many drug treatments per month, errors are virtually inevitable despite several levels of prospective and redundant reviews by nurses, pharmacists, and NICU unit secretaries involved in the drug treatment process. At Primary Children’s NICU, the medication error rate averages 0.04% for nurses and pharmacists and 0.07% for physicians (unpublished observations, 1994). Many errors are inconsequential, whereas others have serious adverse effects. Medication errors incur significant costs, ranging from the obvious ones such as direct patient injury, prolonged hospital stays, and additional corrective treatments to the more subtle costs associated with monitoring and regulation of medication use within hospitals (ASHP, 1995). In a study of 393 malpractice claims reported to the Physician Insurers Association of America, the secondmost common cause of malpractice claims was drug errors (Physician Insurers Association of America, 1993). Among 16 medical specialties with two or more claims, pediatric practice ranked sixth in the number of claims, yet it had the third highest average cost per indemnity. The medications most frequently involved in all claims were antibiotics, glucocorticoids, narcotic or non-narcotic analgesics, and narcotic antagonists. In pediatric practice, the medications most frequently involved were vaccines (diphtheria-pertussis-tetanus) and bronchodilators (theophylline). In the Physician Insurers Association of America review 428 PART VIII Care of the High-Risk Infant (Physician Insurers Association of America, 1993), the five most common causes of drug errors were as follows: ll Incorrect doses ll Medications that were inappropriate for the medical condition ll Failure to monitor for drug side effects ll Failure of communication between physician and patient ll Failure to monitor drug levels The primary opportunity for prevention of these five most common errors rests with the prescribing physician. Additional information and time for communication and documentation may be needed. Prescriptions and drug orders are a means of communicating, but clinicians often devote too little attention to making them legible, clear, and unambiguous (ASHP guidelines on preventing medication errors in hospitals, 1993). Physicians should keep the following recommendations in mind to ensure that their medication orders communicate more effectively: ll Write instructions in full rather than with abbreviations. ll Avoid vague instructions (e.g., “take as directed”). ll Specify exact dose strengths. ll Avoid abbreviations of drug names (e.g., MS could mean morphine sulfate or magnesium sulfate). ll Avoid U as an abbreviation for units, because the U may be mistaken for a 0 (zero). ll Avoid trailing zeroes (e.g., 5.0 mg). ll Use leading zeroes (e.g., 0.5 mg). ll Minimize the use of verbal orders. ll Ensure that prescriptions and signatures are legible, even if it means printing the name that corresponds to the signature. The process for ordering, preparing, dispensing, and administering medications in an ICU with acutely ill patients is often complicated and may contribute directly to errors. The frequency of those errors, however, may be reduced in almost every NICU. Although complex and expensive computerized systems may help reduce medication errors, caregivers can take steps that are completely within their control to reduce medication errors without waiting for changes in the entire pharmacy process within the hospital (ASHP guidelines on preventing medication errors in hospitals, 1993). DRUG EXCRETION IN BREAST MILK The excretion of drugs in breast milk remains a source of confusion and concern for many physicians and families. Newer analytic techniques and more thorough pharmacokinetic studies have improved the available data in this area of neonatal pharmacology. The available data regarding drug exposure of the newborn through human milk have been organized, in decreasing levels of concern, from drugs that are associated with adverse effects on the infant during nursing to those that are of concern pharmacologically to those that have not been associated with problems during nursing. The list of drugs clearly contraindicated during nursing is surprisingly short (American Academy of Pediatrics, Committee on Drugs, 2001). On average, the breastfeeding infant receives approximately 2% to 3% of a maternal dose through milk. Drugs that are organic bases or are lipid soluble may reach higher concentrations in milk than in maternal serum. SUMMARY The extensive drug exposure of the sick newborn in the NICU is dangerous because of the frequency of adverse, sometimes fatal, drug reactions. Unfortunately, in the rapidly changing fetus and newborn, drug therapy is often empiric because of a lack of gestational age–appropriate kinetic data. Methods appropriate for the study of therapeutics in newborns present unique difficulties, but a review by Ward and Green (1988) may provide assistance for investigators. Drug therapy of newborns requires practical application of the principles of pharmacokinetics and pharmacodynamics—which describe the processes of drug absorption, distribution, metabolism, and excretion—to the estimation and individualization of dosages. SUGGESTED READINGS American Academy of Pediatrics: Committee on Drugs: The transfer of drugs and other chemicals into human breast milk, Pediatrics 108:776-789, 2001. ASHP guidelines on preventing medication errors in hospitals: Am J Hosp Pharm 50:305-314, 1993. Buxton ILO: Pharmacokinetics and pharmacodynamics: the dynamics of drug absorption, distribution, action, and elimination, In Brunton LL, Lazo JS, Parker KL, editors: Goodman and Gilman’s the pharmacological basis of therapeutics, ed 11, New York, 2006, McGraw-Hill, pp 1-39. Ingelman-Sundberg M: Genetic polymorphisms of cytochrome P450 2D6 (CYP2D6): clinical consequences, evolutionary aspects and functional diversity, Pharmacogenomics J 5:6-13, 2005. Kaushal R, Bates DW, Landrigan C, et al: Medication errors and adverse drug events in pediatric inpatients, J Am Med Assoc 285:2114-2120, 2001. Kearns GL, Abdel-Rahman SM, Alander SW, et al: Developmental ­pharmacology— drug disposition, action, and therapy in infants and children, N Engl J Med 349:1157-1167, 2003. Kunac DL, Kennedy J, Austin N, et al: Incidence, preventability, and impact of Adverse Drug Events (ADEs) and potential ADEs in hospitalized children in New Zealand: a prospective observational cohort study, Paediatr Drugs 11:153160, 2009. Lacroix D, Sonnier M, Moncion A, et al: Expression of CYP3A in the human liver—evidence that the shift between CYP3A7 and CYP3A4 occurs immediately after birth, Eur J Biochem 247:625-634, 1997. Leeder JS, Kearns GL: Pharmacogenetics in pediatrics. Implications for practice, Pediatr Clin North Am 44:55-77, 1997. McCracken GH: Aminoglycoside toxicity in infants and children, Am J Med 80(Suppl 6B):172-178, 1986. Rowland M, Tozer TN: Clinical pharmacokinetics and pharmacodynamics: concepts and applications, Baltimore, 2010, Williams and Wilkins. Spector R, Park GD, Johnson GF, et al: Therapeutic drug monitoring, Clin Pharmacol Ther 43:345-353, 1988. Wade KC, Wu D, Kaufman DA, et al: Population pharmacokinetics of fluconazole in young infants, Antimicrob Ag Chemo 52:4043-4049, 2008. Ward RM, Lugo RA: Drug therapy in the newborn, In MacDonald MG, Seshia MMK, Mullett MD, editors: Avery’s neonatology: pathophysiology and management of the newborn, ed 6, Philadelphia, 2005, Lippincott Williams and Wilkins, pp 1507-1556. Complete references used in this text can be found online at www.expertconsult.com C H A P T E R 35 Neonatal Pain and Stress: Assessment and Management Dennis E. Mayock and Christine A. Gleason Relief of human suffering is one of the most important goals of all health care providers. Advances in neonatology have significantly improved neonatal morbidity and mortality, but pain, discomfort, and stress remain sad realities for babies in the neonatal intensive care unit (NICU). Assessing, managing, and trying to limit these clinical realities, particularly while caring for critically ill neonates, are challenging and increasingly controversial. Fortunately there has been considerable clinical and laboratory research and much clinical dialogue aimed at developing the best clinical practices in this problematic arena. This chapter describes the developmental biology, history, and public policies that have informed and shaped current clinical practices; it also summarizes relevant clinical and basic research regarding clinical assessment tools and both pharmacologic and nonpharmacologic management approaches. Finally, future directions in this field are discussed. HISTORY AND DEVELOPMENT OF PUBLIC POLICIES ll ll ll ll ll HISTORICAL TIMELINE: NEONATAL PAIN MANAGEMENT Like all challenging medical issues, it is important to know the history and current practice of neonatal pain management to understand what direction research should take in the future. Management of neonatal pain and stress serves as an excellent example of this philosophy; therefore a brief history of neonatal pain management follows, beginning with the isolation of morphine from the opium poppy, which has been used to treat pain since approximately 3500 bc. ll 1806: Morphine (named for Morpheus, god of dreams) isolated from opium poppy by Friedrich Serturner (a pharmacist) and used to treat pain ll 1960s to 1980s: Infants believed to be too immature to feel pain; adverse effects of anesthetics feared; Liverpool method (pancuronium only, no anesthesia) used widely for patent ductus arteriosus (PDA) ligation in premature infants ll 1985: Landmark paper published by Anand et al (1985) describing adverse physiologic effects of Liverpool method and improved outcomes using anesthesia for PDA ligation; anesthesia and postoperative pain medication begin to become more widely used in neonatal care ll 1987: Joint statement issued by the American Academy of Pediatrics (AAP) and American Society of Anesthesiologists (ASA) regarding safety of (and necessity for) operative anesthesia and postoperative analgesia for neonates, regardless of their age or maturity ll 1990s: Use of morphine infusions in the NICU expands from operative and postoperative pain relief to include ll ll ll ll ll preemptive sedation, particularly during mechanical ventilation 1999: Pilot study (NOPAIN) suggesting that neurologic outcome is improved if preemptive morphine infusions are used in a neonate receiving mechanical ventilation 2000: AAP and Canadian Pediatric Society issue joint statement regarding neonatal pain management; primarily directed at surgical anesthesia and postoperative and procedural pain assessment and relief (American Academy of Pediatrics Committees on Fetus and Newborn and on Drugs, Sections on Anesthesiology and on Surgery et al, 2000) 2001: International evidence-based group develops consensus guidelines for prevention and management of neonatal pain, most of which are pharmacologically based 2003: The Joint Commission issues mandate regarding pain assessment and management in all hospitalized patients, including neonates 2004: Cochrane Review (Stevens et al, 2004) supports use of oral sucrose for procedural analgesia 2004: NEOPAIN trial results published (Neurologic Outcomes and Pre-emptive Analgesia in Neonates), concluding that preemptive morphine infusion did not improve outcomes in neonates receiving ventilation and did not relieve procedural pain 2005: Cochrane Review (Bellù et al, 2005) states “There is insufficient evidence to recommend routine use of opioids in mechanically-ventilated newborns.” 2005: Lee et al (2005) conclude that “fetal perception of pain is unlikely before the 3rd trimester” and “little or no evidence addresses the effectiveness of direct fetal anesthetic or analgesic techniques.” 2006: Cochrane Review (Shah et al, 2006) recommends breastfeeding or oral breast milk for neonatal procedural pain 2009: More than 40 infant pain assessment tools available around the world; most were developed for research and have not been validated for clinical use 2009: First national practice guidelines developed in Italy for procedural pain management for NICU patients DEVELOPMENT OF PUBLIC POLICY Since the late 1980s, in response to a public outcry regarding the recognition and management of pain in hospitalized patients, mandates have been promulgated by the U.S. Department of Health and Human Services, The Joint Commission, and other professional organizations. The initial public policy statement regarding pain management for neonates undergoing surgical interventions was issued as a joint communication by the AAP and the ASA 429 430 PART VIII Care of the High-Risk Infant (American Academy of Pediatrics Committees on Fetus and Newborn and on Drugs, Sections on Anesthesiology and on Surgery, 1987; American Society of Anesthesiologists, 1987). This statement made it clear that anesthesia and analgesia could be given relatively safely to neonates despite their age or cortical immaturity. A second influential document was issued by the Acute Pain Management Guideline Panel of the U.S. Agency for Health Care Policy and Research (Acute Pain Management Guideline Panel, 1992; Agency for Health Care Policy and Research Pain Management Guideline Panel, 1992). This statement unequivocally endorsed the need for pain management in neonates. The result of the publication of these two documents was the initiation of research studies into the prevention and amelioration of neonatal pain. The most influential of these documents came from The Joint Commission. The 2003 accreditation standards required health care providers to look across the continuum of life, including the neonatal period, at the complex nature of the pain experience so as to create new foundations for care (Joint Commission International Accreditation Standards for the Care Continuum, 2003). However, these guidelines do not provide specific instructions for assessing or managing pain in neonates. Neonatal caregivers needed to assess and treat perceived neonatal pain and discomfort, but had little research-based evidence on which to base their assessment and therapy. Organizations have commissioned interdisciplinary teams to incorporate regulatory directives and results of scientific investigation into institutional practice guidelines and standards for care (Anand, 2001; Howard et al, 2008a, 2008b; Lago et al, 2009). These guidelines include a patient’s right to regular and systematic assessments of pain, interventions to relieve pain, evaluation of effectiveness of interventions, attention to long-term pain management needs, deleterious effects of unmanaged pain, and educational needs of families and staff members who provide care (American Pain Society, 2006; Bell, 1994; Carrier and Walden, 2001; Howard et al, 2008a, 2008b; Joint Commission on Accreditation of Healthcare Organizations, 2004; Lago et al, 2009). As a result of these public policy initiatives and regulations, new tools for pain assessment and innovative methods to treat pain have been developed and evaluated. However, many old questions remain unanswered and new concerns have been raised. Neonates can suffer both acute and chronic pain. Treatment protocols for acute pain may not be appropriate for chronic pain, because the origin and resultant physiologic status can be quite different. Long-term use of narcotics and other drugs leads to drug tolerance and the need to wean slowly to avoid drug withdrawal. The drugs themselves, as well as drug tolerance and drug withdrawal, may all contribute to adverse effects on brain development and neurodevelopmental outcomes. RECENT SURVEYS OF CLINICAL PRACTICE Surveys of clinical practice have shown that attention to neonatal pain management has improved over the last decade, but they also demonstrate that much more needs to be done. This concern is evidenced by three recent surveys from NICUs. Prestes et al (2005) surveyed four Brazilian units in October 2001 prospectively. They found that, of 91 neonates admitted to the NICU, there was documentation of systemic analgesia in only 25% of 1025 patient days. No specific drug was administered during arterial, venous, and capillary sticks, lumbar punctures, or intubation. Only 9 of 17 surgical patients received any postoperative analgesia (Prestes et al, 2005). Taylor et al (2006) surveyed 10 NICUs in North America to determine their use of postoperative analgesia. Data were collected for 25 consecutive postoperative admissions. Nurses documented pain assessments for 88% of the infants, whereas physicians documented pain assessments for only 9%. The study concluded that “documentation of postoperative pain assessment and management in neonates was extremely variable” and called for development of evidence-based guidelines for postoperative care and education of professional staff. Carbajal et al (2008) prospectively studied 430 infants admitted to tertiary NICUs in Paris between September 2005 and January 2006. Each neonate averaged 115 total procedures (75 of which were considered to be painful) and 16 procedures per day (10 of which were considered to be painful). Of a total of 42,413 painful procedures, 2.1% were performed with pharmacologic therapy; 18.2% with nonpharmacologic therapy; 20.8% with pharmacologic, nonpharmacologic, or both; and 79.2% without specific analgesia. All three of these recent survey studies found that despite current knowledge of pain assessment and treatment methods, most infants still are not provided any specific treatment to alleviate pain and discomfort ONTOGENY AND DEVELOPMENT OF PAIN AND STRESS RESPONSES The sensory system of the neonate, especially the preterm infant, is immature (Figure 35-1). Afferent input from both noxious and non-noxious stimuli terminate in the dorsal horn of the spinal cord in a diffuse manner on multiple cells; this results in the inability to distinguish between noxious and non-noxious stimuli and limits the care provider’s ability to correctly interpret the infants’ behavioral response. In the neonatal rat, separation of sensory input is not completed until 3 to 4 weeks after birth (approximately 1 to 2 years in humans; Beggs et al, 2002); this prevents the newborn from differentiating touch from painful sensory input. The responses of the infant are therefore nonspecific. With repeated painful exposures, infants may lose any discriminatory ability and develop hypersensitive states for long periods of time. This hypersensitivity persists even if non-noxious stimuli are introduced (Evans, 2001; Jennings and Fitzgerald, 1998). The responses are less synchronized in the immature central nervous system are also noted in the immature central nervous system because of underdeveloped myelination and slower synaptic transmission as manifested in longer and more variable latencies (Fitzgerald, 2005; Jennings and Fitzgerald, 1998). Neuronal connections within the cortex appear to form at approximately 22 weeks’ gestation (Kostovic and JovanovMilosevic, 2006), suggesting that higher cortical level pain processing may be limited despite the presence of a behavioral response (Fitzgerald and Walker, 2009). In addition, the neonate lacks sufficient descending modulatory control, thereby limiting their ability to benefit from endogenous control over noxious stimuli compared with adults (Fitzgerald and Koltzenburg, 1986; Hathway et al, 2006). CHAPTER 35 Neonatal Pain and Stress: Assessment and Management 431 Excitatory interneuron 4 2 Lamina II neuron A-fibre input Late embryonic Inhibitory interneuron Epidermis Dermis A-fibre central terminal fields C-fibre central terminal fields 5 Descending input Excitatory interneuron C-fibre input 1 Lamina II neuron A-fibre input Newborn Inhibitory interneuron Peripheral nerve Spinal cord DRG Epidermis Dermis N52 3 trkA Descending input Excitatory interneuron  C-fibre input  Lamina II neuron   Adult Inhibitory interneuron Epidermis Dermis trkA N52 IB4 FIGURE 35-1 1, In early postnatal life, descending fibres are present, but inhibitory and excitatory influences are weak or absent. The connections gradually strengthen, becoming fully functional at the end of the third postnatal week. 2, A-fibres are the first primary afferents to enter the dorsal horn gray matter and are present during the last few embryonic days. Their distribution is diffuse with exuberant, more superficial projections gradually retracting over the first 3 postnatal weeks. C-fibres are present in the dorsal horn during late embryonic stages, but only enter the gray matter 2 to 3 days before birth. Unlike A-fibres, they project to topographically appropriate regions in lamina II of the spinal cord as soon as they enter. C-fibre synaptic connectivity is present, although weak at the time of birth, with connections strengthening over the first 2 postnatal weeks. 3, At birth the majority (approximately 80%) of dorsal root ganglion (DRG) neurons express the nerve growth factor (NGF) receptor trkA. Over the first postnatal week, this population reduces, with approximately half of these neurons losing their trkA expression and beginning to express receptors for glial cell line-derived neurotrophic factor (GDNF) (identifiable as the IB4-binding population). 4, The balance of excitation and inhibition in the superficial dorsal horn develops postnatally, through changes in both local interneuron circuitry and descending fibres. A-fibre input is stronger in the neonate and weakens as the influence of C-fibre input increases. 5, Primary afferent innervations of the skin occurs earlier than central projections. By late embryonic stages, primary afferents of all classes have reached the skin and innervate through the dermis into the epidermis. These projections die back during the immediate perinatal period to leave the full adult situation of dermal innervations present soon after birth. (From Beggs S, Fitzgerald M: Development of peripheral and spinal nociceptive systems. In Anand KJ, Stevens BJ, McGrath PJ, editors: Pain in neonates and infants, ed 3, Philadelphia, 2007, Elsevier, p 15.) 432 PART VIII Care of the High-Risk Infant Noxious stimuli in adults result in the release of inflammatory and trophic factors that activate and sensitize nociceptors in the injured tissue. Such noxious stimuli lead to nociceptive afferent input to the central nervous system, exciting nociceptive circuits in the spinal cord, brainstem, thalamus, somatosensory cortex, cingulate cortex, and amygdale (Tracey and Mantyh, 2007; Woolf and Ma, 2007). However, noxious stimuli in infants do not evoke similar patterns of central nervous system activity as noted in adults (Fitzgerald, 2005). The response to noxious stimuli is more diffuse and less spatially focused in infants. Studies in rats have demonstrated major alterations in neuronal circuitry with maturation, appearing to mirror that seen in humans but in a compressed time period. These animal data may parallel developmental changes that have been noted in humans (Fitzgerald, 2005). Local tissue injury resulting from repeated heel sticks and invasive procedures triggers increased proliferation of nerve endings in surrounding tissues, particularly when this damage occurs early in gestation. As a result, scars (e.g., from heel sticks, old intravenous sites) and surrounding tissues can remain hypersensitive well beyond the neonatal period (Jennings and Fitzgerald, 1998; Reynolds and Fitzgerald, 1995). Pain assessment and management are thus most challenging in preterm infants who are exposed to the stressful NICU environment for long periods (Johnston and Stevens, 1996). In summary, the immature nervous system in preterm infants lacks the ability to discriminate consistently between noxious and non-noxious stimuli, reacts often with similar behavior to a variety of stimuli, lacks the ability to modulate pain responses, and does not consistently manifest signs or symptoms that allow care providers to accurately assess the infant’s level of pain and discomfort. Furthermore, infants cannot verbally report their level of pain. This later challenge remains a significant obstacle to accurate assessment of pain in the newborn and determination appropriate treatment. ASSESSMENT OF NEONATAL PAIN AND STRESS DEVELOPMENT OF BEHAVIORAL RESPONSES TO PAIN Tools for assessment of neonatal pain and stress must be based on an understanding of the normal development of behavioral responses to pain and stress—an infant is not a small adult. Behavioral responses to noxious stimuli in infants are not always predictable because of immaturity of the central nervous system; therefore assessment of pain and response to therapeutic intervention can be similarly unpredictable. Significant structural and functional changes occur in pain pathways during development, many of which continue to take place after birth (see later discussion of ontogeny and development of pain and stress responses). When assessing pain and discomfort in infants, especially those born prematurely, it is important to remember that the patients are nonverbal and cannot say where and how badly they hurt. Indeed, even if they could verbalize such information, their ability to accurately localize and describe the pain would be limited as discrimination between inputs (tactile or nociceptive) is immature. Moreover, their behavioral response repertoire is also limited and is not always predictable. These limitations make the use of assessment tools problematic, especially when trying to determine whether there has been a response to a drug or other intervention. INFANT PAIN SCORES Over the past two decades, more than 40 infant pain scales have been developed (Ranger et al, 2007). Most of these scales were developed for use in clinical trials to assess efficacy of various pharmacologic therapies and have not been validated for general clinical use. However, the use of these scales in clinical practice has led to better health care provider recognition that procedural pain and stress are common. Moreover, it is assumed that this pain and distress can be ameliorated by appropriate use of pharmacologic and nonpharmacologic approaches. However, assessment and treatment of chronic pain and discomfort in infants remains problematic. The pain instruments developed for assessment of acute pain were not designed or validated for the chronic pain and discomfort associated with mechanical ventilatory support, or for use in paralyzed or neurologically compromised infants (Anand, 1998). Furthermore, painful stimuli may be processed at the cortical level in infants without producing detectable behavioral changes (Slater et al, 2008). Therefore the clinician is left to decide whether the touted benefits of opiate sedation during mechanical ventilation in newborns outweigh the possible adverse effects, both acute and long term. The importance of selecting a valid, reliable, and practical pain assessment tool is evident from the current literature. These tools assess the response to noxious stimuli by categorizing the behavioral or physiologic reactions of the infant, or a combination of both (Grunau and Craig, 1987). The behavioral responses can include limb movements, muscle tone, crying, and characteristic facial expressions. The physiologic measures can include heart rate, oxygen saturation, and respiratory rate. Of note, specific facial expression changes have been determined to be the most reliable indicators of pain (brow bulge, eye squeeze, nasolabial furrow, taut lips, and open mouth) whereas crying is the least reliable indicator (Grunau and Craig, 1987; Guinsburg et al, 1997; Stevens et al, 1993). Use of facial expression changes can be challenging when the infant’s face is partially covered with adhesives to secure tubes and lines, or with a phototherapy mask in place. A recent publication lists the current available pain assessment tools that can be used in neonates (Stevens et al, 2007). The most commonly used tools are listed in Table 35-1. Many pain assessment tools are available; however, they have not been systematically reviewed, and there is no specific gold standard for comparison. Because each instrument requires training and practice for optimal use, care providers should choose one that best fits the infant population being assessed to allow for consistency. The optimal approach to pain management should include reducing the frequency of painful procedures, reducing environmental stressors, facilitating neurologic developmental, determining the best technique to minimize the pain and stress 433 CHAPTER 35 Neonatal Pain and Stress: Assessment and Management TABLE 35-1 Infant Pain Scales Measure Age Level Indicators Pain Type Psychometric Properties Unidimensional Behavioral Measures of Infant Pain Baby facial action coding system (Rosenstein and Oster, 1988) Term infants Facial actions based on adaptation from adult work Procedural Interrater reliability (r = 0.65-0.85) Infant body coding system (Craig et al, 1993) Preterm infants (32 wk GA) to term infants Movement of hand, foot, arm, leg, head, torso Procedural Interrater reliability (r = 0.83), face validity, content validity Neonatal facial coding system (Grunau and Craig, 1987) Preterm infants (>25 wk GA) to term infants Brow bulge, eye squeeze, nasolabial furrow, open lips, horizontal mouth, vertical mouth, lips pursed, taut tongue, chin quiver, tongue protrusion Procedural Interrater reliability (r = 0.88), content validity, face validity, construct validity, convergent validity, feasibility Multidimensional Pain Measures in Infants Children’s and infants’ postoperative pain scale (Büttner and Finke, 2000) Birth to 4 years Crying, facial expression, posture of the trunk, posture of the legs, motor restlessness Prolonged (­postoperative) Interrater reliability (<3 yr old subsample; r = 0.64-0.77), internal consistency (<1 yr old subsample; r = 0.96), content validity, construct and concurrent validity demonstrated in older subsample Modified behavioral pain scale (Taddio et al, 1995a) 2-6 months Facial expression, cry, body movement Procedural Interrater reliability (ICC = 0.95), internal consistency (r = 0.55-0.66), test-rest reliability (r = 0.95), content validity, construct validity (p <0.01), concurrent validity (r = 0.68-0.74) Composite Pain Measures in Infants Clinical scoring system (Barrieret al, 1989) 1-7 months Infant sleep during the preceding hour, facial expression, cry, motor activity, excitability, flexion, sucking, tone, consolability Prolonged (­postoperative) Interrater reliability (r = 0.79-0.88), content validity, discriminant validity (p <0.0001) Modified postoperative comfort score (Guinsburg et al, 1998) <32 wk GA (postnatal age of 12-48 h) Sleep, facial expression, sucking, hyperreactivity, agitation, hypertonicity, toe or finger flexion, consolability Prolonged (mechanical ­ventilation) Convergent validity in bedside (p <0.0001) and laboratory video coding (p = 0.02), content validity, divergent validity shown between placebo and analgesic samples (p < 0.05) Echelle Douleur Inconfort Nouveau-né (Debillon et al, 2001) 26-36 wk GA Facial expression, movement, sleep, consolability Prolonged (­postoperative) Interrater reliability (r = 0.59-0.74), content validity, construct validity (p <0.000) Modified postoperative comfort score (Guinsburg et al, 1998) Preterm infants Sleep, facial expression, activity, tone, consolability, cry, sociability Prolonged (­postoperative) Content validity, discriminant validity (p <0.0001) Neonatal pain, agitation and sedation scale (Hummel et al, 2008) <28-35 wk GA (age correction for prematurity) Crying, irritability, behavior state, facial expression, extremities, tone, vital signs (heart rate, respiratory rate, blood pressure, oxygen saturation) Prolonged (mechanical ventilation or postoperative) Preliminary reliability and validity testing in progress Pain assessment tool (Hodgkinson et al, 1994) 27 wk GA to full term Posture, tone, sleep pattern, expression, color, cry, respirations, heart rate, oxygen saturations, blood pressure, nurse perception Prolonged (­postoperative) Interrater reliability (r = 0.85), ­content validity, convergent validity (r = 0.38), concurrent validity (r = 0.76) Scale for use in newborns (Blauer and Gerstmann, 1998) 24-40 wk GA Central nervous system state, breathing, movement, tone, face, heart rate, mean blood pressure Procedural (excluded ­postoperative) Beginning indications of reliability, content validity, discriminant validity (p < 0.05-0.01) Distress scale for ventilated newborn infants (Sparshott, 1996) — Crying, requires increased oxygen, increased vital signs, expression, sleeplessness (Krechel and Bildner, 1995) Neonates 32-60 wk GA Facial expression, body movement, color, heart rate, blood pressure, oxygenation, temperature Crying, requires increased oxygen, increased vital signs, expression, sleeplessness — Face validity, content validity (Duhn and Medves, 2004) Prolonged ­(postoperative) Interrater reliability (r = 0.72), content validity, concurrent validity (r = 0.49-0.73), discriminant validity (p <0.0001), concurrent validity for first 24 h (ICC = 0.340.65; McNair et al, 2004) Continued 434 PART VIII Care of the High-Risk Infant TABLE 35-1 Infant Pain Scales—cont’d Measure Age Level Indicators Pain Type Psychometric Properties Neonatal infant pain scale (Lawrence et al, 1993) Preterm and full term Facial expression, cry, breathing patterns, arm movement, leg movement, state of arousal Procedural Interrater reliability (r = 0.92-0.97), internal consistency (0.87-0.95), content validity, concurrent validity (r = 0.53-0.83) Pain assessment in neonates scale GA (HudsonBarr et al, 2002) 26-47 wk GA Facial expression, cry, breathing patterns, extremity movement, state of arousal, oxygen saturation, increased heart rate Procedural Content validity, concurrent validity (r = 0.93) Modified infant pain scale (Bucholz et al, 1998) 4-30 wk old Sleep during evaluation, facial expression, quality of cry, spontaneous motor activity, excitability and responsiveness to stimulation, finger and toe flexion, sucking, overall tone, consolability, sociability, change in heart rate, change in blood pressure, fall in oxygen saturation Prolonged preceding hour Interrater reliability (postoperative; r = 0.85), content validity, convergent validity in dichotomous rating (p <0.0001) Premature infant pain profile (Stevens et al, 1996) Term and preterm neonates GA, behavioral state, heart rate, oxygen saturation, brow bulge, eye squeeze, nasolabial furrow Procedural Interrater reliability (ICC = 0.930.96), intrarater reliability (ICC = 0.94-0.98), internal consistency (alpha = 0.59-0.76), content validity, construct validity (in preterm neonates, p = 0.0001-0.02; in term neonates, p <0.02), construct validity in clinical setting (p <0.0001), interrater reliability (r = 0.94-0.98; Ballantyn et al 1999), concurrent validity with cry duration (Johnston et al, 1999) Modified from Stevens BJ et al: Assessment of pain in neonates and infants. In Anand KJ, Stevens BJ, MaGrath PJ, editors: Pain in neonates and infants, ed 3, Philadelphia, 2007, Elsevier, pp 70-75. GA, Gestational age. associated with procedures, delegating responsibility for pain assessment and treatment to the bedside nurse staff, and using a balanced multimodal approach to pain control (Allegaert et al, 2009). BEDSIDE NONINVASIVE NEUROIMAGING TO EVALUATE PAIN AND STRESS Several new modalities are being evaluated for their capability to help health care providers recognize pain in neonates. Near infrared spectroscopy might be a helpful technology. In a study of 29 infants between 26 and 36 weeks’ gestation at birth, cortical activation occurred over both somatosensory cortices during unilateral tactile and painful stimuli (Bartocci et al, 2006). Amplitude-integrated electroencephalography also can be used to detect cortical activation in neonates (Toet and Lemmers, 2009). These and other new technologies will be needed to first recognize and then continually assess and manage pain and discomfort in our fragile patients. LONG-TERM CONSEQUENCES OF NEONATAL PAIN AND STRESS Untreated pain and stress has been linked to adverse longterm outcomes. Acute effects include elevations of cortisol, catecholamines, and lactate, hypertension, tachycardia, respiratory instability, glucose instability, and changes in cerebral blood flow. Chronic pain can affect growth, immune function, recovery, and length of hospitalization. In addition, a growing body of evidence has drawn attention to the potential deleterious effects of repeated handling, stress, and pain on long-term memory, social and cognitive development, and neural plasticity (Anand, 1998; Anand et al, 1987, 1989; Evans, 2001; Jennings and Fitzgerald, 1998; Pokela, 1994; Porter et al, 1999; Taddio et al, 1995a). Extrapolation of information from older children and adults can be inappropriate for our most vulnerable infants (Berde et al, 2005; Howard, 2003; Walker, 2008). Pain and stress may alter neurodevelopment (Abdulkader et al, 2008; Anand et al, 1987, 2000). Pain suffered during neonatal intensive care has been associated with adverse long-term outcomes, such as altered pain perception to subsequent immunizations after circumcisions without anesthesia (Taddio et al, 1997), abnormal cortisol responses to stress in later infancy (Grunau et al, 2007), and altered pain responses in childhood (Grunau et al, 1998). There is concern that such hormonal changes might lead to the development of cardiovascular disease and type 2 diabetes in adulthood (Kajantie et al, 2002; Rosmond and Björntorp, 2000). As a result, appropriate use of sedation and analgesia might be beneficial to infants requiring intensive care support; however, a clear delineation of the benefits still requires further study. CHAPTER 35 Neonatal Pain and Stress: Assessment and Management CLINICAL PAIN AND STRESS MANAGEMENT STRATEGIES SURGICAL ANESTHESIA The recognition that surgical intervention in infants results in dramatic physiologic and metabolic changes similar to those noted in adults with significant pain has changed current practice. It is well recognized that infants perceive pain, and surgical intervention without anesthesia and analgesia leads to increased morbidity and excess mortality; however, experience obtained in older children and adults cannot be extrapolated to immature patients. Providing general anesthesia in infants requires an intimate knowledge of the developmental status and function of each organ system. A complete review of this subject is beyond the scope of this chapter. Interested readers are referred to A Practice of Anesthesia for Infants and Children by Spaeth and Kurth (2009) or Anesthesia for Infants and Children by Brett et al (2006). FETAL SURGERY The fetus has increasingly become a candidate for surgical intervention to repair fetal anomalies such as congenital cystic adenomatoid malformation of the lung, sacrococcygeal teratoma, and myelomeningocele, and to treat fetal disorders such as severe anemia secondary to hemolytic diseases. Many fetal interventions are accomplished using general maternal anesthesia, and therefore fetal anesthesia, but certain procedures are attempted with maternal analgesia and local anesthesia. Thus, during such fetal surgical interventions, the fetus requires consideration for pain management to attenuate fetal physiologic and hormonal stress responses (Fisk et al, 2001; Giannakoulopoulos et al, 1999). Anand et al (1987) summarized the available evidence regarding fetal and neonatal nociceptive activity in 1987 and put all care providers on notice that late gestation fetuses and newborns have intact cortical and subcortical centers necessary for pain perception and demonstrate physiologic responses to painful stimuli similar to evidence in adult subjects. A recent review of the literature supports the contention that the fetus perceives pain (Lee et al, 2005). The International Association for the Study of Pain supported the view that the human fetus is capable of reacting to nociceptive stimuli during the second trimester (Anand, 2006). NEONATAL SURGERY Selection of anesthetic agents to be used in neonates must account for the developmental status and function of each organ system and the potential adverse and toxic effects of specific anesthetic agents. Animal models have demonstrated that anesthetic agents can be both neuroprotective and neurotoxic in the immature brain. Frequently used anesthetics act by two principal mechanisms, either by decreasing excitation via N-methyl-d-aspartate (NMDA) receptors (e.g., ketamine, nitrous oxide) or by increasing inhibition via gamma-aminobutyric acid receptors (e.g., benzodiazepines, barbiturates, propofol, etomidate, isoflurane, enflurane, halothane). In the immature rat brain, 435 drugs that act by either of these two mechanisms can induce widespread neuronal apoptosis when given during the period of synaptogenesis (Ikonomidou et al, 1999; Ishimaru et al, 1999). The applicability of these findings to the human infant has been questioned (Soriano and Anand, 2005); however, recent data suggest that some toxicity may occur in human infants (Wilder et al, 2009). Regional anesthesia for neonatal surgery is commonly used for minor interventions. As more experience with regional anesthesia for major surgical procedures is gained, such practice is expanding. Bösenberg (1998) presented results of the use of epidural analgesia for major neonatal surgical interventions in 240 infants weighing between 900 and 5800 g and reported that all infants had effective intraoperative analgesia with only two complications. Advantages of regional anesthesia may include less postoperative apnea, lees need for intubation, and better postoperative pain control. Sethna and Suresh (2007) reviewed this subject in depth recently. POSTOPERATIVE PAIN MANAGEMENT STRATEGIES Pain management after surgical intervention, like acute pain management, requires knowledge of the developmental status and function of each organ system and the potential adverse and toxic effects of specific analgesic agents. Many neonates remain intubated for mechanical ventilation after major surgery. Analgesia can be provided via intravenous administration of an analgesic agent, either by intermittent bolus dosing or by continuous infusion. Another option is continuation of regional analgesia (Bösenberg, 1998). Regional analgesia appears to be effective for pain control, but sedation may also be required in active or vigorous infants (Frumeinto et al, 2000). Regional anesthesia, when used alone, may also reduce the incidence of postoperative apnea in preterm infants (Craven et al, 2003). Research-based evidence of appropriate postoperative pain management in neonates is limited. Taylor et al (2006) surveyed 10 NICUs regarding their postoperative pain assessment and management practices; they found that pain assessment documentation was extremely variable. Nursing documentation was done for most infants, whereas few physicians documented any assessment. Most infants were treated with opioids, benzodiazepines, or both, and some infants (7%) received no analgesia despite recent major surgery (Taylor et al, 2006). Van der Marel et al (2007) evaluated the use of rectal acetaminophen as an adjuvant treatment to continuous morphine infusion in postoperative neonates and could not demonstrate any additional analgesia effect. Recent evidence suggests that neonates require significantly less morphine to control pain and discomfort than older infants do, based on monitoring of pain scale data. Bouwmeester et al (2003a, 2003b) determined that neonates required less morphine for postoperative pain control and that the dose requirement increased with age. Both studies found that morphine, given by bolus or continuous infusion, was equally effective. The later study (2003b) found that mechanical ventilation decreased morphine metabolism and clearance. 436 PART VIII Care of the High-Risk Infant The Association of Pediatric Anesthetists of Great Britain and Ireland commissioned guideline development for pain management after surgery and painful medical procedures (Howard et al, 2008a, 2008b). The level of evidence from the literature was determined and the recommendations were graded, allowing for better interpretation. These guidelines provided evidence-based recommendations and listed best clinical practice points when published evidence was insufficient to make formal recommendations. MECHANICAL VENTILATION The use of mechanical ventilation in neonates with respiratory failure is a fairly common practice in the United States, although decreasingly less than in the 1980s and 1990s. In older children and adults who require mechanical ventilation, sedation is routinely provided—most often with opiates (Gélinas et al, 2004). However, pharmacologic sedation is used less frequently in the adult intensive care unit because of concern regarding adverse cognitive outcomes and longer duration of ventilator support (Izurieta and Rabatin, 2002; Oderda et al, 2007). Extrapolation of evidence from past studies in adult patients led to the routine use of opiate sedation in neonates during mechanical ventilation (Kahn et al, 1998), with limited information regarding safety and efficacy (Anand et al, 2004; Simons et al, 2003b). As noted previously, the ability to assess and treat discomfort and pain is limited, and preemptive use of pharmacologic sedation during mechanical ventilation in newborns, especially preterm infants, remains controversial (Anand and Hall, 2007). Mechanical ventilation in neonates is associated with an increase in hormonal stress responses, including increased cortisol and catecholamine levels (Guinsburg et al, 1998; Quinn et al, 1998). In the past, infants who appeared uncomfortable while receiving ventilatory support demonstrated asynchronous respiratory effort (i.e., “fighting” the ventilator), compromised gas exchange, and altered stress responses (Dyke et al, 1995). Furthermore, pain and stress in newborns receiving mechanical ventilation have been associated with decreased pulmonary compliance, atelectasis, and intrapulmonary shunting (Bolivar et al, 1995). More recently, with the introduction and use of surfactant replacement therapy and synchronized ventilatory technology, some of the problems with fighting the ventilator have been eliminated (Claure and Bancalari, 2009; Keszler, 2009). Clinical studies from the 1990s demonstrated that opiate treatment prevented these adverse effects on neonatal ventilation and reversed the previously described hormonal stress changes (Quinn et al, 1993; Saarenmaa et al, 1999). Opiate sedation has been demonstrated to decrease stress scores in newborns who receive mechanical ventilation (Orsini et al, 1996; Quinn et al, 1993, 1998). In fullterm infants receiving mechanical ventilation, the severity of respiratory failure as assessed by the oxygenation index directly correlated with the need for analgesia and sedation (Aretz et al, 2004). A small randomized trial of routine morphine infusion in preterm infants receiving mechanical ventilation concluded that morphine lacked a “measurable analgesic effect” and there was “absence of a beneficial effect on poor neurological outcome” (Orsini et al, 1996). The larger clinical trial that followed (i.e., NEOPAIN) reported no beneficial effect of preemptive morphine infusions in ventilated preterm infants and an increased incidence of severe intraventricular hemorrhage in preterm infants born at 27 to 29 weeks’ gestation and receiving morphine (Anand et al, 2004). Bhandari et al (2005) found that morphine infusions had no beneficial effects in preterm infants receiving mechanical ventilation. Additional bolus doses of morphine resulted in worse respiratory outcomes and longer requirement for ventilatory support. None of these controlled clinical trials provide evidence that routine narcotic sedation during mechanical ventilatory support in neonates is beneficial. In addition, the well-recognized and common adverse effects of narcotic exposure (e.g., hypotension, feeding intolerance, respiratory depression) are mentioned infrequently in sedation and pain treatment protocols in neonates receiving mechanical ventilation. Future trials have been deemed unethical if they involve withholding sedation from a group of infants. One approach to this dilemma would be to minimize the use of ventilatory support as much as possible. A recent Cochrane Review evaluated the effects of opioid analgesics on pain, duration of mechanical ventilation, mortality, growth, and development in neonates requiring mechanical ventilation (Bellù et al, 2008). The authors found no differences in mortality, duration of mechanical ventilation, and short- and long-term neurodevelopmental outcomes. Preterm infants given morphine took longer to achieve full enteral feeding. If morphine sedation prolongs ventilatory support needs and time to full enteral feeds, then an increase in the risk of complications related to the use of venous lines (bloodstream infections) and parenteral nutrition (cholestasis) should be expected (Menon et al, 2008). Hällström et al (2003) studied risk factors for necrotizing enterocolitis in premature infants and found that the duration of morphine use was the strongest predictor for development of severe necrotizing enterocolitis. The Cochrane Review’s overall conclusion regarding the use of sedation during mechanical ventilation was that “there is insufficient evidence to recommend routine use of opioids in mechanically ventilated newborns” (Bellù et al, 2008). Menon and McIntosh (2008) came to a similar conclusion in their recent review. PROCEDURES Infants undergoing intensive care suffer through many painful procedures, often several times each day. Although new pharmacologic and nonpharmacologic treatment strategies to decrease or eliminate some of this pain and stress have been developed, there is still a need to develop better management techniques. An article by D’Apolito (2006) reviews this issue in detail and outlines where knowledge remains limited. BLOOD SAMPLING AND MONITORING Heel sticks are routinely performed to obtain blood samples in neonates. The most appropriate method for relieving pain from a heel stick is yet to be determined. The heel CHAPTER 35 Neonatal Pain and Stress: Assessment and Management should be warmed to aid blood sampling. Eutectic mixture of local anesthetics (EMLA) cream does not relieve the pain of a heel lance (Stevens et al, 1999b; Taddio et al, 1998). Shah et al (1997) and Larsson et al (1998a, 1998b) demonstrated that neonates experiencing venipuncture had lower pain scores than those who underwent heel stick for blood sampling. In selected neonates, venipuncture should be used preferentially over heel stick. The pain of arterial puncture can be decreased by infiltrating the site with 0.1 to 0.2 mL of 0.5% or 1% lidocaine using the smallest-gauge needle possible (Franck and Gregory, 1993). Buffering the lidocaine with sodium bicarbonate is recommended to decrease the burning caused by lidocaine. In addition, EMLA may reduce the pain of arterial puncture. TRACHEAL INTUBATION The use of premedication to minimize the pain and stress of intubation has been demonstrated to benefit neonates (Simons et al, 2003a). However, concerns about rapid medication availability, ability to maintain the airway, and the ability to provide ongoing ventilatory support continue to cause controversy (Carbajal et al, 2007). Premedications typically include atropine, narcotics for sedation, and muscle relaxants. Atropine abolishes vagal bradycardia. Narcotics attenuate the increases in arterial blood pressure. Muscle relaxants attenuate the increases in intracranial pressure. Combinations decrease the time and number of attempts needed to intubate the infant. When considering the use of medications for intubation, several questions need to be asked: ll Does the infant have adequate vascular assess? ll What is the urgency of intubation need? ll Is the infant known to have a difficult airway? ll When was the last feeding? ll Can the infant be preoxygenated while avoiding gastric distension? If the decision is made to use medications for intubation, typical doses include: ll Atropine 0.02 mg/kg fast intravenous (IV) push ll Fentanyl 2 μg/kg slow IV push ll Vecuronium 0.1 mg/kg IV push Alternative medications may include succinylcholine (1 to 2 mg/kg IV push) or rocuronium (1 mg/kg IV push). The investigational agent sugammadex can reverse rocuronium-induced neuromuscular blockade in less than under 2 minutes. Sugammadex forms a 1:1 complex with steroidal nondepolarizing neuromuscular blockers in the plasma. Although approved for use in Europe, sugammadex is not currently available in the United States; it might render succinylcholine unnecessary. Rocuronium could be administered and if intubation is unsuccessful, paralysis could be immediately reversed with sugammadex (de Boer et al, 2007). Propofol has been used as a premedication for intubation. Being a single agent, it is easier and faster to prepare than three separate drugs. Propofol is a hypnotic agent without anesthetic properties. If combined with remifentanil, there is no need for paralysis in older children (Batra et al, 2004); however, propofol is painful when injected in small veins, and extremely painful if it extravasates. 437 A major advantage is continued spontaneous breathing during the intubation procedure. The dose is 2.5 mg/kg IV; this dose might need to be repeated. Concerns over the use of propofol for intubation in neonates include minimal experience in neonates, uncertain pharmacokinetics and duration of action, and compatibility with precutaneously inserted central catheters lines. Because intubation can raise both blood pressure and intracranial pressure, a short-acting benzodiazepine, such as midazolam, can be beneficial for infants with stable cardiovascular function. Fentanyl can be used as an alternative for infants with compromised cardiovascular function (McClain and Anand, 1996). Any infant who is pharmacologically paralyzed during mechanical ventilation should receive adequate sedation. In addition, the infant should receive pain medication if pain is suggested by the infant’s condition or because of the procedures being performed. CIRCUMCISION The AAP Circumcision Policy Statement (American Academy of Pediatrics Task Force on Circumcision, 1999) states that analgesia must be provided to infants undergoing circumcisions. EMLA cream, dorsal penile nerve block, and subcutaneous ring block are all possible options. The AAP reports that subcutaneous ring block may provide the best analgesia (American Academy of Pediatrics Task Force on Circumcision, 1999). Subcutaneous ring block has been found to be more effective than EMLA or dorsal penile nerve block in other studies (Lander et al, 1997). Dorsal penile nerve block has been found to be more effective than EMLA, but this method is not always available (Lee and Forrester, 1992). EMLA has been established as superior to placebo for pain relief during circumcision (Benini, 1993; Taddio et al, 1997). An effective method for applying EMLA in preparation for circumcision is to apply one third of the dose to the lower abdomen, extend the penis upward gently, pressing it against the abdomen, and then apply the remainder of the dose to an occlusive dressing placed over the penis. This dressing is then taped to the abdomen so that the cream surrounds the penis. Another method is to apply the cream and then place plastic wrap around the penis in a tubelike fashion to direct the urine stream out and away from the cream. Acetaminophen is ineffective for the management of severe pain associated with the circumcision procedure, but it provides some analgesia in the postoperative period. Acetaminophen has been found to decrease pain 6 hours after circumcision (Howard et al, 1994). OTHER INVASIVE PROCEDURES Placement of a central venous catheter requires topical anesthesia with EMLA or infiltration of the skin with lidocaine. In addition, a parenteral opioid, such as morphine or fentanyl, is typically required. Consideration should also be given to regional blocks for central line placement. The pain of a lumbar puncture is compounded by both the needle puncture and the distress caused by the body position required for the procedure. EMLA has been 438 PART VIII Care of the High-Risk Infant shown to decrease the pain of lumbar puncture in children (Halperin et al, 1989). Chest tube insertion requires an intravenous opioid, adequate local analgesia (lidocaine), or both. NONOPIOID ANALGESICS Administering 10 mg/kg may be inadequate for pain control, because this dose is based on antipyretic doseresponse studies. The maximum recommended daily dose is 75 to 90 mg/kg for infants, 45 to 60 mg/kg for term and preterm neonates more than 32 to 34 weeks of postconceptual age, and 25 to 40 mg/kg/day for preterm neonates 28 to 32 weeks of postconceptual age (Berde and Sethna, 2002; Morris et al, 2003). Rectally administered acetaminophen has a longer halflife, but absorption is highly variable because it depends on the individual infant and placement of the suppository. It should also be noted that the suppository may contain all of the drug in its tip and should be divided lengthwise if a partial dose is desired. The analgesic effect of acetaminophen may be additive when the agent is administered with opioids. This coadministration may enable a decrease in the opioid dose and therefore in corresponding opioid side effects; however, demonstration of this potential benefit awaits further study (van der Marel et al, 2007). Nonsteroidal Antiinflammatory Drugs (Indomethacin, Ibuprofen) OPIOID ANALGESICS PHARMACOLOGIC ANALGESIA INTERVENTIONS The severity of the pain, etiology, available administration routes, and consideration of potential side effects should all be evaluated during selection of an analgesic. Once medication administration has begun, careful monitoring for side effects can decrease potential adverse events related to administration of pain medications to infants. A key component of effective pain management is reassessment after a painful intervention, although this is difficult to do with limited pain assessment tools. Nonsteroidal antiinflammatory drugs (NSAIDs) inhibit prostaglandin synthesis by inhibiting the action of cyclooxygenase enzymes. Cyclooxygenase enzymes are responsible for the breakdown of arachidonic acid to prostaglandins. NSAIDs have many physiologic effects, including sleep cycle disruption, increased risk of pulmonary hypertension, cerebral blood flow alterations, decreased renal function by decreasing glomerular filtration rate, alteration in thermoregulatory control, and changes in platelet function. Moreover, development of the central nervous, cardiovascular, and renal systems is dependent on prostaglandins. These adverse effects are particularly worrisome for neonates and infants; however, these drugs are used frequently in the NICU for pharmacologic closure of the PDA. Aside from effects on renal and perhaps mesenteric circulation, which are difficult to separate from the PDA, no clear side effects have been reported. Acetaminophen Acetaminophen is the most widely administered analgesic in patients of all ages. Acetaminophen inhibits the activity of cyclooxygenase in the central nervous system, decreasing the production of prostaglandins, and peripherally blocks pain impulse generation (Arana et al, 2001). Neonates are able to form the metabolite that results in hepatocellular damage (Arana et al, 2001); however, it is inappropriate to withhold acetaminophen in newborns because of concerns of liver toxicity. The immaturity of the newborn’s cytochrome P-450 system may actually decrease the potential for toxicity by reducing production of toxic metabolites (Collins, 1981). Current recommendations are for less frequent oral dosing (every 8 to 12 hours in preterm and term neonates), because of slower clearance times, and higher rectal dosing because of decreased absorption (Arana et al, 2001; van Lingen et al, 1999). Typical oral doses for acetaminophen are 10 to 15 mg/kg every 6 to 8 hours for term neonates and 10 to 15 mg/kg every 4 to 6 hours for infants. Opioids are believed to provide the most effective treatment for moderate to severe pain in patients of all ages. There is a wide range of interpatient pharmacokinetic variability. Opioid dosing depends on the severity of the pain as well as the age and clinical condition of the infant. Opioids should be used in infants younger than 2 months only in a monitored setting such as an intensive or intermediate care unit (Yaster et al, 2003). Some clinicians propose a more conservative recommendation, restricting the use of opioids to monitored settings for any infant younger than 6 months. Morphine Morphine remains the standard for pain management in neonates, although not necessarily because it has been shown to be the most effective analgesic. Morphine is metabolized in the liver by uridine diphosphate glucuronyltransferase into two active metabolites: (1) morphine6-glucuronide (M6G), a potent opiate receptor agonist, and (2) morphine-3-glucuronide (M3G), a potent opiate receptor antagonist. Both metabolites and some unchanged morphine are excreted in the urine. The predominant metabolite in preterm and full-term neonates is M3G. Because of slow renal excretion, the metabolites can accumulate substantially over time (Bouwmeester et al, 2003a, 2003b; Saarenmaa et al, 2000). There is a potential for late respiratory depression because of a delayed release of morphine from less well-perfused tissues and the sedating properties of the M6G metabolite (Anand et al, 2000). Because the predominant metabolite of morphine in infants is M3G, a potent opiate receptor antagonist, using the lowest dose possible to achieve the needed analgesia should be considered. Escalating morphine doses will also increase the levels of M3G in the infant, interfering with the goal of adequate analgesia. Doses as low as 1 to 5 μg/kg/h can provide adequate analgesia, minimizing the risk of accumulation of high M3G levels with CHAPTER 35 Neonatal Pain and Stress: Assessment and Management the metabolite’s prolonged half-life (Bouwmeester et al, 2003a, 2003b). Clearance or elimination of morphine and other opioids is prolonged in infants, because of the immaturity of the cytochrome P-450 system at birth. The rate of elimination and clearance of morphine in infants 6 months and older approaches that in adults. Chronologic age seems a better indicator than gestational age of how an infant metabolizes opioids (Scott et al, 1999; Yaster et al, 2003). Infants are at greater risk for opioid-associated respiratory depression because of their immature responses to hypoxia and hypercarbia. There is an increase in unbound or free morphine and M6G available to reach the brain as a result of the reduced concentration of albumin and alpha1 acid glycoproteins (Houck, 1998). Hypotension, bradycardia, and flushing constitute the response to the histamine release and rapid intravenous administration of morphine. Histamine release may cause bronchospasm in infants with chronic lung disease, although this is not commonly seen (Anand et al, 2000). Morphine sedation may result in extended need for ventilatory support in neonates (Anand et al, 1999; Bhandari et al, 2005). Dosing recommendations currently reflect the wide range of interpatient pharmacokinetic variability. In the past, 0.03 mg/kg of morphine IV was suggested as a starting dose in infants not receiving ventilation (Acute Pain Management Guideline Panel, 1992), whereas 0.05 to 0.1 mg/kg of morphine IV was recommended as an appropriate starting dose in infants receiving ventilation. Recently, much lower doses have been recommended (Anand et al, 2008; Bouwmeester et al, 2003b, 2004; Lynn et al, 2000; Saarenmaa et al, 2000). Titration to the desired clinical effect is required in adjusting both the dose and the frequency of administration. Furthermore, it is important to continually assess need and responses so that dosing can be adjusted both up and down (Allegaert et al, 2009). As the use of morphine for analgesia and sedation in neonates is explored further, it is becoming clear that some of the risks may outweigh the potential benefits (Allegaert et al, 2009; Anand et al, 2008; Black et al, 2008; Carbajal et al, 2005; Nandi et al, 2004; Ng et al, 2003; Ranger et al, 2007). Fentanyl Fentanyl is 80- to 100-fold more potent than morphine and causes less histamine release, making it a more appropriate choice for infants with hypovolemia, hemodynamic instability, or congenital heart disease. Another potential clinical advantage of fentanyl is its ability to reduce pulmonary vascular resistance, which can be of benefit for infants who have undergone cardiac surgery, have persistent pulmonary hypertension, or need extracorporeal membrane oxygenation (Anand et al, 2000). Bolus doses of fentanyl must be administered over a minimum of 3 to 5 minutes to avoid chest wall rigidity, a serious side effect observed after rapid infusion. Chest wall rigidity, which can result in difficulty or inability to ventilate, can be treated with naloxone or a muscle relaxant such as pancuronium or vecuronium. Fentanyl is highly lipophilic. It has a quick onset and relatively short duration of action. Because of fentanyl’s 439 short duration of action, it is typically used as a continuous infusion for postoperative pain. In infants 3 to 12 months of age, total body clearance of fentanyl is greater than that of older children, and the elimination half-life is longer because of its increased volume of distribution (Singleton et al, 1987). Fentanyl has been demonstrated to have a prolonged elimination half-life in infants with increased abdominal pressure (Gauntlett et al, 1988; Koehntop et al, 1986). Because of tachyphylaxis, continuous infusions of fentanyl are often increased to maintain constant levels of sedation and pain management. Infusion dosing can reach substantial levels requiring prolonged withdrawal. A rebound transient increase in plasma fentanyl levels is a phenomenon known to occur after discontinuation of therapy in neonates. It is a result of the accumulation of fentanyl in fatty tissues, which can prolong its effects after continued use; therefore caution must be exercised in the use of repeated doses or a continuous infusion. Oral Opioids Oral methadone can be used to wean infants from long-term opioid use. Methadone is widely used in neonates and children, although there are limited data regarding its efficacy and pharmacokinetics in this population (Chana and Anand, 2001; Suresh and Anand, 1998). The respiratory depressant effect of methadone is longer than its analgesic effect. Methadone is metabolized slowly, and it has a long half-life. Codeine is prescribed at 0.5 mg to 1 mg/kg orally every 4 hours as needed. Scarce data are available to recommend use of codeine in neonates. Most pharmacies supply acetaminophen and codeine in a set formula, consisting of acetaminophen (120 mg) and codeine phosphate (12 mg per 5 mL) with alcohol (7%). The dose prescribed is limited by both the appropriate dose of codeine and the safe dose of acetaminophen. This combination is not recommended in neonates. Oxycodone dosing is 0.05 mg/kg to 0.15 mg/kg orally every 4 to 6 hours as needed. No data are available to recommend oxycodone in neonates. The liquid form is not universally available. MIXED OPIOID AGONIST-ANTAGONIST DRUGS Nalbuphine is a mixed agonist-antagonist opioid receptor drug; therefore its administration in infants of opioidaddicted mothers may precipitate withdrawal. This agent is equianalgesic with morphine. Nalbuphine has a ceiling effect for analgesia. Additional studies are needed regarding the safety and efficacy of nalbuphine use in infants. It is not recommended for use in neonates, although it may be useful during opioid drug withdrawal (Jang et al, 2006). LONG-TERM CONSEQUENCES OF NEONATAL OPIOID EXPOSURE Experimental Animal Studies Perinatal and neonatal opioid exposure in experimental animals is associated with both short- and long-term adverse neurologic effects that should make clinicians ask 440 PART VIII Care of the High-Risk Infant whether the use of such medications with questionable benefits should be used at all. Data from previous studies suggested that perinatal narcotic exposure restricts brain growth, induces neuronal apoptosis, and alters behavioral pain responses later in life (Handelmann and DowEdwards, 1985; Hu et al, 2002; Kirby et al, 1982; Seatriz and Hammer, 1993). One area of particular concern to clinicians is the developing cerebral circulation, which is extremely vulnerable to physiologic perturbations and the effects of drugs (Volpe, 1998). Cerebrovascular effects of drug exposure early in development can have lifelong consequences, including increased risk for stroke (Barker, 2000; Craft et al, 2006; Hanson et al, 2004). The acute effects of exogenous narcotics, including morphine, on the developing cerebral circulation have been described in piglets and include modulation of prostaglandin-induced pial artery dilation during hypoxia, alteration in endothelin production, and increases in endothelin A receptor mRNA expression (Armstead et al, 1990, 1996; Van Woerkom et al, 2004). Endogenous opioids are important regulators of cerebrovascular tone and angiogenesis (Blebea et al, 2000; Gupta et al, 2002; Pasi et al, 1991; Poonawala et al, 2005). Exposure to morphine in fetal sheep and neonatal rats permanently alters cerebrovascular control mechanisms (Mayock et al, 2005, 2006). We have also shown permanent neurobehavioral and neuropathologic changes in a rodent model of neonatal stress and morphine exposure (Boasen et al, 2009; McPherson et al, 2007; Vien et al, 2009). These and other animal studies demonstrate short- and longterm effects of neonatal morphine exposure, which is not surprising because opioid receptor-mediated signaling likely has a role in several aspects of early brain development (Durrmeyer et al, 2010). However, the clinical relevance of these animal studies regarding the long-term effects of neonatal opioids is difficult because of species differences in the timing of brain development, the development of opiate receptors and major neurotransmitter systems, and the pharmacokinetics of administered opioids. Overall, it is difficult but necessary to simulate the premature infant’s NICU experience in an animal model (Durrmeyer et al, 2010). Clinical Studies Clinical studies addressing the short- and long-term effects of prolonged opiate use in neonates are limited. The few that exist are contradictory and confounded by illness severity. Bergman et al (1991) described reversible encephalopathic changes in neonates receiving long-term sedative and narcotic infusions. MacGregor et al (1998) demonstrated no adverse neurodevelopmental outcomes in a small group of newborns who received morphine for a median of 5 days. Roźe et al (2008) presented 5-year neurodevelopmental outcomes in very low-birthweight (VLBW) infants exposed to prolonged sedation or analgesia (defined as greater than 7 days of sedative or opioid drugs). They found that exposed VLBW infants had more severe or moderate disability at 5 years (42%) compared with those not exposed (26%), but after adjusting for gestational age and propensity score (as a way to ensure that treatment effects are only compared between infants who are equally likely to receive that treatment), the association was no longer significant. Preterm infants (23 to 32 weeks’ gestation at birth) evaluated at 36 weeks postconceptual age in the NEOPAIN study already demonstrated neurobehavioral abnormalities if exposed to morphine during ventilatory support (Rao et al, 2007). TOPICAL AND LOCAL ANESTHETICS Lidocaine reduces the pain and stress of venipuncture and IV catheter placement (Larsson et al, 1998a, 1998b; Long et al, 2003). EMLA cream has been demonstrated to reduce the pain of circumcision. Place the cream on the area where anesthesia is desired and then cover it with an occlusive dressing for 1 hour before the procedure. Longer application times provide deeper local anesthetic penetration, but can lead to toxicity. There is a slight risk of methemoglobinemia with the use of EMLA cream in infants and patients who are G6PD deficient. A rare occurrence, methemoglobinemia can occur when hemoglobin is oxidized by exposure to prilocaine. EMLA should not be used in patients with methemoglobinemia or in infants younger than 12 months who are also receiving methemoglobinemia-inducing drugs, such as acetaminophen, sulfonamides, nitrates, phenytoin, and class I antiarrhythmics. (Refer to Table 35-2 for recommended maximum doses of EMLA cream by age and weight.) A study of 30 preterm infants found that a single 0.5-g dose of EMLA applied for 1 hour did not lead to a measurable change in methemoglobin levels (Taddio et al, 1995b). A systematic review concluded that EMLA diminishes the pain during circumcision. Limited efficacy was noted with pain from venipuncture, arterial puncture, and percutaneous venous line placement. EMLA was not found to diminish pain from heel lancing (Taddio et al, 1998). Oral sucrose or glucose may be as effective as EMLA for venipuncture (Abad et al, 2001; Gradin et al, 2002). SEDATIVES Benzodiazepines Benzodiazepines such as lorazepam and midazolam are sedatives which activate gamma-aminobutyric acid receptors and should not be used in place of an appropriate EMLA Cream: Recommended Maximum Dose by Age and Weight TABLE 35-2 Body Weight (kg) Maximum Total EMLA Dose (g) Maximum Application Area (cm2) Maximum Application Time (h) Birth to 3 mo or <5 1 10 1 3-12 mo and >5 kg 2 20 4 Age Data from Taketomo CK, Hodding JH, Kraus DM, editors: Pediatric dosage handbook, 2001-2002, ed 8, Hudson, Ohio, Lexi-Comp Inc, p 595. EMLA, Eutectic mixture of local anesthetics. CHAPTER 35 Neonatal Pain and Stress: Assessment and Management pain medication, because this class of medication has no analgesic effect. For painful procedures, an analgesic must be used in conjunction with the benzodiazepine. Benzodiazepines are administered to decrease irritability and agitation in infants and to provide sedation for procedures. In ventilated infants, benzodiazepines can help to avoid hypoxia and hypercarbia from breathing out of sync with the ventilator although, as noted for opioids, new synchronized infant ventilators make this clinical problem less likely. When given as continuous infusions, dosing often escalates rapidly to maintain apparent sedation resulting in need for prolonged weaning. Use of such medications has been associated with abnormal neurologic movements in both preterm (Lee et al, 1994) and term infants (Chess and D’Angio, 1998). In rats, prenatal exposure to diazepam results in long-term functional deficits and atypical behaviors (Kellogg et al, 1985); exposure of 7-day-old mice to diazepam induces widespread cortical and subcortical apoptosis (Bittigau et al, 2002); and midazolam potentiates pain behavior, sensitizes cutaneous reflexes, and has no sedative effect in newborn rats (Koch et al, 2008). Whether these data can be extrapolated to human infants is unknown, but clinicians have reason to be concerned and should use these drugs with caution in the NICU. Dexmedetomidine Dexmedetomidine is a potent and relatively selective α2 adrenergic receptor agonist indicated for the short-term sedation of patients in intensive care settings, especially those receiving mechanical ventilatory support. The drug is administered by either bolus doses for short procedural sedation (1 to 3 μg/kg) or continuous intravenous infusion (0.25 to 0.6 μg/kg/h). Because dexmedetomidine does not produce significant respiratory depression, it has been used for procedural interventions in spontaneously breathing infants (Barton et al, 2008; Chrysostomou et al, 2009). As neonatologists become more familiar with dexmedetomidine, its use may increase (O’Mara et al, 2009); however, short- and long-term safety and effectiveness information needs to be assessed in infants, as has been initiated in adults (Pandharipande et al, 2007; Riker et al, 2009). NONPHARMACOLOGIC ANALGESIC INTERVENTIONS Nonpharmacologic interventions for prevention or relief of neonatal pain and stress are numerous and widely publicized in the medical literature. These interventions have been used either as the sole method of pain control or in combination with pharmacologic interventions. No one would argue with the statement that neonatal intensive care is associated with stress, pain, and discomfort. Because opioid analgesia and sedation have not been proved to be efficacious and may possibly be harmful, alternative methods of pain and stress relief need to be evaluated for efficacy and safety. A variety of approaches have been investigated. As stated clearly by Golianu et al (2007), “These therapies may optimize the homeostatic mechanisms of the infant, thereby mitigating some of 441 the adverse consequences of untreated pain, as well as facilitating healthy physiologic adaptions to stress.” However, widespread adoption of specific techniques is not consistent. Nonnutritive sucking with pacifiers reduces pain responses to heel prick, injections, venipuncture, and circumcision procedures (Sexton and Natale, 2009; Shiao et al, 1997; South et al, 2005). Infant massage has been demonstrated to decrease plasma cortisol and catecholamine levels in preterm infants (Acolet et al, 1993; Kuhn et al, 1991). Maternal skin-to-skin contact (also termed kangaroo care) is associated with greater physiologic stability and reduced responses to acute pain (Bergman et al, 2004; Fohe et al, 2000; Gray et al, 2000; Johnston et al, 2003; Ludington-Hoe and Swinth, 1996). Kangaroo care can decrease Neonatal Infant Pain Scale scores after vitamin K injections (Kashaninia et al, 2008). Maternal rocking has been shown to diminish neonatal distress (Jahromi et al, 2004). Breastfeeding reduces the physiologic and behavioral responses to acute pain and stress in neonates and has been recommended as the first line of treatment (Osinaike et al, 2007; Shah et al, 2006). The Neonatal Individualized Developmental Care and Assessment Program (NIDCAP) systematically changes a protocol-based model of nursing care to a relationshipbased approach (Als et al, 1994). There is a significant body of empiric evidence that use of the NIDCAP approach improves the clinical and neurodevelopmental outcomes of preterm infants (Als and Gilkerson, 1997; Brown and Heermann, 1997; Kleberg et al, 2002; Westrup et al, 2000; Wielenga et al, 2007), but recent metaanalyses of published studies could not demonstrate any improvement in long-term outcomes (Jacobs et al, 2002; Symington and Pinelli, 2006). Another approach is multisensory stimulation of preterm infants undergoing painful procedures. This approach entails simultaneous gentle massage, soothing vocalizations, eye contact, smelling a perfume, and sucking on a pacifier. This technique was associated with analgesia and calming of the infants in several reports from one unit (Bellieni et al, 2001, 2002, 2007). Music therapy may reduce the behavioral and physiologic responses to acute procedural pain (Hartling et al, 2009). Oral sucrose (versus intragastric) reduces pain behavior in preterm and term infants and is used widely (Stevens et al, 1997, 2004). The mechanism of oral sucrose analgesia is believed to be the sweet taste stimulation of endogenous opioid release (Shide and Blass, 1989). Of all methods and techniques discussed, oral sucrose has been the most widely used. As more data regarding the limitations of pharmacologic treatment are published, consideration of nonpharmacologic interventions will likely become more important and commonplace. IATROGENIC DRUG DEPENDENCE, TOLERANCE, WITHDRAWAL A clear distinction must be made between opioid or benzodiazepine dependence, tolerance, and addiction. Physical dependence is demonstrated by the need to continue 442 PART VIII Care of the High-Risk Infant the administration of the drug to prevent signs or symptoms of physical withdrawal. Tolerance is a reduction in the drug effects after repeated administration, or the need to increase the dose to achieve the same clinical effect. Addiction is compulsive drug-taking behavior (Gutstein and Akil, 2001). Infants are not capable of becoming psychologically addicted, but they clearly develop tolerance and dependence. DRUG WEANING CONSIDERATIONS Baseline pain and withdrawal scores should be obtained before beginning the drug weaning process, and infants should be reassessed every 2 to 4 hours for signs of withdrawal. In addition, when an opioid dosage is being tapered, the infant should be assessed for the presence of pain a minimum of every 4 hours. If an infant is receiving both an opioid and a benzodiazepine, it is prudent to taper and stop only one class of medication at a time. Typically, a weaning schedule is 10% of the total initial dose every other day. Many patients can tolerate a relatively large initial decrease in dose, but subsequent decreases may need to be smaller. Environmental stressors should be eliminated or reduced whenever possible. It should be noted that the potential onset of withdrawal symptoms varies according to the half-life of the opioid or benzodiazepine and the half-life of active metabolites, which may be much longer than that of the parent compound (Tobias, 2000). PALLIATIVE CARE Palliation, as defined in the New Oxford American Dictionary, is “making a disease (or its symptoms) less severe or unpleasant without removing the cause.” Most deaths in neonatal intensive care units occur after withdrawal of supportive measures, when the burden of care is extreme and the likelihood of recovery is remote (Cook and Watchko, 1996; Wall and Partridge, 1997). Traditionally, initiation of palliative care occurs when further medical interventions are no longer curative and death of the infant is expected. However, the AAP and the World Health Organization both support palliative care that starts early in the course of an illness and works with other therapies to prolong life. “The goal of palliative care is the achievement of the best quality of life for the patients and their families, consistent with their values, regardless of the location of the patient” (World Health Organization, 1998). Both the AAP (American Academy of Pediatrics Committee of Bioethics and Committee on Hospital Care, 2000) and the World Health Organization (1998) have provided guidelines for palliative care for infants that emphasize optimizing quality of life until death occurs and eliminating of the notion of euthanasia. Such care is an active process that includes familycentered care with attention to physical, emotional, and spiritual issues (Papadatou, 1997). Several good reviews of neonatal end-of-life care have been published (De Lisle-Porter and Podruchny, 2009; Institute of Medicine of the National Academies, 2003; Moro et al, 2006). Box 35-1 provides a comprehensive listing of the many aspects of end-of-life care that should be considered for the dying neonate. Parents and other family members should be actively involved in palliative care. Palliative treatments focus on relief from clinical symptoms such as dyspnea, pain, agitation, vomiting, and seizures, in addition to maintaining dignity and providing warmth. Palliative care should also include attention relieving psychological stress in caregivers such as loneliness, depression, anxiety, grief, and separation from family and loved ones. Family members should be prepared for the end of life. They need to be provided with a description of the dying process that includes mention of the potential for gasping respiratory efforts and an estimate of the length of the dying process. Opportunities to create memories for the family should be offered, such as pictures, footprints, locks of hair, hats, or blankets. Consideration of to the place of death is important. Should it occur in the hospital or at home with hospice care? Unfortunately, many deaths occur in busy NICUs with little privacy. Placement of the infant and the family on a postpartum ward with other parents and their healthy infants is also suboptimal. Ideally, for deaths likely to occur in the hospital, the infant and family should be placed in a private room with enough space for all family members to be present to support the parents in their grieving process. Bereavement and follow-up discussions should be planned with the family. Care providers should discuss autopsy, discuss follow-up discussion times, and provide ongoing bereavement support to the family for as long as needed. SUMMARY Recognition and treatment of pain and discomfort in the neonate remain challenging issues. Despite significant progress in the understanding of human neurodevelopment, pharmacology, and more careful attention to the care of sick infants, there is still have much to learn. Because protecting and comforting these patients are important, and because external regulatory forces have required intervention to minimize distress, what is known about adult patients must be applied to infants. Some good has come from these endeavors, but errors have been made along the way. It is important to minimize pain and distress in these patients to avoid more aggressive interventions. Such care will minimize the risks of adverse effects on neurodevelopment. Learning to provide good care without doing harm should be the goal. Nonpharmacologic methods of pain and distress control should be explored further. When pharmacologic intervention is necessary for pain control, use the least amount of drug that controls the pain. Escalation of drug doses may, in fact, be adding to the problem. As newer techniques and medications are introduced to clinical practice, it must be demonstrates that such additions achieve the goal of pain control. It must also be demonstrates that neonatal interventions are safe over the lives of the patients. Better tools are needed to help optimize the outcomes for infants. CHAPTER 35 Neonatal Pain and Stress: Assessment and Management 443 BOX 35-1 Guideline for End-of-Life Care of the Neonate I. Patient population Any infant in the NICU diagnosed with a medical condition incompatible with life. II. Purpose The guideline is for use by the bedside nurse in the NICU when faced with caring for an infant at the end of life. It provides the bedside nurse with alternative ways to encourage parents to perform normal parenting activities, such as bathing, holding, and picture taking with their infant, in the time surrounding his death. These activities may be the only, and sometimes last, opportunity the mother and father will have to parent their infant. III. Background and rationale Many NICUs have policies in place relating to bereavement and postmortem care of the neonate, but few have policies or guidelines on palliative care (Sudia-Robinson, 2003). Many nurses have not received sufficient training to support families or themselves through this bereavement process. Many nursing programs lack or have only recently added curricula regarding end-of-life care (Romesberg, 2004). This lack of protocols, along with the limited education available to professionals, serves to decrease the chance that infants and their families will receive the end-of-life care they deserve. Palliative care consists of three components: (1) pain and comfort management, (2) assistance with end-of-life decision making, and (3) bereavement support (March of Dimes, 2010). IV. Procedure A. Inclusion criteria: any infant in the NICU with a diagnosis of nonviability or a fatal condition should be considered for inclusion in this guideline. These medical conditions may fall into three categories: (1) newborns born at the edge of viability with extremely low birthweight (<500 g) and <24 weeks’ gestational, (2) newborns with complex or multiple congenital anomalies, and (3) newborns not responding to intensive care interventions and experiencing deterioration despite medical efforts to save them (Catlin and Carter, 2002). For these infants, the focus of care should change from curative to palliative. B. Predeath: any intervention should focus on supporting the parents during their infant’s end-of-life care. Some nurses may be uncomfortable with their own feelings and have trouble communicating with parents at this time. Communication, however, is vital, and the words spoken should serve to validate the infant’s life and death (Capitulo, 2005). Communication also helps the parents understand their infant’s condition and offers them the opportunity to make choices concerning the extent of their participation in his end-of-life care (Workman, 2001). 1. Communication a. The following are suggestions for successful communication techniques to use with parents: (1) Active listening (2) Open communication (3) Total presence (4) Uninterrupted time for parents to express their wishes (5) Nonverbal gestures such as touch b. Refer to the infant by name c. Inform parents of what to expect while their infant is dying (see, hear, smell, feel) d. Inform parents of their options for disposition of their infant’s remains (1) Transfer to morgue (2) Pick up by mortuary e. Use language that does not confuse the parents. (1) Use definite words such as death, die, and dying instead of euphemisms such as not doing well or passing away (2) Words such as good, stable, or better could be misinterpreted to mean the infant could improve or survive 2. Environment a. Do not restrict the number of visitors b. Provide privacy and comfort (1) Low lights (2) Decreased noise and activity 3. Religion a. Ask parents about any religious preference b. Notify clergy of the parents’ choice c. Offer baptism, blessing, anointing, or prayer 4. Parental activity a. Provide a stuffed animal b. Keep the family involved (1) Decisions (medical and physical care) (2) Handprints, footprints (3) Kangaroo care, holding (4) Bathing (5) Dressing (6) Changing diaper (7) Any other parenting acts (8) Taking pictures of the infant and family members at any time during this process C. Active dying: the time period when the infant is dying offers special challenges for both the nurse caring for the dying infant and his family members. Pain management for the infant during the extubation and dying process is of extreme importance. Parents must be informed of what they may see, hear, and experience during this time. The family and infant should be offered a comfortable, private area for this process. 1. Pain management a. Maintain intravenous access b. Medicate the infant appropriately c. Assess pain routinely and frequently after the withdrawal of life support 2. Communication a. Inform the parents what to expect (see, hear, smell, feel) b. Inform the parents that their baby may not die immediately after removal of the endotracheal tube c. Be available for any questions or concerns the parents may have 3. Environment a. Do not restrict the number of visitors b. Provide privacy and comfort (1) Low lights (2) Decreased noise and activity 4. Religion a. If ceremonies or rituals have not been performed before now, offer any of the following: (1) Clergy of choice (2) Blessing, baptism, anointing, orprayer b. Parental activity (1) Offer the opportunity to hold the baby during extubation (2) Encourage the parents to notify the nurse if they believe their baby is in pain and needs assistance Continued 444 PART VIII Care of the High-Risk Infant BOX 35-1 Guideline for End-of-Life Care of the Neonate—cont’d D. After death: once the infant has died, creating lasting memories is integral to the healing of the family. Regardless of whether the family participates in their infant’s death care, no steps should be eliminated. When an infant dies, there are few, if any, memories of their time together, and parents will cherish any memento of their infant’s life (Jansen, 2003). It is the nurse’s responsibility to help ensure that as many memories as possible are created (Capitulo, 2005). 1. Communication a. Offer the parents opportunities to be involved with all aspects of after-death care. b. If parents decline, respect their decision, but gently remind them that this will be their only chance to perform these tasks. c. If parents decline, offer them the opportunity to help the nurse with the infant’s care. d. Reassure the family that the nurse is available at any moment if assistance is required. 2. Environment a. A room with a bed and chairs for family and friends to be alone with their infant should be provided. b. Complimentary snacks and beverages should be made available. c. Family time with the infant should be unrushed and unlimited. 3. Parental activities a. Encourage the parents to hold and dress their infant. b. Offer the parents the opportunity to bathe their infant if this has not already been done. c. Offer the parents the opportunity to have family pictures taken at this time. 4. Memory makers: memory boxes should contain items that have meaning and have been a part of the infant’s hospitalization. Objects that seem trivial to others will have profound significance for the family. They will be the only concrete evidence to the parents and family that their infant was alive. The nurse needs to be creative. There is no second chance to create memories for the family (Catlin and Carter, 2002). a. Identification band b. Name card c. Blood pressure cuff d. Oxygen saturation probe e. Thermometer f. Respiratory paraphernalia g. Measuring tape with infant’s length marked and name and date written on it h. Polaroid or digital photographs pictures i. Stuffed animal j. Pacifier k. Diaper l. Lock of hair m. Handprints, footprints n. Swaddling blanket (do not discard blanket if soiled) o. Clothing (include postmortem gown, hat, socks, T-shirt; do not wash if soiled) p. Ring q. Mold of hand, foot, or both r. Cord clamp s. Phototherapy mask t. Dried washcloth and soap from bath, placed in a small plastic bag to preserve soap fragrance u. Black mourning (remembrance) pin 5. Final departure of parents: ensure that the parents do not leave with “empty arms” a. Memory box b. Stuffed animal c. Baby’s blanket d. Pictures e. Baby’s belongings V. Nursing responsibilities A. Prepare the infant’s body for transportation to the hospital morgue per hospital protocol. B. Complete the necessary paperwork. C. Ensure that all tasks are completed per hospital policy. Adapted from De Lisle-Porter M, Podruchny AM: The dying neonate: family-centered end-of-life care, Neonatal Netw 28:75-83, 2009. NICU, Neonatal intensive care unit. SUGGESTED READINGS Allegaert K, Veyckemans F, Tibboel D: Clinical practice: analgesia in neonates, Eur J Pediatr 168:765-770, 2009. American Academy of Pediatrics Committees on Fetus and Newborn and on Drugs, Sections on Anesthesiology and on Surgery; Canadian Paediatric Society Fetus and Newborn Committee: Prevention and management of pain and stress in the neonate, Pediatrics 105:454-461, 2000. Anand KJ, Anderson BJ, Holford NH, et al: Morphine pharmacokinetics and pharmacodynamics in preterm and term neonates: secondary results from the NEOPAIN trial, Br J Anaesth 101:680-689, 2008. Anand KJ, Hall RW, Desai N, et al: Effects of morphine analgesia in ventilated preterm neonates: primary outcomes from the NEOPAIN randomised trial, Lancet 363:1673-1682, 2004. Carbajal R, Lenclen R, Jugie M, et al: Morphine does not provide adequate analgesia for acute procedural pain among preterm neonates, Pediatrics 115:14941500, 2005. Carbajal R, Roussit A, Danan C, et al: Epidemiology and treatment of painful procedures in neonates in intensive care units, J Am Med Assoc 300:60-70, 2008. De Lisle-Porter M, Podruchny AM: The dying neonate; family-centered end-oflife care, Neonatal Netw 28:75-83, 2009. Durrmeyer X, Vutskits L, Anand KJ, et al: Use of analgesic and sedative drugs in the NICU: Integrating clinical trials and laboratory data, Pediatr Res 67:117127, 2010. Fitzgerald M: The development of nociceptive circuits, Nature Rev Neurosci 6:507520, 2005. Howard R, Carter B, Curry J, et al: Quick reference summary of recommendations and good practice points, Pediatr Anesth 18(Suppl 1):4-13, 2008b. Lago P, Garetti E, Merazzi D, et al: Guidelines for procedural pain in the newborn. For the Pain Study Group of the Italian Society of Neonatology, Acta Paediatr 98:932-939, 2009. Stevens BJ, Riddell RR, Oberlander TE, et al: Assessment of pain in neonates and infants, In Anand KJ, Stevens BJ, McGrath PJ, editors: Pain in neonates and infants, ed 3, Philadelphia, 2007, Elsevier, pp 67-90. Complete references used in this text can be found online at www.expertconsult.com P A R T I X Immunology and Infections C H A P T E R 36 Immunology of the Fetus and Newborn Calvin B. Williams, Eli M. Eisenstein, and F. Sessions Cole Understanding the contribution of the newborn infant’s immunologic response to neonatal disease requires a review of the complex immunologic environment of pregnancy and the developmentally regulated changes in fetal and neonatal immunity. The contrasting functions of the fetal, neonatal, and maternal immunologic responses (i.e., preservation of fetal well-being as an allogenic graft versus adequate immunologic protection in a nonsterile extrauterine environment) are regulated by a host of incompletely understood developmental and genetic mechanisms. The diversity and importance of these mechanisms are suggested by the heterogeneity and frequency of the infectious problems encountered in newborns. Differences in immunologic responsiveness between adults and newborns should not be considered defects or abnormalities. Just as the ductus arteriosus, a cardiopulmonary necessity in the intrauterine environment, closes at different rates in different infants, human fetal and newborn infant immunologic response mechanisms are developmentally and genetically programmed to change from graft preservation to identification and destruction of invading pathogens at different rates. MATERNAL AND PLACENTAL IMMUNOLOGY Although pregnancy is a natural process that is essential for the propagation of all mammalian species, from an immunologic perspective it poses a great dilemma. The fetus represents a hemiallogenic graft that expresses paternal antigens, which under ordinary circumstances would be rejected as foreign tissue. The fact that such rejection ordinarily does not occur cannot be explained by systemic suppression of the maternal immune system. Most studies of maternal immune function during pregnancy have not shown significant abnormalities (AagaardTillery et al, 2006) and are consistent with the clinical observation that pregnant women are not at increased risk for opportunistic infection. Furthermore, maternal acceptance of the fetus does not occur exclusively as a result of physical separation between immunogenic fetal and maternal tissues. Trophoblastic tissue expressing paternal antigens comes into direct contact with maternal immune cells, and fetal cells can readily be detected in maternal peripheral blood (Price et al, 1991). Conversely, maternal leukocytes and somatic cells are present in significant numbers in the fetus and may persist into adult life in a phenomenon known as microchimerism (Bianchi et al, 1996; Lo et al, 1996; Maloney et al, 1999). As a result, mechanisms must exist whereby antigen-specific responses against paternal antigens either are not initiated or are specifically suppressed. Beginning with the seminal studies of Peter Medawar more than 50 years ago, significant progress has been made in unraveling the immunologic mechanisms underlying maternal-fetal tolerance (Seavey and Mosmann, 2008; Trowsdale and Betz, 2006). Currently it appears that no single factor can explain the commensal immunologic relationship that exists between mother and fetus throughout pregnancy. Rather, several distinct but complementary innate and adaptive immune mechanisms contribute to this process. Before fertilization, both systemic and local immunologic changes occur in the endometrium to favor acceptance of the fetal allograft. For example, progesterone induced by ovulation can reduce the proliferation of T cells and inhibit antigen-presenting cells (Ehring et al, 1998; Miyaura and Iwata, 2002). In addition, as discussed later, T cells with immunosuppressive properties accumulate cyclically in the endometrium during normal menses. Semen has contrasting immunomodulatory effects on the female reproductive tract and immune system, respectively; however, seminal plasma induces expression of proinflammatory cytokines, including granulocyte-monocyte colony-stimulating factor (GM-CSF), interleukin (IL) 6, IL-8, and the chemokine monocyte chemotactic protein 1 by the uterine endometrium (Sharkey et al, 2007). These molecules help to induce a local inflammatory cell infiltrate containing macrophages, neutrophils, and dendritic cells, which may help to break down endometrial mucin and thereby facilitate adherence of the blastocyst to the endometrium (Dekel et al, 2010). GM-CSF also has embryotrophic effects that promote fetal growth and viability. Knockout mice lacking GM-CSF have normal implantation rates, but have 15% to 25% smaller litters (Robertson et al, 1999). However, semen contains immunosuppressive factors that teleologically are necessary to inhibit initiation of an immunologic reaction against sperm antigens. Therefore seminal plasma can induce secretion of the immunosuppressive cytokine IL-10 by monocytic cell lines (Denison et al, 1999) and is rich in transforming growth factor (TGF) β-family cytokines (Tremellen et al, 1998). TGF-β1 is essential for inducing local differentiation of certain forms of immunoregulatory T cells (Chen et al, 2003; Keskin et al, 2007). 445 446 PART IX Immunology and Infections Inflammatory mediators and cytokines produced by uterine tissues have a key role in embryonic implantation. Prostacyclin is expressed by uterine tissues by day 5 of gestation and is required for blastocyst adherence and subsequent placental decidualization (Lim et al, 1999). The IL-6 family of cytokines, which includes leukemia inhibitory factor (LIF), IL-11, and IL-6 itself, is also essential for placental implantation. This family of cytokines is defined by their receptors, which share a common signaling component, GP130, in addition to a cytokine-specific alpha chain. LIF is expressed by uterine endometrium, beginning during the mid-secretory phase of the menstrual cycle (Cullinan et al, 1996), probably in response to progesterone (Eijkemans et al, 2005). Implantation does not occur in LIF-deficient mice (Stewart, 1992). Interleukin-11, a cytokine that possesses hematopotic and antiinflammatory properties, has also been shown to be essential for implantation, albeit at a later stage than LIF. Placenta of mice deficient in the alpha chain of the IL-11 receptor fail to undergo decidualization, resulting in infertility (Robb et al, 1998). IL-6 itself is expressed by endometrial epithelial and stromal cells, mainly during the secretory phase of menses (Perrier d’Hauterive et al, 2004). Reduced expression of IL-6 mRNA has been noted in secretoryphase endometrium of women with recurrent miscarriages (Jasper et al, 2007). This finding was attributed to a possible role of IL-6 in facilitating placental decidualization and trophoblast development. In summary, an orchestrated series of specific proinflammatory and tolerogenic mechanisms support implantation and early fetal allograft survival. STRATEGIES FOR IMMUNOLOGIC EVASION AT THE MATERNAL-FETAL INTERFACE One possible way in which the hemiallogeneic fetus might avoid rejection is to evade detection by the maternal immune system. This approach clearly is not the entire explanation for tolerance of the fetal allograft. Indeed, fetal antigenic maturity in eliciting a maternal immunologic response has been well documented (Beer and Billingham, 1976; Billington, 1987). The view that maternal tolerance is an active process is further supported by observations that, at least in mice, maternal T lymphocytes specific for paternal antigens in particular are activated during pregnancy. In addition, the fact that paternally derived tumor cells are not rejected by pregnant female mice supports a central role for antigen-specific systemic immune inhibition (Tefuri et al, 1995); therefore, tolerance appears to be at least in part a metabolically active process. Nevertheless, several unique features of the maternal-fetal interface undoubtedly contribute to attenuation of antifetal immunity (Table 36-1). One such mechanism is altered expression of major histocompatibility complex (MHC) proteins. MHC class II molecules are necessary for CD4+ T cell responses, which contribute both effector and regulatory functions needed for graft rejection. MHC class II molecules are not ordinarily constitutively expressed, except by certain ­antigen-presenting cells; however, proinflammatory stimuli such as interferon (IFN)-γ are capable of inducing MHC class II expression on most cell types. In contrast, fetal trophoblastic tissues do not express MHC class II TABLE 36-1 Selected Molecules of Immunologic Significance at the Fetal-Maternal Interface Molecule MHC class II Expression Pattern Functional Consequence Not expressed by trophoblast, either constitutively or inducibly Impaired CD4+ T cell activiation MHC Class I Molecules HLA-A, HLA-B Not expressed by trophoblast Impaired CD8+ T cell activation HLA-C Expressed by trophoblast during 1st trimester Modulation of NK cell function HLA-G Expressed by trophoblast Inhibition of T cell and dendritic cell function; modulation of T cell and dendritic cell function Fas ligand (CD178) Expressed by placental decidual and glandular epithelium and by fetal trophoblast Induces leukocyte apoptosis PDL-1 Expressed on placental deciduas basalis Immune inhibition Cofactor for induction of regulatory T cells Galectins Expressed by trophoblast early in pregnancy Induction of T cell apoptosis HLA, Histocompatibility leukocyte antigen; MHC, major histocompatibility complex; NK, natural killer; PDL, programmed cell death ligand. proteins, either in the resting state or in response to IFN-γ (Peyman et al, 1992). MHC class II gene silencing in trophoblasts is at least partially the result of the transcriptional repression of class II transactivator, a necessary transcription factor for class II gene expression (Murphy and Tomasi, 1998; Peyman et al, 1992). More importantly, human trophoblasts also do not express conventional MHC class I human leukocyte antigen (HLA) A or HLA-B molecules. This fact undoubtedly contributes to reduced alloantigenic recognition at the fetal-maternal interface. Human trophoblasts express HLA-C, principally during the first trimester of pregnancy (King et al, 1996) and two nonclassic HLA molecules— HLA-E and HLA-G. The role of the latter in pregnancy has been studied extensively. HLA-G is a nonpolymorphic molecule encoded by a single gene that encodes seven alternatively spliced transcripts and exists as a protein in both membrane-bound and soluble forms. Its expression is primarily limited to fetal tissue, but it can be detected on some adult tissues, and its expression can be induced on other cells in response to infection, inflammation, or malignant transformation (Carosella et al, 2008). Unlike classical MHC molecules, HLA-G does not have a significant role in stimulating CD8+ T cells via the T cell receptor complex. Rather, the principal function of HLA-G molecules expressed by the trophoblast appears to be modulating the activity of natural killer (NK) cells. HLA-G has also been shown to have other immunomodulatory properties, including inhibition of cytotoxic T cell activity, inhibition of alloproliferative responses by CD4+ T cells, CHAPTER 36 Immunology of the Fetus and Newborn and modulation of dendritic cell maturation and function (Carosella et al, 2008). These data reveal that the unique MHC class I molecule expression pattern on fetal trophoblast constitutes an intricate mechanism for orchestrating the activity of immune cells. Several additional interesting molecules expressed at the maternal-fetal interface confer unique immunologic properties to this environment. Fas ligand (FasL) is expressed in both maternal and fetal components of the uteroplacental unit throughout gestation. Activated T cells express the Fas receptor, which delivers an apoptotic (death) signal when bound by FasL. Implicit in these observations is the hypothesis that expression of FasL limits the reciprocal migration of activated fetal and maternal T cells. This idea is strengthened further by experiments examining the homozygous matings between generalized lymphoproliferative disease (gld) mice, which make a nonfunctional FasL. Pregnant generalized lymphoproliferative disease (gld) mice demonstrate leukocyte infiltration and necrosis at the decidual-placental border, with many resorption sites and small litters (Hunt et al, 1997). ­Progesterone-induced blocking factor is an immunomodulatory molecule released in response to progesterone by trophoblastic cells (Anderle et al, 2008). Its properties include indirectly suppressing NK cell function and inducing bias of CD4+ T cells toward Th2-type cytokine secretion (Szekeres-Bartho and Wegmann, 1996). The galectins represent a family of molecules expressed at the maternal-fetal interface with immunoregulatory properties. These glycoproteins regulate the immune response by recognition of cell surface glycans on immune cells. Galectins are expressed in human placenta primarily by the syntrophoblast early in pregnancy (Than et al, 2009). Upon cell surface contact, galectins downregulate the cellular immune response, in part by inducing programmed cell death (apoptosis) of T lymphocytes (Liu and Rabinovich, 2010). A series of murine investigations points to a role for local metabolic factors in suppressing maternal T cell alloresponses in preventing spontaneous fetal loss. When monocytes are induced to differentiate into macrophages in vitro using macrophage colony-stimulating factor, they become inhibitors of T cell proliferation rather than activators. This inhibition was shown to be the result of selective degradation of tryptophan by the inducible enzyme indoleamine 2,3-dioxygenase (IDO). Serum tryptophan levels fall during pregnancy, possibly in response to IDO expression by syncytiotrophoblast cells, which make the enzyme as early as 7.5 days after coitus. Given the localization of IDO to the maternal-fetal interface and the immunosuppressive effects of tryptophan depletion, it seemed plausible to investigators that this pathway had a role in survival of the fetal allograft. This hypothesis was supported by experime nts with either normal or recombinase activating gene knockout mice, which lack T and B cells, together with the IDO inhibitor 1-methyl tryptophan. Normal mice carrying allogenic fetuses lose all their concepti by 11.5 days after coitus when they are treated with the inhibitor. In contrast, the delivery of healthy litters in RAG-deficient mice treated with 1-methyl-tryptophan demonstrates an immunologic basis for the observation (Munn et al, 1998). More 447 recently, an additional tolerogenic mechanism influenced by tryptophan metabolism has been suggested by studies of the indirect influence of the IDO pathway on differentiation and activation of regulatory T lymphocytes (Baban et al, 2009; Chen et al, 2008). However, the lack of fetal rejection observed in female, genetically deficient IDO mice suggests that IDO may be necessary, but not sufficient, in maintaining tolerance during pregnancy (Baban et al, 2004). Another possible mechanism through which the maternal immune system might be rendered relatively indifferent to the presence of fetal alloantigens concerns antigen presentation within the uterine environment. Dendritic cells (DCs) are a type of antigen-presenting cell that is critical for cellular and humoral immune responses. Collins et al (2009) recently reported that the migration of DCs from decidual tissue to draining uterine lymph nodes is impaired compared with myometrial DCs. Therefore placental entrapment of DCs might represent a way by which presentation of fetal antigens to the maternal immune system is retarded, thereby potentially inhibiting initiation of allogenic immune responses. NK cells found in the placental decidua are referred to as decidual NK cells (dNK cells) to distinguish them from NK cells found in maternal blood and those found in the nonpregnant uterine endometrium (Yagel, 2009). These cells comprise 70% of decidual lymphocytes, express a CD56+CD16– phenotype (Manaster et al, 2008), and have a unique gene expression profile which indicates that dNK cells are a unique cell population (Koopman et al, 2003). Furthermore, despite expressing high amounts of granzyme and other molecules necessary for mounting a cytotoxic response, these cells exhibit reduced cytotoxicity compared with peripheral blood NK cells (Kopcow et al, 2005) and were noncytotoxic to trophoblastic cells. In contrast to NK cells found in peripheral blood and tissues that have important innate immune functions, including surveillance against infection and tumor cells, it is not certain whether dNK cells have any role in immunity during pregnancy. Rather, they have been shown to have an essential role in promoting trophoblast invasion into the decidua and placental vascularization. These cells accomplish these tasks by secreting cytokines with effects on the placental vasculature including IFN-γ, chemokines that affect trophoblastic growth, and angiogenic factors including vascular endothelial growth factor, placental growth factor, angiopoietin-2, and NKG5 (Hanna et al, 2006). To facilitate activation of dNK cells to perform their essential tissue remodeling functions, while at the same time avoiding overactivation that could possibly lead to cytotoxic damage to the trophoblast, their activity must be tightly modulated by a series of interactions between several activating and inhibitory receptors and their respective ligands (Manaster and Mandelboim, 2010). The nonclassic class I MHC molecules are among these dNK cell ­receptor ligands. For example, a principal physiologic ligand of β2-microglobulin–associated HLA-G is the inhibitory receptor leukocyte immunoglobulin (Ig)-like receptor 1 (Ponte et al, 1999). Free heavy chain HLA-G also binds to this receptor in a noninhibitory fashion, and this interaction could represent a feedback mechanism to avoid overinhibition (Gonen-Gross, 2010). 448 PART IX Immunology and Infections HLA-C is a major ligand for the different killer immunoglobulin-like receptors (KIRs), some of which are activating and others which are inhibitory. Amino acid allotypes at position 80 of HLA-C confer preferential binding to different forms of KIR, which in turn may result in clinically significant differences in dNK cell activation. Various combinations of maternal KIR and fetal HLA-C genes have been shown to confer differential risk for preeclampsia, as a result of differences in dNK cell-mediated placental vascular changes (Hiby et al, 2004). A final innate immune condition necessary for successful completion of pregnancy is the absence of placental activation of complement. Fetal death occurs in mice that are genetically deficient in the complement regulatory protein Crry, accompanied by placental complement deposition and inflammation (Xu et al, 2000). Although there is no direct homolog of Crry in humans, complement inhibition has been shown to be essential for normal pregnancy in a murine model of the antiphospholipid syndrome, an autoimmune condition characterized by thrombosis, thrombocytopenia, and recurrent fetal loss. In this model, fetal injury occurs as a result of placental inflammation initiated by local dysregulation of complement proteins. Both complement activation and fetal loss can be prevented administering anticoagulants with complement-inhibitory properties such as heparin, but not by anticoagulants lacking complement-binding properties (Girardi et al, 2003, 2004). Some but not all human clinical interventional studies using anticoagulants with complement-binding properties to prevent fetal loss in the antiphospholipid syndrome have shown positive results (Di Nisio et al, 2005). CYTOKINE SECRETION BY T CELLS DURING PREGNANCY Bias in T cell cytokine secretion toward Th2-type cytokines and away from Th1-type cytokines has been proposed as an immunologic condition necessary for maintaining healthy pregnancy. Excessive exposure to Th1-type cytokines is clearly detrimental to the fetus. The Th1 cytokines IFN-γ and tumor necrosis factor-α have been shown to inhibit the growth of trophoblasts and to inhibit embryonic and fetal development (Haimovici et al, 1991). Th1 cytokines terminate normal pregnancy when injected into pregnant mice (Chaouat et al, 1990). Th2-type cytokines contribute to implantation as discussed previously, and to maintenance of pregnancy as suggested by LIF expression throughout pregnancy by placental leukocytes, decidua, and chorionic villous cells (Sharkey et al, 1999). LIF production by maternal T cells from healthy pregnancies is also positively associated with IL-4 production and inversely with IFN-γ (Piccinni et al, 1998). These studies have been interpreted to support the notion that the balance between Th1 and Th2 cytokines is important in maintaining pregnancy. However, it has been indicated that mice lacking all Th2-type cytokines (deleted by multiple genetic mutations) undergo normal allogenic pregnancies with normal litter sizes (Fallon et al, 2002). Based on this result, some investigators have concluded that Th1/Th2 cytokine shifts is an epiphenomenon that is irrelevant to maternal-fetal tolerance (Trowsdale and Betz, 2006). ROLE OF REGULATORY T CELLS IN PREGNANCY The mechanism whereby adaptive allogenic responses against paternal antigens are actively suppressed during pregnancy was largely unexplained until the discovery of a distinct lineage of T lymphocytes with dominant immunosuppressive properties (Brunkow et al, 2001). This type of regulatory T (Treg) cell is characterized by expression of a lineage-specific transcription factor, Foxp3. Upon activation via their antigen-specific T cell receptor, Treg cells are capable of suppressing immune effector cells, including dendritic cells and T effector cells through a variety of mechanisms (Tang and Bluestone, 2008), thereby preventing a fatal form of autoimmune disease throughout life (Kim et al, 2007). The ontogeny of Treg cells is discussed more fully in the section on adaptive immunity. The possible role of Treg cells during pregnancy was first investigated in mice. These experiments disclosed that Treg cell numbers are expanded in peripheral lymphoid tissues, and this expansion is not dependent on the presence of alloantigen. Particularly high numbers of Treg cells are found at the maternal-fetal interface, where they comprise 30% of all CD4+ T cells. Treg cells are not required for pregnancy in syngeneic mice, in which the fetus does not display any foreign paternal antigens. In contrast, in allogenic pregnancy in which the fetus expresses paternal antigens, Treg cell depletion resulted in pregnancy failure (Aluvihare et al, 2004). In another series of experiments examining an abortion-prone strain of mice, allogeneic pregnancy loss could be prevented by adoptive transfer of Treg cells from healthy mice (Zenclussen et al, 2005). These data provide a plausible mechanism whereby the adaptive immune system can actively suppress cellular immune responses directed against paternal alloantigens. Another line of evidence from studies performed in mice provides additional indirect support for an essential role of Treg cells in maintaining fetal tolerance in mice. Programmed death-1 (PD-1) is an inhibitory receptor expressed on activated lymphocytes. There are two programmed death ligands (PDLs) for the PD-1 receptor, PDL-1 and PDL-2. PDL-1 is expressed on the decidua basalis, the portion of the placenta abutting the fetal trophoblast, and by Treg cells. Blockade of PDL-1 using antibodies in allogenic pregnancy results in T cell–mediated fetal loss, indicating an essential role for this molecule in preventing rejection of the fetal allograft (Wang et al, 2008). This effect of anti–PDL-1 antibody is abrogated when Treg cells are depleted. Furthermore, in PDL-1–deficient mice, suppressive activity of Treg cells was found to be impaired (Habicht et al, 2007). Finally, an independent series of experiments has demonstrated an important role for PD-1–PDL-1 interaction in Treg cells (Krupnick et al, 2005). These data support a critical role for PD-1–PDL-1 interactions in generating functional Treg cell populations necessary for successful allogenic pregnancy. Studies of Treg cells in humans further support an essential role for this cell type in initiating and maintaining healthy pregnancy. Treg cells are increased in the peripheral blood of women during the late follicular and luteal phases of the menstrual cycle. This increase is believed to occur in response to estradiol (Polanczyk et al, 2004; Tai et al, 2008). Soluble factors in seminal fluid may contribute to further CHAPTER 36 Immunology of the Fetus and Newborn local expansion of these cells at the time of copulation (Robertson et al, 2009). Human Treg cells selectively migrate to the placental decidua under the influence of soluble factors including human chorionic gonadotropin (Schumacher et al, 2009). Most but not all studies have shown that Treg cell numbers are increased in peripheral blood throughout pregnancy (Mjosberg et al, 2009; Somerset et al, 2004). Alterations in Treg cells have been associated with certain pathologic conditions during pregnancy. Although preeclampsia does not appear to be associated with reduced numbers of Treg cells in peripheral blood, the ratio of Treg cells relative to CD4+ T cells bearing proinflammatory phenotypic markers, such as Th17 cells, is abnormal in preeclamptic women (Santner-Nanan et al, 2009). In addition, toxoplasma infection in pregnant mice is associated with reduced numbers of Treg cells in peripheral lymphoid tissues (Ge et al, 2008). This observation raises the possibility that embryopathy associated with prenatal infection is caused in part by the maternal immune system, because of the breakdown of maternal-fetal tolerance as a result of altered Treg cell differentiation. A fascinating permutation of the role of Treg cells during pregnancy concerns the effects of these cells on the developing fetal immune system. As already noted, fetal lymphoid tissues are microchimeric—that is, they contain small but significant numbers of maternally derived cells. Because not all maternal antigens are inherited, it would be expected that the fetal immune system would reject these cells. Recently it was demonstrated that maternal cells expressing noninherited maternal antigens (NIMAs) are tolerated by the fetal immune system as a result of active suppression by fetal Treg cells (Mold et al, 2008). In this study, Treg cell numbers were not found to be increased in the fetal thymus, but were markedly increased in number in fetal lymphoid tissues. It is therefore likely that these cells represent an induced rather than a natural Treg cell population, which differentiates in peripheral lymphoid tissues under the influence of TGF-β. It was also shown in this study that small numbers of these NIMA-specific Treg cells persist into childhood, but can be reactivated using appropriate stimuli. This property of Treg cells has important clinical implications in the transplantation setting. Bone marrow and solid organ transplants are known to be more successful when either the donor or recipient expresses NIMA (Burlingham, 2009; Dutta et al, 2009; van Rood et al, 2009). This finding is explained by restimulation of dormant, NIMA-specific Treg cells by transplanted cells expressing these antigens, thereby contributing to active tolerance of the graft. In addition to their potential relevance for developing future clinical strategies for improving transplantation outcomes while reducing the need for pharmacologic immune suppression (Burlingham, 2009), these findings also show that some of what is learned by the fetal immune system in utero is retained into childhood, and perhaps throughout life. DEVELOPMENTAL FETAL-NEONATAL IMMUNOLOGY The newborn infant, especially the preterm infant, is at increased risk for developing a considerable spectrum of opportunistic infections, including Candida spp., herpes simplex, and cytomegalovirus. Developmental and 449 genetic differences between adults and infants in immunologic responsiveness account for much of this enhanced susceptibility (Adkins et al, 2004; Levy, 2007), although under certain circumstances neonates can mount adultlevel responses (Marchant et al, 2003; Sarzotti et al, 1996). Research has focused on the molecular, cellular, and functional definitions of these differences. This discussion focuses on those developmental aspects of innate and adaptive immunity known to be important for fetal or neonatal responsiveness to infection. INNATE IMMUNITY During fetal adaptation from the sterile intrauterine environment to neonatal colonization of mucosal surfaces, the innate immune system shields the newborn from infection while helping to orchestrate the acquisition of protective adaptive immune responses (Levy, 2007). These innate mechanisms include protective barriers like the vernix caseoa, which contains antimicrobial peptides and fatty acids (Tollin et al, 2005), developmentally controlled regulation of toll-like receptor signaling (Fusunyan et al, 2001; Lotz et al, 2006), expression of acute phase reactants (Jokic et al, 2000; Levy 2007) and complement proteins, and alterations in neutrophil and monocyte function (Forster-Waldl et al, 2005; Levy et al, 2004, 2006a, 2006b). Importantly, functional maturation of innate immunity allows for colonization with commensal organisms while reducing potentially dangerous inflammatory responses. Complement The complement system, a principal component of the innate immune response, consists of more than 40 plasma, cell surface, and regulatory proteins that interact dynamically to regulate multiple physiologic functions, including resistance to pyogenic infections, interaction between adaptive and innate immunity, and elimination of immune complexes, products of inflammatory injury, and apoptotic self cells (Carroll, 2004; Walport, 2001a; Zipfel and Skerka, 2009). Components of the complement system recognize and lyse bacteria, opsonize microorganisms nonspecifically, release anaphylatoxins, solubilize immune complexes, and induce B‑cell proliferation and differentiation. Activation of the complement cascade can occur via three pathways—classic, lectin, or alternative. The activation steps in these pathways have been reviewed recently (Thiel, 2007; Zipfel and Skerka, 2009). Several characteristics of the complement cascade are important for the fetal–neonatal immunologic response. First, whereas the specificity of classic pathway activation results from interaction of antigens with antibodies of several isotypes, activation of the alternative and lectin pathways is antibody-independent and can be initiated by structures such as endotoxin and polysaccharides frequently expressed by pathogenic organisms. For the fetus or infant who has not yet produced antigen‑specific IgG for immunologic recognition, the alternative and lectin pathways may be critical for triggering the effector functions of the complement cascade (Kielgast et al, 2003; Simister, 2003; Stossel et al, 1973; Zilow et al, 1997). Second, the enzymatic activation 450 PART IX Immunology and Infections of the complement cascade permits rapid amplification of its functions: deposition of a single immunoglobulin molecule or C3b fragment can generate enzymatic cleavage of thousands of later‑acting components and thus multiple complement activities (Carroll, 2004; Walport, 2001a, 2001b; Zipfel and Skerka, 2009). In addition, the alternative pathway can be amplified via a positive feedback activation mechanism, because C3b, an activation product of the alternative pathway C3 convertase, is a component of this convertase (Janssen et al, 2006). Because of the importance of antibody‑independent recognition for the immunologic responsiveness of the fetus and infant, the positive amplification loop of the alternative pathway is critical for rapid generation of complement effector functions without specific immunologic recognition. Third, the continuous activation of the alternative pathway requires rigorous regulation in the fetus to avoid tissue damage during organ remodeling (Zipfel et al, 2007). Finally, the contributions of the lectin pathway to fetal–neonatal complement activation and fetal well-being are still under investigation. Complement activation via any of the three pathways occurs in four main steps: initiation of complement activation, C3 convertase activation and amplification, C5 convertase activation, and terminal pathway activity or the assembly of the membrane attack complex (Zipfel and Skerka, 2009) (Figure 36-1). Although the classic and lectin pathways may be triggered by recognition of ligands via complement activating soluble pattern recognition molecules (e.g., IgG or IgM antibodies, mannan-binding lectin [MBL], or the ficolins), the alternative pathway is spontaneously and constantly activated on biological surfaces in plasma and in most or all other body fluids (Pangburn and Muller-Ederhard, 1984). The spontaneous activation of the alternative pathway readily initiates amplification and requires rigorous regulation (Zipfel et al, 2007). Complement activation results in assembly of the first enzyme in the cascade, C3 convertase, via early‑acting components of the classical (C1, C4, and C2), alternative (factor B, factor D, and C3), or lectin (MBL, L-ficolin, M-ficolin, H-ficolin, MBL-associated serine proteases [MASP-1, -2, -3]), and a smaller MBL-associated protein (MAp19) pathway. C3 convertase is assembled via highly specific and limited proteolysis. Proteolytically activated components form specific enzymatic complexes composed of classical (C2b and C4b), alternative (C3bBb), or lectin (C2b and C4b) pathway components, which proteolytically cleave C3 into C3a and C3b (see Figure 36-1). These two endopeptidases (C3bBb and C4bC2b) have identical substrate specificities; each cleaves the single peptide bond Argine77–Serine78 of the alpha chain of C3 (Janssen et al, 2006). The rates of formation and dissociation of both C3 convertases are regulated by multiple soluble (e.g., factor H, factor I, C4b‑binding protein) and membrane‑associated proteins (e.g., membrane cofactor protein, delay accelerating factor) (Zipfel and Skerka, 2009). During the first phase of complement activation, small (8 to 10 kD) peptides are released by proteolytic cleavage from the second (C2a), third (C3a), and fourth (C4a) components of complement. These fragments have anaphylatoxin and antimicrobial activities and can establish chemotactic gradients for effector cells (Nordahl et al, 2004). If C3b is deposited close to the site of generation and surface-bound convertases are formed, complement cascade activation can be amplified (Gros et al, 2008). If C3b fragments coat microbial or apoptotic surfaces, they opsonize the particles for phagocytosis. Although C3b deposition on the surface membrane of intact self cells can be prevented by regulators that block further complement activation, C3b deposition on microbial surfaces or on modified self cells that lack such regulators amplifies complement activation (Zipfel and Skerka, 2009). The role of non–inflammation-inducing, C3b-mediated removal of Alternative pathway Classical pathway C3 C1 Initiation C3 Lectin pathway MASP C4 and C2 C3 convertase C3bBb C3a C5 convertase Inflammation C5a Terminal pathway C4bC2b C3b C3bBbC3b Opsonization C4bC2bC3b C5 C5b C5b–C9 Terminal complement complex (lysis) FIGURE 36-1 Summary of complement activation via the classic, alternative, and lectin pathways. (From Zipfel PF, Skerka C: Complement regulators and inhibitory proteins, Nat Rev Immunol 9:729-740, 2009.) CHAPTER 36 Immunology of the Fetus and Newborn cells during fetal organ remodeling has not been explored in detail. If complement pathway activation continues beyond C3 cleavage, the binding of a second C3b molecule to either C3 convertase generates the alternative (C3bBbC3b) or classical or lectin (C4bC2bC3b) C5 convertases (see Figure 36-1). Both enzymes proteolytically cleave C5 to C5a and C5b. C5a is a powerful anaphylatoxin, and the C5b fragment can initiate activation of the terminal pathway (Ward, 2009). The C5a-induced inflammation and the C5b-initiated terminal pathway seem to be separately regulated (Zipfel and Skerka, 2009). Binding of C5b to either of the active C5 convertases initiates directed, nonenzymatic assembly of the terminal pathway components C5, C6, C7, C8, and C9 to form the membrane attack complex (see Figure 36-1). The assembly and conformational changes of these soluble, hydrophilic proteins generate pores composed of lipophilic, membrane-inserting complexes (the membrane attack complex) that ultimately cause cell lysis by disrupting the osmotic gradient between the intracellular and extracellular environments (Morgan, 1999). Studies of fetal and neonatal complement have focused on quantification of serum concentrations of individual components, examining maternal–fetal transport of these proteins, assessing specific effector functions of the classical and alternative pathways, and investigating contributions of complement activation to common neonatal diseases. In humans, Gitlin and Biasucci (1969) reported detectable concentrations of C3 (1% of adult levels) and C1 inhibitor (20% of adult levels) by immunochemical methods as early as 5 to 6 weeks’ gestation. By 26 to 28 weeks’ gestation, both C3 and C1 inhibitor concentrations increased to 66% of adult levels. Since these studies, multiple investigators have demonstrated that functionally and immunochemically measured classical and alternative pathway protein concentrations in cord blood increase with advancing gestational age and that they are only 50% to 75% of adult concentrations at full‑term gestation (Sonntag et al, 1998; Wolach et al, 1997). Although cord blood lectin pathway component concentrations are lower than those in older children and adults, the correlation between MBL and gestational age has not been consistently observed (Hilgendorff et al, 2005; Kielgast et al, 2003; Kilpatrick et al, 1996; Lau et al, 1995; Swierzko et al, 2009; Thiel et al, 1995). The important roles of complement regulatory proteins, decay‑accelerating factor, membrane cofactor protein, and CD59 have also prompted examination of the ontogeny of these proteins in the human fetus (Simpson et al, 1993). On the basis of studies of genetically determined, structurally distinct complement variants in maternal and cord serum, no transplacental passage from mother to fetus of C3, C4, factor B, or C6 has been observed (Colten et al, 1981; Propp and Alper, 1968). The presence of detectable amounts of C2 and C1 inhibitor in cord blood, but not in the sera of mothers with genetic deficiencies of these proteins, suggests that fetal–maternal transport of these components does not occur. Regulation of complement effector functions in the fetus and newborn infant has not been as extensively examined. Opsonization of invading microorganisms without 451 specific immunoglobulin recognition requires activation of the alternative or lectin pathways. For infants born prematurely or without organism‑specific maternal IgG, alternative or lectin pathway activation provides a critical mechanism for triggering complement effector functions (Maruvada et al, 2008; Super et al, 1989; Swierzko et al, 2009). For example, Stossel et al (1973) demonstrated opsonic deficiency in 6 of 40 cord sera examined because of decreased factor B concentrations, despite normal C3 and IgG levels. The functional contribution of the classic pathway to neonatal effector functions has been assessed through the use of cord blood–mediated opsonophagocytosis by adult polymorphonuclear leukocytes of group B streptococci type Ia (Edwards et al, 1983). This serotype may be opsonized by classical pathway components in the absence of specific antibodies and thus permits evaluation of the function of classical pathway activation. In 8 of 20 neonatal sera examined, decreased bactericidal activity was detected and correlated with significantly lower functional activity of C1q and C4. These studies did not determine whether this decrease was mediated by an inhibitor of function or by an intrinsic change in functional activity of these components in neonatal sera. The contribution of the lectin pathway to immunosusceptibility of the newborn infant has been suggested by studies of both MBL concentrations and pathway activity (Kielgast et al, 2003; Kilpatrick et al, 1996; Sumiya et al, 1991; Super et al, 1989; Swierzko et al, 2009; Thiel et al, 1995). Complement regulatory proteins (e.g., C4b-binding protein and factor H) also contribute to neonatal immunosusceptibility as suggested by the failure of neonatal serum to stop invasion by group B streptococci and Escherichia coli into human brain microvascular endothelial cells (Maruvada et al, 2008). The importance of the terminal complement component C9 for cytolysis of multiple isolates of E. coli was suggested by in vitro experiments in which killing of E. coli by neonatal serum samples was limited by C9, but not by other classical pathway components (Lassiter et al, 1992, 1994). Although lower serum concentrations of classical, alternative, and lectin pathway complement proteins can contribute to enhanced susceptibility of infants to systemic infection, other complement functions important for fetal and neonatal well‑being, but not related to antimicrobial response, can contribute to reduced capacity to activate the classical and alternative pathways. For example, reduced serum concentration of C4b‑binding protein (8% to 35% of pooled adult plasma levels), which is a critical regulator of classical pathway C3 convertase activity, has been noted in fetal and neonatal sera (Fernandez et al, 1989; Malm et al, 1988; Melissari et al, 1988; Moalic et al, 1988). Lower C4b‑binding protein concentration increases the functional anticoagulant activity of protein S, with which it complexes and thereby contributes to decreased coagulation function of the fetus and newborn. Consideration of functions besides immunologic effector functions may be important in furthering the current understanding of the developmental regulation of complement component production. The contribution of complement activation to tissue injury has been investigated in several common neonatal diseases, including neonatal hypoxic ischemic encephalopathy, necrotizing enterocolitis, meconium aspiration 452 PART IX Immunology and Infections syndrome, and intrauterine growth restriction and fetal loss (Girardi and Salmon, 2003; Girardi et al, 2006; Lassiter, 2004; Mollnes et al, 2008; Schlapbach et al, 2008; Schultz et al, 2005). Concern has also been raised that unregulated complement activation can occur in selected infants who undergo extracorporeal membrane oxygenation therapy (Johnson, 1994; Kozik and Tweddell, 2006). Complement activation is an important regulator of multiple functions of the host immunologic response. Further study of the fetus and newborn infant will be aimed at understanding the developmental and genetic regulation of immunologic and nonimmunologic functions of this important group of plasma and cell surface proteins. Natural Killer Cells NK cells are a component of the innate immune response and represent approximately 10% to 15% of all peripheral blood lymphocytes. NK cells are present in the spleen, lungs, and liver, but rarely are found in lymph nodes and thoracic duct lymph (Cerwenka and Lanier, 2001). Interestingly, NK cells are the major type of lymphocyte found in the maternal decidual tissue, where they represent up to 70% of all lymphocytes (King et al, 1996). NK cells are distinguished from other lymphocytes by their morphology, function, and expression of distinct surface molecules. Expression of the cell surface markers CD16 (FγRIII) and CD56 (nerve cell adhesion molecule-1) can be used to identify the NK population by analytical flow cytometry. Mature NK cells appear larger and more granular than T or B cells (Cooper et al, 2001) Mature NK cells are also distinguished by the presence of both activating and inhibitory receptors that are used to selectively identify and kill virally infected cells and tumors (Biassoni et al, 2001). NK receptors recognize MHC class I molecules on target cells, resulting in signals that suppress NK cell function. Target cells that are deficient in or lack MHC class I activate NK cell function, resulting in the release of lysosomal granules. These granules contain serine proteases, perforin, and TGF-β, which disrupt the target cell membrane and induce an inflammatory response. Studies show significantly lower NK cell activity in fetuses and neonates compared with adults (Georgeson et al, 2001; Kadowaki et al, 2001). NK cells are derived from a common hematopoietic progenitor that retains T and B cell developmental potential (Boos et al, 2008). NK cells first make their appearance in fetal liver as early as 6 weeks’ gestation. Committed CD34+CD56– NK progenitors have been identified in the fetal thymus, bone marrow, and liver. In humans, there are several subsets of NK cells with distinct cell surface markers and function. For example, the CD56high CD16– subset is activated by IL-2 and also expresses CCR7 and CD62L, which allows these cells to traffic through lymph nodes (Di Santo, 2006). This subset is similar to NK cells that develop in the mouse thymus through an IL-7- and GATA3-dependent pathway and, in both species, is characterized by low cytotoxicity and enhanced cytokine production (Boos et al, 2008; Vosshenrich et al, 2006). In the human neonate, the NK population is immature; only half of all NK cells express CD56, and the NK cytolytic activity is lower (Dominguez et al, 1993). This functional reduction in NK activity has been proposed as a factor that contributes to the severity of neonatal herpes simplex virus infections. Profound defects in NK cell activity result in familial hemophagocytic lymphohistiocytosis, a disease characterized by fever, hepatosplenomegaly, cytopenia, hyperferritinemia, and hemophagocytosis. Familial hemophagocytic lymphohistiocytosis arises from mutations in genes that encode proteins involved in the granule-­exocytosis pathway and is a fatal disorder without bone marrow transplantation (Jordan and Filipovich, 2008; Orange, 2006). NK cell receptors are fundamentally different from the T cell receptor (TCR) and B cell receptor (BCR). NK receptor gene expression does not require gene segment rearrangement, and the receptors are not clonally distributed. Instead, NK cells use an array of stimulatory and inhibitory receptors to regulate their cytolytic functions (Lanier, 2008). A cluster of 10 or more genes encoding KIRs was located on human chromosome 19q13.4 (Biassoni et al, 2001; Lanier, 1998). Each of these type I glycoproteins recognizes a different allelic group of HLAA–, HLA-B–, HLA-C–, or HLA-G– encoded proteins, and each KIR is expressed by only a subset of NK cells. Another family of immunoglobulin-like receptor genes termed ILT is present near the KIR locus at 19q13.3. These receptors are not as restricted as the KIRs and bind multiple HLA class I molecules. A third inhibitory receptor gene locus has been identified on chromosome 12p12-p13. These genes encode a C-type lectin inhibitory heterodimeric receptor called CD94/NKG2 that binds HLA-E. Importantly, those KIRs, ILT receptors, and CD94/NGK2 molecules with long cytoplasmic tails and two immunoreceptor tyrosine-based inhibitory motifs (ITIMs) function as inhibitory receptors. Upon phosphorylation, the two ITIMs recruit and activate the Src homology domain 2 (SH2)-containing phosphatases, which turn off the kinase-driven activation cascade (Ravetch and Lanier, 2000). The KIR family member KIR2DL4 is distinct from other KIRs in structure and distribution. KIR2DL4 binds HLA-G, has a single ITIM in the cytoplasmic tail and a lysine in the transmembrane region, which allows association with adaptor proteins. This inhibitory receptor was found on all dNK cells in the placenta at term, but not on circulating maternal NK cells. This observation suggests that expression of KIR2DL4 is induced during pregnancy (Rajagopalan and Long, 1999). Other KIR or members of the C-type lectin superfamily serve as activating receptors (Moretta et al, 2001). These receptors lack the long cytoplasmic tail of the inhibitory receptors and therefore do not contain ITIMs. Instead, they have a charged amino acid in the transmembrane region that allows the receptor to associate with the adaptor molecule DAP12 (Lanier et al, 1998). This adaptor contains an immunoreceptor tyrosine-based activation motif (ITAM) that allows these receptors to activate NK cells. The physiologic role of these HLA class I–specific activating receptors remains unknown. Another group of NK cell–activating receptors has been identified and termed natural cytotoxicity receptors (Moretta et al, 2001). These proteins (NKp46, NKp30, NKp44) are immunoglobulin superfamily members with little similarity to one another or to other NK receptors. They are highly specific for NK cells and appear to interact with non-HLA molecules. CHAPTER 36 Immunology of the Fetus and Newborn 453 CD244 (2B4) is a member of the signaling lymphocyte activation molecule family of receptors expressed on all human NK cells (Ma, 2007). Upon interaction with the ligand CD48 on target cells, NK signaling proceeds via interactions between the ITSM (switch motif) in the cytoplasmic tail of CD244 and one of two SH2 domaincontaining adaptor proteins, SAP and EAT-2. SAP interactions result in activation, as evidenced in humans with X-linked lymphoproliferative disease, which is caused by loss-of-function mutations in the SAP linker. In the absence of SAP, interactions with EAT-2 may be inhibitory (Lanier, 2008). infants and systemic neutropenia motivated several investigators to give neutrophil replacement therapy to neutropenic, infected infants (Christensen et al, 1980). Although this approach has been successful in some cases, the results have not been uniformly beneficial (Cairo, 1987; Cairo et al, 1984; Menitove and Abrams, 1987; Stegagno et al, 1985). Similarly, the use of granulocyte CSF or GM-CSF to treat infants with suspected infection has generally not shown a reduction in mortality (Carr et al, 2003). This heterogeneity emphasizes the importance of individualizing immunologic interventions for the developmental stage of the infant and the invading microorganism being treated. Polymorphonuclear Neutrophils Monocytes and Macrophages As observed for T and B lymphocytes, neonatal polymorphonuclear neutrophils (PMNs) are present at early stages of gestation, but their functional capacities are different from those of adult PMNs. Progenitor cells that are committed to maturation along granulocyte or macrophage cell lineages (granulocyte-macrophage colony‑forming units) are detectable in the human fetal liver between 6 and 12 weeks’ gestation in proportions comparable to those observed in adult bone marrow (Christensen, 1989). Human fetal blood has detectable granulocyte-macrophage colony‑forming units from 12 weeks’ gestation to term (Christensen, 1989; Liang et al, 1988). Although these progenitor cells are detectable in the fetus and newborn infant, developmental differences between adult and mature neonatal PMNs have been demonstrated—in signal transduction, cell surface protein expression, cytoskeletal rigidity, rolling adhesion, microfilament contraction, transmigration oxygen metabolism, intracellular antioxidant mechanisms, and extracellular trap formation (Carr, 2000; Henneke and Berner, 2006; Hill, 1987; Levy, 2007; Ricevuti and Mazzone, 1987; Yost et al, 2009). The severity of functional differences correlates with the maturity of the infant and begins to decrease within the first few weeks after birth (Carr, 2000). Besides intrinsic differences in PMN function, induction of specific functions and maturation of these cells are developmentally regulated by the availability in the microenvironment of specific inflammatory mediators and growth factors (Christensen, 1989; Vercellotti et al, 1987). For example, an activation product of the fifth component of complement, C5a, is a chemoattractant at sites of inflammation. Low concentrations of C5 in neonatal sera might not permit establishment of chemoattractant gradients at sites of inflammation in newborns comparable to those in adults. Differences between adult and fetal– neonatal PMN functions may thus reflect intrinsic cellular differences required for fetal well‑being and differences in the availability or activity of substances that regulate PMN function. The recognition that systemic bacterial infection in newborns is frequently accompanied by profound neutropenia prompted the investigation of neutrophil kinetics in infected infants (Christensen et al, 1980, 1982; Santos, 1980). These studies have suggested diverse, developmentally specific regulatory mechanisms required for mobilization of the neutrophil response to infection. The lack of neutrophil precursors in bone marrow aspirates of infected Cells committed to phagocyte maturation (granulocyte or monocyte-macrophage) are detectable in the human fetal liver by 6 weeks’ gestation and in peripheral fetal blood by 15 weeks’ gestation. Unlike granulocytes, whose tissue half‑life is hours to days, macrophages migrate into tissues and reside for weeks to months. In a tissue‑specific fashion, these cells regulate availability of multiple factors, including proteases, antiproteases, prostaglandins, growth factors, reactive oxygen intermediates, and a considerable repertoire of cytokines. The importance of macrophages in the neonatal response to infectious agents has been documented in multiple studies. For example, increased antibody response and protection from lethal doses of Listeria monocytogenes were induced in newborn mice by administration of adult macrophages (Lu et al, 1979). Functional differences in chemotaxis, phagocytosis, and toll-like receptor signaling between adult and neonatal cells have been observed and most likely result from both intrinsic fetal–neonatal monocyte-macrophage characteristics, such as reduced IL-12p70 production, and from nonmacrophage factors (e.g., decreased production of the lymphokine IFN-γ) (English et al, 1988; Kollmann et al, 2009; Levy, 2007; Stiehm et al, 1984; van Tol et al, 1984). Inducible expression of individual complement proteins by lipopolysaccharide (LPS), a constituent of gram‑negative cell walls, has also been shown to differ between adult and neonatal monocyte–macrophages (Strunk et al, 1994; Sutton et al, 1986). This difference suggests that, although signal transduction mediated by LPS, LPS‑induced transcription, and accumulation of mRNAs, which direct the synthesis of the third component of complement and factor B, are comparable in adult and neonatal cells, a translational regulatory mechanism does not permit these important inflammatory proteins to be synthesized by LPS‑induced neonatal cells. This observation emphasizes the fact that fetal–neonatal monocytes–macrophages can have functions developmentally distinct from those of adult cells. For example, in utero production of growth factors and removal of senescent cells during tissue remodeling may be critical to fetal development (Kannourakis et al, 1988). Concurrent induction of these functions and immunologic effector functions in fetal monocyte–macrophages would potentially elicit nonspecific inflammation in actively remodeling tissues. Besides having antibacterial functions, neonatal monocyte–macrophages contribute to tissue‑specific regulation of the microenvironment in individual organs. For 454 PART IX Immunology and Infections example, considerable attention has focused on the contributions of these cells to antioxidant defenses and to regulation of protease-antiprotease balance. Because of the importance in tissue injury and repair, tissue and injury‑specific treatment by appropriately targeted and primed monocyte–macrophages may provide therapeutic options for treating a spectrum of problems, from oxygen toxicity in the lung to hemorrhage in the brain. ADAPTIVE IMMUNITY Lymphocytes play multiple critical roles in the adaptive immune responses. Major lymphocyte lineages identified by cell surface and functional criteria include T, B, and NK cells (Kawamoto and Katsura, 2009). All three types develop from CD34+CD38dim hematopoietic stem cells found in the fetal liver and bone marrow (see Figure 36-1). CD34 is also expressed on early committed progenitors, which differentially express other markers usefully in lineage determination. Some early T cell progenitors retain myeloid potential, suggesting a revision of the classic model base on the presence of a common lymphoid progenitor (Bell and Bhandoola, 2008; Wada et al, 2008). The process of lymphocyte differentiation is best viewed as a progressive narrowing of differentiation potential based on the sequential expression of specific transcriptional regulators (Mansson et al, 2010). T Lymphocytes T lymphocytes or T cells develop in the thymus, which is formed from the third branchial cleft and the third or fourth brachial pouch. Thymic lobes are generated when tissue from these sites moves caudally to fuse in the midline. Each lobe can be divided into three regions based on structure and function: the cortex, the corticomedullary junction, and the medulla. The thymic cortex is composed of specialized epithelial cells that express MHC class I and class II molecules and mediate the early stages of T cell maturation. This complex developmental process (Figure 36-2) begins when multipotent CD1a–, CD5–, CD34+, and Transcription factors Ikaros Surface receptors Notch-1 HSC CD38low stem cells enter the thymus at the corticomedullary junction and migrate into the outer cortex (Awong et al, 2009; Res et al, 1996). Maturation continues as cells migrate back through the cortex toward the corticomedullary junction. The first committed T cells express low levels of CD4 and show DNA rearrangement at the TCR δ gene locus. These events are followed by low levels of CD8 expression and by rearrangement at the TCR-β locus, which generates a pre-TCR on the cell surface (Spits et al, 1998). The pre-TCR is composed of several polypeptides, including a TCR-β chain and the invariant pre-Tα chain. Expression of the pre-TCR serves as an important developmental checkpoint (beta selection; von Boehmer and Fehling, 1997). Cells that do not successfully rearrange their TCR-β chain cannot express the pre-TCR and die by apoptosis (Falk et al, 2001). The remaining cells generate antigen-specific receptors when the TCR-α chain replaces pre-Tα. These intermediate- to late-stage progenitors also express both the CD4 and CD8 coreceptors and are called double positives. The minor population of δγT cells follows a similar developmental pattern. Small double-positive cells (CD4+, CD8+, TCRlow) constitute approximately three fourths of all thymocytes, reflecting the critical and rate-limiting developmental stage that follows. Further maturation requires that the unique TCR on each thymocyte interact with self-peptide–MHC complexes expressed on the surface of thymic epithelial cells. Several factors control the outcome of this interaction, including the strength and timing of the TCR signal generated and the nature of the antigen-presenting cell (Hogquist et al, 2005). The TCR and associated CD3 signaling complex translate subtle differences in the affinity of ligand binding into cell fate decisions through induced conformational changes in the CD3 signaling complex (Dave, 2009; Gil et al, 2005; Hayes, 2005), differential utilization of mitogen-activated protein kinase signaling pathways (Sohn et al, 2008; Sugawara et al, 1998), and subcellular compartmentalization of mitogen-activated protein kinase proteins (Daniels et al, 2006). Heterogeneity among thymic epithelial cells and the peptides that they HES-1 GATA-3 Sox4, HEB, NFATc E2A, STAT5 c-Kit, IL-7Rc DN1 DN2 TCF-1-Lef-1, NF-B, p53 Frizzled-receptor, Death-receptor, pre-TCR DN3 DN4 TCR DP SP CD4CD8 or CD4CD8 -selection Hematopoletic stem cells CD44CD25 CD44CD25 CD44CD25 Commitment CD44CD25 CD4CD8 SP Pre-TCR TCR checkpoint checkpoint FIGURE 36-2 Transcription factors and checkpoints in thymic selection. TCR, T cell receptor. (From Wu L, Strasser A: “Decisions, decisions”: betacatenin-mediated activation of TCF-1 and Lef-1 influences the fate of developing T cells, Nat Immunol 2:823-824, 2001.) 455 CHAPTER 36 Immunology of the Fetus and Newborn present also contributes to the complexity of the selection process (Klein et al, 2009). In the affinity-based model of thymic selection (see Figure 36-2), thymocytes that are unable to interact with thymic cortical epithelial cells fail to generate a survival signal and die by apoptosis, the default pathway for most developing cells. Weak interactions generate survival signals and continued development of positive selection. Studies in mice show that the result of positive selection is a thymocyte population containing a high frequency of cells that are broadly cross-reactive for multiple MHC alleles (alloreactivity) and for multiple peptides (Huseby et al, 2005). For many positively selected cells, subsequent high-avidity interactions with bone marrow–derived antigen-presenting cells in the thymic medulla result in activation-induced apoptosis or negative selection (Gong et al, 2001). Negative selection effectively culls from the positively selected pool of cells with cross-reactive TCR and is the primary mechanism for the elimination of self-reactive T cells (Huseby et al, 2003). For a small subset of thymocytes, high-affinity interactions with an agonist ligand can also induce a transcription factor, Foxp3, that is associated with regulatory T cell development rather than cell death (Jordan et al, 2001; Relland et al, 2009). Following avidity-based selection, mature T cells that express either the CD4 or CD8 coreceptor and a single TCR heterodimer are found in the thymic medulla just before the egress into the peripheral circulation. In the human embryo, the first naive, mature T cells appear at approximately 11 to 12 weeks of embryonic development. Thymopoiesis continues for many years thereafter, and integrity of this process is essential for a healthy immune system. Thymectomy early in life results in a substantial loss of naive T cells and in an oligoclonal memory T cell compartment (Prelog et al, 2009; Sauce et al, 2009). The decay process is accelerated by chronic CMV infection, resulting in an immunosenescent T cell phenotype similar to that seen in elderly individuals and associated with increased morbidity and mortality (Wikby et al, 2006). The final phase of T cell development is independent of the thymus and involves peripheral (lymph nodes, spleen- and gut-associated lymphoid tissue) homeostatic mechanisms. These poorly understood events control the expansion of clones recognizing specific antigens and the development of T cell memory. Peripheral repertoire selection begins with migration of lymphocytes from the blood into lymphatic tissue. Naive T cells express the adhesion molecule L-selectin (CD62L) that interacts with peripheral-node addressins, a group of sialomucins present on high endothelial venules. Lymphocyte attachment, coupled with the shear forces produced by blood flow, results in rolling of the lymphocytes along the endothelial cell surface. Signals mediated by CC-chemokine receptor 7 (CCR7) molecules on the surface of rolling lymphocytes result in their tight binding to endothelial cells via integrins and in transendothelial migration (Forster et al, 2008). After entering the lymph nodes, movement of lymphocytes is also controlled by chemokines and their receptors (Worbs et al, 2007). CCR7 signaling directs cells to the T cell areas, whereas CXCR5-mediated signaling controls movement to B cell follicles (Reif et al, 2002). After activation, changes in chemokine receptor expression control the mobilization and function of lymphocytes within the lymph nodes. T cell activation requires a complex molecular cascade that results in reorganization of signaling molecules in the membrane into an “immunological synapse” and in signal transduction (Bromley et al, 2001). Many of the important biochemical events in this process have been described. TCR engagement by an appropriate MHC–peptide ligand results in phosphorylation of components of the CD3 complex. The CD3 complex is composed of the αβTCR and γ, δ, ε and ζ chains. These later molecules all contain specific amino acid sequences called ITAMs, which serve as molecular targets for the tyrosine kinases fyn and lck. The ζ chain is thought to be the most critical component and is found as a homodimer. Each ζ chain contains three ITAMs, which will bind the tyrosine kinase Zap-70 when sequentially phosphorylated. Appropriate phosphorylation of ζ results in a downstream cascade that involves the tyrosine phosphorylation of multiple cellular substrates including phospholipase Cγ1, the guanine nucleotide exchange factor Vav, and the adaptor protein Shc. Ultimately the genetic program of the cells is altered, leading to the transcription of genes for cytokines, cytokine receptors, and transcription factors (Cantrell, 2002). After activation, CD4+ T cells differentiate into Th1, Th2, or Th17 effector cells or become induced Treg cells (Figure 36-3) (Murphy and Reiner, 2002; Zhu et al, 2010). Each of these CD4+ T cell types is defined by prototypical transcription factors and by secretion of a characteristic profile of cytokines in response to antigenic stimulation. The cytokine milieu in the local environment during antigen presentation is a primary factor influencing the developmental fate of a naive T cell following activation. IL-12 promotes Th1 cell development by a signal transducer and activator of transcription 4–dependent mechanism and nTreg Thymus Foxp3 CD4 CD8 Foxp3 Naive T IL-12 T-bet Runx3 Th1 IFN-, IL-2 Autoimmune disease Anti-tumor immunity iTreg IL-2, RA, TGF- Foxp3 IL-6, TGF- Th17 IL-4 RORt Th2 GATA-3 IL-17 Autoimmune disease Microbial immunity IL-4, 5, 10, 13 Parasite infection Allergy FIGURE 36-3 Cytokines and transcription factors associated with T cell differentiation. IFN, Interferon; IL, interleukin; iTreg, inducible regulatory T cell; nTreg, natural regulatory T cell; TGF, transforming growth factor. (From Sakaguchi S, Yamaguchi T, Nomura T, et al: Regulatory T cells and immune tolerance, Cell 133:775-787, 2008.) 456 PART IX Immunology and Infections results in expression of the transcription factor T-bet. Th1 cells secrete IL-2, IFN-γ, lymphotoxin alpha, and tumor necrosis factor-α. Th1 responses are generally proinflammatory. Th1 cytokines act synergistically to lyse virally infected cells and activate antigen-presenting cells as well as granulocytes. In addition to being the signature cytokine of Th1 cells, IFN-γ also promotes the development of the Th1 phenotype and inhibits the development of Th2 cells (Lighvani et al, 2001). Human neonatal T cells are biased against Th1 polarization relative to adult T cells (Levy, 2007). This bias has been linked to reduced IL12(p35) gene expression by dendritic cells (Goriely et al, 2001, 2004) and IL-4–dependent apoptosis of Th1 cells after reexposure to antigen (Li et al, 2004). Poor Th1 function is associated with impaired killing of intracellular pathogens and a reduction in vaccine responsiveness (Levy, 2007). IL-4–mediated activation of the transcription factor STAT6 is required for Th2 cell development (Kaplan et al, 1996). STAT6 activation drives expression of another transcription factor, GATA3, which promotes Th2 cell differentiation. Th2 cells produce IL-4, IL-5, IL-10, and IL-13. IL-4 induces immunoglobulin heavy-chain class switching to the IgE isotype and promotes the development of the Th2 phenotype. IL-5 is an eosinophil growth factor. Thus, Th2 cells are thought to promote antibody production and the allergic response by multiple mechanisms (Ouyang et al, 2001). Human fetal T cells are biased toward Th2 polarization (Prescott et al, 1998), and their abundant representation in the neonate contributes to the Th2 skewing of neonatal responses (Adkins et al, 2004). The cytokines TGF-β and IL-6 activate STAT3, which induces the expression of the transcription factor RORγt and Th17 cell differentiation (Littman and Rudensky, 2010). Th17 cells, which are named for their secretion of the cytokine IL-17A and related family members, are proinflammatory cells associated with the clearance of intracellular pathogens and protection from infection by a number of bacterial and fungal species (Weaver et al, 2007). Abnormalities in Th17 cell development and function are also linked to immunodeficiency and autoimmunity. For example, humans with mutations in molecules affecting Th17 cell differentiation have the hyper-IgE syndrome (Al Khatib et al, 2009; de Beaucoudrey et al, 2008), and genes associated with Th17 cell differentiation are associated with increased susceptibility to Crohn’s disease (Brand, 2009). Activation of naive T cells requires a second signal, and the most potent costimulatory molecule is CD28. The ligands for CD28 are B7-1 and B7-2, which are present on the surface of antigen-presenting cells. Cyotoxic T-lymphocyte antigen 4 (CTLA-4), a molecule expressed only on activated T cells, also binds B7-1 and B7-2 and functions as a negative regulator of T cell activation. Experiments with knockout mice demonstrate that CD28 is the major costimulatory receptor and that signaling through CTLA-4 attenuates CD28-dependent responses (Alegre and Thompson, 2001; Subudhi et al, 2005). A maturational delay in the development of costimulatory function that contributes to reduced T cell responses in the neonatal period has been suggested (Orlikowsky et al, 2003), and the expression of costimulatory molecules is further suppressed by treatment with dexamethasone (Orlikowsky et al, 2005). Control of the clonal proliferation seen in a primary immune response involves a number of inhibitory mechanisms, including Fas/FasL–mediated cell death, ligation of the cell surface molecule PD-1, and regulatory T cell– mediated suppression (Haribhai et al, 2007; Lettau et al, 2009; Riley, 2009; Sakaguchi et al, 2008). These mechanisms reset the peripheral immune system and thereby maintain adequate clonal diversity. A small number of activated T cells survives, and these cells differentiate into memory cells (Wakim and Bevan, 2010). Memory cells are defined by function (i.e., more rapid response) and by expression of certain cell surface markers. The transcription factors T-bet and eomesodermin influence memory development, probably by controlling expression of IL-7R and IL-15R (Intlekofer et al, 2005; Pearce et al, 2003). These receptors are linked to cell survival. Early memory T cells are L-selectinlow, CD45ROhigh, and CD44high. Late memory cells can become L-selectinhigh, which is a surface marker seen in naive T cells. The ultimate number of memory cells has been shown in mice to reflect the initial antigenic load and clonal burst size. Memory T cells allow the secondary response to be more rapid and more potent, characteristics that form the basis for vaccinations. The memory phenotype is not seen early in life, but increases with age. The precise mechanisms that generate and maintain memory T cells have not been determined, although it is known that the persistence of CD8+ memory T cells in mice does not depend on the persistence of specific antigen (Surh and Sprent, 2002; Williams and Brady, 2001). In one study, preterm and term infants developed comparable memory T cell responses after vaccination (Klein et al, 2010). Immunocompetent T cells capable of responding to foreign lymphocytes in the mixed lymphocyte reaction are found in the fetal liver at 5 weeks’ gestation. Before 8 weeks’ gestation, lymphocytes are not detectable in the fetal thymus. After 8 weeks, lymphoid follicles, T lymphocytes, and Hassall’s corpuscles can be identified. By 12 to 14 weeks’ gestation, T lymphocytes can be found in the fetal spleen (Timens et al, 1987). By 15 to 20 weeks’ gestation, the fetus has readily detectable numbers of peripheral T lymphocytes. The first T cell proliferative responses that occur are to mitogens and can be measured at approximately 12 weeks’ gestation. By 15 to 20 weeks’ gestation, antigen-specific responses can be detected (Adkins, 1999; Garcia et al, 2001). Fetal growth has been shown to influence T, B, and NK lymphocyte counts by 3% to 6% per week increase of gestational age (Duijts et al, 2009). Maturation of immune responses continues throughout the first year of life, a fact that affects vaccination schedules (Gans et al, 1998). A low level of thymopoiesis has been shown to continue into adulthood (Kennedy et al, 2001). Regulatory T Cells Natural regulatory T cells develop as a distinct lineage in the thymus dedicated to maintaining self tolerance (Sakaguchi et al, 2008). Induced Treg cells arise in the periphery from conventional CD4+ T cells activated in the presence of TGF-β (Curotto de Lafaille and Lafaille, CHAPTER 36 Immunology of the Fetus and Newborn 2009). Expression of the forkhead-winged helix transcription factor Foxp3 ultimately identifies these cell types and is essential for the acquisition of suppressive effector function (Zheng and Rudensky, 2007). Treg cells are required for the maintenance of immunologic tolerance, as illustrated by the autoimmunity that arises after neonatal thymectomy and by the fatal autoimmune lymphoproliferative disease that develops shortly after birth in mice and humans deficient in Foxp3 (Bennett et al, 2001; Brunkow et al, 2001; Chatila et al, 2000; Sakaguchi et al, 1995; Wildin et al, 2001). Treg cells have risen to the forefront of research in immunology because of their essential role in maintaining immunologic tolerance. Preclinical animal models have demonstrated that adoptive transfer of Treg cells can prevent or cure diabetes, experimental allergic encephalomy­ elitis (multiple sclerosis), inflammatory bowel disease, lupus, arthritis, and graft-versus-host disease (Hori et al, 2002; Kohm et al, 2002; Morgan et al, 2005; Mottet et al, 2003; Scalapino et al, 2006; Tang et al, 2004; Tarbell et al, 2004; Taylor et al, 2002). In humans, removal of Treg cells in vitro enhances the proliferation of T cells in response to self-antigens (Danke et al, 2004). Defects in Treg cell function and number have been described in a number of different human autoimmune diseases, including diabetes, multiple sclerosis, rheumatoid arthritis, and juvenile idiopathic arthritis (Baecher-Allan and Hafler, 2006). Clinical trials using the adoptive transfer of Treg cells after allogenic hematopoietic stem cell transplantation as therapy for graftversus-host disease have begun in Germany and are planned in the United States (Roncarolo and Battaglia, 2007). B Lymphocytes B cells are lymphocytes that upon activation give rise to terminally differentiated, immunoglobulin-secreting plasma cells. Immunoglobulins form the humoral arm of the immune system and provide the main form of protection against many pathogens. B cell development can be divided into embryonic and adult phases. The embryonic phase begins in the fetal liver at the same gestational age as T cell development, follows a similar time course, and uses similar developmental strategies. The adult phase occurs in the bone marrow and continues throughout life. In both phases, B cell progenitors are selected at key developmental checkpoints for the presence of a functional BCR rearrangement and for the absence of BCR self-reactivity. Those progenitors with self-reactive receptors are either eliminated or generate new antigen receptors by continued gene segment rearrangements that are not self-reactive. Both B cell development and peripheral B cell homeostasis require signal transduction through the BCR. Although the B cell compartment is well formed before birth, diversification of the antibody repertoire and several important antibody responses are not developed until long after the neonatal period (Hardy and Hayakawa, 2001; Rohrer et al, 2000; Rolink et al, 2001; Yankee and Clark, 2000). B cell development begins before 7 weeks’ gestation in the human fetal liver, and pre-B cells in various stages of maturation are seen by 8 weeks’ gestation. By 8 to 10 weeks’ gestation, CD34+ hematopoietic stem cells are also found in the bone marrow. The early phase of 457 B cell development, like T cell development, is antigen independent. The first committed progenitors (pro/preB-I) express CD34, terminal deoxynucleotidyl transferase (TdT), recombinase-activating genes (RAG) and CD19 (Li et al, 1993). They lack immunoglobulin gene rearrangement (Allman et al, 1999). The next lineage is marked by heavy chain (H) diversity to joining (DH to JH) gene segment rearrangements followed by variable to diversity/joining segment rearrangement. Terminal deoxynucleotidyl transferase is used during this process to insert nucleotides between the segments to create additional diversity at a region that encodes the VH portion of the antibody-binding site. Late cells in this stage express an invariant surrogate light chain (Kitamura et al, 1992). The normal light chain genes remain in germline configuration. The first developmental checkpoint in B cell development requires expression of surrogate light chain together with Ig-α and Ig-β on the cell surface as the pre-BCR, marking the transition to the pre–BII stage. Expression of the pre-BCR is essential for positive selection of cells that have a functional μH chain rearrangement and for expansion of the pre–B cell lineage (Gong and Nussenzweig, 1996). Cells with a nonfunctional H chain rearrangement or with H chains that assemble poorly with surrogate light chain are unable to progress. VL to JL rearrangement takes place next, and the newly expressed polymorphic L chain protein replaces the surrogate light chain. The completed BCR is antigen-specific and contains the mIgM molecule. A second developmental checkpoint occurs here, and those B cells with self-reactive receptors are eliminated or edited (Hartley et al, 1991). Only 10% to 20% of the immature B cells survive this negative selection and migrate to the spleen, which they enter through the terminal branches of the central arterioles. Once in the spleen, they rapidly differentiate into mIgM+, mIgD+, B220+ mature B cells that enter the recirculating pool of B lymphocytes (Hardy and Hayakawa, 2001). When mature B cells contact antigen through the BCR, a signal is transduced that promotes further growth and differentiation into surface Ig+ memory cells and plasma cells (Calame, 2001). BCR signaling involves activation of the tyrosine kinases Syk and Btk, which are also involved in pre-BCR signal transduction. Certain mutations in Btk result in a failure of pre-BCR signaling and lead to X-linked agammaglobulinemia. Many B cell responses use the ability of B cells to capture antigen at vanishingly low concentrations, process the antigen into small peptide fragments, and then present the antigen to Th cells in the context of MHC class II molecules. Antigen-specific T cells then form conjugates with antigen-presenting B cells. The T cell membrane protein gp39 interacts with CD40 on B cells and triggers B clonal expansion, cytokine responsiveness, and isotype switching. B lymphocytes with surface IgM are first found in the human fetal liver at 9 weeks’ gestation and in the fetal spleen at 11 weeks’ gestation (Owen et al, 1977; Timens et al, 1987). Antigen‑specific antibody production can be detected in the human fetus by 20 weeks’ gestation. Fetal spleen cells can synthesize in vitro IgM and IgG by 11 and 13 weeks’ gestation, respectively. 458 PART IX Immunology and Infections TABLE 36-2 Characteristics and Functions of Immunoglobulins Molecular Weight (kD) Serum Concentration (g/L) Serum Half-life (Days) Neutralization Opsonization IgG1 150 10 21 ++ +++ IgG2 150 5 21 ++ (+) IgG3 170 1 7 ++ ++ IgG4 150 0.5 21 ++ + IgA1 160 3 7 ++ + IgA2 160 0.5 7 ++ + IgM 900 2 5 + + IgD 180 0.03 3 – – IgE 190 0.00003 3 – – Adapted from Mix E, Goertsches R, Zett UKL: Immunoglobulins: basic considerations, J Neurol 253:V9-V17, 2006. Ig, Immunoglobulin; NK, natural killer. Immunoglobulins Immunoglobulin G Immunoglobulins are a heterogeneous group of proteins that are detectable in plasma and body fluids and on the surfaces of mucosal barriers and B lymphocytes. Although these proteins have multiple, diverse functions, they are classified as a family of proteins because of their capacity to act as antibodies—that is, to recognize and bind specifically to antigens. The rapid advances in understanding molecular structure and regulation, genetic diversity, and differences in functions of immunoglobulins have been reviewed recently (Bengten et al, 2000; Mix et al, 2006). The functions of immunoglobulins relevant to fetal and neonatal immunity are summarized in Table 36-2. There are five known classes of immunoglobulins: IgG, IgM, IgA, IgE, and IgD. Human IgM circulates as a pentamer or hexamer and IgA as a dimer. Multimers are formed in association with an additional J-chain (or joinchain). Functions of individual immunoglobulin classes are different but overlapping (see Table 36-2) (Mix et al, 2006). The prototype immunoglobulin molecule consists of a pair of identical heavy (H) chains that determine the immunoglobulin class in combination with a pair of identical light (L) chains (Figure 36-4). Disulfide bonds and electrostatic forces link the chains. Each immunoglobulin molecule contains two N-terminal, identical domains with antigen‑binding activity (Fab). The Fab fragment consists of variable (VH and VL) and constant (CH and CL) regions on both the H and L chains. The Fab domains function to bind antigenic epitopes via complementarity determining regions. A third fragment (i.e., F crystalline [Fc]) is devoid of antibody activity, but mediates immunoglobulin effector functions (Table 36-3). The principal functions of the Fc fragment include: receptor-mediated phagocytosis (IgG1/3 and IgA), cytotoxicity (IgG1/3), release of inflammatory mediators (IgE), receptor-mediated transport through mucosa (IgA and IgM) and placenta (IgG1/3), and complement activation (IgG1/3 and IgM). The five different isotype classes of human immunoglobulins (IgG, IgM, IgA, IgD, and IgE) are defined structurally by differences in the Fc fragments. Within isotypes, there are four IgG subclasses (IgG1, IgG2, IgG3, and IgG4) and two IgA subclasses (IgA1 and IgA2). IgG is the most abundant immunoglobulin class in human serum and accounts for more than 75% of all antibody activity in this compartment. Its monomeric form circulates in plasma, has a molecular mass of approximately 155 kD, and in adults constitutes approximately 45% of total body IgG in the extravascular compartment. The human conceptus is able to produce IgG by 11 weeks’ gestation (Gitlin and Biasucci, 1969; Martensson and Fudenberg, 1965). The importance of the contribution of IgG to immunologic function is illustrated by the clinical problems encountered in individuals with genetically disrupted IgG production (hypogammaglobulinemia). These patients have recurrent infections if they do not receive treatment with immunoglobulin replacement therapy (Yong et al, 2008). Several investigators have observed that infants in whom group B streptococcal sepsis develops have low concentrations of type‑specific IgG, which has prompted attempts at acute or prophylactic treatment with immunoglobulin replacement therapy (Stiehm et al, 1987). Although successful in some trials, replacement therapy in newborns has not proved as efficacious as in individuals with genetically determined hypogammaglobulinemia (Cohen-Wolkowiez et al, 2009; Noya and Baker, 1989). This difference might be caused in part by fetal and neonatal IgG synthesis being regulated by both developmental and genetic mechanisms (Cates et al, 1988; Lewis et al, 2006). The kinetics of IgG placental transport suggest both passive and active transport mechanisms (Saji et al, 1999; Simister, 2003). Because IgG transport begins at approximately 20 weeks’ gestation, preterm infants are born with lower IgG concentrations than are their mothers or infants born at term. The full‑term infant has a complete repertoire of adult, maternal IgG antibodies. Thus, provided that relevant maternal IgG has been transported to the fetus, newborns are generally immune to many viral and bacterial infections (e.g., measles, rubella, varicella, group B streptococcus, and E. coli) until transplacentally acquired antibody titers decrease to biologically nonprotective concentrations at 3 to 6 months of age (Pentsuk and van der Laan, 2009). The regulation of IgG production in CHAPTER 36 Immunology of the Fetus and Newborn Epithelial Transport Complement Activation Placental Transport Sensitization for Killing by NK Cells Sensitization of Mast Cells and Basophils Strong classic, alternative – +++ ++ + Classical, alternative – + – – Strong classical, alternative – ++ ++ + Alternative – (+) – – Alternative +++ – – – Alternative +++ – – – Strong classical + – – – Alternative – – – – – – – – +++ Idiotopes H L V re ari gi ab on le V s Idiotopes s s L s s Constant region CH 3 s s s s Constant region CH 4 Carbohydrates Hinge s s s s s s s s s s Constant region CH 2 s s s Constant region CH 2 s s s t an st C L on n C gio re s s V re ari gi ab on le V s Constant region CH 3 s s C re on gi sta on n C t s t an 1 st H on C C ion g re Light chain Heavy chain re Con gi s on ta C nt H 1 e bl L ria n V Va gio re s s Paratope with 3 complementaritydetermining regions (CDRs) s s Carbohydrates Constant region CH 4 s e bl H ria n V Va gio re N-terminal 459 C-terminal FIGURE 36-4 Structure of monomeric immunoglobulin. (From Mix E, Goertsches R, Zett UK: Immunoglobulins: basic considerations, J Neurol 253(Suppl 5):V9-V17, 2006.) preterm infants has been a topic of study for more than 60 years (Ballow et al, 1986; Bauer et al, 2002; Bonhoeffer et al, 2006; Dancis et al, 1953; Meffre and Salmon, 2007; Schroeder et al, 1995). Although adults with antibody deficiency syndromes have a higher rate of infections when IgG concentrations are less than 300 mg/dL, the serum IgG concentrations of many preterm infants decrease to less than 100 mg/dL apparently without consequences. These observations suggest that preterm infants have additional immunologic protective mechanisms or that the protective capacity of humoral immunity is not accurately assessed by serum IgG concentrations alone in preterm infants. Studies also suggest that reduced endogenous secretion of IgG by newborns may result from reduced ability of neonatal B cells to undergo immunoglobulin isotype switching, because of decreased or ineffective expression of the ligand for the B cell surface protein CD40 on activated cord blood T cells (Brugnoni et al, 1994; Fuleihan et al, 1994). 460 PART IX Immunology and Infections TABLE 36-3 Severe Combined Immunodeficiency Classification Mechanisms Mutated Genes Premature cell death ADA AR T, B, NK Defective cytokinedependent survival signaling γc X-L T, NK Defective VDJ rearrangement Defective pre-TCR and TCR signaling Inheritance Affected Cells multiple effector functions of this cascade. Second, its pentameric structure provides conformational flexibility to accommodate multivalent ligand binding. Third, because of its localization in the vascular compartment and its high efficiency in complement activation, IgM has a prominent role in clearance from serum of invading microorganisms. Immunoglobulin A JAK3 AR T, NK IL7RA AR T RAG1 or RAG2 AR T, B Artemis AR T, B CD3 δ, ξ, ε AR T CD45 AR T Adapted from Cavazzana-Calvo et al: Gene therapy for severe combined immunodeficiency: are we there yet? J Clin Invest 117:1456-65, 2007. AR, Autosomal recessive; NK, natural killer; TCR, T cell receptor; X-L, X-linked. IgG functions in host defenses in several ways. It can neutralize a variety of toxins in plasma by direct binding. As described previously in the section on complement, after antigen binding IgG can activate the complement cascade via interaction with early‑acting, classical pathway complement components. The Fc portion of IgG can interact with cell surface receptors on mononuclear phagocytes and polymorphonuclear leukocytes and thereby promote clearance of immune complexes and phagocytosis of particles or microorganisms—a process known as opsonization. Finally, the presence of IgG on specific target cell antigens (e.g., tumors or allogeneic transplant tissues) can mediate antibody‑dependent cellular cytotoxicity, a mechanism through which lymphocyte subpopulations bearing Fc receptors can acquire the capacity to recognize non–self-antigens. Immunoglobulin M IgM represents approximately 15% of normal adult immunoglobulin. IgM circulates in serum as a pentamer of disulfide‑linked immunoglobulin molecules joined by a single cross‑linking peptide (J-chain). The size of IgM (molecular mass >900 kD) restricts its distribution to the vascular compartment. Although the antibody‑binding affinity of monomeric IgM is low, the multivalent structure of the molecule provides high pentameric antibody avidity. IgM synthesis has been detected in the human conceptus at 10.5 weeks’ gestation (Rosen and Janeway, 1964). Because maternal‑fetal transport of IgM does not occur, elevated (>20 mg/dL) concentrations of IgM in the fetus or newborn may be suggestive of intrauterine infection or immunologic stimulation (Alford et al, 1969; Stiehm et al, 1966). However, because it is technically difficult to distinguish IgM molecules with specificity for individual organisms, diagnosis of infections by analysis for specific IgM antibody remains of limited usefulness. IgM is important for fetal and neonatal host defenses for several reasons. First, the IgM molecule is the most efficient of any immunoglobulin isotype in activation of the classical pathway of complement. As a result, it can trigger Although IgA accounts for approximately 10% of serum immunoglobulins, it is detectable in abundance in all external secretions (Brandtzaeg, 2010). In serum, IgA is present as a monomer (molecular weight, 160 kD), whereas in secretions it exists as a dimer (molecular weight, 500 kD) attached to a J chain identical to that found in IgM. In addition to this structural difference, mucosal IgA is attached to an additional protein called the secretory component. This protein is a proteolytic cleavage fragment of the receptor involved in the secretion of polymeric IgA onto mucosal surfaces and into bile. Secretory IgA produced locally on mucosal surfaces by plasma cells is thus readily distinguishable from serum IgA. Although the amount of IgA produced daily is not rigorously quantified, it is estimated to exceed immunoglobulin production of all isotypes combined. Despite its relative abundance, unlike IgM and IgG, IgA cannot activate the classical pathway of complement nor effectively opsonize for phagocytosis particles or microorganisms. Although IgA is detectable on the surface of human fetal B cells at 12 weeks’ gestation, adult concentrations of serum and secretory IgA are not achieved until approximately 10 years of age. Because serum IgA is not transplacentally transferred in significant amounts, IgA is almost undetectable in cord blood. Colostrum‑derived secretory IgA can provide a source of IgA in both gastrointestinal tract and other secretions for the newborn infant (Brandtzaeg, 2010). Unlike other immunoglobulin isotypes, amino acid sequences of the hinge region of the IgA‑2 subclass confer partial resistance to bacterial proteases. IgA is thus more resistant than other immunoglobulin isotypes to proteolytic effects of gastric acidity. Although considerable investigation suggests that passive immunization with IgA occurs with breastfeeding in humans, the overall importance of IgA in host defenses is currently not well characterized (Bailey et al, 2005). Immunoglobulin E The serum concentration of IgE is undetectable by standard immunochemical techniques and accounts for ­approximately 1/10,000 of the immunoglobulin in adult serum. It circulates in the monomeric form (molecular weight, 190 kD). Structurally, IgE lacks a hinge region. It is produced by most lymphoid tissues in the body, but in greatest amounts in the lung and gastrointestinal tract. It is not secreted, and its appearance in body fluids generally occurs only with induction of inflammation. IgE cannot activate complement or act as an effective opsonin. Its bestknown role is as a mediator of immediate hypersensitivity reactions. Specifically, antigen‑specific IgE triggers mast cell degranulation with resultant bronchoconstriction, tissue edema, and urticaria via interactions with IgE receptors on the mast cell surface. Because of the presence of IgE in lung secretions and its potential importance in mediating CHAPTER 36 Immunology of the Fetus and Newborn allergic pulmonary and gastrointestinal reactions, considerable interest has been recently focused on the use of serum IgE concentrations to identify premature infants at risk for developing reactive airways disease or in the diagnosis of gastrointestinal hypersensitivity reactions (Bazaral et al, 1971; Herz, 2008; Jarrett, 1984; Kagan, 2003). Immunoglobulin D Although IgD is found in trace quantities in adult human serum and has neither complement‑activating activity nor the capacity to opsonize particles or microorganisms, approximately 50% of cord blood lymphocytes exhibit IgD on their cell surface (Colten and Gitlin, 1995; Haraldsson et al, 2000; Preud’homme et al, 2000). These pre–B lymphocytes express surface IgM and IgD simultaneously. Because of its wide distribution on B cells, IgD may have an important role in primary antigen recognition for the fetus and newborn infant. Immunoglobulin Therapy Because of low serum immunoglobulin concentrations, a concurrent greater susceptibility to infection in preterm infants, and successful use of immunoglobulin replacement in patients with genetically based hypogammaglobulinemia, several investigators have proposed prophylactic, intravenous administration of immunoglobulin (IVIG) to prevent antibody deficiency and nosocomial infection (Baker et al, 1992; Clapp et al, 1989; Cohen-Wolkowiez et al, 2009; Fanaroff et al, 1994; Kinney et al, 1991; Lewis et al, 2006; Magny et al, 1991; Weisman et al, 1994). Immunoglobulin therapy in very low-birthweight neutropenic infants with bacterial sepsis and shock has been tried with some success, when the potential benefits might outweigh the immunomodulatory properties of IVIG (Christensen et al, 2006). However, well-designed, placebo-controlled studies have failed to show consistent benefit of this strategy using different preparations of immunoglobulin, different treatment groups, and different dosage regimens (Ohlsson and Lacy, 2004a). These results suggest that serum immunoglobulin concentrations do not predict immunoglobulin function and immunologic response as accurately in the newborn infant as in the older child or adult. In infants with sepsis, the results of immunoglobulin therapy have been difficult to interpret because of the complexities in study design associated with enrolling acutely ill infants and the small numbers of enrolled infants (Ohlsson and Lacy, 2004b). A recent metaanalysis of randomized, controlled studies that examined the effects of polyvalent immunoglobulin for treatment of sepsis or septic shock in newborn infants (12 trials, 710 newborn infants) suggested a reduction in mortality (Kreymann et al, 2007). However, heterogeneous study quality led the Cochrane reviews to suggest that there is still insufficient evidence to recommend routine use of this strategy. The availability of results from the large cohort recruited for the International Neonatal Immunotherapy Study (www.npeu.ox.ac.uk/inis) should provide a more definitive recommendation (INIS Study Collaborative Group, 2008). Other interventions including administration of colony-stimulating factors, antistaphylococcal antibodies, probiotics, glutamine supplementation, 461 recombinant human protein C, and lactoferrin have been reviewed recently (Cohen-Wolkowiez et al, 2009). Besides prophylactic or acute treatment of systemic bacterial infection, the therapeutic scope of monoclonal and polyclonal immunoglobulin therapy has expanded over the last 10 years. For example, although controversy still exists concerning specific populations of infants for whom prophylaxis is appropriate, administration of anti–respiratory syncytial virus monoclonal antibody has become standard of care for prevention of respiratory syncytial virus (RSV) infection in high-risk infants and for management of nosocomial outbreaks of RSV (Fitzgerald, 2009; Halasa et al, 2005; Kurz et al, 2008; Wu et al, 2004). In addition, administration of intravenous IgG is a safe and effective immunomodulatory strategy for reducing the need for exchange transfusion in hemolytic disease of the newborn if the total serum bilirubin is rising despite intensive phototherapy (American Academy of Pediatrics Subcommittee on Hyperbilirubinemia, 2004; Smits-Wintjens et al, 2008; Walsh and Molloy, 2009). Intravenous immunoglobulin is also indicated for treatment of immune thrombocytopenia (Bussel and Sola-Visner, 2009). The administration directly into the ocular vitreous of antibody against vascular endothelial growth factor to bind and inactivate vascular endothelial growth factor has been shown in early studies to reduce the need for laser therapy for retinopathy of prematurity (Mintz-Hittner and Best, 2009). The historically promising use of oral immunoglobulin for preventing necrotizing enterocolitis has not proved effective in a recent Cochrane review of three trials (n = 2095 infants) (Foster and Cole, 2004). In view of the continued expansion of neonatal diseases that are potentially treatable with immunoglobulin, the use of immunoglobulin therapy must continue to be evaluated not only in terms of immediate benefits, but also in terms of long-term effects on immunologic function. SPECIFIC IMMUNOLOGIC DEFICIENCIES The most common reason for increased immunologic susceptibility to infection in newborns, besides prematurity, is iatrogenic immunosuppression caused by administration of corticosteroids for treatment or prevention of bronchopulmonary dysplasia. Although pulmonary and neurodevelopmental benefits have been attributed to this therapy (Abman and Groothius, 1994; Cummings et al, 1989), caution has developed concerning possible long-term, adverse neurodevelopmental effects of steroid administration (Bancalari, 2001; Barrington, 2002; Committee on the Fetus and Newborn, 2002). Although the mechanisms that lead to steroid-induced amelioration of pulmonary disease have not been elucidated completely, the pulmonary inflammatory response as measured by the concentrations of the anaphylatoxin C5a, leukotriene B4, IL-1, elastase‑α1‑proteinase inhibitor, and the number of neutrophils has been shown to be attenuated in infants receiving steroid treatment (Groneck et al, 1993). Whereas shorter courses of steroids can reduce side effects, including immunosusceptibility, the availability of nebulized steroids may provide effective antiinflammatory therapy with minimum toxicity (Cole et al, 1999). 462 PART IX Immunology and Infections There are more than 130 recognized primary immunodeficiency diseases (Notarangelo, 2010). The physician should attempt to differentiate infants with specific genetically regulated immunologic deficiencies from those with developmentally regulated, environmentally induced, or infection-related susceptibility to microbial invasion (Rosen, 1986; Rosen et al, 1984). Documenting a full family history during an antenatal visit can be helpful in identifying relatives removed by as many as two to four generations with histories suggestive of primary immunodeficiency disease, as most arise from single-gene defects. Genetic testing is widely available for many primary immunodeficiency diseases, although de novo mutations are common (Notarangelo, 2010). Several single-gene defects resulting in blocks in T and B cell development are shown in Figure 36-5. Several disorders arising from these defects can manifest in the first months of life. SEVERE COMBINED IMMUNODEFICIENCY Severe combined immunodeficiencyt (SCID) is a rare category of diseases that affect development of both B cells and T cells. The estimated frequency is between 1 per 70,000 and 1 per 100,000 live births (Buckley et al, 1997; Stephan et al, 1993). Before identifying the genes involved in producing the SCID phenotype, investigators classified patients with the disorder by analyzing the number, cell surface proteins, and function of circulating lymphocytes. The most common subgroups are T-B- SCID, T-B+SCID, and adenosine deaminase deficiency. T-B+ SCID accounts for more than half of all SCID cases, and both X-linked and autosomal recessive forms exist (Fischer, 2000). Approximately 80% of T-B+ SCID myeloid precursor cases involve males with the X-linked form, which results from deleterious mutations in the common gamma chain (γc) shared by the IL-2, IL-4, IL-7, IL-9, and IL-15 cytokine receptors (see Figure 36-5) (Noguhi et al, 1993). Defects in the γc protein abrogate the development of T cells and NK cells and lead to B cell dysfunction despite normal B cell numbers. IL-7 receptor signaling is essential for T cell development within the thymus, and the deficiency of this cytokine accounts for the absence of T cells (Cao et al, 1995; DiSanto et al, 1995; Peschon et al, 1994). IL-15 is implicated in NK cell development, and reductions in IL-4 receptor signaling are thought to contribute to poor B cell function seen in γc deficient patients (Mrozek et al, 1996). Nevertheless, the precise mechanisms involved in the NK cell developmental defect and the B cell functional defect are largely unknown. The less common autosomal recessive form of T-B+SCID is primarily due to mutations in the Janus-associated tyrosine kinase JAK3 (Macchi et al, 1995; Russell et al, 1995). This kinase is expressed in cells of hematopoietic lineage and associates with γc as part of the cytokine receptor signaling cascade. The T-B+NK- immunophenotype of these patients is identical to that seen in γc deficiency. Rarely, patients with autosomal recessive T-B+ SCID have NK cells. In two such patients, a defect was found in the IL-7Rα chain (Puel et al, 1998). This finding was predicted by the phenotype of the IL7R knockout mouse, and the data imply that the IL-7 receptor is not essential for NK cell development. Clinically, the X-linked and autosomal recessive forms of T-B+ SCID are indistinguishable. Affected infants often appear healthy at birth. In the neonatal period a morbilliform rash, probably the result of attenuated PERIPHERY pro NK NK  coronin-1A THYMUS  AK2 HSC Orai-1 Stim1 CLP pro T JAK-3 IL-7R CD3 , ,  TAP1,2 Tapasin ZAP70 CD8  RAG-1,2 Artemis Cernunnos DNA ligase 4 DNA-PKcs CD4 MHC cl. II def. mature B CD40, CD40L, pre B immature B pro B AID, UNG, RAG-1,2 TACI, ICOS, Artemis, Cernunnos CD19 DNA ligase 4, DNA-PKcs switched B cell BTK, mu heavy chain, 5, Ig, Ig, BLNK, LRRC8 FIGURE 36-5 Blocks in T and B cell development associated with severe combined immunodeficiency and other primary immunodeficiencies. NK, Natural killer. (From Notarangelo LD: Primary immunodeficiencies, J Allergy Clin Immunol 125:S182-S194, 2010.) CHAPTER 36 Immunology of the Fetus and Newborn graft‑versus‑host disease from transplacental passage of maternal lymphocytes, may be the only symptom of SCID (Rosen, 1986). Over the first several months of life, as acquired maternal antibody levels drop, failure to thrive and undue susceptibility to infection become universal features. Intractable diarrhea, pneumonia, and persistent thrush, especially oral thrush, constitute the triad of findings most frequently seen in infants with this disease (Stephan et al, 1993). Mutations in the gene encoding adenosine deaminase (ADA), an enzyme in the purine salvage pathway, account for approximately 20% of all cases of SCID. Mutations in another enzyme in the purine salvage pathway, purine nucleoside phosphorylase, are found in approximately 4% of patients with SCID (Fischer, 2000; Fischer et al, 1997). The ADA-deficient phenotype is variable, and neonatal onset, delayed onset, and partial forms have been described (Santisteban et al, 1993). In neonatal-onset ADA deficiency, patients have a profound T, B, and NK cell lymphopenia, with a clinical presentation similar to T-B+ SCID (Hirschhorn, 1990). Although ADA is normally present in all mammalian cells, life-threatening disease in ADA deficiency is limited to the immune system. Less severe manifestations include flaring of the costochondral junction as seen on the lateral chest radiograph (termed rachitic rosary) and pelvic dysplasia. Other findings associated with ADA deficiency include hepatic and renal dysfunction, deafness, and cognitive problems (Fischer, 2000; Rogers et al, 2001). Patients who lack ADA accumulate deoxy-adenosine triphosphate in red blood cells and lymphocytes, and the concentration correlates with disease severity. The ADA substrates, adenosine and deoxyadenosine, are found at increased levels in the serum and urine. Studies in ADA-deficient mice suggest that developing thymocytes are particularly sensitive to these metabolic derangements and that the few T cells that mature have signaling defects (Blackburn et al, 2001). The precise mechanism by which the accumulation of ADA substrates results in the immunologic pathology seen in humans remains unclear (Yamashita et al, 1998). The remaining 20% of SCID patients lack mature T and B cells, but have functional NK cells (T-B-NK+). These patients exhibit an autosomal recessive pattern of inheritance and have defects in the VDJ recombination machinery (Schwarz et al, 1996). Somatic recombination of variable, diversity, and joining DNA segments is a critical step in the development of B and T cells. A rearranged IgM heavy chain and TCR-β chain form components of the pre–B and pre–T receptors, which provide essential survival signals for developing lymphocytes. In the absence of recombination, these lymphocyte precursors do not receive a survival signal, die, and produce the T-BSCID phenotype. NK cells do not require somatic cell recombination and survive. In the recombination process, DNA cleavage is mediated by RAG1 and RAG2, which are the recombination-activating proteins that recognize specific sequences flanking V, D, and J segments. The two proteins act in concert to introduce a double-strand DNA break and leave hairpin-sealed coding ends. Recruitment of a DNA-dependent protein kinase and other components of the general DNA repair machinery completes the process. 463 Mutations in the genes encoding both RAG1 and RAG2 have been identified in patients with T-B-NK+ SCID (Cornero et al, 2000) and in some patients with Omenn’s syndrome (Cornero et al, 2001). Omenn’s syndrome is clinically characterized by failure to thrive, diarrhea, erythroderma, alopecia, hepatosplenomegaly, and lymphadenopathy. This concurrence of RAG mutations implies that factors other than RAG mutations also contribute to the Omenn’s syndrome phenotype. Finally, a small subset of T-B-NK+ SCID patients also have increased sensitivity to ionizing radiation, which suggests mutations that affect the DNA repair mechanism (Fischer, 2000). This supposition is strengthened by murine and equine SCID in which specific defects in the DNA-dependent protein kinase involved in VDJ recombination and DNA repair have been identified. The diagnosis of SCID is often suggested by an opportunistic or unusually severe infection in the setting of profound lymphopenia (<1000 lymphocytes per mm3) (Gennery and Cant, 2001). Only 10% of patients with SCID have lymphocyte counts in the normal range. T cells detectable in peripheral blood of affected infants shortly after birth may be either maternal T cells or circulating thymocytes. The thymus gland is not seen on chest radiographs. Histologically, the gland is composed of islands of endodermal cells that have not become lymphoid and contain no identifiable Hassall’s corpuscles. Lymphocyte subpopulation analysis using monoclonal antibodies to cell surface markers and analytical flow cytometry is the most important confirmatory test. As noted previously, each phenotypic pattern suggests a specific diagnosis and molecular defect. Other useful tests include measurements of ADA and purine nucleoside phosphorylase activity in red blood cells and isohemagglutinins as a marker of specific IgM production. Quantitative immunoglobulin levels are not particularly helpful in the diagnosis of neonatal SCID, because most IgG is maternal in origin and IgA and IgM levels are often low in the neonatal period. Once the immunophenotype has been established, a precise molecular diagnosis should be obtained and genetic counseling should be provided. Females carrying deleterious mutations in the IL2RG gene (X-linked SCID) can be identified by nonrandom inactivation of the X chromosome in lymphocytes. Prenatal diagnosis is available by gene identification from a chorionic villous biopsy. In the absence of definitive prenatal genetic testing, all newborn siblings of patients with known SCID, and infant male first cousins of patients with known or suspected X-linked SCID, should be considered at high risk for this disorder and investigated promptly after birth. An effective strategy for detecting newborn infants with SCID with statewide newborn screening was first implemented in Wisconsin (Routes et al, 2009). The test is based on the detection of T cell recombination excision circles by PCR and also identifies patients with lymphopenia resulting from other causes (Baker et al, 2009; Chan and Puck, 2005). When a fetus at risk for a genetic form of SCID is identified, treatment should begin in the delivery room and should be coordinated with antenatal diagnostic interventions. Specifically, cord blood samples should be obtained for white blood cell count and differential, lymphocyte subset determinations, karyotype (if not 464 PART IX Immunology and Infections performed antenatally), mitogen stimulation studies, and immunoglobulin measurements. Because the majority of children with genetic SCID do not become ill within the 1st week of life, care in an incubator should be provided, and staff members should observe strict handwashing technique. SCID is a pediatric emergency and is invariably fatal if untreated. Most untreated patients die in the 1st year of life. Treatment of SCID begins with aggressive antibiotic and antiviral therapy for infections, intravenous immunoglobulin replacement, and prophylaxis for Pneumocystis carinii infection. Besides greater susceptibility to opportunistic infections, these infants are also susceptible to development of graft‑versus‑host disease, either before birth as a result of engraftment of maternal T lymphocytes or after birth as a result of the engraftment of T lymphocytes present in transfused blood products (Pollack et al, 1982; Thompson et al, 1984). Therefore infants in whom SCID is suspected should receive only irradiated blood products and should not be given live viral vaccines. Good and colleagues performed the first successful bone marrow transplantation (BMT) in a SCID patient (Gatti et al, 1968). Allogeneic BMT, preferably from an HLAmatched sibling, is now the standard treatment for most types of SCID (Buckley, 2000). In these transplants, conditioning regimens are not required for engraftment. Recipients usually become chimeric, with only T and NK cells of donor origin. B cell function is frequently deficient and many patients continue to require monthly therapy with IVIG. Survival rates for this type of BMT exceed 90%. When no HLA-identical donor is available, haploidentical transplants have been used successfully, although overall survival is lower with such transplants (Fischer, 2000). The rates of engraftment after haploidentical transplantation of T-B-NK+ SCID patients are low; this is likely due to host NK cell function, and some form of preconditioning may be indicated in this subset of patients. In addition to BMT, ADA deficiency can be treated with enzyme replacement. This treatment involves weekly injections of ADA coupled to polyethylene glycol. Response, consisting of decreasing deoxyadenosine triphosphate levels and increasing T cell numbers, is seen in most patients within weeks. Finally, gene therapy based on vector-mediated transfer of therapeutic gene into autologous hematopoietic stem cells has been used to treat patients with ADA and γc SCID and patients with chronic granulomatous disease. Results in more than 30 patients showed engraftment and clinical benefit, but were limited by transient expression and by insertional mutagenesis (Aiuti and Roncarolo, 2009). WISKOTT‑ALDRICH SYNDROME Wiskott‑Aldrich syndrome (WAS), another form of primary immunodeficiency, is characterized by severe eczema, thrombocytopenia, increased risk of malignancy, and susceptibility to opportunistic infection. Other manifestations, which may be present in the newborn period, include petechiae and bruises, bloody diarrhea, and hemorrhage after procedures. In infants with any of these clinical findings, abnormal low mean platelet volume in a complete blood count report is an important clue to possible WAS. WAS is inherited as a sex‑linked recessive trait (Rosen, 1986). In untreated cases, children survive longer than infants with SCID (median survival, 5.7 years). T lymphocytes in affected patients are decreased in number and diminished in function. Reductions in platelet size and thrombopoiesis are also noted. Affected children can be treated with BMT, and the 5-year survival rate after HLA-identical sibling BMT is approximately 90% (Filipovich et al, 2001). Although transplantation corrects the T cell defects, thrombocytopenia persists. As for children with SCID, all blood products given to children with WAS should be irradiated before administration to avoid T cell engraftment and graft‑versus‑host disease. The gene responsible for the WAS was cloned in 1994 (Derry et al, 1994). It is composed of 12 exons and encodes a 502 amino acid cytoplasmic protein (WAS) expressed in all hematopoietic cells. Evaluation by flow cytometry for expression of this protein by lymphocytes can be used to diagnose WAS. Other mammalian WAS family members include a more widely expressed N-WAS and three WAVE/Scar isoforms. These proteins have multiple domains and are important regulators of actin polymerization (Symons et al, 1996); they have a common carboxyterminal region through which they activate Arp2/3, an actin-nucleating complex involved in actin assembly and cytoskeletal structure. The amino termini are distinct and allow the proteins to couple Arp2/3 activation to a wide variety of different intracellular signals. More than 340 mutations in WAS have been described, which have profound effects on cell motility, signaling, and apoptosis (Rengan and Ochs, 2000). DiGEORGE SYNDROME The embryologic anlage of the thymus gland and the parathyroid gland is the endodermal epithelium of the third and fourth pharyngeal pouches. When normal development of these structures is disturbed, thymic and parathyroid hypoplasia can occur (Rosen, 1986). Infants with this disorder, DiGeorge syndrome, can exhibit abnormalities of calcium homeostasis during the neonatal period (hypocalcemia and tetany) and variable T cell deficits, which appear to depend on the presence and number of small, normal‑appearing ectopic thymic lobes. In addition, these infants have congenital, conotruncal cardiac defects, low‑set ears, midline facial clefts, hypomandibular abnormalities, and hypertelorism. The availability of methodology for performing gene dosage studies and more refined cytogenetic techniques (fluorescence in situ hybridization) has permitted the description of a contiguous gene syndrome that includes the DiGeorge phenotype (Hall, 1993). The DiGeorge syndrome is known to result from varying sized deletions on chromosome 22q11 in more than 90% of patients (Markert, 1998; Shprintzen, 2008). This deletion has been linked to several other diagnostic labels, including velocardiofacial (Shprintzen) syndrome, Cayler syndrome, and Opitz G/BBB syndrome. Collectively, these are referred to as the 22q11 deletion syndromes. The cardiac anomalies associated with the 22q11 deletion syndrome are variable, but usually involve the outflow tract and the derivatives of the branchial arch arteries. These defects include interrupted CHAPTER 36 Immunology of the Fetus and Newborn aortic arch type B, truncus arteriosus, and tetralogy of Fallot (Carotti et al, 2008). Children with the 22q11 deletion syndrome also exhibit a higher incidence of receptive-expressive language difficulties, cognitive impairment, and behavioral problems including psychotic illness (Jolin et al, 2009; Scambler, 2000). The cardiovascular defects have been shown to be the result of haploinsufficiency of TBX1 (Merscher et al, 2001). In the nursery, identification of infants with congenital conotruncal abnormalities or unexplained, persistent hypocalcemia should prompt consideration of this syndrome. Thymic implants can partially correct the immunologic deficits (Markert et al, 2007). IMMUNIZATION MATERNAL IMMUNIZATION Immunization before or during pregnancy has been effective in preventing several specific neonatal infections including diphtheria, pertussis, tetanus, hepatitis B, and rabies (ACOG Technical Bulletin Number 160, 1993; Hackley, 1999; Immunization Practices Advisory Committee, 1991; Stevenson, 1999). For example, in developing countries, immunization during pregnancy with tetanus toxoid is a cost‑effective method for preventing neonatal tetanus and for providing up to 10 years of protection for infants (Gill, 1991; Schofield, 1986; Vandelaer, 2003). The benefits are substantial for both mother and infant from induction before or during pregnancy of maternal IgG antibody that can be transferred to the fetus and protect both the fetus and mother against postpartum morbidity and mortality (Healy and Baker, 2006; Insel et al, 1994; Linder and Ohel, 1994). However, immunization during pregnancy is biologically distinct from immunization of nonpregnant individuals. Vaccine epitopes can be shared with vital fetal or placental tissues; therefore vaccination can lead to unanticipated maternal or fetal morbidity. Maternal immunization can induce an antibody response in the fetus, as has been demonstrated with tetanus toxoid (Gill et al, 1983) and thereby induce potentially undesirable immunologic side effects (e.g., immunologic unresponsiveness or tolerance) in the infant. Nevertheless, the increase in availability of potentially protective, transplacentally transferred IgG through active maternal vaccination prompted the Institute of Medicine to recommend the establishment of a program of active immunization to control early-onset and late-onset group B streptococcal disease in both infants and mothers (Institute of Medicine and National Academy of Sciences, 1985). Efforts to develop safe, effective vaccines for protection from group B streptococcal infections have encountered the same difficulties with immunogenicity and safety observed in the development of other vaccines that induce protection from polysaccharide‑encapsulated organisms (Baker et al, 1988; Noya and Baker, 1992). The availability of conjugate vaccines, which include group B streptococcal polysaccharide antigens covalently linked either to tetanus toxoid or to a protein in the membrane of group B streptococci (beta C protein) (Madoff et al, 1994; Wessels et al, 1993; Yang et al, 2007) have shown some promise. 465 The indications for active vaccination during pregnancy rest on assessment of maternal risk of exposure, the maternalfetal-neonatal risk of disease, and the risk from the immunizing agents. (ACOG Technical Bulletin Number 160, 1993; Hackley, 1999; Stevenson, 1999). In general, immunization with live viral vaccines during pregnancy is not recommended (Box 36-1). Preferably, immunizations with live viral vaccines are performed before pregnancy occurs. However, rare instances may occur in which live viral vaccine administration is indicated. For example, if a pregnant woman travels to an area of high risk for yellow fever, administration of that vaccine might be indicated because of the susceptibility of the mother and the fetus, the probability of exposure, and the risk of the mother and fetus from the disease. More common examples in the United States include influenza and polio virus vaccination. If a chronic maternal medical condition would be adversely affected by influenza, active immunization may be indicated during pregnancy. Similarly, if imminent exposure to live polio virus in an unprotected woman is anticipated, live oral polio virus vaccine may be used during pregnancy. If immunization can be completed before the anticipated exposure, inactivated polio virus vaccine can be given. A summary of recommendations for immunizations during pregnancy is provided in Box 36-1. For the pediatrician, maternal immunization represents an important preventive intervention. Breastfeeding does not adversely affect immunization, and inactivated or killed vaccines pose no special risk for mothers who are breastfeeding or for their infants. Maternal immunizations with vaccines against polysaccharide‑encapsulated BOX 36-1 Summary of Recommendations for Immunization during Pregnancy LIVE VIRUS VACCINES Influenza (LAIV)—contraindicated ll Measles—contraindicated ll Mumps—contraindicated ll Rubella—contraindicated ll Yellow fever—safety not established (travel to high‑risk areas only) ll Varicella—contraindicated ll INACTIVATED VIRUS VACCINES Hepatitis A—consider if high risk of exposure ll Influenza—recommended ll Rabies—consider if indicated ll INACTIVATED AND RECOMBINANT BACTERIAL VACCINES Cholera—to meet international travel requirements ll Meningococcal polysaccharide vaccine—consider if indicated ll Plague—selective vaccination of exposed persons ll Typhoid—safety not established ll Pneumococcal polysaccharide vaccine—consider if indicated ll Tetanus‑diphtheria—consider if indicated ll Hepatitis B—consider if indicated ll Meningococcal conjugate vaccine—safety not established ll POOLED IMMUNE SERUM GLOBULINS ll ll Hepatitis A—postexposure prophylaxis Measles—postexposure prophylaxis Adapted from Centers for Disease Control and Prevention: Guidelines for Vaccinating Pregnant Women (May, 2007). Available at www.cdc.gov/vaccines/pubs/preg-guide.htm. LAIV, Live attenuated influenza vaccine. 466 PART IX Immunology and Infections organisms (e.g., Haemophilus influenzae type b and group B streptococci) can decrease morbidity and mortality from these diseases during the first 3 to 6 months of the infant’s life (Amstey et al, 1985; Baker et al, 1988; Colbourn et al, 2007; Walsh and Hutchins, 1989). The benefit of decreasing the risks of development of hepatocellular carcinoma, cirrhosis, and chronic active hepatitis from perinatal transmission of hepatitis B through prenatal screening and active and passive immunization of the infant is substantial (Arevalo and Washington, 1998). The implications of maternal vaccination during pregnancy for preterm infants have not been studied. The anthrax attacks of September 2001 have raised the possibility of the need for preexposure or postexposure immunization programs that may include pregnant women. Currently available data tend to suggest that the anthrax vaccine licensed since 1970 in the United States has no detrimental effects on pregnancy and does not increase adverse birth outcomes, although caution is advised when administering this vaccine during the first trimester (Inglesby, 2002; Ryan et al, 2008; Wiesen and Littell, 2002). A second infectious agent that might be used in a bioterrorist attack is variola virus (smallpox). Although eradicated in 1977, this virus remains a concern because of its lethality, especially in pregnancy (Enserink, 2002; Suarez and Hankins, 2002). Pregnancy is a contraindication to smallpox immunization (Wharton et al, 2003); however, if an intentional release of smallpox virus should occur, pregnant women should be immunized because of their high risk of mortality if unprotected (Suarez and Hankins, 2002). INFANT IMMUNIZATION Advances in understanding the developmental regulation of immunity have suggested that immunization during the neonatal period offers important advantages (Lawton, 1994). The recommendations of the American Academy of Pediatrics for immunization of infants can be found at http://aapredbook.aappublications.org/resources/IZ Schedule0-6yrs.pdf (Anonymous, 2002; Committee on Infectious Diseases et al, 2002). Immunization against H. influenzae type b should be considered for infants who are discharged from intensive care nurseries at or after 2 months of age. For the preterm infant, different clinical approaches to immunization are used by practitioners, including decreasing the dose of immunogen, postponing the first immunization until a corrected age of 2 months, or waiting for an arbitrary weight to be achieved by the infant (e.g., 4.5 kg). However, the American Academy of Pediatrics recommends administering full‑dose diphtheria, tetanus, and pertussis immunization beginning at 2 months of age. These recommendations (i.e., that no correction needs to be made for prematurity when initiating routine immunization in preterm infants) have been supported by longitudinal evaluation of serum antibody response in preterm infants (Bernbaum et al, 1989; Conway et al, 1993). The Advisory Committee on Immunization Practices (ACIP) of the United States Public Health Service recommends universal immunization of infants to protect against perinatal transmission of the hepatitis B virus (HBV) and chronic HBV infection (Mast et al, 2005). This recommendation has been endorsed by the Committee on Infectious Diseases of the American Academy of Pediatrics. Chronic HBV infection occurs in approximately 90% of infected infants and is associated with hepatocellular carcinoma and cirrhosis leading to end-stage liver disease. All medically stable infants born to women with a negative test result for hepatitis B surface antigen (HBsAg) and weighing more than 2000 g at birth should begin a hepatitis immunization schedule in the newborn period before hospital discharge. Only single-antigen hepatitis B vaccines should be used for the birth dose (Mast et al, 2005). All infants regardless of gestational age at the time of birth, and whose mothers have a positive test result for HBsAg, should receive passive immunization with hepatitis B immunoglobulin and active immunization with a single-antigen HBV vaccine within 12 hours of birth. Repeated vaccinations are given at 1 to 2 months and again at 6 months of age. For premature infants weighing less than 2000 g and born to HBsAg-positive mothers, the birth dose is not counted as part of the vaccine series. Special efforts should be made to complete the hepatitis B vaccination schedule within 6 to 9 months in populations of infants with high rates of childhood hepatitis B infection (Peter, 1994). For premature infants with birthweights of less than 2000 g born to HBsAg‑negative women, vaccination can be delayed until just before discharge. These infants do not need routine serologic testing for anti-HBsAg after the third dose of vaccine. The infant whose mother has a negative test result for HBsAg, but has received active or passive immunization during pregnancy because of exposure to hepatitis B, should receive no treatment as long as the mother was HBsAg‑negative at the time of birth. For current HBV vaccinations recommendations, the reader is directed to the Centers for Disease Control and Prevention Web site: www.cdc.gov/hepatitis/ HBV/VaccChildren.htm. In 1999, the U.S. Agency for Toxic Substances and Disease Registry raised concern about the possibility that administration of thimerosal-containing vaccines, including the hepatitis B vaccines, might lead to exposure to mercury that exceeded federal guidelines (Clark et al, 2001). Since March 2000, HBV vaccines used in the United States have not contained thimerosal. These yeast‑derived recombinant HBV vaccines contain 10 to 40 μg/mL of HBsAg protein, have excellent safety records, induce minimum adverse reactions, and are highly immunogenic (Greenberg, 1993). Multivalent vaccines are available, but are not recommended for the birth dose. Guidelines for hepatitis immunization of infants have been developed for implementation for term and preterm infants. However, vaccine responses in the extremely preterm infant whose mother is HBsAg‑positive, a population seen with increasing frequency because of the coincidence of intravenous drug abuse and carriage of HBsAg, have not been studied. Short-term and long-term neonatal immunologic protection after fetal exposure to maternally administered vaccines against anthrax or smallpox, the mass use of which might be prompted by a bioterrorist attack, has not been studied. Similarly, guidelines for neonatal immunization against these agents are not currently available. Before the eradication of smallpox, smallpox vaccine was routinely administered to older infants and children. No data CHAPTER 36 Immunology of the Fetus and Newborn or experience is available on immunization of newborn infants with anthrax vaccine. For postexposure prophylaxis in newborn infants, both chemotherapy (amoxicillin, ciprofloxacin, or doxycycline) and immunization should be considered. Clinicians faced with decisions concerning immunization strategies for pregnant women and newborn infants after a bioterrorist attack should consult the Web site for the Centers for Disease Control and Prevention (www.cdc.gov). SUGGESTED READINGS Chase NM, Verbsky JW, Routes JM: Newborn screening for T-cell deficiency, Curr Opin Allergy Clin Immunol 10:521-525, 2010. D’Argenio DA, Wilson CB: A decade of vaccines: Integrating immunology and vaccinology for rational vaccine design, Immunity 33:437-440, 2010. Koga K, Mor G: Toll-like receptors at the maternal-fetal interface in normal pregnancy and pregnancy disorders, Am J Reprod Immunol 63:587-600, 2010. 467 Leber A, Teles A, Zenclussen AC: Regulatory T cells and their role in pregnancy, Am J Reprod Immunol 63:445-459, 2010. Levy O: Innate immunity of the newborn: basic mechanisma and clinical correlates, Nat Rev Immunol 7:379-390, 2007. M’Rabet L, Vos AP, Boehm G, et al: Breast-feeding and its role in early development of the immune system in infants: consequences for health later in life, J Nutr 138:782S-1790S, 2008. Mor G, Cardenas I: The immune system in pregnancy: a unique complexity, Am J Reprod Immunol 63:425-433, 2010. PrabhuDas M, Adkins B, Gans H, et al: Challenges in infant immunity: implications for responses to infection and vaccines, Nat Immunol 12:189-194, 2011. Trowsdale J, Betz AG: Mother’s little helpers: mechanisma of maternal-fetal tolerance, Nat Immunol 7:241-246, 2006. Willems F, Vollstedt S, Suter M: Phenotype and function of neonatal DC, Eur J Immunol 39:26-35, 2009. Zaghouani H, Hoeman CM, Adkins B: Neonatal immunity: faulty T-helpers and the shortcomings of dendritic cells, Trends Immunol 30:585-591, 2009. Zhu J, Yamane H, Paul WE: Differentiation of effector CD4 T cell populations, Annu Rev Immunol 28:445-489, 2010. Complete references and supplemental color images used in this text can be found online at www.expertconsult.com C H A P T E R 37 Viral Infections of the Fetus and Newborn and Human Immunodeficiency Virus Infection during Pregnancy* Mark R. Schleiss and Janna C. Patterson Viral Infections of the Fetus and Newborn Viral infections of the fetus and newborn infant are common and underrecognized. Given the urgency of identifying invasive bacterial disease in the neonate, the identification of viral infections is often relegated to a matter of secondary importance. However, identifying viral infections can also be a matter of great urgency, because antiviral agents are available for many of the most common infections. Accordingly, an appropriately high index of clinical suspicion of a neonatal viral infection can be lifesaving. Moreover, identification of viral disease in the newborn period may be of great prognostic significance, particularly for neurodevelopmental issues. Making a diagnosis of a neonatal viral infection can help to direct and focus the anticipatory management of the child’s pediatrician during the course of well-child care. This chapter reviews the epidemiology, pathogenesis, diagnosis, and short- and long-term clinical management of many of the more common viral infections encountered in neonates. One of the great challenges in the evaluation of viral disease in the newborn is the ascertainment of the timing of acquisition of infection. Some infections can be acquired either in utero or in the early postnatal period. In this chapter, congenital infection is defined as any infection acquired in utero. Perinatal infections are defined as those acquired intrapartum, typically during the labor and delivery process. Postnatal infections are acquired in the post-partum period and are defined as infections acquired after delivery through the first month of life. In some situations, correctly identifying the timing of acquisition of infection can have substantial consequences, not only for the management of the infant, but for the long-term prognosis. Figure 37-1 outlines the most common timing of acquisition of neonatal viral infections with an emphasis on the relative importance of the many viruses that can be acquired congenitally, perinatally, or postnatally. There is considerable overlap across categories for some viral infections that can be transmitted at any of these time points, and this will be considered on a pathogen-by-pathogen basis. GENERAL DIAGNOSTIC APPROACH Clinicians caring for newborns have long recognized that there are some common clinical manifestations that suggest the presence of a congenital or perinatal viral *In writing this chapter, the authors excerpted portions of the chapters Identification, Evaluation, and Care of the Human Immunodeficiency Virus–Exposed Neonate by Karen P. Beckerman and Viral Infections of the Fetus and Newborn by Erica S. Pan, F. Sessions Cole, and Peggy Sue Weintrub from the previous edition of this book. 468 infection. These manifestations include evidence of intrauterine growth restriction, hydrops fetalis, hepatomegaly, splenomegaly, pneumonitis, bone lesions, rashes, and hematologic abnormalities (Box 37-1). Because the incidence of congenital viral infections is high (Alpert and Plotkin, 1986), it is appropriate for clinicians to have a high index of suspicion in any newborn with signs or symptoms. However, caution should be taken in efforts to lump all neonatal viral infections into a single diagnostic category. An unfortunate example of this manner of thinking is the continued use of the terminology TORCH titers in the diagnostic approach to a symptomatic neonate. The TORCH acronym, first coined by Nahmias et al (1971), stands for toxoplasmosis, other infections, rubella, cytomegalovirus (CMV), and herpes simplex virus (HSV). Numerous variants of this acronym have been suggested over the past four decades (Ford-Jones and Kellner, 1995; Kinney and Kumar, 1988; Ronel et al, 1995; Tolan, 2008). Unfortunately numerous clinical laboratories continue to offer the “TORCH panel,” typically consisting of serologic tests for toxoplasmosis, HSV, CMV, and rubella. This acronym has outlived its usefulness (Lim and Wong, 1994), and should be discarded from clinical parlance, based on the following considerations: ll Measurements of immunoglobulin G antibody titers virtually always simply reflect transplacental maternal antibody and provide little information of relevance to the infant’s infection status. With the exception of the identification of antibodies to Treponema pallidum, which is always of interest and significance, antibodies against the other members of the TORCH panel are of little diagnostic significance. ll Congenital and perinatal infection can occur with HSV and CMV, even in the face of preconception maternal immunity. Thus the finding of antibody in a TORCH titer is neither diagnostic of infection nor reassuring in regard to protection against that infection. ll Highly sensitive virologic and molecular tools are available to identify virtually all pathogenic viruses. These tools include standard culture and nucleic acid identification techniques, typically based on polymerase chain reaction (PCR) amplification of viral nucleic acids. Such studies can facilitate rapid pathogen-specific diagnosis; therefore the diagnosis of neonatal viral disease should depend on diagnostic virology, not serology. ll Most importantly, the use of the TORCH acronym vastly underemphasizes the great diversity of viral pathogens that have been associated with infection in the newborn. A list of the myriad of viral pathogens that have been reported to cause congenital infection and disease is included in Box 37-2. A discussion of CHAPTER 37 Viral Infections of the Fetus and Newborn and Human Immunodeficiency Virus Infection during Pregnancy Prenatal Perinatal/intrapartum Cytomegalovirus Parvovirus B19 Varicella-zoster virus Rubella LCMV* HSV** Parechovirus EBV*** HHV 6, 7 HSV Hepatitis B, C Enterovirus Varicella-zoster virus Cytomegalovirus Adenovirus Parechovirus Postnatal Respiratory syncytial virus Enterovirus Rotavirus Cytomegalovirus Varicella-zoster virus Hepatitis Adenovirus Influenza * Lymphocytic choriomeningitis virus ** Herpes simplex virus *** Epstein Barr virus FIGURE 37-1 Relative importance of neonatal viral infections related to timing of acquisition of infection. Viruses listed in declining relative order of importance relative to prenatal, perinatal (intrapartum), and postnatal timing of typical infection. Some neonatal virus infections (e.g., cytomegalovirus) can be substantial causes of disease whether acquired during gestation or postpartum, whereas others (e.g., respiratory syncytial virus) are typically acquired in the postnatal period. EBV, Epsteinn–Barr virus; HSV, herpes simplex virus; HHV, human herpes virus; LCMV, Lymphocytic choriomeningitis virus BOX 37-1 C  linical Features Commonly Associated With Congenital Viral Infections in Neonates ll ll ll ll ll ll ll ll ll ll ll ll ll ll ll ll ll ll Intrauterine growth restriction Nonimmune hydrops fetalis Echogenic bowel (prenatal ultrasound) Hepatosplenomegaly Jaundice (>20% direct-reacting bilirubin) Hemolytic anemia Purpura, ecchymoses and petechiae Skeletal defects (“celery-stalking”) Microcephaly and hydrocephaly Intracranial calcification Neuronal migration defects Pneumonitis Myocarditis Cardiac abnormalities Chorioretinitis Keratoconjunctivitis Cataracts Glaucoma many of the more unusual viral causes of congenital infection is beyond the scope of this chapter, but this table underscores the diversity of agents associated with fetal infection. A recent travel history or recent emigration might suggest consideration of some of these more unusual agents. Rather than rely on a large battery of serologic tests, the clinician can usually narrow the differential diagnosis of a suspect neonatal or congenital viral infection with the history and physical examination, followed by the use of focused, specific diagnostic studies. Important questions include: What was the overall health of the mother during her pregnancy? What is her age and marital status (e.g., young, unmarried women have a higher risk of 469 BOX 37-2 V  iral Pathogens Reported to Cause Congenital Infections Adenoviridae Adenovirus serogroup 3 Arenaviridae ll Lymphocytic choriomeningitis virus ll Lassa fever virus Bunyaviridae ll Bunyamwera serogroup (Cache Valley virus) ll La Crosse encephalitis virus Flaviviridae ll Hepatitis C virus ll Japanese encephalitis virus ll West Nile virus ll St. Louis encephalitis virus ll Yellow fever virus ll Dengue virus Hepadenoviridae ll Hepatitis B virus Herpesvirinae ll Herpes simplex viruses 1 and 2 ll Varicella zoster virus ll Cytomegalovirus ll Epstein–Barr virus ll Human herpesviruses 6 and 7 Orthomyxoviruses and paramyxoviruses ll Influenza ll Measles Parvoviridae ll Human parvovirus B19 Picornaviridae ll Poliovirus ll Coxsackievirus ll Enteric cytopathic human orphan virus ll Parechovirus ll Hepatitis A virus Retroviridae ll Human T-lymphotropic viruses 1 and 2 ll HIV Togavirinae ll Mumps ll Western equine encephalitis virus ll Venezuelan equine encephalitis virus ll acquiring primary genital HSV infection). What is her immunization history? Has she had chickenpox? Did she have other common childhood viral infections? What part of the world is she from? Are there potential animal exposures (e.g., exposure to cat litter or consumption of undercooked meat might suggest toxoplasmosis; exposure to rodents might suggest lymphocytic choriomeningitis virus)? Does she have other children and, if so, what are their ages, overall health status, and histories of group day care attendance? Have there been recent illnesses in the household? What time of year is it? (For example, respiratory syncytial virus and enterovirus infections have characteristic seasonality in temperate zones.) What are her occupational exposures? Did she have any symptomatic infectious illnesses during pregnancy? The answers to these types of questions, considered in the context of the infant’s physical examination, can direct the next steps in establishing a definitive etiologic diagnosis. 470 PART IX Immunology and Infections TABLE 37-1 Clinical Findings in Selected Congenital and ­ erinatal Infections That Suggest a Specific P ­Diagnosis Congenital Infection Findings Rubella Cataracts, cloudy cornea, pigmented retina; petechiae with “blueberry muffin” rash; bone defects with longitudinal bands of demineralization (“celery stalking”); cardiovascular malformations (patent ductus arteriosus, pulmonary artery stenosis); sensorineural hearing loss; hydrops Cytomegalovirus Microcephaly with periventricular calcifications; chorioretinitis; petechiae with thrombocytopenia; jaundice; sensorineural hearing loss; bone abnormalities; abnormal dentition, hypocalcified enamel Herpes simplex virus Skin vesicles, keratoconjunctivitis, acute central nervous system findings (seizures), hepatitis, pneumonitis Parvovirus B19 Hydrops, ascites, hepatomegaly, ventriculomegaly, hypertrophic myocardiopathy, anemia Varicella zoster virus Limb hypoplasia, dermatomal scarring in cicatricial pattern, gastrointestinal tract atresia Lymphocytic choriomeningitis virus Hydrocephalus, chorioretinitis, intracranial calcifications Some of the classic presentations of the more common perinatal viral infections are reviewed in Table 37-1. It should be noted that there can be considerable overlap of these clinical features across the different infectious categories listed; for example, the “blueberry muffin” rash of congenital rubella syndrome may be indistinguishable from that of congenital CMV, and both syndromes can include sensorineural deafness. The presence of brain calcifications is similarly nonspecific, and this finding should always suggest a differential diagnosis that includes CMV, toxoplasmosis, lymphocytic choriomeningitis virus, and the recently described parechoviruses. Because neuroradiologic studies cannot reliably distinguish these entities, definitive diagnostic virology is necessary. Specific viral pathogens, their basic virology, the clinical manifestations of diseases they cause in the newborn, management strategies, and prospects for prevention are considered on a pathogen-specific basis in the remainder of this chapter. HERPESVIRIDAE Currently there are eight recognized human herpesviruses, which are subdivided into three categories based on aspects of viral biology, pathogenesis, and clinical presentations. These categories are the α-herpesviruses, consisting of HSV-1, HSV-2, and varicella-zoster virus (VZV); the β-herpesviruses, which include CMV and the roseola viruses HHV-6 and HHV-7; and the γ-herpesviruses, which include Epstein–Barr virus (EBV) and Kaposi’s sarcoma herpesvirus (KSHV), reviewed in Schleiss (2009). Remarkably, all of these agents have been implicated in varying degrees as causes of clinically important congenital and perinatal infections, although these associations are less well studied for the γ-herpesviruses. HERPES SIMPLEX VIRUS INFECTIONS HSV-1 and HSV-2 are highly related viruses. Although classically HSV-1 has been identified as a cause of oral infections (gingivostomatitis and pharyngitis), and HSV-2 has been implicated as the most common virus associated with genital herpes, in recent years these distinctions have become blurred. The greatest risk for the newborn is in the context of a first-time episode of maternal genital HSV infection occurring during pregnancy. The entity of neonatal herpes is reviewed in the following section, along with current management approaches for this infection. VIROLOGY, EPIDEMIOLOGY, AND CLINICAL MANIFESTATIONS OF HSV DISEASE HSV-1 and HSV-2 demonstrate a high degree of similarity, both at the molecular level and in their clinical manifestations. The degree of genetic relatedness of these two viruses is approximately 45%, and the genome structures and morphology of the virion (virus particle) are virtually identical (Kieff et al, 1972). HSV-1 and HSV-2 are both acquired predominantly at mucosal surfaces and require intimate contact for transmission. After primary infection in epithelial cells, intraaxonal trafficking of viral DNA to the dorsal route ganglia results in the establishment of latency. The latent state is characterized by the cessation of virtually all gene transcription, except for the latencyassociated transcript (LAT), which is expressed in the dorsal route ganglia even as the virus maintains a quiescent state (Steiner et al, 2007; Taylor et al, 2002). The function of the latency-associated transcript is unknown, but it might involve a novel RNA-mediated mechanism, because there does not appear to be a protein product associated with the transcript (Umbach et al, 2008). After a number of triggers, including ultraviolet radiation, stress, and immunosuppression, the virus reactivates at the level of the dorsal route ganglia and initiates a cascade of viral transcription that leads to the production of infectious virus, which can traffic via the axon to the cutaneous surface or ocular surface, producing lesions (Toma et al, 2008). The recrudescence of HSV lesions, usually manifest as vesicular or ulcerative lesions at the site of primary infection, can in turn lead to person-to-person transmission, including maternal-fetal and maternal-infant transmission. Importantly, clinically evident lesions need not be present for person-to-person transmission of infection to susceptible individuals, because asymptomatic or subclinical shedding of virus is well documented, particularly in the setting of genital herpes. A wide variety of disease syndromes are associated with primary and recurrent HSV infection. Classically HSV-1 has been described as causing disease “above the belt,” whereas HSV-2 is associated with disease “below the belt.” The finding of HSV-2 antibodies in seroprevalence studies is generally viewed as indicative of genital herpes, and in the pregnant patient it can be considered to be diagnostic of genital herpes. The most common disease associated with HSV infection is herpetic gingivostomatitis, characterized CHAPTER 37 Viral Infections of the Fetus and Newborn and Human Immunodeficiency Virus Infection during Pregnancy typically by perioral and intraoral lesions involving the pharyngeal mucosa. HSV infection of the oropharynx may present in adolescence as herpetic pharyngitis (McMillan et al, 1993); it can be indistinguishable from other causes of pharyngitis. Other common manifestations of HSV infection include primary cutaneous infection, which can manifest as herpes gladiatorum or as a herpetic whitlow (Johnson, 2004; Wu and Schwartz, 2009). In children with atopic dermatitis, cutaneous HSV infection can be associated with eczema herpeticum, a serious illness that can be associated with systemic symptom (Bussman et al, 2008). All of these cutaneous manifestations of HSV could put a newborn at risk for acquisition of infection if care is not taken to protect the infant. HSV encephalitis is unlikely to be transmitted person-to-person; it is the most common sporadic cause of viral encephalitis in North America, and it can be associated with primary infection or reactivation of latent infection. It is most commonly associated with HSV-1 (Baringer, 2008). The most important manifestation of HSV disease is maternal genital herpes. Genital herpes can be associated with either HSV-1 or HSV-2. HSV is the most common cause of genital ulcerative disease in the developed world, and the prevalence has increased steadily in recent year (Fleming et al, 1997). Genital herpes is characterized by blisters, ulcers, or crusts on the genital area, buttocks, or both. Typically, symptomatic disease manifests with a mixture of vesicles, ruptured vesicles with resulting ulcers, and crusted lesions. Systemic flulike symptoms such as headache, fever, and swollen glands can accompany an outbreak of genital herpes, particularly during primary infection. Other symptoms include dysuria, urinary retention, vaginal or penile discharge, genital itching, burning or tingling, and groin sensitivity. Genital lesions vary in number, are painful in nature, and if untreated persist for up to 21 days (Whitley et al, 1998). It has been recognized in recent years that many individuals with genital herpes are asymptomatic and unaware of their status (Wald et al, 2000). Therefore a negative maternal history of HSV should not dissuade the clinician from considering the possibility of neonatal herpes in an infant with compatible signs and symptoms. Patients with recurrent symptomatic episodes continue to shed virus in between episodes, even after lesions have healed and crusted over (Leone, 2005). This information has important implications for the continued evolution of the HSV epidemic in the United States. Because individuals with asymptomatic genital herpes shed virus frequently in the absence of lesions, all HSV-2 seropositive individuals are probably at risk for transmitting infection in setting of sexual activity, labor and delivery, and intimate contact. NEONATAL HERPES For the neonatologist, the most important category of HSV-associated disease is that of neonatal herpes. In the United States, the reported range of neonatal herpes ranges from 1 in 2500 births to 1 in 8000 births (Corey and Wald, 2009; Kimberlin, 2005; Whitley, 2004). Approximately 70% to 85% of neonatal herpes simplex infections are caused by HSV-2 (Kimberlin et al, 2001). The majority of cases occur in infants born to women who were recently 471 infected, rather than to women with histories of recurrent genital herpes. Primary infection late in pregnancy poses a higher risk of transmission to the infant than do primary infections occurring before or early in pregnancy, suggesting that the evolution of a maternal antibody response confers some measure of protection for the infant (Brown et al, 2003; Caviness et al, 2008a). Primary genital herpes infection in a pregnant mother results in an attack rate of 33% to 50% for her infant, whereas recurrent maternal infection results in a 1% to 3% attack rate (Arvin, 1991; Brown et al, 1997; Prober et al, 1987). Overall, it is estimated that approximately 22% of pregnant women have genital herpes (HSV-2 seropositivity), and 2% of women will acquire HSV during pregnancy (Brown et al, 2005). Approximately 90% of these women are undiagnosed, because they are asymptomatic or have other subtle symptoms that are incorrectly attributed to other vulvovaginal disorders. Strategies for preventing neonatal HSV infection can be optimized by identifying women with genital lesions at the time of labor for cesarean delivery, prescribing antiviral suppressive therapy as appropriate (see Treatment and Outcomes, later), and minimizing invasive intrapartum procedures for women in whom the diagnosis of genital herpes cannot be excluded. Rarely, cases of intrauterine infection have been described; these are often associated with overwhelming primary maternal infection and usually result in fetal demise. On occasion, placentitis is also observed (Baldwin and Whitley, 1989; Chatterjee et al, 2001; Florman et al, 1973; Hutto et al, 1987; Vasileiadis et al, 2003). Although intrauterine transmission is possible, the vast majority of HSV infections in newborns are acquired intrapartum, related to the presence of virus in the maternal genital track. Neonatal acquisition of HSV from an individual other than the mother, such as from a recurrent oropharyngeal lesion, is unusual (Light, 1979; Linnemann et al, 1978; Yeager et al, 1983). Neonatal HSV has been described as a complication of ritual circumcision (Rubin and Lanzkowsky, 2000). Approximately 85% of infants with neonatal HSV acquire infection from the maternal genital tract at the time of delivery, but only 15% to 30% of mothers who give birth to infants with neonatal HSV infection have a known history of genital HSV (Whitley, 1988). Intrapartum interventions that have the risk of penetrating fetal skin, such as scalp electrode monitoring, increase the risk of transmission to the infant (Golden et al, 1977; Parvey and Ch’ien, 1980). Neonatal herpes can have devastating long-term consequences, making early recognition of paramount importance. Most infants with perinatal or postnatal HSV infection are normal at birth. Illness typically develops after 3 days of age; therefore the presence of skin lesions, oral ulcers, and other signs and symptoms in the first 72 hours of life should suggest diagnoses other than HSV. Premature infants appear to be at greater risk, possibly because of reduced transplacental transfer of protective antibody. Approximately 40% to 50% of affected infants are less than 36 weeks in gestational age (Whitley, 1988). Although a history of maternal cervical, vaginal, or labial lesions should be sought when neonatal HSV is being considered in the differential diagnosis, overt herpetic disease in the maternal genital tract is evident in approximately 472 PART IX Immunology and Infections • Sepsis syndrome • Hepatitis, DIC, pneumoia • Skin vesicles may be absent • High mortality even with therapy • Survivors may be free of CNS disease • Disease limited to skin, eyes, mucous membranes • May progress to CNS or disseminated disease • Frequent cutaneous recurrence in first year of life – increased risk of poor neurodevelopmental outcome Disseminated disease SEM disease CNS disease • Skin vesicles are absent in ~ of cases • Severe seizures common • High incidence of CNS imaging abnormalities and neurologic sequelae FIGURE 37-2 Characteristic presentations of neonatal herpes simplex virus (HSV) infection. Approximately 45% of neonatal HSV manifests as skin, eye, or mucous membrane disease (SEM disease); 25% as disseminated disease; and 30%, central nervous systems (CNS) disease. Characteristic features of each subtype of neonatal HSV are listed (arrows). Disease may span categories; for example, infants with SEM disease may progress to disseminated or CNS disease, and infants with CNS disease may develop skin vesicles later in hospital course, although up to one-third of infants with CNS disease never have cutaneous ­manifestations. one third of patients (Overall, 1994). In the remaining two thirds, infection is presumably via asymptomatic maternal genital tract shedding of virus. Infection can manifest in newborns in one of three forms: disease limited to the skin, eye, or mucous membrane disease (SEM); disease involving the central nervous system (CNS); or disseminated HSV infection, frequently manifesting as a sepsislike syndrome, with pneumonia, hepatitis, and viremia (Figure 37-2) (Kimberlin, 2005; Whitley, 2004). There can be overlap in these syndromes; for example, an infant with disseminated disease may initially have only skin lesions. The relative proportion of infants with disseminated disease has been declining in recent years, probably because earlier recognition and treatment of SEM disease have resulted in more timely intervention with antiviral therapy. Disseminated disease usually begins toward the end of the 1st week of life. Skin vesicles may be an early sign, but they are entirely absent in almost half of patients. The scalp should be inspected carefully, particularly near the site of insertion of fetal scalp electrodes, because such lesions are easy to overlook. Systemic symptoms, although initially insidious in onset, progress rapidly. Poor feeding, lethargy, and fever may be accompanied by irritability or seizures if the CNS is involved. These symptoms are followed rapidly by jaundice, hypotension, disseminated intravascular coagulation, apnea, and shock. This form of disease is indistinguishable at its onset from both neonatal enterovirus infection and bacterial sepsis. HSV infection should be considered in the differential diagnosis of infants who have fever during the first 2 weeks of life, because fever can herald the onset of systemic disease. Localized disease may begin somewhat later, with most cases appearing in the 2nd to 3rd weeks of life. When the CNS is the primary site of infection, the skin or eyes may be involved: up to one third of infants with neonatal CNS disease will never have skin lesions during their clinical course. The infants are lethargic, irritable, and tremulous, and seizures are common and difficult to control. Other less common but potentially localized or disseminated findings are keratoconjunctivitis, chorioretinitis, and pneumonitis, which can manifest as a focal infiltrate or as diffuse bilateral disease. Supraglottitis, intracranial hemorrhage, aseptic meningitis, and fulminant liver failure have been described (Abzug and Johnson, 2000; Erdem et al, 2002; Greenes et al, 1995; Kohl, 1994, 1999; Schlesinger and Storch, 1994). Less common presentations of neonatal herpes include hydrops fetalis (Anderson and Abzug, 1999) and laryngitis (Vitale et al, 1993). Diagnosis The cornerstone of the diagnosis of neonatal HSV is virologic detection; HSV serology is of little use. HSV-1 and HSV-2 are both easily recovered by culture of clinical samples. In disseminated disease, virus is present in blood, conjunctivae, respiratory secretions, and urine; it is also present in the central nervous system in approximately half of patients. In SEM disease, the virus can usually be found at the site of disease (i.e., within a vesicle). Definitive microbiologic diagnosis requires growth of the virus in tissue culture, or detection of viral nucleic acid by PCR. HSV can also be presumptively identified by immunofluorescence of infected cells using HSV-specific antibodies; such tests are commercially available and should be used in lieu of the outdated and insensitive Tzanck smear. Viral culture has the added advantage of allowing detection in the clinical virology laboratory of other agents that can mimic HSV disease in neonates, such as enteroviruses (see Enteroviruses later), because most diagnostic viral culture systems will support the growth of a variety of pathogens. When neonatal herpes is suggested, viral cultures of the throat, conjunctiva, blood, stool or rectum, and urine should be obtained, as should scrapings of vesicular, pustular, and ulcerative skin lesions. Of these sites, skin and conjunctival HSV cultures have the highest yield (Kimberlin et al, 2001). All infants with presumed neonatal HSV disease should undergo lumbar puncture, even if SEM disease is the only observed clinical manifestation. In some infants, CNS infection may be present but subclinical, and the finding of HSV DNA in the cerebrospinal fluid (CSF) has important therapeutic and prognostic implications. Blood should be sent for viral blood culture or PCR, because viremia is CHAPTER 37 Viral Infections of the Fetus and Newborn and Human Immunodeficiency Virus Infection during Pregnancy a common finding in neonatal HSV infection (Diamond et al, 1999; Stanberry et al, 1994). Usually if the CNS is involved in the setting of neonatal HSV disease, evaluation of the CSF reveals a lymphocytosis, red blood cells, normal or high protein level, and low or normal glucose level. In addition to viral culture, CSF, blood, skin lesions, and other specimens should be analyzed by PCR for the presence of HSV genome (Ryan and Kinghorn, 2006). Caution should be taken with interpretation of a negative CSF PCR result: many infants with neonatal HSV disease will not have CNS involvement, so a negative CSF PCR considered in isolation does not exclude the diagnosis of neonatal HSV. PCR of DNA extracted from the dried newborn blood spot has been reported as a way to retrospectively identify HSV infection in young infants (Lewensohn-Fuchs et al, 2003). Some experts recommend obtaining a second CSF specimen for evaluation at the end of antiviral therapy, and it has been reported that persistence of HSV DNA may be a poor prognostic factor and an indication for continuing antiviral therapy (Kimberlin et al, 1996b; Malm and Forsgren, 1999; Mejías et al, 2009). In disseminated disease, transaminase elevations consistent with hepatocellular injury are typically present; in severe disease, fulminant hepatitis with hepatic necrosis may be observed. Infants with CNS disease should undergo neuroradiographic imaging with computed tomography (CT) or magnetic resonance imaging (MRI). Early in the course of illness, imaging may demonstrate nonspecific lack of graywhite matter junction differentiation and general signs of encephalitis. Later CT findings include dilated ventricles, parenchymal echogenicity, cystic degeneration, and intracranial calcifications (O’Reilly et al, 1995), whereas MRI images demonstrate a variable appearance (Vossough et al, 2008). Neonatal HSV-2 encephalitis can be multifocal or limited to only the temporal lobes, brainstem, or cerebellum. Deep gray matter structures are involved and hemorrhage is observed in more than half of patients. In approximately 20% of patients, lesions are seen only by diffusion-weighted imaging. In 40% of patients, watershed distribution ischemic changes are also observed in addition to areas of presumed direct herpetic necrosis. In contrast to the classical description in older patients, neonatal HSV encephalitis is not usually confined to the temporal lobes. Electroencephalography should also be considered in all infants with CNS involvement or with disseminated disease, to evaluate for seizures. Electroencephalography findings will be abnormal in approximately 80% of such patients (Kimberlin et al, 2001). Treatment and Outcomes The cornerstone of the treatment of neonatal HSV disease is the nucleoside analog known as acyclovir. The development of acyclovir in the 1980s was a watershed event in the management of HSV infection. The use of acyclovir and other antivirals is summarized in Table 37-2. Although acyclovir had an efficacy similar to a previously used antiviral agent (vidarabine) in a controlled trial, acyclovir has emerged as the drug of choice because of its greater ease of administration and its highly favorable toxicity profile. The current recommendation of the American Academy of Pediatrics (AAP) Committee on Infectious Diseases is that 473 neonatal HSV infections should be treated with high-dose acyclovir at a dose of 60 mg/kg/day intravenously, divided into three doses given every 8 hours, for either 14 days for infants with SEM disease or 21 days for infants with disseminated or CNS disease. Support for this recommendation was derived from a comparison of the results from an earlier study using standard-dose acyclovir (30 mg/kg/ day) for 10 days (Kimberlin et al, 2001). There is no role for oral acyclovir in the management of neonatal HSV, and no role for topical acyclovir in neonatal SEM disease. Herpetic keratoconjunctivitis should receive topical ophthalmic antiviral therapy along with parenteral treatment. In addition to antiviral therapy, appropriate supportive care is essential, with anticipatory management targeting complications of neonatal HSV disease such as seizures, pneumonitis, and hepatic insufficiency. A beneficial role of intravenous immunoglobulin (IVIG) has been inferred from animal models of neonatal HSV disease, in which passive antibody is clearly beneficial (Bravo et al, 1996), but the absence of controlled studies precludes any recommendations in infants. Even with timely institution of antiviral therapy, the prognosis following neonatal HSV infection is guarded. The mortality rates in infants with localized CNS disease range from 4% to 14% with antiviral therapy, and most survivors have long-term neurologic sequelae. Risk factors for increased morbidity and mortality from CNS infection include prematurity and seizures upon initiation of therapy (Kimberlin et al, 2001; Kohl, 1999). Infants with disseminated disease have a high mortality rate without antiviral therapy; 80% die, and most survivors have serious neurologic sequelae (Whitley, 1988). Intravenous antiviral therapy has decreased mortality of disseminated disease to approximately 30%, and approximately 80% of surviving infants have a normal neurologic outcome (Corey and Wald, 2009). Antiviral therapy has decreased the mortality of neonatal CNS infection from approximately 50% to 6%, but more than 70% of survivors have sequelae (Corey and Wald, 2009; Engman et al, 2008; Freij and Sever, 1988). Lethargy at initiation of therapy has been associated with a higher mortality rate in neonates with disseminated HSV infection (Kimberlin et al, 2001). Infants with skin involvement often have recurrent crops of skin vesicles for several years. In an infant younger than 6 months, readmission to the hospital for diagnostic evaluation and administration of intravenous acyclovir is appropriate when cutaneous recurrences are observed. It is postulated that recurrent cutaneous lesions may be associated with subclinical CNS reactivation, and it has been proposed that treatment of these recurrences with acyclovir will improve long-term outcome. Management of asymptomatic neonates potentially exposed to HSV in the birth canal is controversial. This situation sometimes arises when a maternal perineal lesion is discovered after vaginal delivery. Some experts recommend neonatal surface cultures and administration of prophylactic antivirals, but there is little evidence to support this approach. The most important variable informing clinical management is probably the maternal history. Term infants exposed to HSV in the birth canal and born to women with long-standing histories of recurrent genital herpes are at low risk (<1%) of infection and, if the infant is 474 PART IX Immunology and Infections TABLE 37-2 Commonly Used Antiviral Agents in Neonatology Practice Dose, Route of Administration, ­Duration of Therapy Antiviral Agent Indication Acyclovir Neonatal HSV 60 mg/kg/day, dosing every 8 h IV; 21 days for disseminated infection or CNS disease; 14 days for SEM disease Comments Monitor CBC twice weekly; adjust dosage for renal insufficiency Oral suppression following neonatal HSV Efficacy for improving neurodevelopmental outcomes unproven; 10 to 20 mg/kg/dose twice daily; duration of therapy, 12 months Neutropenia observed in 1⁄2 to 2⁄3 of infants; lesions while receiving suppressive therapy should suggest acyclovir-resistant strain VZV infection 15 mg/kg every 8 h for minimum of 5 to 7 days; longer courses may be needed for severe ­end-organ disease (pneumonia, hepatitis) VZV postexposure prophylaxis 10 mg/kg PO four times per day; treat for at least 7 days (from 7 days after the earliest exposure until 14 days after the last exposure) Use in conjunction with VariZIG Trifluridine 1% HSV ophthalmic disease Apply as eye drops; 1 drop every 2 h to affected cornea while awake; maximum 9 drops per day Manage in consultation with ophthalmologist Ganciclovir Congenital CMV; acquired CMV 12 mg/kg/day, dosing every 12 h IV; duration of therapy is 6 weeks for prevention of hearing loss; shorter courses of therapy (14 to 21 days) are reasonable for serious end-organ disease Efficacy against CMV associated hearing loss in controlled trial; benefits of shorter courses of therapy unknown; neutropenia observed in 63% of patients in controlled trial; adjust dose for renal insufficiency; consider G-CSF if continued therapy desired in setting of neutropenia Valganciclovir Congenital CMV 32 mg/kg/day divided twice daily; no data available for duration or efficacy of oral formulation; CASG is currently performing a controlled clinical trial of 6 months oral suppression Valine ester (prodrug) of ganciclovir; similar toxicity profile as GCV; long-term suppressive therapy not well studied in infants; theoretical concerns of carcinogenesis, gonadal toxicity Lamivudine Hepatitis B, HIV For children 3 months to 16 years of age, ­recommended dose is 4 mg/kg, up to 150 mg per dose, twice daily Chronic hepatitis B infection; also used for HIV therapy Interferon α2b Hepatitis B, ­hepatitis C 3 to 6 million international units/m2 3 times weekly; up to 24 months duration; combined with oral ribavirin for hepatitis C Chronic hepatitis B infection; no data in neonates; chronic hepatitis C infection when administered with ribavirin; systemic side effects (fever, flu-like symptoms, anorexia); leucopenia; thyroid autoantibodies Pegylated interferon α2b Hepatitis B, ­hepatitis C 1.5 μg/kg once per week; no information on dosing in children <2 yr old Chronic hepatitis B; administer in conjunction with ribavirin for hepatitis C; systemic side effects (fever, flulike symptoms, anorexia); leucopenia; thyroid autoantibodies; side effects less common with pegylated formulations Adefovir Hepatitis B Not recommended in children <10 years of age; 0.3 mg/kg once daily PO in children 2 to 6 yr old has favorable pharmacokinetics Chronic hepatitis B infection; no safety or efficacy data in infants or young children Ribavirin (oral) Hepatitis C 15 mg/kg/day PO; no information on dosing in children <2 years of age Hemolytic anemia; teratogenic in animal models Ribavirin (aerosol) Respiratory ­syncytial virus Standard ribavirin aerosol therapy is 6 g per 300 mL water for 18 h daily; short-duration therapy, 6 g per 100 mL water given for a period of 2 h 3 times per day Not indicated for use with mechanical ­ventilator; conjunctivitis; bronchospasm Amantadine Influenza A FDA-approved dosage for children 1 to 9 yr old for treatment and prophylaxis is 4.4 to 8.8 mg/kg/day, not to exceed 150 mg/day in children <9 yr old Not approved for use in children <1 yr old; influenza B resistant; influenza A; H1N1 resistant Rimantidine Influenza A Administered in one or two divided doses at a dosage of 5 mg/kg/day, not to exceed 150 mg/ day in children <9 yr old Not approved for use in children <1 yr old; influenza B resistant; influenza A H1N1 resistant Oseltamivir Influenza A, ­influenza B For children <15 kg, recommended dose is 30 mg PO twice daily for 5 days (treatment) or 10 days (prophylaxis) Not typically recommended in children <1 yr old; FDA-approved guidelines for emergency use of oseltamivir in pediatric patients younger than 1 year in 2009 in response to H1N1 influenza pandemic Zanamivir Influenza A, influenza B Recommended dose of zanamivir for treatment of influenza is two inhalations (one 5-mg blister per inhalation for a total dose of 10 mg) twice daily (approximately 12 h apart) Not recommended in children <5 yr old CASG, Collaborative Antiviral Study Group; CBC, complete blood cell count; CNS, central nervous system; FDA, U.S. Food and Drug Administration; G-CSF, granulocyte colonystimulating factor; HSV, herpes simplex virus; IV, intravenous; PO, by mouth; SEM, skin, eye, or mucous membranes; VariZIG, varicella-zoster immunoglobulin; VZV, varicellazoster virus. CHAPTER 37 Viral Infections of the Fetus and Newborn and Human Immunodeficiency Virus Infection during Pregnancy asymptomatic, observation without antiviral therapy is sufficient. If the maternal HSV history is unclear, the infant is premature, or other obstetric complications or risk factors are present (e.g., prolonged rupture of membranes, maternal fever, signs or symptoms of chorioamnionitis, fetal scalp electrode monitoring), then empiric antiviral therapy is warranted. After the appropriate specimens are collected and sent for viral culture and PCR analysis, the infant should receive parenteral acyclovir (60 mg/kg/day) pending the results of diagnostic virology studies. There is no role for topical or oral formulations of acyclovir in this setting. Similarly controversial is the question of whether empiric acyclovir therapy should be administered to neonates who are readmitted for evaluation of febrile illness in the first 30 days of life. Standard practice in most children’s hospitals is to rule out sepsis in this setting by administering broad-spectrum antibiotics pending the result of diagnostic cultures of blood, urine, and spinal fluid. Some experts recommend the use of empiric acyclovir in this setting (Long, 2008), citing data that indicate that the prevalence of neonatal HSV infection is similar to that of invasive bacterial infection in this setting, and that infants may exhibit fever and sepsis-like syndrome as the only manifestations of HSV infection (Caviness et al, 2008b). Other experts recommend a more selective approach, based on analysis of history, risk factors, and laboratory and radiographic analyses, such as liver function tests and chest radiograph (Kimberlin, 2008). Additional studies will be required to ascertain whether acyclovir should be included in the empiric antimicrobial regimen of all neonates who are evaluated for fever in the first few weeks of life. In addition to its critical role in the management of the acute clinical syndromes of neonatal HSV infection, acyclovir may be of benefit when administered as long-term suppressive therapy in the first 6 months of life. The Collaborative Antiviral Study Group (CASG) has reported results of controlled studies evaluating long-term oral suppressive acyclovir therapy after neonatal HSV aimed at both prevention of recurrent skin lesions and neurologic sequelae. The observation driving this study is that recurrent skin lesions (more than three episodes) within the first 6 months of life predict an adverse neurologic prognosis, possibly because recurrent skin lesions are a surrogate marker for subclinical reactivation events in the CNS (Whitley, 1991). Phase I and II trials demonstrated fewer cutaneous recurrences in the treatment group, but did not have enough subjects to analyze the effect on neurologic outcomes (Kimberlin et al, 1996a). Whether reducing cutaneous recurrences in the 1st year of life will result in improved neurodevelopmental outcome remains unknown (Gutierrez and Arvin, 2003). Given the favorable safety profile of acyclovir, the use of long-term suppression after an episode of neonatal herpes is probably an appropriate management strategy, and it is likely to reduce the need for repeat hospitalization when lesions reappear. Any theoretical benefit of acyclovir must be considered against the potential risk of neutropenia, which has been reported in 20% to 50% of neonates receiving long-term acyclovir therapy. Although readily reversible with discontinuation of drug, this concern necessitates frequent monitoring of the absolute neutrophil count. The emergence of 475 acyclovir-resistant HSV strains is also a potential concern (Kimberlin et al, 1996a; Oram et al, 2000). That long-term oral suppressive therapy may be beneficial for improving long-term neurodevelopmental outcomes owing to neonatal HSV infection was suggested by a 2-year pilot study of oral suppressive therapy in a cohort of 16 infants (Tiffany et al, 2005). In this uncontrolled study, all children were independently mobile, free of seizures, and had normal vision and speech development at the time of final neurodevelopmental assessment. Prospects for Prevention Even with prompt recognition and therapy, the risk of long-term morbidity with neonatal HSV infection, particularly neurodevelopmental morbidity, remains significant. Ideally, prevention of neonatal HSV infection would be the best approach to solving the problem of HSV-induced long-term neurologic injury. Recent developments in the approach to HSV infections among women of childbearing age include consideration of maternal screening programs and the use of antiviral agents in women at risk for transmission of infection. The availability of reliable kits that allow specific serodiagnosis of HSV-1 and HSV-2 infections has enabled the identification of asymptomatic women with genital herpes who can be counseled appropriately during pregnancy (Sauerbrei and Wutzler, 2007). Although routine seroscreening of pregnant women is not currently considered a standard of care, screening in some instances, including situations where high maternal anxiety exists, seems justified. There is no evidence that routine administration of suppressive antivirals during pregnancy can improve outcomes or reduce disease in newborns (Sheffield et al, 2006), although this approach has become common in many obstetric practices (Hollier and Wendel, 2008). One benefit of suppressive antiviral therapy during pregnancy for a woman with a history of genital herpes may be a reduced likelihood of cesarean section, predicated on the assumption that suppressive therapy will prevent reactivation of HSV (Brown et al, 2003; Hollier and Wendel, 2008). Strategies for preventing neonatal herpes can also target prevention of intrapartum transmission by caesarean section. If a mother has active genital herpes simplex infection at the time of delivery, and if the membranes are either intact or have been ruptured for less than 4 hours, both the AAP and the American College of Obstetricians and Gynecologists (ACOG) recommend cesarean delivery. The role of intrapartum antivirals or immunoglobulin has not been studied in this setting. Even with careful histories and meticulous physical examination during labor and delivery, many at-risk deliveries cannot be predicted or identified, because so many HSV-2–seropositive women are asymptomatic, do not know that they have genital herpes, and may unknowingly shed virus at the time of delivery. Infants in whom HSV infection is known or highly suspected should be put in isolation with contact precautions, and skin lesions should be covered. Finally, any health care provider with active herpetic whitlow or other skin lesions should not have direct patient care responsibilities for neonates. Ultimately, prevention of neonatal HSV infection could be conferred by the development of a vaccine. Several 476 PART IX Immunology and Infections phase I, II, and III clinical trials are ongoing to investigate the utility of various HSV vaccines to prevent genital infections (Stanberry and Rosenthal, 2005). A double-blind, randomized trial of an HSV-2 glycoprotein-D-subunit vaccine given with alum adjuvant and 3-O-deacylatedmonophosphoryl lipid A, was reported in a study of individuals whose regular sexual partners had a history of genital herpes. The primary end point was the occurrence of genital herpes disease. The vaccine was efficacious in women who were seronegative for both HSV-1 and HSV2, but it was not efficacious in women who were seropositive for HSV-1 and seronegative for HSV-2 at base line. The vaccine had no efficacy in men, regardless of serostatus (Stanberry et al, 2002). The basis for this discrepancy remains unexplained, but may relate to anatomic differences in how primary infections are established in men and women, or sex-related differences in immune response. It remains to be studied what effect widespread HSV-2 vaccination might have on the incidence or severity of neonatal HSV infection. VARICELLA-ZOSTER VIRUS Varicella-zoster virus (VZV) is a member of the α-herpesvirus subfamily of the Herpesviridae. Like the related HSV-1 and HSV-2, VZV can infect neurons where it can establish latent infection. Primary varicella infection, commonly known as chickenpox, usually results in a fever and a characteristic vesicular exanthem. The illness often includes other systemic symptoms, such as headache and malaise. Reactivation from latency, which can occur decades after the primary VZV infection, is usually referred to as zoster or shingles. Zoster is characterized by a painful vesicular rash in a dermatomal distribution, and in older patients can lead to postherpetic neuralgia. The neonatologist may encounter consequences of maternal VZV infection in two different clinical presentations. In congenital varicella, VZV is transmitted to the fetus in the first or second trimester of pregnancy, where it can produce a number of teratogenic consequences (Auriti et al, 2009; Laforet and Lynch, 1947; Smith and Arvin, 2009; Srabstein et al, 1974). In contrast, neonatal varicella occurs in the setting of primary maternal varicella acquired late in the third trimester, and the affected infant can exhibit symptoms and signs in the neonatal period. This section reviews both of these presentations of VZVrelated disease in infants. Epidemiology of Maternal and Perinatal Varicella-Zoster Virus Before the advent of routine childhood vaccination against chickenpox, it was estimated that the VZV seroprevalence in women of childbearing age was greater than 95%, because of the formerly ubiquitous nature of this infection. In this era, primary VZV infections during pregnancy occurred with a frequency of 5 to 7 per 10,000 pregnancies in the United States (Balducci et al, 1992; Brunell, 1992). The major concern in the setting of primary maternal VZV infection is the risk for congenital varicella syndrome (CVS). It is not yet clear whether the risk for CVS has been ameliorated by the widespread use of VZV vaccine. For pregnant women with primary varicella infection, the transmission rate to the fetus is estimated to be approximately 25%. Only a subset of infected fetuses exhibit symptomatic disease. Approximately 100 cases of CVS have been reported in the literature (Sauerbrei and Wutzler, 2000). The risk of symptomatic intrauterine VZV infection after maternal varicella occurring during the first 20 weeks of pregnancy is approximately 1% to 2% (Enders et al, 1994; Paryani and Arvin, 1986; Pastuszak et al, 1994; Siegel, 1973). Symptomatic disease appears to be more common in female infants (Sauerbrei and Wutzler, 2000). CVS does not appear to occur in the setting of maternal zoster. A prospective study of infants born to 366 mothers with a clinical history of zoster during pregnancy found no infants with CVS (Enders et al, 1994). Maternal infection in the third trimester is not associated with CVS, presumably because this falls outside the time frame when teratogenicity can occur. Maternal infection just before or after delivery poses a high risk for neonatal varicella. Before the advent of VZV vaccination, neonatal varicella was encountered much more commonly that CVS. For infants in which maternal illness begins 5 days or less before delivery or up to 2 days after delivery, the infant attack rate is 17% to 31% (Brunell, 1992; Feldman, 1986; Meyers, 1974). Because the incubation period for varicella is between 10 and 21 days, cases beginning in the first 10 days of life are considered to have been acquired in utero. Pathogenesis and Clinical Manifestations As noted, CVS is acquired from a maternal primary varicella infection that occurs during the first or second trimester. The virus is thought to be transmitted transplacentally during the viremia that precedes or accompanies the rash of chickenpox. Ascending infection from cervical infection has been proposed as a potential mechanism of transmission, but is probably much less common (Sauerbrei and Wutzler, 2000). Some of the congenital malformations associated with CVS may be a consequence of zosterlike virus reactivation events in the infected fetus rather than the direct effects of the primary viral infection. This explanation is supported by the common finding of unusual cicatricial rashes in dermatomal distributions in the newborn with CVS. Other clinical manifestations include asymmetric muscular atrophy with limb hypoplasia, low birthweight, neurologic abnormalities (cortical or spinal cord atrophy, seizures, microcephaly, encephalitis, Horner syndrome), and ophthalmologic abnormalities (chorioretinitis, microphthalmia, atrophy, and cataracts; Brunell, 1992; Feldman, 1986; Sauerbrei and Wutzler, 2000). Gastrointestinal abnormalities are reported in 15% to 23% of cases; findings include duodenal stenosis, dilated jejunum, small left colon, intestinal atresia or bands, and hepatic calcifications (Alkalay et al, 1987; Jones et al, 1994). Immature fetal cell–mediated immune response may explain the short latency period and the inadequate protection from the consequences of episodes of reactivation in utero (Higa et al, 1987; Kustermann et al, 1996). Pathology reports have noted destruction of neural tissue with residual dystrophic calcifications, chronic active inflammation in nonneural tissues surrounding viral inclusions, and evidence of CHAPTER 37 Viral Infections of the Fetus and Newborn and Human Immunodeficiency Virus Infection during Pregnancy chronic placental villitis (Bruder et al, 2000; Petignat et al, 2001; Qureshi and Jacques, 1996). Maternal varicella near term or immediately postpartum can lead to neonatal varicella. Neonatal varicella is usually caused by maternal chickenpox acquired during the last 3 weeks of pregnancy. Infection may be transmitted perinatally by transplacental viremia (most cases) or by ascending infection from the birth canal; it can also be treated postnatally by the aerosol route or direct contact with infectious lesions. Serious postnatal infection acquired from maternal varicella via breastfeeding has not been reported. Neonatal varicella may develop despite the administration of varicella-zoster immunoglobulin (VariZIG) to the infant at birth. Transplacentally transmitted infections occur in the first 10 to 12 days of life, whereas chickenpox after that time is most likely acquired by postnatal infection. The clinical presentation differs markedly for cases in which maternal rash began 5 days or more before delivery from those in which maternal illness occurred from 5 days before to 2 days after delivery. When maternal varicella is noted more than 5 days before delivery, neonatal disease usually begins within the first 4 days of life and is typically mild. In contrast, neonatal varicella in an infant whose mother develops varicella from 5 days before until 2 days after delivery has a high risk of morbidity and mortality (Isaacs, 2000; Tan and Koren, 2006). In this second group, neonatal disease typically begins between 5 and 10 days after delivery, and a fatal outcome has been reported in 23% to 30% of cases (Brunell, 1966; Sauerbrei and Wutzler, 2001). When the disease appears in this setting, it closely resembles varicella in the immunodeficient or immunosuppressed host. Recurrent crops of skin vesicles develop over a prolonged period. Typical presenting signs are fever, hemorrhagic rash, and visceral dissemination with involvement of the liver, lung, and brain. Secondary bacterial infection may occur. Prospective studies of the long-term outcomes of CVS have not been performed, and there is probably a wide continuum of disease. Among the 96 infants reviewed by Sauerbrei and Wutzler (2000), 14 (15%) had clinical signs of zoster during early infancy; these infants had been exposed to varicella between 8 and 24 weeks’ gestation. Some experts consider early zoster infections as one criterion for diagnosis of CVS. Between 1% and 2% of infants without clinical evidence of CVS, but whose mothers had chickenpox in the second and third trimesters, develop zoster in the first few weeks of life; if this occurs, consideration should be given to ophthalmologic evaluations to rule out the possibility of CVS (Enders et al, 2004; Sauerbrei and Wutzler, 2000). Overall mortality rates for CVS are estimated at 30%. Deaths occur in the first few months of life, usually secondary to severe pulmonary disease (Pastuszak et al, 1994; Sauerbrei and Wutzler, 2000). Diagnostic Studies Prenatal diagnosis—by means of quantification of varicella-specific IgM on fetal blood obtained by cordocentesis or through PCR analysis of chorionic villi, fetal blood, and amniotic fluid—has been attempted (Cuthbertson et al, 1987; Kustermann et al, 1996; Mouly et al, 1997). Although the yield of PCR on amniotic fluid or fetal 477 blood is higher than that of serologic analysis of fetal IgM or viral cultures, large prospective studies of the correlation of such findings with clinical outcomes have not been performed. Reported prenatal ultrasonographic findings include polyhydramnios, hydrops, progressive intrauterine growth restriction, microcephaly, limb hypoplasia, and liver hyperechogenicities (Petignat, 2001; Pretorius et al, 1992). Abnormal ultrasonographic findings might not develop in a fetus for at least 5 weeks after maternal infection (Kerkering, 2001). Serologic studies can be performed postnatally on infants (Paryani and Arvin, 1986). Suspected congenital infection with varicella can be confirmed by the finding of persistent VZV IgG beyond the presumed duration of passive transfer of maternal antibodies (at least 6 to 7 months). Detection of fetal IgM can also confirm infection, but it is less useful because approximately one fourth of infants reported with classic CVS have positive VZV IgM titer values (Enders et al, 1994; Sauerbrei and Wutzler, 2000). Scrapings of skin lesions, as with herpes simplex infections, can show large multinucleated cells when stained with Wright or Giemsa stain (Tzanck smears), but this procedure is no longer recommended, because highly sensitive direct fluorescent antibody (DFA) tests are readily available and are more sensitive and specific (Chan et al, 2001; Coffin and Hodinka, 1995). VZV can also be detected from skin lesions by PCR (Leung et al, 2010). Although VZV can be detected by DFA staining of cells or PCR of DNA extracted from skin and visceral lesions, these tests in infants with suspected congenital infection are often extremely low yield compared with infants and children who have acquired varicella infection. Neuroradiographic demonstration of intracranial calcifications has been reported with CVS, but this is not a common finding (Kerkering, 2001). Treatment No controlled studies examining the effect of antiviral therapy to prevent or treat congenital varicella syndrome have been conducted. Some experts have recommended oral acyclovir for pregnant women with varicella, especially during the second and third trimesters (AAP Committee on Infectious Diseases, 2010a). Intravenous treatment with acyclovir for the infected pregnant woman is recommended for patients with serious complications of varicella, particularly pneumonia. Anecdotal observations suggest a potential effect of acyclovir on the progression of eye disease in CVS (Sauerbrei and Wutzler, 2000). However, there are no recommendations for the use of acyclovir or immunoglobulin for treatment or prevention of CVS. For neonatal varicella, treatment with intravenous acyclovir, 60 mg/kg/day divided into doses every 8 hours, is recommended, particularly for infants at the highest risk of adverse outcomes (i.e., the infant born to a woman who develops varicella from 5 days before until 2 days after delivery; see Table 37-2). Prevention If a susceptible pregnant woman has a significant exposure to varicella, administration of VariZIG to her and her newborn infant should be considered seriously. These 478 PART IX Immunology and Infections BOX 37-3 C  andidates for VariZIG* after Significant Exposure† to VaricellaZoster Infection in Perinatology and Neonatology Practice ll ll ll ll Pregnant women without evidence of immunity Newborn infant whose mother developes chickenpox within 5 days before delivery or within 48 hours after delivery Hospitalized premature infant (≥28 weeks’ gestational age) whose mother has no history for chickenpox or serologic evidence of prior infection Hospitalized preterm infants (<28 weeks’ gestational age or birthweight ≤1000 g) regardless of the maternal history of varicella or varicella-zoster virus serostatus *Varicella-zoster immunoglobulin (VariZIG) is given intramuscularly at a recommended dose of 125 units per 10 kg body weight, up to a maximum of 625 units. If VariZIG is unavailable, immunoglobulin (400 mg/kg, intravenously) can be substituted. †Significant exposures include: household exposure; face-to-face indoor play; face-to-face contact in hospital setting with infectious staff member, patient, or visitor; intimate contact (hugging or touching) with patient with active zoster lesions. recommendations are summarized in Box 37-3. Infants of mothers in whom varicella develops from 5 days before to 2 days after delivery should receive VZIG as soon as possible (AAP Committee on Infectious Diseases, 2010a). VariZIG is given intramuscularly at a recommended dose of 125 units per 10 kg bodyweight, up to a maximum of 625 units. If VariZIG is unavailable, IVIG (400 mg/kg) can be substituted. Passive immunoprophylaxis has been shown to prevent chickenpox in exposed older children (Brunell et al, 1969), but does not always prevent neonatal disease (Reynolds et al, 1999). Approximately 50% of exposed infants treated with immunoglobulin can still develop varicella, but the disease is often attenuated, and approximately 10% have severe disease (Hanngren et al, 1985). For healthy term infants exposed postnatally to varicella, including infants whose mother’s rash began more than 48 hours after delivery, VariZIG is not generally indicated. VariZIG is not indicated for an infant whose mother has zoster. Breastfeeding is not contraindicated. Follow-up of infants exposed to VZV and treated with VariZIG can include consideration of serologic testing (enzyme immunoassay, latex agglutination, or indirect fluorescent antibody staining for IgG) to determine whether asymptomatic infection has elicited immune protection. Some experts recommend repeated administration of VariZIG after a repeated exposure of an infant in whom varicella did not develop more than 3 weeks after administration of the initial dose of VariZIG, although the risk to these infants is less well defined. Infants receiving VariZIG should also be placed in respiratory isolation for 28 days or until discharge, because administration of VariZIG can prolong the incubation period. In the event of a significant varicella exposure in a nursery situation, infants whose mothers have no history of chickenpox and who have undetectable antivaricella antibody titers should be considered candidates for VariZIG. All exposed infants less than 28 weeks of gestational age or birthweight less than 1000 g, regardless of maternal history, should receive VariZIG (see Box 37-3) (AAP Committee on Infectious Diseases, 2010a). The decision to use VariZIG in the premature infant greater than 28 weeks’ gestation should be predicated on maternal history of chickenpox or serologic evidence of protection. As of late 2010, VariZIG is available under an investigational new drug protocol and can be obtained by calling FFF Enterprises (800-843-7477; www.fffenterprises. com/Products/VariZIG.aspx). Acyclovir has also been recommended by some experts for postexposure prophylaxis in the setting of a nursery outbreak of VZV (Hayakawa et al, 2003; Shinjoh and Takahashi, 2009). A suggested dose is 10 mg/kg by mouth, four times per day for 7 days. All exposed health care professionals without evidence of immunity should be excused from patient contact from day 8 to 21 after exposure to an infectious patient, or to day 28 if the individual has received VariZIG (AAP Committee on Infectious Diseases, 2010a). The best means of prevention is to follow current recommendations for universal varicella immunization of all children at 12 to 15 months of age, as well as vaccinating all susceptible adolescents and adults at high risk of exposure to varicella. VZV vaccine is available in three formulations: a monovalent formulation for children and young adults, a combination vaccine given with the measles-mumps-rubella vaccine in children and young adults, and a monovalent formulation for adults older than 60 years for prevention of herpes zoster (shingles). There is no evidence that CVS occurs after exposure to varicella vaccine during pregnancy. From March 17, 1995, through March 16, 2005, 981 women were enrolled in a pregnancy registry for women exposed to varicella vaccine (Shields et al, 2001; Wilson et al, 2008). Pregnancy outcomes were available for 629 prospectively enrolled women. Among the 131 live births to VZV-seronegative women, there was no evidence of congenital varicella syndrome. Nonetheless it is recommended that adolescents and women of childbearing age should avoid pregnancy for at least 1 month after immunization. CYTOMEGALOVIRUS CMV infection is ubiquitous in the general population and generally produces few if any symptoms in the immunocompetent infant, child, or adult. The mild nature of primary infection in most persons belies the severe nature of CMV-induced illness in those with impaired, suppressed, or immature immune systems, including infected newborns. Among the perinatally acquired viral infections in the developed world, CMV imposes the largest economic burden and produces the greatest long-term neurodevelopmental morbidity. This section focuses on the epidemiology, pathogenesis, diagnosis, and therapeutic management of maternal, congenital, and perinatal CMV infections. Epidemiology In retrospect, the first description of congenital CMV disease was in 1904, when Ribbert observed the large inclusion-bearing cells that represent the typical histopathologic finding of CMV end-organ disease in a stillborn infant. In 1920 a viral cause was proposed for the “cytomegaly” seen in tissue sections of these inclusion-bearing cells (Goodpasture and Talbot, 1921), and it would be several more decades before the ubiquitous nature of this virus CHAPTER 37 Viral Infections of the Fetus and Newborn and Human Immunodeficiency Virus Infection during Pregnancy Recurrent infection 0.5% to 3% transmission rate Primary infection Congenital cytomegalovirus infection 85% to 90% asymptomatic at birth 30% to 50% transmission rate 10% to 15% symptomatic at birth Asymptomatic Symptomatic 10% to 15% of asymptomatic infants have sequelae 60% to 90% of symptomatic infants have sequelae Sequelae • Sensorineural hearing loss • Mental retardation • Seizure disorders • Cerebral palsy • Visual deficits • Developmental delay FIGURE 37-3 Profiles of congenital cytomegalovirus epidemiology, infection, and outcome. Transmission rates to the fetus are highest in the setting of primary maternal infection (up to 50%), although 0.5% to 3% of women with preconception immunity may nonetheless transmit cytomegalovirus (CMV) because of reinfection or reactivation of latent infection. Among all infants with congenital CMV infection, regardless of maternal immune status during pregnancy, approximately 10% to 15% have symptoms or signs at birth (e.g., microcephaly, chorioretinitis, hepatosplenomegaly, petechiae, purpura, thrombocytopenia, hepatitis, seizures, pneumonitis). Symptomatic infants have the highest risk of neurodevelopmental sequelae, although any infant with congenital CMV is potentially at risk for sequelae. Among asymptomatic congenitally infected infants with sequelae, the most common manifestation is sensorineural hearing loss, which may not be present at birth. and the depth and breadth of the pathology it produces would be elucidated. In the developed world, CMV transmission occurs in 0.5% to 2% of all live births, making it the most common congenital viral infection (Demmler, 1991). Approximately 40,000 infants are born with CMV infection every year in the United States. Consequently, congenital CMV has a much greater impact than other more commonly recognized causes of birth defects in newborns. More than 8000 children per year will be permanently disabled by congenital CMV infection (­Figure 37-3). This number is higher than those affected by other, better known childhood conditions such as Down syndrome, fetal alcohol syndrome, and spina bifida (Ross et al, 2006). Rates of congenital CMV infection tend to parallel those of maternal seropositivity and vary substantially across populations (Mustakangas et al, 2000). Although the lifetime risk of acquiring CMV infection is high, approaching 90% by the eighth decade of life (Staras et al, 2006), seroprevalence is substantially lower among women of childbearing age. Seronegative women are therefore at risk for acquiring primary infections during pregnancy. Primary infections pose an increased risk of transmission to the fetus, and possibly a higher risk of sequelae. Every year in the United States, approximately 27,000 seronegative women acquire a primary CMV infection (Colugnati et al, 2007). 479 Seroprevalence rates for CMV vary significantly globally and are generally inversely correlated with socioeconomic status. CMV seroprevalence is greater in childhood in developing countries. In developed countries, seroprevalence tends to be higher in blacks and Hispanics, low-income groups, and persons without higher education. Young maternal age, single marital status, and non-­Caucasian race are associated with higher rates of congenital CMV infection. Women with increased occupational exposure to young children (including daycare providers) appear to be at elevated risk for primary CMV infections (Adler, 1989; Pass et al, 1986, 1990; Stagno and Britt, 2006; Stagno and Cloud, 1994). Health care providers in contrast are not at increased risk for acquisition of a primary CMV infection (Dworsky et al, 1983). Pathogenesis Morphologically and at the genome level, CMV is the largest pathogen (Schleiss, in press). There are approximately 160 known CMV genes; although based on its coding potential, CMV has the potential to encode more than 250 open reading frames. The mechanisms by which CMV injures the fetus are complex and involve a complex interplay of viral gene products, maternal immune response, and placental biology. CMV encodes genes modifying the cell cycle, cellular apoptosis mechanisms, inflammatory responses, and evasion of host immune responses. The pathogenesis of fetal injury in the setting of congenital CMV infection is incompletely understood. The pathogenesis of disease associated with acute CMV infection has been attributed to lytic virus replication, with end-organ damage occurring either secondary to virus-mediated cell death or from pathologic host immune responses targeting virus-infected cells (Britt, 2008; Schleiss, in press). Factors that contribute to fetal injury include the timing of infection relative to the gestational age of the fetus (Pass et al, 2006), the maternal immune status (Fowler et al, 1992), the extent of associated placental injury (Fisher et al, 2000), the magnitude of the viral load in the amniotic fluid (Lazzarotto et al, 2000), the induction of host genes occurring in response to infection (Challacombe et al, 2004), and possibly the genotype of the particular strain of CMV infecting the fetus (Arav-Boger et al, 2006). The relative contribution of maternal, placental, and fetal compartments in the pathogenesis of disease remains incompletely defined. Much of the injury that CMV produces in the newborn may be caused by placental insufficiency and not by viral infection of the fetus (Schleiss, 2006a). Delivery of oxygen, substrate, and nutritional factors to the fetus is impaired for a CMV-infected placenta. Moreover, CMV infection of the placenta might contribute to intrauterine growth restriction and fetal injury via induction of proinflammatory cytokines and modulation of normal trophoblast gene expression (Chan and Guilbert, 2005; Chou et al, 2006; La Torre et al, 2006; Maidji et al, 2007; Yamamoto-Tabata et al, 2004). Identified as a site for major pathogenic virus-induced injury, the placenta is increasingly considered to be a target for novel therapeutic interventions, such as CMV hyperimmune globulin for women with primary CMV infection during pregnancy (La Torre et al, 2006). 480 PART IX Immunology and Infections Congenital CMV infection appears to be pathologically more severe in the setting of primary maternal infection, which confers a 40% to 50% risk of intrauterine transmission during gestation, versus the 0.5% to 2% transmission risk in women with preconceptional immunity (Fowler et al, 2003). Congenital CMV infection in previously immune mothers appears to be related to reinfection with new strains of virus, with variations in epitopes of virally encoded proteins that may correlate with decreased maternal immune (Boppana et al, 1999, 2001). Many of the pathologic manifestations of congenital CMV infection that reflect visceral organ involvement (hepatitis and pneumonitis) are also observed in immunocompromised adults with disseminated CMV disease. The developing fetal brain is also highly susceptible to CMV-induced injury. CNS injury may be attributable to interference with important cellular functions of the stem and progenitor cells in the brain. The pathogenesis of CMV infection in the CNS seems to be strongly related to perturbations in neural migration, neural death, cellular compositions, and the immune system of the brain (Cheeran et al, 2009). In infants with severe symptomatic congenital CMV infection, histopathologic evidence of viral dissemination is commonly found in the brain, ear structures, retina, liver, lung, kidney, and endocrine glands (Bissinger et al, 2002). Distinctive features include large cells with large nuclei containing oval inclusions separated from the nuclear membrane by a clear zone, which gives them a characteristic “owl’s eye” appearance. Inclusion-bearing epithelial cells have been described in the semicircular canals, vestibular membrane, cochlea, and other structures of the ear. In addition, temporal bone anomalies with cochlear, vestibular, and auditory canal defects have been noted in association with hearing loss in affected infants (Davis, 1969; Myers and Stool, 1968). Clinical Presentation, Sequelae, and Prognosis Prenatal ultrasonography provides clues to the possible diagnosis of fetal CMV infection. Findings include intrauterine growth restriction, microcephaly, ventriculomegaly, periventricular calcifications, echogenic bowel, polyhydramnios, pleural effusion, pericardial effusion, hepatosplenomegaly, intrahepatic calcifications, pseudomeconium ileus, and placental enlargement (Guerra et al, 2008; Nelson and Demmler, 1997). Fetal hydrops is also a common finding (Sampath et al, 2005). However, the sensitivity of ultrasound to detect congenital CMV infection is poor given that the majority of congenitally infected infants are asymptomatic. In a study of 600 pregnant women with primary CMV infection, abnormal ultrasound findings were detected in 51 of 600 (8.5%) pregnancies and in 23 of 154 (14.9%) fetuses in which congenital infection was documented. The positive predictive value of an abnormal ultrasound finding that predicted symptomatic congenital infection in women with primary CMV infection was only 35.3% when fetal infection status was unknown, compared with 78.3% when congenital CMV infection was confirmed (Guerra et al, 2000, 2008). Signs and symptoms are apparent at birth in 10% to 15% of all children with congenital CMV infection. Table 37-1 outlines the clinical manifestations of symptomatic CMV infection. Each year in the United States, approximately 40,000 babies are born with congenital infection; of these, 4000 to 6000 have symptomatic disease (see Figure 37-3) (Sharon and Schleiss, 2007). Infection in the symptomatic infant can involve any organ and manifests along a spectrum from mild illness to severe disseminated multiorgan system disease. The mortality rate of symptomatic congenital CMV disease in the 1st year of life is estimated to be greater than 10% (Boppana et al, 1992). Clinical features include jaundice, hepatosplenomegaly, lethargy, respiratory distress, seizures, and petechial rash. Infants with symptomatic disease are often premature and small for gestational age. A wide spectrum of disease can be observed, including hemolysis, bone marrow suppression, hepatitis, pneumonitis, enteritis, and nephritis. Bone abnormalities have been described (Alessandri et al, 1995). Common laboratory abnormalities include thrombocytopenia, anemia, abnormal levels of liver enzymes (particularly elevated transaminases), and elevated conjugated bilirubin levels. End-organ involvement is often inferred on a clinical basis; it can be confirmed by isolation of the virus or viral DNA from tissue or by the histopathologic demonstration of characteristic inclusion bodies, positive in situ hybridization with CMV gene probes, or immunofluorescence using appropriate CMV-specific antibodies from tissue biopsies. Of particular concern are the CNS pathologies observed with symptomatic congenital CMV infection. These pathologies include meningoencephalitis, calcifications, microcephaly, neuronal migration disturbances, germinal matrix cysts, ventriculomegaly, and cerebellar hypoplasia (Cheeran et al, 2009). CNS disease is usually characterized by at least one of the following signs and symptoms: lethargy, microcephaly, intracranial calcifications, hypotonia, seizures, hearing deficit, or an abnormal eye examination finding, such as chorioretinitis or optic atrophy. The clinical finding of microcephaly implicates the CNS as a site of infection with CMV, and abnormal findings on cranial imaging studies corroborate its involvement (Figure 37-4). Long-term neurodevelopmental disabilities are observed in 50% to 90% of children who are symptomatic at birth. In contrast, long-term neurodevelopment injury is strikingly less likely in congenitally infected infants who are asymptomatic at birth: when it does occur, it is typically limited to hearing deficits. If there is a control for sensorineural hearing loss (SNHL), the intellectual development of asymptomatic congenitally infected infants appears to be normal (Conboy et al, 1986, 1987; Stagno et al, 1982b). Among symptomatic congenitally infected infants, longterm sequelae can include microcephaly, hearing loss, motor deficits (paresis or paralysis), cerebral palsy, mental retardation, seizures, ocular abnormalities (chorioretinitis, optic atrophy), and learning disabilities (Cheeran et al, 2009; Sharon and Schleiss, 2007). The incidence of hearing loss among children with congenital CMV infection ranges from 10% to 15% of those who are asymptomatic at birth to up to 60% of those who are symptomatic as newborns (Pass, 2005). SNHL can be progressive and fluctuating in both asymptomatic and symptomatically congenitally infected infants (Rosenthal et al, 2009). Among asymptomatic congenitally infected CHAPTER 37 Viral Infections of the Fetus and Newborn and Human Immunodeficiency Virus Infection during Pregnancy A B C D 481 FIGURE 37-4 Magnetic resonance image abnormalities in infants with congenital cytomegalovirus (CMV) infection with neurologic manifestations. A, T1 axial flair image of neonate with congenital CMV diagnosed with CNS malformation in utero by prenatal ultrasound, demonstrating severe hydrocephalus and cortical dysplasia and polymicrogyria. B, Axial T1 image of infant with symptomatic congenital CMV infection demonstrating ventriculomegaly and periventricular enhancement (arrow). C and D, T1 flash axial (C) and sagittal (D) images of a symptomatic, congenitally infected infant demonstrating ventriculomegaly, polymicrogyria, and porencephalic cyst (arrow). infants, SNHL tends to be high frequency in nature; it ranges in severity from a unilateral, mild hearing deficit to severe, bilateral, profound deafness. CMV-induced SNHL is a dynamic, evolving lesion; it may be present at birth or can appear later in childhood. Delayed-onset hearing loss usually occurs before 4 years of age (Dahle et al, 2000; Fowler et al, 1997; Rivera et al, 2002; Williamson et al, 1992), but has been reported to evolve and progress through 6 years of age and beyond. The pathogenesis of the SNHL is incompletely understood. SNHL appears to be related to an inflammatory labyrinthitis, possibly potentiated by the fact that the labyrinth is a site of immune privilege, allowing virus to persist in this compartment long after it has been cleared from the systemic compartment (Schleiss and Choo, 2006). Although some temporal bone and cochlear abnormalities have been described in a small case series of hearing-impaired infants with congenital CMV infection (Bauman et al, 1994), a more comprehensive analysis indicated that CMV-associated SNHL was not associated with structural abnormalities, such as enlarged vestibular aqueduct syndrome (Pryor et al, 2005). Logistic regression analysis of the clinical and laboratory parameters observed in 180 children with symptomatic congenital CMV infection showed that the presence of petechiae and intrauterine growth restriction were independently associated with the development of hearing loss (Rivera et al, 2002). A retrospective review of symptomatic congenitally infected infants identified neurologic and radiologic sequelae in 81% of affected patients. There was a significant correlation between the severity of the initial pure-tone audiometry and the development of progressive hearing loss, in addition to a significant correlation between a less severe final pure-tone audiometry and the presence of cerebral palsy (Madden et al, 2005). All congenitally infected infants, regardless of the results of functional hearing assessment at birth, should be monitored prospectively for SNHL by an audiologist using ageappropriate diagnostic tools. For severe SNHL, cochlear implantation has been used with success (Lee et al, 2005; Yoshida et al, 2009). Diagnosis and Infant Assessment Congenital CMV infection is best diagnosed by detection of virus, either through culture techniques or via PCR, in samples collected within the first 2 to 3 weeks of life. The timing of these samples is important because subsequent viral isolation may represent neonatal infection acquired in the birth canal or after exposure to breast milk (Schleiss, 2006b). Urine and saliva are the clinical samples of choice for viral culture. Specimens are typically inoculated onto a monolayer of human fibroblasts. The high viral titer in 482 PART IX Immunology and Infections these samples usually allows virus detection by cytopathic effect within 1 to 3 days; however, 2 to 4 weeks may be required (Revello and Gerna, 2002). Rapid virus isolation can be achieved using monoclonal antibodies to CMVspecific early protein with low-speed centrifugation of the clinical isolate onto the monolayer of fibroblasts growing on cover slips inside shell vials. This “shell vial” assay has a high sensitivity and specificity, and it allows the confirmation of diagnosis within 24 hours of inoculation (Gleaves et al, 1984; Rabella and Drew, 1990). PCR can readily amplify CMV DNA from various clinical samples, including urine, cerebrospinal fluid, blood, plasma, saliva, and biopsy material (Revello and Gerna, 2002). If primer selection and amplification conditions are carefully chosen, PCR can yield results comparable to a standard tissue culture test. Obvious advantages of PCR over culture are the small sample requirement, the short time requirement for test results, and the ability to use frozen specimens for diagnosis. The magnitude of the systemic viral load in the congenitally infected infant can be a predictor of neurodevelopmental prognosis. In some studies, the magnitude of viral DNA in blood (DNAemia) assessed by quantitative PCR correlated with an increased risk of long-term sequelae, including SNHL (Lanari et al, 2006). Quantitative PCR assays, particularly of blood, should be ordered in the evaluation of congenitally infected infants and may be useful in making decisions about which infants to treat with antiviral therapy. Serodiagnosis of congenital CMV is problematic. In congenital CMV infection, antibody production by the infected fetus begins in utero; however, serodiagnosis of congenital infection is complicated by the presence of maternal IgG antibodies that cross the placenta. Although a negative antibody titer in infant and maternal sera provides sufficient evidence to exclude the diagnosis of congenital CMV infection, positive titers in the newborn by no means confirm congenital infection. The presence of IgM antibodies to CMV in cord or neonatal blood is highly specific and in principle represents fetal antibody response, but many IgM commercial detection assays cross-react with IgG antibodies. Therefore a positive IgM titer has limited sensitivity in diagnosing congenital CMV infection and should not take the place of viral culture or PCR (Revello et al, 1999b). A recent development in molecular diagnostics for congenital CMV infection has been the use of dried blood spots (DBSs) as a source of CMV DNA for PCR-based detection (Barbi et al, 2006; Scanga et al, 2006; Yamagishi et al, 2006). DBS screening is amenable to long-term storage, so diagnosis can be made retrospectively, even after several years. The test might be useful in the future for implementation of widespread population-based newborn screening for congenital CMV infection (Dollard et al, in press; Kharrazi et al, 2010). However, recent evidence from a large multicenter study suggests that DBS is not sufficiently sensitive for diagnosis of congenital CMV compared with detection in other bodily fluids such as saliva and urine (Boppana et al, 2010). An alternative approach would be to analyze the DBSs obtained from infants with failed newborn hearing screens, because congenital CMV infection can be identified in approximately 3% of such cases (Choi et al, 2009). This approach would allow early detection and more timely intervention; however, it would fail to identify those congenitally infected infants who pass the newborn hearing screening but later develop SNHL. Cranial ultrasonography, head CT, and brain MRI are used to detect brain lesions associated with congenital CMV infection. CNS anomalies can be detected in some infants in utero. Fetal MRI can also detect abnormalities, including microcephaly and cortical anomalies, even when the ultrasound is normal; this appears to be the preferred modality for diagnosis of fetal CNS involvement (see F ­ igure 37-4; Benoist et al, 2008; Doneda et al, 2010). Because the finding of CNS disease is a potential harbinger of permanent sequelae, diagnostic CNS imaging is warranted in all suspected cases of congenital infection (Ancora et al, 2007; Boesch et al, 1989; Boppana et al, 1997; Kylat et al, 2006). Any of the standard imaging modalities is valuable in assessing CNS involvement. Ultrasound, because of its convenience, is an appropriate initial study and is particularly valuable and sensitive in detecting periventricular calcifications and lenticulostriate vasculopathy associated with mild to moderate ventricular dilatation. MRI provides important additional information, particularly the presence of associated polymicrogyria, hippocampal dysplasia, and cerebellar hypoplasia (de Vries et al, 2004); therefore the staged sequential use of ultrasound and MRI is probably the preferred approach to CNS imaging in this setting. Careful initial hearing evaluation and longitudinal monitoring for SNHL is required in all infants with documented congenital CMV infection, given that this complication is the most common late sequela of congenital CMV infection. Early recognition of SNHL and institution of appropriate interventions (speech–language therapy, centers for deafness education, and cochlear implants) can markedly improve the developmental, social, and language skills of a child with hearing impairment. Any child born with congenital CMV infection, whether symptomatic or not, deserves careful and recurring evaluation for cognitive delays, visual impairment, or motor disabilities. Follow-up evaluation should include a multidisciplinary team approach involving a pediatric infectious diseases specialist, pediatric otolaryngologist, and child behavioraldevelopmental specialist, in addition to a physical therapist, ophthalmologist, and neurologist as needed. Treatment: Antiviral Intervention in the Newborn and the Pregnant Patient Experience with the use of antiviral therapies against CMV in immunosuppressed patients, particularly solid organ and hematopoietic stem cell recipients and patients with HIV infection, is considerable (Boeckh and Ljungman, 2009; Razonable, 2005). Several studies in recent years have extended this experience to infants and children and have clearly indicated a benefit of antiviral therapy against CMV infection in several settings. In a phase 3, randomized, nonblinded controlled trial of ganciclovir for newborns with congenital CMV disease (Kimberlin et al, 2003), a group of infants with virologically confirmed congenital CMV received a 6-week course of ganciclovir (6 mg/kg every 12 hours). The primary endpoint was improved hearing or retention of normal hearing. Significant differences CHAPTER 37 Viral Infections of the Fetus and Newborn and Human Immunodeficiency Virus Infection during Pregnancy between treated and untreated groups were noted, including a statistically higher likelihood of normal or improved hearing at 6 months of age in treated infants compared to controls. Although almost two thirds of the treated infants had neutropenia, it was reversible when antiviral therapy was halted. In a follow-up assessment, infants with symptomatic congenital CMV involving the CNS who received intravenous ganciclovir therapy had fewer developmental delays at 6 and 12 months, using the Denver Developmental Screening test, compared with untreated infants (Oliver et al, 2009). Because the use of ganciclovir can promote improved hearing and neurodevelopmental outcomes, antiviral therapy should be considered for all infants with congenital CMV infection with any evidence of CNS involvement (microcephaly, abnormal CNS imaging study, positive CSF for CMV DNA, chorioretinitis, or evidence of SNHL). Ganciclovir should also be used in any infant with severe or life-threatening end-organ CMV disease, whether acquired via congenital infection or via a postnatal route (see Table 37-2) (Schleiss and McVoy, 2004). CMV retinitis in the congenitally infected infant can represent a particularly problematic management issue; it has been reported that up to 6 months of antiviral therapy may be required to control chorioretinitis in the symptomatic congenitally infected infant (Shoji et al, 2010). All infants with documented congenital infection should have an ophthalmologic evaluation. If chorioretinitis is present, it should be managed in consultation with an ophthalmologist and infectious diseases expert. Although ganciclovir improves neurodevelopmental outcomes for symptomatic infants, it is not clear whether ganciclovir holds the promise of improving outcomes in infants with asymptomatic congenital CMV infection. Because some of these infants are at risk to progress to SNHL, clinical trials are warranted to explore whether antiviral therapy could prevent development of hearing loss. Other investigational studies including the assessment of the therapeutic potential in infants of the prodrug of ganciclovir, valganciclovir, are needed. The potential role of valganciclovir in long-term (suppressive) therapy of CMV in congenitally infected infants is currently under investigation. The CASG is currently conducting a clinical trial of 6 weeks versus 6 months of valganciclovir (www.casg.uab.edu; Nassetta et al, 2009). A number of case reports and small case series have reported the safety, tolerability, and possibly efficacy of long-term oral therapy with valganciclovir in the setting of congenital CMV infection (Amir et al, 2010; Hilgendorff et al, 2009; Lombardi et al, 2009; Yilmaz Çiftdogan and Vardar, 2011), but the uncontrolled nature of these reports precludes any recommendations for oral therapy at this time. The prospect of treating the pregnant patient to prevent transmission of CMV to the fetus is an area of active investigation. Ganciclovir has demonstrated teratogenic risk in some studies (Schleiss and McVoy, 2004); although this has never been demonstrated in humans, it has limited research in this area. A case report of the use of oral ganciclovir in a pregnant liver transplant patient did not show any evidence of teratogenicity (Pescovitz, 1999). Ganciclovir has been demonstrated to cross the placenta, and therefore could theoretically be used to treat CMV infection in utero (Brady et al, 2002). An observational study of 20 483 women with 21 fetuses, with confirmed congenital CMV infection treated with oral valacyclovir, demonstrated placental transfer of valacyclovir with measurable concentrations in the amniotic fluid and a subsequent reduced viral load in the fetal blood (Jacquemard et al, 2007). There have been several case reports of treatment of congenital CMV infection in utero with oral, parenteral, or intraamniotic ganciclovir with varying degrees of success (Miguelez et al, 1998; Puliyanda et al, 2005; Revello et al, 1993, 1999a). Although it is probably safe, prenatal treatment of fetal CMV infection with ganciclovir is currently not supported by the available data; further study with a randomized controlled trial is needed. Passive immunization with CMV human immunoglobulin (HIG) has been studied for the in utero treatment and prevention of congenital CMV infection. CMV HIG is a pooled, high-titer immunoglobulin preparation derived from donors with high levels of CMV antibody. Nigro et al (2005) completed a prospective study of CMV HIG for the treatment of pregnant women with primary CMV infection, including some women with confirmed fetal CMV (Nigro et al, 2005). The women were enrolled in the therapy group if they had an amniocentesis and confirmed congenital CMV infection, as evidenced by a positive PCR in the amniotic fluid; they were enrolled in prevention group if they did not have an amniocentesis. In the therapy group, only 1 in 31 of the treated mothers delivered an infant with congenital CMV disease, compared to 7 in 14 mothers who were not treated with HIG. In the prevention group 6 of 37 mothers receiving HIG delivered infants with congenital CMV, compared to 19 of 7 mothers who did not receive treatment. Overall, there was a statistically significant reduction of risk for congenital CMV infection with HIG therapy. In a subsequent study, three fetuses treated with HIG had resolution of their ultrasonographically detected cerebral abnormalities; in contrast the two untreated fetuses had persistence of their cerebral abnormalities (Nigro et al, 2008). In addition to fetal effects, CMV HIG has been demonstrated to affect the placenta. In pregnancies treated with HIG, significant reductions in placental thickness have been demonstrated (La Torre et al, 2006). The reduction in placental thickness with HIG treatment suggests that at least part of the beneficial effect of treatment is mediated at the level of the placenta (Schleiss, 2006a). Randomized controlled trials of HIG for the treatment and prevention of congenital CMV infection are needed. Until such data are available, clinicians could consider treatment with CMV HIG in pregnant patients with confirmed fetal CMV infection. Natal Acquisition of CMV Infection: Implications for the Premature Infant In addition to congenital infection, CMV can produce disease in the newborn infant after natal acquisition; this can occur via one of three mechanisms: (1) transmission in the birth canal during vaginal delivery after exposure to infectious cervicovaginal secretions, (2) through ingestion of breast milk, and (3) via blood transfusion. Of these potential mechanisms, the most common is via breast milk (Schleiss, 2006c), with transmission in the birth canal occurring less commonly. It has long been recognized that 484 PART IX Immunology and Infections CMV is shed by the cervix (Montgomery et al, 1972) and excreted in the breast milk of seropositive women (Hayes et al, 1972; Stagno, 1982a). The risk of CMV transmission in infants who are breastfed by seropositive women shedding virus in their breast milk has been reported to be between 58% and 69% (Dworsky et al, 1983; Stagno et al, 1980). CMV infection acquired in the postnatal period in healthy term infants by this route typically is asymptomatic, only rarely producing any morbidity. There is no convincing evidence that acquisition of CMV via breast milk leads to any adverse neurodevelopmental sequelae (Kurath et al, 2010). In a study of CMV transmission through breastfeeding, all the infants who acquired CMV infection had normal neurodevelopment at a mean follow-up of 51 months of age (Vollmer et al, 2004). Although the safety of breastfeeding has been established in term infants in CMV-seropositive women, controversy exists regarding the safety of breastfeeding low-birthweight, premature infants. Studies in low-birthweight and very low-birthweight (VLBW) preterm infants yield conflicting results regarding to the risk of developing symptomatic infection after breast milk acquisition of CMV (Schleiss, 2006c). In one study of VLBW premature infants (<32 weeks’ gestation, <1500 g) exposed to CMV via breast milk, virus transmission occurred in 33 of the 87 exposed infants (Maschmann et al, 2001). Approximately half of these infants were ill and exhibited symptoms such as hepatopathy, neutropenia, thrombocytopenia, and sepsislike deterioration. Conflicting reports of early postnatal CMV infection in preterm infants in highly immune populations have suggested that symptomatic CMV infection acquired by breast milk is rare (Mussi-Pinhata et al, 2004; Kurath et al, 2010). Proposed efforts to reduce the infectivity of breast milk from seropositive mothers have included freezing breast milk at –20°C, Holder pasteurization, and short-term pasteurization (Hamprecht et al, 2004). Of these methods, freezing is the most studied and most likely to maintain the salutary immunologic properties of breast milk. Some experts recommend freeze-thawing all breast milk before feeding the VLBW premature infant if the mother is known to be CMV seropositive, or if her CMV serostatus is unknown. Although freezing breast milk may lower the incidence of postnatally acquired CMV infection, it does not entirely eliminate the risk (Maschmann et al, 2006). It is presumed that VLBW premature infants are at increased risk because they possess fewer transplacentally acquired antibodies against CMV than do term babies, and thus are more likely to develop disease upon infection. It has recently been shown that treating the VLBW infant with IVIG appeared to reduce the likelihood of transmission of CMV by breast milk (Capretti et al, 2009), although this is not currently considered to be an indication for the use of IVIG in the premature infant. Further evidence is necessary to make recommendations regarding what, if any, interventions are appropriate in low-birthweight, preterm infants receiving breast milk from CMV-seropositive mothers. CMV can be transmitted through blood transfusion, and transfusion-associated infections were at one time a major problem in the neonatal intensive care setting (Adler, 1986; Adler et al, 1983, 1984; Prober et al, 1981). Two approaches are currently used to decrease the risks of transfusion-associated CMV, leukocyte reduction, and directed transfusion of CMV-negative blood products (Lamberson et al, 1988). Although leukocyte reduction has had a dramatic effect on the risk of transfusion-associated CMV, reports are conflicting in the literature regarding the question of whether this intervention is completely effective at eliminating the risk of transfusion-transmitted CMV (Allain et al, 2009; Fergusson et al, 2002; Vamvakas, 2005). A recent survey of the American Association of Blood Banks physician membership revealed that 65% of those responding believed that leukocyte-reduced and CMV-negative blood components were equivalent in their ability to prevent transfusion-associated transmission of CMV (Smith et al, 2010). Leukoreduction should be adequate for preventing the overwhelming majority of transfusion-associated CMV infections in the newborn intensive care unit. However, despite the American Association of Blood Banks survey results on attitudes toward leukocyte-reduced blood products, fetuses and neonates are more likely to receive CMV-negative products compared with other groups receiving transfusions, and additional studies are warranted to ascertain what risk may exist for premature infants receiving leukocyte-reduced products from CMV-seropositive donors. Prevention One important strategy for addressing the problem of congenital CMV is the education of women of childbearing age about the risks of transmission and strategies for prevention. Child care providers (including daycare workers, special education teachers, and therapists) appear to have a higher risk of occupational exposure to CMV, because of extensive contact with infants and young children (Adler, 1989; Pass et al, 1986, 1990). Education of the potential occupational risk in this group is essential. In contrast to these child care providers, health care workers who appropriately use routine infection control practices are not at increased risk of CMV acquisition. CMV infections in pregnant women are typically clinically “silent.” Like most healthy individuals, more than 90% of pregnant women with primary CMV infections have no symptoms. When symptoms occur, they are nonspecific and vague, often described as a flulike syndrome. Potential manifestations include fever, fatigue, headache, myalgia, lymphadenitis, and pharyngitis, but these are the exception and not the rule. Because most maternal CMV infections are asymptomatic, a major goal is education of all women of childbearing age on hygienic practices (ACOG Practice Bulletin, 2002; Jeon et al, 2006; Ross et al, 2006). Hygienic strategies are important because the saliva and urine of infected children are significant sources of CMV infection among pregnant women. Strategies include washing hands whenever there is contact with a child’s saliva or urine, not sharing food, utensils, or cups, and not kissing a child on the mouth or cheek (Anderson et al, 2007; Cannon and Davis, 2005). It is also essential that women become better educated on the importance of CMV infection. A survey of women in 2005 showed that only 14% of women knew what CMV was, but most believed that preventative measures for an infection that could harm an unborn baby would generally be acceptable (Ross et al, 2008). The CHAPTER 37 Viral Infections of the Fetus and Newborn and Human Immunodeficiency Virus Infection during Pregnancy effectiveness of educating pregnant women on methods to prevent CMV transmission has been demonstrated (Adler et al, 1996). In a study in which seronegative mothers with a child in group day care were instructed on measures to prevent CMV transmission, pregnant mothers had a significantly lower rate of CMV infection when compared with nonpregnant mothers attempting conception (Adler et al, 2004). In addition, a study in France recently demonstrated a lower CMV seroconversion rate after counseling pregnant women on hygienic measures (Vauloup-Fellous et al, 2009). Prenatal maternal screening for CMV antibodies is controversial. Because women who are CMV-immune can be reinfected with new strains that can then be transmitted to the fetus, with subsequent sequelae (Boppana et al, 1999, 2001), the finding of a positive preconception titer for CMV IgG antibody may provide a false sense of reassurance and decrease a pregnant patient’s motivation to engage in careful hygienic practices. A recent study evaluated three screening strategies and suggested that universal maternal screening for CMV could be a cost-effective strategy if a treatment were available that could achieve a 47% reduction in disease (Cahill et al, 2009). Ultimately the control of congenital CMV could be realized by the development of an effective vaccine. No CMV vaccines are currently licensed; however, because of the enormous economic impact of congenital CMV, the Institute of Medicine has identified a CMV vaccine as the highest-level priority for new vaccine development (excluding HIV vaccines) for the United States (Stratton et al, 2001). The CMV envelope glycoprotein B has been the most studied subunit vaccine candidate for this purpose, because it is a target of neutralizing antibody in all CMVseropositive individuals. Results of a phase II, placebo-controlled, randomized, double-blind trial of the recombinant CMV envelope glycoprotein B with MF59 adjuvant were published recently (Pass et al, 2009). In this study, three doses of the CMV vaccine or placebo were administered at 0, 1, and 6 months to healthy women within 12 months postpartum. Women in the vaccine group were less likely than the placebo group to become infected with CMV (p = 0.02) with 18 of 225 women becoming infected with CMV in the vaccine group, compared to 31 o 216 in the placebo group. This report is the first to document efficacy of any CMV vaccine in a clinical trial and should help to accelerate the pace of future vaccine development and testing. Other clinical trials include evaluation of live, attenuated CMV vaccines and the testing of subunit vaccines targeting other key proteins involved in the humoral and cellular immune responses to CMV infection (Schleiss, 2008). HUMAN HERPESVIRUS 6 AND 7 Human herpesvirus (HHV) 6 was isolated in tissue culture in 1986 from peripheral blood leukocytes of patients with both lymphoproliferative disorders and HIV infection (Bernstein and Schleiss, 1996; Schleiss, 2009). The virus was eventually shown to have a tropism for T cells, and molecular studies revealed homology with CMV. These findings suggested that the virus belonged to the family Herpesviridae. For several years after discovery, its role in disease was unclear, but it is now known to be the 485 major etiologic agent of roseola infantum (exanthem subitum) and has been implicated in other clinical syndromes. HHV-6 is a prototypical β-herpesvirus, with a doublestranded DNA genome contained within an icosahedral capsid, surrounded by an outer envelope. HHV-6 is subclassified as either variant A or B, based on differences in nucleotide sequence, restriction enzyme profile, and reactivity with monoclonal antibodies. HHV-6B is the subtype typically associated with exanthem subitum (Yamanishi et al, 1988). HHV-6 and HHV-7 are ubiquitous in nature, and typically cause infection in the first 2 years of life. HHV-7 is highly related to HHV-6 and, like HHV-6, is responsible for roseola infantum (Tanaka et al, 1994). HHV-7 is a β-herpesvirus, structurally and molecularly similar to CMV and HHV-6. It was first isolated from CD4+ T cells of a healthy individual. The high degree of homology with HHV-6 creates difficulty in interpretation of serologic assays, which are largely investigational in nature and not generally useful for clinical practice, because there is considerable cross-reactivity of antibodies between HHV-6 and HHV-7 proteins. As with HHV-6, infection with HHV-7 appears to be ubiquitous, although infection appears to be acquired somewhat later in life than is HHV-6 (Caserta et al, 1998; Suga et al, 1997). Approximately 40% to 45% of children have antibodies to HHV-7 by 2 years of age, and 70% of children are seropositive by 6 years of age. Like other β-herpesviruses, HHV-7 can be found in the saliva, suggesting a route for person-toperson transmission. Based on its identification in cervical secretions, HHV-7 also has the potential for perinatal transmission (Okuno et al, 1995). It is of interest to examine the evidence that suggests that HHV-6, like CMV, can cross the placenta and infect the fetus in utero, given the extensive molecular similarities between these two viruses. In one early study, HHV-6 DNA was detected by PCR in approximately 25% of pregnant seropositive women at the time of delivery, and in approximately 1% of cord blood samples, suggesting the possibility of congenital transmission (Dahl et al, 1999). Examination of 305 cord blood samples in another study identified HHV-6 DNA by PCR in 1.6% of infants (Adams et al, 1998). Interestingly, this potential vertical transmission rate of 1% to 2% is similar to congenital infection rates commonly reported for CMV. Congenital HHV-6 transmission was first definitively reported in a study of 5638 cord bloods; 57 samples (1%) had HHV-6 DNA by PCR, but none had HHV-7 (Hall et al, 2004). Of note, these infections were all asymptomatic, and the HHV-6 genome variant in one third of congenital infections was HHV6A, in contrast to the more common variant HHV6B, which is encountered in virtually all postnatal infections. Although the rate of congenital HHV-6 transmission at approximately 1% is highly similar to that observed for congenital CMV, the mechanisms of transmission are different. It has been shown that vertical transmission of HHV-6 most often occurs (90% of cases) because of the germline passage of chromosomally integrated HHV-6 (Hall et al, 2008). The mode of inheritance of HHV-6 genome in this form of vertical transmission appears to be exclusively maternal (Hall et al, 2010). HHV-6 is unique among the human Herpesviridae regarding its capacity to integrate into the host chromosome and mediate germline transmission of 486 PART IX Immunology and Infections viral genome from mother to fetus. Remarkably, HHV-6 is the first functional virus of any type reported to be passed through the human germline. The clinical consequences of such transmission, and the differences in germline transmission of viral genome and the more common postnatal acquisition of HHV-6 in early childhood, remain unknown. That HHV-6 genome appears to integrate into the telomere—a chromosomal component important in cellular aging and in cancer—suggests that there may be interesting long-term consequences associated with vertical germline transmission of this virus (Arbuckle et al, 2010; Nacheva et al, 2008) that remain to be elucidated. HHV-6 and HHV-7 DNA has been found to be present in cervical secretions, placenta, and peripheral blood mononuclear cells of pregnant women (Caserta et al, 2007). Whether intrapartum transmission of these viruses can occur in the birth canal during delivery remains unknown. HHV-6 DNA can also be found in breast milk. Perinatal transmission via this mechanism has been postulated (Joshi et al, 2000), but has not been demonstrated. KAPOSI’S SARCOMA HERPESVIRUS AND EPSTEIN–BARR VIRUS In 1994 a novel herpesvirus, HHV-8 or KSHV, was identified in patients with AIDS-associated Kaposi’s sarcoma (KS; Chang et al, 1994). This virus was assigned to the γ-herpesvirus family of the Herpesviridae, based on its molecular and sequence similarity to the other prototypical γ-herpesvirus, Epstein–Barr virus. Subsequent studies have linked KSHV to both AIDS-associated KS and the endemic forms of KS that are prevalent in elderly Mediterranean men. Little is known about routes of transmission of KSHV in nonpregnant patients, let alone whether it can be transmitted perinatally. The clinical significance of acquiring primary KSHV infection outside of the HIV setting is also unclear. Serologic evidence from sexually transmitted disease clinics suggests that sexual contact is a risk factor for acquiring KSHV. The epidemiology of primary HHV-8 infection appears to vary considerably worldwide. Indeed, the routes of acquisition of infection and mechanisms responsible for person-to-person transmission remain uncertain. After the virus was initially discovered, the unique role it seemed to play in inducing malignant disease in HIV-infected patients suggested that the primary route of transmission of HHV-8 was through sexual contact, particularly among gay men. However, more recent evidence suggests that other routes of infection exist, including transmission by saliva (Pica and Volpi, 2007). A recent cross-sectional study of the seroprevalence of HHV-8 in children and adolescents in the United States indicated a prevalence of approximately 1% (Anderson et al, 2008). There appears to be considerable regional variation in prevalence in the United States. In a population of children in south Texas, the seroprevalence was 26%, strongly suggesting that nonsexual modes of transmission predominate (Baillargeon et al, 2002). In Sub-Saharan Africa, prevalence in children is even higher, approaching 60% in some studies (Sarmati, 2004). There are reports of infections in infants that suggest possible vertical transmission, but congenital infection has not been demonstrable by PCR techniques (Sarmati et al, 2004). HHV-8 can also be transmitted by blood transfusion (Hladik et al, 2006); this observation would suggest that transplacental transmission is at least theoretically feasible. In a study of 89 KSHV-seropositive women, 13 mothers (14.6%) had KSHV DNA detected in their peripheral blood mononuclear cells; 2 of 89 samples drawn at birth from infants born to these mothers had KSHV DNA detectable within their peripheral blood mononuclear cells. These findings suggest that KSHV can be transmitted perinatally, but infrequently (Mantina et al, 2001). As serologic and nucleic acid–based diagnostics tests become more widely available, a better assessment of the worldwide seroepidemiology of HHV-8 infection and an increased understanding of its modes of transmission will be achievable. Most primary infections with HHV-8 are probably asymptomatic, although the clinical course of primary symptomatic HHV-8 infection in immunocompetent children has been described (Andreoni et al, 2002). In this study, fever and rash were noted with primary infection. The rash first appeared on the face and gradually spread to the trunk, arms, and legs. It initially consisted of discrete red macules that blanched with pressure and eventually became papular. An upper respiratory tract infection appeared as a secondary symptom in most children, and a lower respiratory tract infection appeared as a secondary symptom in one third of symptomatic children. Additional information on the epidemiology and modes of transmission of this pathogen, particularly in the prenatal and intrapartum period, is needed. There is a minimal amount of information available about prenatal and perinatal modes of transmission of EBV. EBV is the causative agent of infectious mononucleosis and is associated with nasopharyngeal carcinoma, Burkitt’s lymphoma, and lymphoproliferative disease in immunocompromised patients (Schleiss, 2009). Primary EBV infection during pregnancy appears to be rare. In a prospective study, susceptibility to EBV infection in 1729 pregnant women was evaluated by screening for EBV antibodies. Fifty-eight subjects (3.4%) had no detectable EBV antibody and were presumably susceptible. Of the 54 women who agreed to participate in this study, none acquired EBV antibody during pregnancy (Le et al, 1983). It is not clear whether transplacental passage of EBV in seropositive pregnant women occurs. Because EBV can be acquired by blood transfusion, such a mode of transmission is feasible. It has been postulated that high-titer antibodies cross the placenta and protect the fetus from hematogenous transmission of virus in women who reactivate EBV during pregnancy (Purtilo and Sakamoto, 1982). A solitary case report describes the occurrence of severe EBV disease in a premature infant, born at 28 weeks’ gestation, who was examined on the 42nd day of life with hepatosplenomegaly, hemolytic anemia, thrombocytopenia, and atypical lymphocytosis (Andronikou et al, 1999). In another study, the potential for EBV vertical transmission from a seropositive mother to her child was evaluated in 67 pregnant women by nested PCR (Meyohas et al, 1996). Two of 67 neonates were positive for EBV DNA, suggesting that mother-to-child transmission of free EBV or of maternal EBV-infected cells can occur during pregnancy in the setting of latent EBV infection, but no clinical consequences CHAPTER 37 Viral Infections of the Fetus and Newborn and Human Immunodeficiency Virus Infection during Pregnancy of these putative congenital infections were reported. In a study of six pregnant women with evidence of primary EBV infection, four pathologic births were observed: one spontaneous abortion, two premature babies, one of whom died, and one stillborn with multiple malformations (Icart et al, 1981). However, the assessment of primary infection was drawn solely from the serologic profile (including the presence of anti–early antigen antibodies) and the relationship of EBV infection to adverse pregnancy outcome was undefined, and it was unclear whether intrauterine EBV infection was present. In another study, placentas and some fetuses were studied in five cases of pregnancy interruption caused by maternal infectious mononucleosis in early gestation (Ornoy et al, 1982). Decidual lesions, consisting of perivasculitis and necrotizing deciduitis, were noted, and endovasculitis, perivasculitis, and occasional vascular obliteration were found in villi, as well as mononuclear and plasma cell infiltrates. Two fetal hearts exhibited evidence of myocarditis. These observations suggested a pathogenic role for placental and fetal EBV infection, but no direct virologic evidence of EBV was obtained, and these observations have not been duplicated. Accordingly a recent review concluded that maternal EBV infection did not impose a serious threat to pregnancy outcome (Avgil and Ornoy, 2006) and that no strong evidence for teratogenicity or fetal injury existed. HUMAN PARVOVIRUS B19 Parvovirus B19, a small, single-stranded DNA virus, is the only member of the parvovirus family that causes human disease. The virus was identified in 1975 (Cossart et al, 1975) and was first linked to a disease in 1981—aplastic crisis in children with sickle cell anemia (Pattison et al, 1981). Primary infection with parvovirus B19 is commonly known as fifth disease or erythema infectiosum; it is classically described as a childhood exanthem with a “slapped cheek” appearance (Anderson et al, 1984). The agent is less ubiquitous than other common childhood exanthematous illnesses, such as HHV-6; therefore many individuals reach adulthood with no prior evidence of infection, placing women of childbearing age at potential risk. Considerable interest in the role of this virus in hydrops fetalis (nonimmune) and fetal aplastic crisis has evolved since the first cases of fetal death associated with maternal parvovirus B19 infection were reported in the 1980s (Brown et al, 1984; Kinney et al, 1988). Epidemiology Parvovirus B19 infection is global in nature. It is common in childhood and continues at a low rate throughout adult life. One study identified an annual seroconversion of 1.5% in women of childbearing age unrelated to their occupation (Koch and Adler, 1989). The peak incidence of erythema infectiosum is in the late winter and early spring. Periodic epidemics at intervals of a few years are typical. The virus is spread by respiratory droplet (Anderson and Cohen, 1987), by blood products (especially pooled clotting factor concentrates; Jordan et al, 1998), and transplacentally (Ergaz and Ornoy, 2006). Approximately 50% to 80% of adults in the United States are seropositive 487 for human parvovirus B19 (Anderson, 1987; Vyse et al, 2007). A significant proportion of childbearing women are thus susceptible to infection (Markenson and Yancey, 1998; Yaegashi et al, 1998). Preconception seroprevalence to parvovirus B19 ranges from 24% to 84% (Ergaz and Ornoy, 2006). During pregnancy the risk of acquiring parvovirus B19 infection is low, ranging from 0% to 16.5% in different studies (Ergaz and Ornoy, 2006). The risk of primary maternal infection is higher during epidemics, with reported seroconversion rates ranging between 3% (Kerr et al, 1994) and 34% (Woernle et al, 1987). It is estimated that one fourth to half of maternal parvovirus infections result in transmission of infection to the fetus (Alger, 1997; Gratacos et al, 1995; Koch et al, 1998). The vast majority of pregnancies are unaffected (Berry et al, 1992; Sheikh et al, 1992). The risk of adverse fetal outcome is increased if maternal infection occurs during the first two trimesters of pregnancy (Skjoldebrand-Sparre et al, 2000), particularly before 20 weeks’ gestation. There are conflicting reports regarding the prognosis once fetal infection has been established. A longitudinal study of fetal morbidity and mortality in more than 1000 women with primary parvovirus B19 infection in pregnancy demonstrated a risk of fetal hydrops of 3.9% and a risk of fetal death of 6.3%; fetal death was observed only if maternal infection occurred before the 20th week of gestation (Enders et al, 2004). A recent retrospective analysis of intrauterine parvovirus B19 infection at a single site suggested that the rate of adverse fetal outcome is much higher than previously appreciated, with fetal hydrops and demise occurring in greater than 10% of pregnancies (Beigi et al, 2008), although the total number of cases reported in this series was low; this primarily represented a referral population to a tertiary care center. Pathogenesis The most common mode of transmission of parvovirus B19 is via a respiratory route. Typically, once the virus establishes infection, viremia occurs, followed by mild systemic symptoms such as fever and malaise. Viremia is short-lived, lasting only 1 to 3 days, and the characteristic immunemediated rash develops 1 to 2 weeks later. Once the rash appears, an individual is no longer infectious. Arthropathy caused by parvovirus B19 is common; it is observed more frequently in adults with primary infection than in children. It typically manifests late in the course of illness with acute onset of arthralgias or frank arthritis involving the hands, knees, wrists, and ankles. The symptoms usually subside within 1 to 3 weeks, although approximately 20% of affected women have persistent or recurring arthropathy for months to years (Woolf et al, 1989). Potential pathogenic mechanisms involve the recognized affinity of parvovirus B19 for progenitor erythroid cells of bone marrow. The blood group P antigen is a main cellular receptor for parvovirus B19, and it is found on red blood cells and on placental trophoblast cells (Jordan et al, 2001). The P antigen is also expressed on fetal cardiac myocytes, enabling parvovirus B19 to infect myocardial cells (Rouger et al, 1987) leading to myocarditis (von Kaisenberg et al, 2001). Myocarditis induced by parvovirus B19 can contribute to high-output cardiac failure, and the 488 PART IX Immunology and Infections myocardial inflammation and subendocardial fibroelastosis may also contribute to fetal hydrops (Morey et al, 1992). Fetal infection most likely occurs hematogenously via the placenta during maternal viremia. Parvovirus B19 infection in utero causes a pronormoblast arrest, which leads to fetal anemia, nonimmune hydrops, and sometimes progressive congestive heart failure (Ergaz and Ornoy, 2006; Kinney et al, 1988). The fetus is especially susceptible to adverse consequences of red blood cell infection, secondary to the intrinsic short fetal erythrocyte life span and rapidly expanding blood volume, especially during the second trimester. It has also been postulated that parvovirus B19 infection leads to cytotoxicity and subsequent anemia by inducing apoptosis of infected red blood cells (Yaegashi et al, 1999, 2000). The NS1 protein of parvovirus B19 induces cell death by apoptosis in erythroidlineage cells by a pathway that involves caspase 3, whose activation may be a key event during parvovirus-induced cell death (Moffatt et al, 1998). The NS1 protein also has a key role in the arrest of infected cells at the G1 phase of the cell cycle prior to apoptosis induction (Chisaka et al, 2003). In addition to erythroid precursors and myocytes, other organs appear to be involved in fetal parvovirus B19 infection. Fetal brain infection has been reported. Neuropathologic findings in the infected hydropic fetus include perivascular calcifications, primarily in the cerebral white matter, as well as multinucleated giant cells. Viral DNA has been demonstrated in the brain and liver (Isumi et al, 1999). Data also suggest that the maternal cell–mediated immune response at the placental level contributes to the pathogenesis of congenital infection. In one study, placentas from women whose pregnancies were complicated by parvovirus B19 infection had increased infiltration of CD3 T cells and elevated levels of interleukin 2 (Jordan et al, 2001). Clinical Spectrum Parvovirus B19 infection causes erythema infectiosum, or fifth disease, in normal hosts, aplastic crisis in patients with hemolytic disorders, and chronic anemia in immunocompromised hosts. A substantial proportion of infected adult women may also have arthropathy in association with parvovirus B19 infection (Woolf et al, 1989). Maternal symptoms have been present in up to two thirds of documented cases of nonimmune hydrops fetalis associated with parvovirus B19 infection (Yaegashi et al, 1998). The major clinical presentation of parvovirus B19 infection in the fetus is hydrops fetalis. Various estimates suggest that human parvovirus B19 infection contributes from 10% to 27% of cases of nonimmune hydrops fetalis (Essary et al, 1998; Markenson and Yancey, 1998; Yaegashi et al, 1994). Nonimmune hydrops fetalis was the main complication in 0.9% to 23% of pregnancies among proven maternal infections with parvovirus B19 (Ergaz and Ornoy, 2006). A risk of 7.1% for hydrops fetalis has been described for pregnant women who acquire parvovirus B19 infection between 13 and 20 weeks’ gestation (Enders et al, 2004). Parvovirus B19 does not appear to have a major role in intrauterine fetal demise in the absence of hydrops fetalis (Riipinen et al, 2008). In addition to hydrops fetalis, in recent years there have been increasing numbers of case reports of neurologic and ophthalmologic anomalies associated with fetal parvovirus B19 infections. Although some studies undertaken to examine the association between fetal infection and congenital anomalies failed to reveal any associations (Kinney et al, 1988; Mortimer et al, 1985), parvovirus B19 has increasingly been recognized as a cause of neuronal migration defects (Pistorius et al, 2008). The role of parvovirus B19 in neurodevelopmental injury has not been fully explored. There have been at least three case reports of fetal encephalopathy associated with in utero infection with parvovirus B19 (Alger, 1997). Parvovirus B19 has recently been recognized as a cause of CNS injury in older children, including encephalitis, meningitis, stroke, and peripheral neuropathy (Douvoyiannis et al, 2009). Consideration should be given to performing brain imaging studies in infants with symptomatic in utero parvovirus B19 infection. Such infants need careful neurodevelopmental follow-up. Few data are available regarding long-term outcomes of infants infected in utero. Two prospective studies in the United Kingdom of approximately 300 congenitally exposed infants found the risk of major congenital or developmental abnormality to be less than 1% (Miller et al, 1998). Parvovirus B19 has been implicated in some cases of congenital anemia (Heegaard and Brown, 2002). In one series of 11 children with a diagnosis of Diamond-Blackfan syndrome, 3 of 11 bone marrow aspirates revealed evidence of parvovirus B19 DNA. All three of these children, but none of the parvovirus B19 PCR-negative cases, underwent spontaneous remission (Heegaard et al, 1996). In light of these data, all infants undergoing evaluation for congenital anemias should be evaluated for the possibility of parvovirus B19 infection. Laboratory Evaluation In primary care, the diagnosis of human parvovirus B19 infection is most commonly made clinically through recognition of the characteristic rash. Serologic confirmation is necessary in high-risk situations, such as after a significant exposure of a pregnant woman to a child with erythema infectiosum. Both radioimmunoassays and enzyme-linked immunosorbent assays are available for detection of human parvovirus B19–specific IgG and IgM antibodies (Kinney and Kumar, 1988). Presence of anti– parvovirus B19 IgM in fetal blood or amniotic fluid may confirm fetal infection, but may be detected in only one fifth of infected fetuses (Torok et al, 1992). False-positive results of parvovirus B19 IgM testing have been reported, including cross-reactions with anti-rubella IgM (Dieck et al, 1999). Monitoring of the pregnant patient with a primary parvovirus B19 infection is an important clinical problem. In the context of a human parvovirus B19 infection in a symptomatic, pregnant woman, elevated or rising weekly measurements of maternal alpha-fetoprotein suggest fetal infection, and rising concentrations may be a marker for an increased risk for hydrops fetalis (Carrington et al, 1987). However, some studies have failed to demonstrate any association between the magnitude of the elevation of the α-fetoprotein and the severity of fetal anemia (Simms et al, 2009). Serial fetal ultrasonographic evaluations of infected CHAPTER 37 Viral Infections of the Fetus and Newborn and Human Immunodeficiency Virus Infection during Pregnancy pregnant women are recommended to evaluate and follow for fetal hydrops or intrauterine demise. The findings of echogenic bowel, ascites, pleural or pericardial effusion, or scalp edema are considered to be important markers of fetal infection and pathology. Middle cerebral artery Doppler to evaluate for fetal anemia may be another useful prospective surveillance tool, because fetal anemia can be detected using this technique before fetal hydrops is evident (Feldman et al, 2010). Other techniques to diagnose fetal parvovirus B19 infection have been studied and used, but most are not readily available. Virus can be cultured from tissue in suspension cultures of bone marrow cells from persons with hemolytic anemias, but it is difficult to isolate; therefore this method is not feasible for prenatal or postnatal diagnosis. Electron microscopy and histology have permitted visualization of parvovirus in fetal blood, ascitic fluid, tissue, and amniotic fluid, but the utility and sensitivity of these evaluations have not been well studied (Markenson and Yancey, 1998). A number of commercial PCR assays are available, and these can be performed on serum, amniotic fluid, or fetal tissue. Presumptive diagnosis can also be made based on finding IgM antibody in the maternal and fetal blood. PCR for parvovirus B19 DNA or in situ hybridization studies can be performed using maternal blood, amniotic fluid, cord blood, or fetal tissues. The detection of B19 DNA in maternal blood appears to have the best diagnostic sensitivity for identifying maternal B19 infection, and new-generation EIA and IgG avidity assays appear to hold promise for improved serodiagnosis during pregnancy (Enders et al, 2006, 2008). Treatment Spontaneous resolution of fetal hydrops with normal neonatal outcome has been reported in approximately one third of cases (Humphrey et al, 1991; Rodis et al, 1998a; Sheikh et al, 1992). Because two thirds of fetuses do not recover without intervention, fetal transfusion is usually recommended (Boley and Popek, 1993; Brown et al, 1994). The earlier fetal transfusion is attempted, the more likely it is to be successful. Cordocentesis allows precise assessment of the magnitude of fetal anemia, which can then be corrected by blood transfusion, typically using packed red blood cells. Using this approach, outcomes have been favorable in most reported series, even among severely anemic fetuses. In one report, packed red cell transfusion was performed in 30 patients with fetal anemia (hemoglobin values ranging from 2.1 to 9.6 g/dL). The overall survival rate was 83.8% (Schild et al, 1999). In a report of 13 cases with severe hydrops fetalis who received intrauterine transfusion, 11 of 13 survived (84.6%), whereas all the nontransfused fetuses with severe hydrops fetalis died (Enders et al, 2004). Prognosis Mortality rates for fetal hydrops resulting from all nonimmune causes continue to exceed 50%, even with current aggressive therapies and intensive care (Huang et al, 2007; Wy et al, 1999). Hydrops secondary to parvovirus B19 seems to have a better outcome than that from other causes (Enders et al, 2004; Ismail et al, 2001). Although 489 few long-term prospective studies of infants born to mothers with documented primary parvovirus B19 infection have been conducted, most report normal developmental outcome. One case-control study involving approximately 200 mother–infant pairs found no differences in frequency of developmental delay between infants born to women with confirmed primary parvovirus B19 infection during pregnancy and infants born to mothers with evidence of preconceptional immunity (Rodis et al, 1998b). Other studies have also suggested a favorable long prognosis in children born to women with primary parvovirus B19 infections during pregnancy (Miller et al, 1998). Prevention If a pregnant woman has a significant exposure to an infectious case of parvovirus B19, counseling should be provided regarding the potential risk of infection. Anti–parvovirus IgM and IgG serologic analyses and serum PCR should be performed; if they show evidence of primary infection, then serial fetal ultrasonographic evaluations should be performed. Postexposure passive immunization with immunoglobulin is not currently recommended because the period of maternal viremia has passed by the time the diagnosis of acute parvovirus B19 infection is made (Boley and Popek, 1993). Although high-dose IVIG has been used to attempt to prevent hydrops fetalis during pregnancy in the setting of acute infection (Selbing et al, 1995), treatment with this modality in the pregnant woman or the neonate has not been shown to improve fetal outcomes; therefore it is not routinely recommended. Human IgG monoclonal antibodies with potent neutralizing activity have been generated, and these are suggested as candidates for the development of immunotherapeutic approaches for individuals chronically infected with parvovirus B19 virus or for acutely infected pregnant women (Gigler et al, 1999), but these interventions are not commercially available. There has been limited progress in the development of a candidate parvovirus B19 vaccine. Phase 1 studies of a recombinant vaccine based on baculovirus-produced capsids have been conducted, and this vaccine was found to have a favorable safety and immunogenicity profile (Bansal et al, 1993). Efforts are under way to research and develop vaccines to prevent parvovirus B19 infections using other expression systems (Lowin et al, 2005). Pregnant health care providers should be counseled about the potential risks to their fetus from parvovirus B19 infections and should, at a minimum, wear masks and use standard droplet precautions when caring for immunocompromised patients with chronic parvovirus B19 infection or patients with parvovirus B19–induced aplastic crises. Some hospitals exclude pregnant health care providers from caring for these high-risk patients, but this issue remains controversial. RUBELLA Rubella virus is an enveloped, single-stranded, positive sense RNA virus belonging to the family Togaviridae. Although togaviruses are typically vector-borne infections, rubella is the notable exception, being transmitted instead by respiratory droplet. Humans are the only known natural 490 PART IX Immunology and Infections host for rubella virus. Rubella infection, commonly known as the German measles, usually results in a mild illness with an accompanying exanthem in adults and children; however, rubella produces serious consequences in pregnant patients, in whom fetal infection can lead to serious anomalies. An ophthalmologist named Norman Gregg offered the first description of congenital rubella syndrome (CRS) in 1941 while investigating an epidemic of neonatal cataracts (Gregg, 1941). Not until the global pandemic of 1964 to 1965, however, were the multiple teratogenic manifestations of CRS fully appreciated and the permanent neurodevelopmental consequences for newborns fully recognized. The capacity to grow the virus in tissue culture led rapidly to the development of a vaccine and subsequent a reduction in the incidence of CRS in the United States and other developed countries. However, CRS is still encountered in the developing world, and unfounded concerns about the safety of measles-mumps-rubella vaccine have set the stage for potential reemergence of these diseases (Omer et al, 2009). Therefore a working knowledge of CRS remains highly relevant to the care of newborns. EPIDEMIOLOGY Since the development of rubella vaccine in 1969, the incidence of rubella in the United States has decreased dramatically. The annual incidence of rubella cases has dropped 99%, from 58 per 100,000 population in 1969 to less than 0.5 per 100,000 population in 1997 to 1999 (Danovaro-Holliday et al, 2001). CRS cases in the United States have demonstrated a similar dramatic decline, and in 2004 the Centers for Disease Control and Prevention (CDC) concluded, against the background of few or no reports of rubella activity from the 50 states and the virtual absence of reported CRS, that endemic rubella had been eliminated from the United States (Centers for Disease Control and Prevention, 2005; Plotkin, 2006; Reef et al, 2006). The persistence of rubella in Latin America serves as the source for a small number of importation cases of rubella that are recognized in the United States annually (Castillo-Solórzano et al, 2003). In parts of the world where routine rubella immunization is unavailable or not used, rubella and CRS continue to be common. Rubella is still an important and potentially preventable cause of birth defects globally, with more than 100,000 cases of CRS occurring annually in developing countries. Maternal rubella infection that occurs in the period of time from 1 month before conception through the second trimester of pregnancy may be associated with transmission of infection and, depending on the timing of infection, disease in the infant. The frequency of congenital infection after maternal rubella with a rash is 70% to 85% if infection occurs during the first 12 weeks of gestation, 30% to 54% during the first 13 to 16 weeks of gestation, and 10% to 25% at the end of the second trimester (Miller et al, 1982; South and Sever, 1985). The classic findings of congenital rubella are most typically associated with the onset of maternal infection during the first 8 weeks of gestation (Miller et al, 1982). Both the risk of fetal infection and severity of disease decline after the first trimester, and the risk of any teratogenic effect is extremely low after 17 weeks’ gestation (Lee and Bowden, 2000). PATHOGENESIS Rubella virus is transmitted via respiratory droplets. Once the oral or nasopharyngeal mucosae are infected, viral replication occurs in the upper respiratory tract and nasopharyngeal lymphoid tissue. The virus then spreads contiguously to regional lymph nodes and hematogenously to distant sites. Fetal infection is believed to occur as a consequence of maternal viremia. The mechanism by which rubella infection of the fetus leads to teratogenesis has not been fully determined, but the cytopathology in infected fetal tissues suggests necrosis, apoptosis, or both, as well as inhibition of cell division of precursor cells involved in organogenesis (Atreya et al, 2004; Lee and Bowden, 2000). The rubella replicase protein p90 interacts with a cellular cytokinesis-regulatory component (i.e., the citron-K kinase) in the process leading to tetraploidy and cell cycle arrest. In tissue culture, a number of unusual manifestations of rubella replication have been observed, including mitochondrial abnormalities and disruption of the cytoskeleton. Characteristic markers of apoptosis such as DNA fragmentation, nuclear chromatin condensation, and annexin V staining can be observed following rubella infection in cell culture. Cytoplasmic inclusions have also been reported in some cell lines. Microarray analysis following rubella infection in cell culture demonstrated upregulation of cytokines and interferon, suggesting that the induction of inflammatory responses may serve as another possible mechanism of injury induced by rubella (Adamo et al, 2008). Fetuses infected with rubella demonstrate cellular damage in multiple sites and a noninflammatory necrosis in target organs, including eyes, heart, brain, and ears (Lee and Bowden, 2000). CLINICAL SPECTRUM The peak incidence of endemic rubella is in the late winter and early spring months. Up to 50% of primary infections are asymptomatic. The period of maximal communicability extends from a few days before until 7 days after onset of the rash. Often a prodrome of mild systemic symptoms precedes the rash by 1 to 5 days. Viremia can be detected as early as 9 days before the onset of rash. Lymphadenopathy, which can precede a rash, often is present in the posterior auricular or suboccipital region. The rash classically begins on the face and spreads caudally to the trunk and extremities. Symptoms generally last up to 3 days, and the incubation period ranges from 14 to 21 days. Transmission by breastfeeding in a case of postpartum acquisition of infection has been described (Klein et al, 1980). Typical illness in adults and children with acquired rubella infection consists of an acute generalized maculopapular rash, fever, and arthralgias, arthritis, or lymphadenopathy. Conjunctivitis is also common. Encephalitis (1:5000 cases) and thrombocytopenia (1:3000 cases) are complications. Infants with congenital rubella are usually born at term, but often are small for gestational age. The most common isolated sequela is hearing loss (Miller et al, 1982; Ueda et al, 1979). The next most common findings are heart defects, cataracts, low birthweight, hepatosplenomegaly, and microcephaly. The triad of deafness, cataracts, and congenital heart disease constitutes the classic syndrome. CHAPTER 37 Viral Infections of the Fetus and Newborn and Human Immunodeficiency Virus Infection during Pregnancy In addition, systemic illness can occur and be characterized by purpura, hepatosplenomegaly, jaundice, pneumonia, and meningoencephalitis; 45% to 70% of infants have cardiac lesions, including patent ductus arteriosus, peripheral pulmonic stenosis, and valve abnormalities (Reef et al, 2000; Schluter et al, 1998). Recently, an extensive review of published reports of cardiovascular disease in the setting of congenital rubella syndrome demonstrated that branch pulmonary artery stenosis is the most commonly identified isolated lesion (Oster et al, 2010). A variety of other signs may be present. Additional ocular findings are pigmentary retinopathy, microphthalmia, and strabismus. The skin lesions have been described as resembling a blueberry muffin and represent extramedullary dermal hematopoiesis; an identical rash can be observed in congenital CMV infection (Avram et al, 2007; Bowden et al, 1989; Brough et al, 1967; Mehta et al, 2008). The clinical manifestations of CRS vary to some extent depending on the timing of fetal infection. In a prospective study following pregnant women with confirmed rubella infection by trimester, a full range of rubella-associated defects (including congenital heart disease and deafness) were observed in nine infants infected before the 11th week. Thirty-five percent of infants (9 of 26) infected between 13 and 16 weeks’ gestation had deafness alone (Miller et al, 1982). Cataracts typically occur secondary to maternal rubella infection occurring before day 60 of pregnancy; heart disease is found almost exclusively when maternal infection is before the 80th day (i.e., first trimester). Disease manifestations that may have their onset after birth (late-onset disease) include: a generalized rash with seborrheic features that may persist for weeks, acute or chronic interstitial pneumonia, abnormal hearing resulting from presumed labyrinthitis, central auditory imperception, and progressive rubella panencephalitis (Franklin and Kelley, 2001; Phelan and Campbell, 1969; Reef et al, 2000; Sever et al, 1985). A higher than expected incidence of autoimmune diseases, such as thyroid disorders and diabetes mellitus, have also been reported years after the diagnosis of congenital rubella (Forrest et al, 2002; Gale, 2008; McEvoy et al, 1988; Reef et al, 2000). Infants with late-onset disease have demonstrated immunologic abnormalities, including dysgammaglobulinemia or hypogammaglobulinemia (Hancock et al, 1968; Hayes et al, 1967; Soothill et al, 1966). Hyper-IgM syndrome has also been described in association with autoimmune disease in a child with CRS (Palacin et al, 2007). Other studies have demonstrated dysfunction of cellular immune responses in children with CRS (Fuccillo et al, 1974; South et al, 1975; Verder et al, 1986). The pathogenesis of these rubella-related syndromes is poorly understood. Psychiatric disturbances, including some with features of autism spectrum disorders, have been observed decades after CRS (Hwang and Chen, 2010). LABORATORY EVALUATION It is common obstetric practice to screen all pregnant women for rubella antibodies, and because of the devastating consequences of CRS, this is a reasonable policy to continue, even in the setting of eradication of endemic rubella in the United States. Women who are exposed to 491 rubella should be screened for evidence of previous immunity; if none is found, they should be tested for rubella IgG and IgM antibodies. A positive IgM titer or a rise in paired IgG titers is indicative of recent infection. Women with such findings should also be evaluated to try to determine the likely gestational age at time of infection in order to assess the potential risk to the fetus. The laboratory diagnosis of congenital rubella can be made definitively only during the 1st year of life, unless the virus can be recovered later from an affected site, such as the lens. Diagnosis can be made with any one of the following four criteria: 1. Positive anti-rubella IgM titer, preferably determined with enzyme immunoassays, but indirect assays are acceptable 2. A significant rise in rubella IgG titer between acute and convalescent measurements 2 to 3 weeks apart or the persistence of high titers longer than expected from passive maternal antibody transfer 3. Isolation of rubella virus cultured from nasal, blood, throat, urine, or CSF specimens (throat swabs have the best yield) 4. Detection of virus by reverse transcriptase PCR in specimens from throat swabs, CSF, or cataracts obtained from surgery (Centers for Disease Control and Prevention, 2001) An infected infant can excrete the virus for many months after birth despite the presence of neutralizing antibody and, thus, may pose a hazard to susceptible individuals. In rare cases the virus can be recovered after 1 year of age. An exception to this rule is the cataract, in which the virus can remain for as long as 3 years. In late-onset disease, the virus can also be found in affected skin and lung. Other laboratory findings are thrombocytopenia, hyperbilirubinemia, and leukopenia. Radiographic findings include large anterior fontanel, linear areas of radiolucency in the long bones (i.e., celery stalking), increased densities in the metaphyses, and irregular provisional zones of calcification (Chapman, 1991; Reed, 1969). The radiographic changes seen in rubella are not pathognomonic of the disease, but resemble those seen in other congenital viral infections, including congenital cytomegalovirus infection (Alessandri et al, 1995). TREATMENT AND PROGNOSIS There is no specific therapy for congenital rubella. Initially the infant may need general supportive care, such as administration of blood transfusion for anemia or active bleeding, seizure control, and phototherapy for hyperbilirubinemia. Long-term care requires a multidisciplinary approach consisting of occupational and physical therapy, close neurologic and audiologic monitoring, and surgical interventions as needed for cardiac malformations and cataracts. The consequences of fetal rubella infection may not be evident at birth. In one study of 123 infants with documented congenital rubella, 85% of cases were not diagnosed until after discharge from the nursery (Hardy, 1973). Communication disorders, hearing defects, some mental or motor retardation, and microcephaly by 1 to 3 years of age were among the major problems that were discovered 492 PART IX Immunology and Infections after the newborn period. A predisposition to inguinal hernias was also noted. Longitudinal studies of somatic growth show that most infants with congenital rubella remain smaller than average throughout infancy, but grow at a normal rate. Stunting of growth was more common after rubella infection in the first 8 weeks of pregnancy than after later infection. Even in the absence of mental retardation, neuromuscular development is commonly abnormal. A study of neurodevelopmental outcomes in 29 affected children without mental retardation found that 25 had other abnormalities; hearing loss, difficulties with balance and gait, learning deficits, and behavioral disturbances were found in more than half of the affected children (Desmond et al, 1978). PREVENTION There is no effective antiviral therapy against congenital rubella infection, so the most useful practice is to ensure that women who are considering pregnancy are immune. The Advisory Committee on Immunization Practices recommends screening of all pregnant women for rubella immunity and postpartum vaccination of those who are susceptible (Centers for Disease Control and Prevention, 2001). Immunity to rubella appears to confer almost complete protection against CRS. Rare cases of documented subclinical maternal reinfection with rubella have been reported (Morgan-Capner et al, 1991; Saule et al, 1988). In rare cases, maternal reinfection can lead to CRS (Banerji et al, 2005). Approximately 20 cases of CRS after maternal reinfection have been reported in the literature, and none have caused symptomatic CRS when the known reinfection occurred after 12 weeks’ gestation (Bullens et al, 2000). Live attenuated rubella virus vaccine is safe and effective, although the duration of immunity is uncertain. It is typically administered in the United States in a trivalent formulation in combination with measles-mumps-rubella vaccine or in a quadrivalent combination with measlesmumps-rubella-varicella vaccine. The vaccine is recommended for children at 12 to 15 months of age and at 4 to 5 years of age. It is also recommended for women of childbearing age in whom results of both a hemagglutination inhibition antibody test and a pregnancy test are negative. Although no cases of symptomatic congenital rubella infection have been reported as a consequence of vaccination during pregnancy in the more than 500 cases monitored, vaccination is not recommended during pregnancy because of the theoretical hazard to the fetus (Josefson, 2001; Nasiri et al, 2009; Tookey, 2001). A mild rubella-like illness is sometimes seen after immunization, with arthralgia occurring 10 days to 3 weeks after injection. If a woman is found to be susceptible, vaccine should be administered during the immediate postpartum period before discharge. Breastfeeding is not a contraindication to postpartum immunization. Immunization in the postpartum period has rarely produced polyarticular arthritis, neurologic symptoms, and chronic rubella viremia (Tingle et al, 1985). The problem of management of the pregnant woman who is exposed to rubella or who contracts the disease should be resolved after the known risks are weighed. If serum antibody is detectable at the time of exposure, the fetus is probably protected. If no antibody is detectable, additional serum samples at 2 to 3 weeks after exposure and again at 4 to 6 weeks after exposure should be obtained. These samples can be run concurrently with the first serum to ascertain whether infection has occurred (i.e., seroconversion). There is no evidence that administration of immunoglobulin prevents rubella infection or viremia. Moreover, administering immunoglobulin may provide an unwarranted sense of security, because infants with congenital rubella have been born to women who received immunoglobulin shortly after exposure. Regardless of these limitations, administration of immunoglobulin may be considered if the pregnancy is not terminated. Immunoglobulin can reduce the likelihood of fetal infection, but will not eliminate the risk; therefore the CDC does not routinely recommend the use of immunoglobulin in a pregnant woman for postexposure prophylaxis unless she does not wish to terminate the pregnancy under any circumstances (Centers for Disease Control and Prevention, 2001). Decisions about the termination of pregnancy should be made only after maternal infection has been proved and should also account for the risk of rubella-associated damage to the fetus, which is highest when maternal infection occurs during the first 8 weeks of pregnancy. LYMPHOCYTIC CHORIOMENINGITIS VIRUS Lymphocytic choriomeningitis virus (LCMV) is a member of the family Arenaviridae. Rodents are the primary reservoir, particularly mice and hamsters. Like other arenaviruses, LCMV has a bisegmented negative-strand RNA genome. The S segment encodes for the virus nucleoprotein and glycoprotein, whereas the L segment encodes for the virus polymerase (L) and Z protein. Susceptible rodents are infected asymptomatically in utero, can harbor chronic infection, and excrete virus in urine, feces, saliva, nasal secretions, milk, and semen for life. Typically mice remain asymptomatic, although hamsters may demonstrate viremia and viruria with variable symptoms (Jahrling and Peters, 1992). Sequelae of human exposure to LCMV range from asymptomatic infection to nonspecific, flulike symptoms; a proportion of infections have neurologic manifestations. LCMV was first described as a cause of congenital infection in England in 1955 in a 12-dayold infant (Komrower et al, 1955) and later in the United States (Barton and Mets, 2001; Barton et al, 1993, 2001; Larsen et al, 1993). Because LCMV has only recently been recognized as a source of congenital infection, it is likely underdiagnosed (Jamieson et al, 2006). EPIDEMIOLOGY Human seroprevalence ranges between less than 1% and 10% worldwide and varies extensively with geographic region (Ambrosio et al, 1994; Childs et al, 1991; Marrie and Saron, 1998; Stephensen et al, 1992). One study noted a higher prevalence in women (Marrie and Saron, 1998). Lower socioeconomic status and older age are associated with higher seroprevalence. Studies conducted in the 1940s through the 1970s found that approximately 8% to 11% of cases of aseptic meningitis and encephalitis were CHAPTER 37 Viral Infections of the Fetus and Newborn and Human Immunodeficiency Virus Infection during Pregnancy associated with LCMV infection (Meyer et al, 1960; Park et al, 1997b). In temperate climates, human exposure is more common during the fall and winter, when rodents move indoors. Outbreaks have been reported in laboratory personnel working with hamsters and mice (Dykewicz et al, 1992; Hinman et al, 1975; Vanzee et al, 1975). ­Multiple outbreaks associated with pet hamsters have also been reported in the United States (Biggar et al, 1975; Maetz et al, 1976) and Europe (Brouqui et al, 1995; ­Deibel et al, 1975); however, congenital infection is relatively rare. A total of 54 cases have been diagnosed worldwide since the first case in 1955, and 27 of those occurred in the United States (Barton and Mets, 2001; Greenhow and Weintrub, 2003). There has been an increased recognition of congenital LCMV in recent years; 34 of the cases described in a review were reported since 1993 (Jamieson et al, 2006). The true frequency of congenital LCMV infection is unknown, because there is no active surveillance. As with other congenital infections, there may be a wide spectrum of disease, including asymptomatic and subclinical or nonspecific infections. PATHOGENESIS Humans acquire LCMV infection from aerosolized particles, bites, or fomite contact with virus excreted from rodents (Jahrling and Peters, 1992). Human-to-human horizontal transmission by organ transplantation has been documented (Fischer et al, 2006). In this report, abdominal pain, altered mental status, thrombocytopenia, elevated aminotransferase levels, coagulopathy, graft dysfunction, renal failure, seizures, and either fever or leukocytosis were variably present within 3 weeks after transplantation. Seven of the eight recipients died, 9 to 76 days after transplantation. The pathogenesis of LCMV infection is poorly understood, although it is likely an immunopathologic process mediated by the host CD8+ T-cell response (Craighead, 2000). It is also postulated that the high rate of spontaneous mutations that arise during LCMV replication allows both variability in pathogenicity and a mechanism of escape from humoral response during the initial phase of infection (Ciurea et al, 2001). Like other arenaviruses, LCMV replicates either at the site of infection or in corresponding lymph nodes; this localized replication is followed by viremia. It is thought that during the viremic stage, the virus travels to parenchymal organs and the CNS. Pathologic findings include lymphocytic infiltration and extramedullary hematopoiesis. In two congenitally infected infants for whom neuropathology results were available, cerebromalacia, glial proliferation, and perivascular edema were reported (Barton et al, 1993). In adult mice inoculated intracranially with LCMV, subsequent viral proliferation in the ependyma, meninges, or both have been described, with injury mediated by CD8+ T cells (Doherty et al, 1990). This viral infiltration, along with the host inflammatory response, can lead to aqueductal stenosis and subsequent hydrocephalus. Studies in a murine model suggest that cytokine-chemokine cascades mediate much of the neuropathology observed in the setting of CNS infection (Christensen et al, 2009). 493 CLINICAL SPECTRUM It is estimated that asymptomatic or mild LCMV infections occur in approximately one third of patients infected; however, the classic presentation of LCMV infection is a nonspecific, flulike, or mononucleosis-like illness that is often biphasic. Symptoms are fever, malaise, nausea, vomiting, myalgias, headache, photophobia, pharyngitis, cough, and adenopathy. After defervescence and resolution of these constitutional symptoms, a second phase of CNS disease may develop. Neurologic manifestations occur in approximately one fourth of infectious episodes and vary from aseptic meningitis to meningoencephalitis. Transverse myelitis, Guillain-Barré syndrome, and deafness have also been reported. Other manifestations are pneumonitis, arthritis, myocarditis, parotitis, and dermatitis. Recovery may take months, but usually occurs without sequelae (Craighead, 2000). When LCMV infection occurs during pregnancy, maternal symptoms typically appear during the first and second trimesters, but only 50% to 60% of mothers of infants diagnosed with LCMV congenital infection recall having symptoms (Wright et al, 1997). Known maternal exposure to rodents is reported in approximately one fourth to one half of cases (Barton and Mets, 2001). Usually, exposed women are from rural settings, come from lower socioeconomic settings with substandard housing conditions, or have pet rodents in the home (i.e., hamsters). The complete spectrum of disease secondary to congenital LCMV infection is still uncertain, although chorioretinitis and hydrocephalus are the predominant characteristics reported among children diagnosed with congenital LCMV infection. Chorioretinitis is present in more than 90% of cases. Other ocular findings are chorioretinal scars, optic atrophy (usually bilateral), nystagmus, esotropia, exotropia, leukocoria, cataracts, and microphthalmia (Barton and Mets, 2001; Brezin et al, 2000; Enders et al, 1999). Some ophthalmologic findings resemble those described in the lacunar retinopathy of Aicardi syndrome (Wright et al, 1997), a finding of interest insofar as Aicardi syndrome can, like congenital CMV or LCMV infection, produce intracranial calcifications. A wide range of neurologic defects are described, including microencephaly, encephalomalacia, chorioretinitis, porencephalic cysts, neuronal migration disturbances, periventricular infection, and cerebellar hypoplasia (Bonthius et al, 2007). Congenital LCMV should be considered in the differential diagnosis of neonatal hydrocephalus (Schulte et al, 2006). Most infants are born at term, and birthweights are generally appropriate or large for gestational age. Thirty-five percent to 40% of infants reported with congenital LCMV have had microcephaly or macrocephaly at birth (Barton and Mets, 2001). Systemic symptoms are rare, although hepatosplenomegaly and jaundice have been noted. Other individual case report findings are pes valgus, dermatologic findings consistent with staphylococcal scalded-skin syndrome (Wright et al, 1997), spontaneous abortion (Biggar et al, 1975), and intrauterine demise secondary to hydrops fetalis (Enders et al, 1999; Meritet et al, 2009). It is important to note that, because systemic symptoms are typically minimal at birth, the diagnosis of congenital LCMV infection often is not considered until an affected infant is a few 494 PART IX Immunology and Infections months of age, when microcephaly, macrocephaly, visual loss, or developmental delay may be noted. LABORATORY EVALUATION Serology is the most reliable and feasible method to diagnose LCMV. In most reports, the diagnosis was established by testing of the infant’s serum, CSF, or both; in some, maternal serum testing was the key. Testing all three fluids provides the most information. Because of the low baseline population seroprevalence, positive titers for LCMV are much more useful for diagnosis than detection of antibodies to microbes such as CMV and toxoplasmosis. There is a commercially available immunofluorescent antibody test that detects both IgM and IgG for LCMV. It has better sensitivity than the complement fixation and neutralizing antibody tests that also have been used (Lehmann-Grube et al, 1979; Lewis et al, 1975). Complement fixation titers generally do not rise until more than 10 days after onset of infection, but immunofluorescent antibody results may be positive within the first few days of illness (Deibel et al, 1975). The CDC also has an enzyme-linked immunosorbent assay test for IgM and IgG; it may be more useful for diagnosis in an older child because it can detect increased IgG later than and persistent IgG for longer than the immunofluorescent antibody test. Some studies have found antibody as late as 30 years after suspected exposure. Virus can be cultured in Vero cell lines or inoculated into newborn mice, but use of these methods is uncommon. Reverse transcriptase–PCR has been used in serum and CSF to diagnosis LCMV and as a surveillance tool; it may become more available in the future (Enders et al, 1999; McCausland and Crotty, 2008; Park et al, 1997a). Information about routine laboratory data in patients with a diagnosis of congenital LCMV infection is minimal, but thrombocytopenia and hyperbilirubinemia have been reported. CSF findings are variable. Up to one half of cases demonstrate a mild increase in white blood cell count (up to 64 cells/μL in one case series of 18 infants), the serum protein concentration may be normal or mildly elevated, and the serum glucose concentration may be normal or mildly decreased (Wright et al, 1997). Among infants reported in whom neuroradiographic imaging was performed, 89% (17 of 19) had hydrocephalus or periventricular calcifications. Flattened gyri, lissencephaly, and schizencephaly have been reported (Barton and Mets, 2001), compatible with a role for LCMV in fetal neuronal migration defects (Bonthius et al, 2007). TREATMENT Ribavirin has been used for management of other arenavirus infections and inhibits LCMV growth in vitro (Géssner and Lother, 1989). Although novel approaches are being used to develop antivirals against LCMV and other pathogenic arenaviruses (de la Torre, 2008), there are currently no recommendations for the use of antiviral agents against these viruses. In the outbreak associated with solid organ transplantation, one affected recipient received ribavirin and reduced levels of immunosuppressive therapy and survived (Fischer et al, 2006). PROGNOSIS Because congenital LCMV infections have been recognized relatively recently and the existing data come from case reports, there may be a wider spectrum of disease than is currently appreciated. The proportion of asymptomatic infected infants is unknown. For the 25 infant cases reported, estimated mortality rates are 16% to 35% between birth and 21 months of age (Barton and Mets, 2001; Wright et al, 1997). Among infants who survived, 84% (32 of 38) have neurologic sequelae, including spastic quadriparesis, mental retardation, developmental delay, seizures, and visual loss (Barton and Mets, 2001). Sensorineural hearing loss is less common and has been reported in only two infants (Barton and Mets, 2001; Wright et al, 1997). Some ophthalmologists suggest that among patients with developmental delay and visual loss consistent with chorioretinitis, LCMV congenital infection may be underdiagnosed. A study in Chicago of prospectively diagnosed patients with chorioretinitis and patients in a home for severely mentally retarded children with chorioretinal scars found six children with elevated LCMV titers and normal toxoplasmosis, CMV, rubella, and HSV titers (Mets et al, 2000). Two children with chorioretinal scars and elevated LCMV titers have been reported in France (Brezin et al, 2000). PREVENTION Public health officials and clinicians should be aware that (1) wild, laboratory, and pet rodent exposure can lead to intrauterine infection with LCMV virus and (2) congenital infection has been associated with potentially devastating ophthalmologic and neurologic sequelae. Pregnant women need to be educated about the potential risks of exposure to infected rodent excreta and instructed to avoid rodents and rodent droppings whenever possible. Obstetricians and neonatologists should seek a history of pet or wild rodent exposure for counseling purposes and to aid in the evaluation of infants with unexplained CNS pathologies. The diagnosis of LCMV should be considered in all unexplained cases of infant hydrocephalus. There are currently no vaccines approved by the U.S. Food and Drug Administration (FDA) for the prevention of arenavirus disease, although candidate multivalent arenavirus vaccines capable of providing T cell–mediated protection against a variety of pathogenic arenaviruses are currently in development (Botten et al, 2010). ENTEROVIRUSES Enteroviruses are single-stranded, positive-sense RNA viruses. These viruses belong to the family Picornaviridae (pico means very small in Spanish). The enteroviruses of humans include polioviruses 1, 2, and 3, coxsackieviruses A and B (named after Coxsackie, New York the city where these viruses were first identified and characterized), and the echoviruses (echo is an acronym for enteric cytopathic human orphan). Of these, poliovirus infection has historically been responsible for the greatest morbidity in infants. Severe, often fatal poliovirus disease used to occur with great frequency in infants infected in the perinatal period, CHAPTER 37 Viral Infections of the Fetus and Newborn and Human Immunodeficiency Virus Infection during Pregnancy with a high incidence of residual paralysis in survivors (Bates, 1955; Cherry, 2005). Fortunately, poliovirus infections have become rare in the developed world, because of the widespread implementation of effective immunizations. However, neonatal diseases associated with coxsackieviruses and echoviruses, for which there are no vaccines, remain common and can be associated with serious morbidity and occasional mortality. Typically acquired from a maternal source, these agents are associated with a wide range of clinical syndromes in the neonatal intensive care unit, including CNS infection, myocarditis, and a sepsislike syndrome. Enteroviruses are also responsible for a large number of hospital readmissions for evaluation of febrile syndromes in infants younger than 2 months. EPIDEMIOLOGY Enterovirus infections are seasonal, occurring most commonly during the summer and autumn in temperate climates. The incidence varies from year to year, with outbreaks sometimes caused by a single coxsackievirus or echovirus serotype and sometimes by several (Sawyer et al, 1994). In older children, enteroviruses are transmitted by the fecal-oral route and are typically associated with a variety of febrile syndromes, including febrile exanthematous syndromes, aseptic meningitis, pneumonia with or without pleural effusion, and myocarditis. Disease in newborn infants is relatively uncommon, but reflects the frequency of infection in the general population (Krajden and Middleton, 1983). Enteroviruses may also be associated with nosocomial outbreaks. Nursery and obstetric clinic outbreaks of both coxsackievirus B (Bhambhani et al, 2007; Brightman et al, 1966; Rantakallio et al, 1970) and echovirus infections (Chen et al, 2005; Jankovic et al, 1999; ­Nagington et al, 1978) have been reported and associated with severe, and sometimes fatal, illnesses. Enteroviral infections account for at least one third of neonatal febrile admissions for suspected sepsis and for between half and two thirds of all admissions during peak enteroviral season (Byington et al, 1999; Dagan, 1996). Neonatal aseptic meningitis is also frequently caused by enteroviral infections. In a review of neonatal meningitis seen over a 15-year period in Galveston, Texas, enterovirus was the most common cause of meningitis in newborn infants older than 7 days of age (Shattuck and Chonmaitree, 1992). Enteroviruses, along with other viruses, have been implicated as a potential cause of sudden infant death syndrome (SIDS), possibly from myocarditis or pulmonary infection (Grangeot-Keros et al, 1996; Shimizu et al, 1995), but this association has been controversial. In one study of SIDS victims, a comprehensive assessment was undertaken to attempt to identify potential viral infection of the myocardium. Overall, 62 SIDS victims and 11 controls were studied. Enteroviruses were detected in 14 (22.5%), adenoviruses in 2 (3.2%), Epstein–Barr viruses in 3 (4.8%), and parvovirus B19 in 7 (11.2%) cases, whereas control group samples were completely negative for viral nucleic acid (Dettmeyer et al, 2004). However, an evaluation of histopathologic features and PCR analysis from 24 SIDS cases failed to demonstrate any association with viral infection (Krous et al, 2009), leaving this putative association unclear. 495 ETIOLOGY AND PATHOGENESIS Neonates can acquire enteroviral infections secondary to in utero transmission, intrapartum transmission during labor and delivery, or postnatally. Intrauterine infections appear to occur via transplacental spread, secondary to maternal viremia, and this mode of transmission appears to be responsible for up to 22% of cases of neonatal enteroviral infection (Kaplan et al, 1983; Modlin, 1986). Enteroviruses have been implicated as a cause of fetal demise (Johansson et al, 1992; Konstantinidou et al, 2007; Nielsen et al, 1988). Intrapartum or postnatal transmission is more common than transplacental transmission. The dominant mode of transmission of serious neonatal infection is through contact with maternal blood, fecal material, or vaginal or cervical secretions, most likely during or shortly after delivery (Hawkes and Vaudry, 2005; Jenista et al, 1984). After acquisition of infection, viremia ensues in the infant, leading to a variety of end-organ diseases. Virus strain or serotype appears to be a factor in predicting the severity of illness, as does the quantity of the inoculum. The presence or absence of transplacental maternally derived antibody also dictates the severity of disease in the infected infant. Severe disease occurs in the absence of type-specific antibody, a scenario that is more likely if the maternal infection is acquired near the time of delivery. Breastfeeding appears to provide a relative degree of protection against acquisition of infection in neonates (Sadeharju et al, 2007). One of the most comprehensive analyses of the incidence of neonatal enterovirus disease came from a prospective study in Rochester, N.Y. This study demonstrated that approximately 13% of all newborns tested positive for enterovirus from throat or stool cultures during a typical season (June to October; Jenista et al, 1984). Although this result represented a remarkably high incidence, 79% of these infections were asymptomatic. The most common clinical findings in the 21% of symptomatic infections were lethargy and fever. These infants were typically admitted to hospital for an evaluation to rule out sepsis, making enterovirus infection a more common reason for hospitalization owing to an infectious disease than group B streptococcus, herpes simplex virus, and cytomegalovirus infections combined. It appears that any of the non–polio enteroviruses can cause disease in the newborn infant. A variety of clinical syndromes are associated with certain enteroviruses. A retrospective chart review of 24 neonatal enteroviral infections in Toronto, Canada, found that 10 infants died, 12 had aseptic meningitis, and 5 had myocarditis (Krajden and Middleton, 1983). Of the 24 isolates, 7 were echovirus, 15 were coxsackievirus B, one was coxsackievirus A, and one was nontypable. In infants requiring hospitalization with acute enterovirus disease, coxsackievirus B is associated primarily with myocarditis and aseptic meningitis (Kibrick and Benirschke, 1958). Recent reports have identified coxsackievirus B1 as an emerging cause of life-threatening myocarditis and other severe, fatal syndromes in neonates (Verma et al, 2009; Wikswo et al, 2009). Echoviruses are associated with severe nonspecific febrile illnesses with disseminated intravascular coagulation (Nagington et al, 1978), aseptic meningitis (Cramblett et al, 1973), or 496 PART IX Immunology and Infections hepatitis (Modlin, 1980). With both coxsackievirus and echovirus, nonspecific febrile illnesses, with or without an exanthem, are commonly observed. Enterovirus 71 is particularly notable for its etiologic role in epidemics of severe neurologic diseases in children (Chen et al, 2010). Nursery-based outbreaks of this neurovirulent enterovirus have also been described (Huang et al, 2010). Except for transplacental infections, the portal of entry for enterovirus is via the oral or respiratory route. After replication in the pharynx and the gastrointestinal tract, virus seeds the tonsils, cervical and mesenteric nodes, and Peyer’s patches. The pathogenesis of enterovirus disease stems from the ensuing viremia, which can lead to infection of the heart, CNS, liver, pancreas, adrenal glands, skin, mucous membranes, and respiratory tract. This feature of enterovirus replication and dissemination helps to explain the diverse disease manifestations that may be observed following infection. The humoral immune response is associated with recovery from systemic and end-organ infection, although virus may continue to be shed in the stool for several weeks. Failure to clear enteroviral infection, particularly from the CNS, should suggest an underlying humoral immunodeficiency, although this would not typically be observed in the neonatal period, when maternally derived Ig is present in the newborn (McKinney et al, 1987; Misbah et al, 1992). Depending on the extent of disease, extensive end-organ pathology may be observed. In coxsackievirus B infections, myocardial necrosis, and inflammation may be seen that are patchy or diffuse, with extensive infiltration by lymphocytes, mononuclear cells, histiocytes, and polymorphonuclear leukocytes. Similar infiltrates are seen in the meninges in both coxsackievirus and echovirus aseptic meningitis. Brainstem encephalitis can be observed with enterovirus 71 infection, often in association with pulmonary edema (Wang and Liu, 2009). When liver or adrenal glands are involved, there is usually extensive hemorrhage as well as inflammation and necrosis; such fatal fulminant infections are often associated with echovirus 11 (Mostoufizadeh et al, 1983). CLINICAL SPECTRUM When neonatal disease is acquired vertically from the mother, the infant is typically asymptomatic at birth, although premature delivery is more common. The mother may be febrile at this time or may have a history of recent high fevers and gastrointestinal symptoms. Fever, anorexia, and vomiting develop in the baby after an incubation period of 1 to 5 days. The onset of illness occurs in the first week of life in more than 50% of affected infants (Krajden and Middleton, 1983). At that point, the clinical evolution depends on the infecting virus and the extent of end-organ involvement. In most instances, the disease is mild and self-limited. Symptomatic infections may be characterized by rash, aseptic meningitis, hepatitis, and pneumonia. A review of 29 infants younger than 2 weeks with enteroviral infections reported that 5 of 29 infants had severe multisystem disease, and all survived (Abzug et al, 1993). In another study in Salt Lake City, 1779 febrile infants younger than 90 days and undergoing evaluation for sepsis were enrolled; 1061 had enterovirus testing, and 214 (20%) were enterovirus positive (57% from blood, and 74% from CSF). The mean age of infants with enterovirus infection was 33 days; 91% were admitted, and 2% required intensive care (Rittichier et al, 2005). These observations underscore the generally benign and self-limited nature of neonatal enterovirus disease; however, morbidity can be substantial in severe disseminated disease. A viral sepsis syndrome—characterized by disseminated intravascular coagulation, refractory hypotension, and death—may occur in the setting of severe disease. Typically these severe infections are acquired in the immediate perinatal period. The mother is commonly symptomatic and may have been empirically treated with broad-spectrum antibiotics for possible chorioamnionitis. A maternal history of suspected chorioamnionitis in the absence of positive bacterial cultures should suggest this association, particularly during the typical “enterovirus season” observed in temperate climates. If myocarditis is present, congestive heart failure is often severe. Some infections, particularly those with echoviruses, are characterized by a rampant and overwhelming hepatitis (Modlin, 1980). Others exhibit primarily pulmonary disease or gastrointestinal involvement including diarrhea and necrotizing enterocolitis (Lake et al, 1976). Intracranial bleeding ranging from small to massive, severe hemorrhage has also been reported as a complication of neonatal enteroviral infection (Abzug, 2001; Abzug and Johnson, 2000; Swiatek, 1997). Other rarely associated findings include disseminated vesicular rash, dermal hematopoiesis, and hemophagocytic syndrome (Barre et al, 1998; Bowden et al, 1989; Sauerbrei et al, 2000). The severity of CNS infection is similarly variable. Enteroviruses, particularly enterovirus 71, can produce overwhelming meningoencephalitis, sometimes with cranial nerve signs. It is more common, however, to see moderate or mild meningitis characterized by temporary irritability, lethargy, fever, and feeding difficulty. Acquired postnatal enteroviral disease in infants is characterized primarily by high fever, irritability, lethargy, or poor feeding. One fourth of infected infants develop diarrhea or vomiting with or without an erythematous maculopapular rash. Conjunctivitis has also been observed. Respiratory tract symptoms are less common (Dagan, 1996). As noted, there is significant overlap between enteroviral and bacterial neonatal infections, and the two syndromes are difficult to distinguish; therefore many febrile infants will be admitted to the hospital and treated with broad-spectrum antibiotics and, if the CNS is involved, with acyclovir for possible bacterial sepsis and neonatal HSV infection. Such treatment seems unavoidable, except in circumstances in which enterovirus infection can be diagnosed rapidly and definitively. Duration of illness varies from less than 24 hours to longer than 7 days, but generally lasts 3 to 4 days. LABORATORY EVALUATION Viral culture from stool or rectal swab, nasopharyngeal swab, blood, buffy coat, urine, or CSF represents the gold standard diagnosis. Stool or rectal swab cultures can remain positive for several weeks following the initial infection, underscoring the importance of fecal-oral transmission in the epidemiology of these infections. Recently, CHAPTER 37 Viral Infections of the Fetus and Newborn and Human Immunodeficiency Virus Infection during Pregnancy reverse transcriptase PCR assays have become standardized and are available from many commercial and reference laboratories. Serologic tests for enteroviruses have been reported, but are less useful than culture and PCR (Swanink et al, 1993). Laboratory evaluation is predicated on the clinical syndrome and the end organs involved. Infants with aseptic meningitis typically have moderate CSF pleocytosis, which can be either lymphocytic or polymorphonuclear, but may lack pleocytosis even in the setting of documented CNS infection (Seiden et al, 2010). Accordingly, during periods of active enterovirus infection in the community, CSF should be sent for PCR even if a pleocytosis is absent. Thrombocytopenia, elevated transaminase levels, hyperbilirubinemia, hyperammonemia, hematologic abnormalities consistent with disseminated intravascular coagulation, anemia, peripheral leukocytosis, and abnormal chest radiographs are among other potential laboratory findings. When myocarditis is a diagnostic possibility, echocardiogram and electrocardiogram are indicated and may reveal diminished LV function, or dysrhythmias. Liver biopsy may be warranted in cases of fulminant hepatic failure (Abzug, 2001). TREATMENT The cornerstone of treatment of neonatal enteroviral disease is supportive care. Myocarditis and heart failure can be treated with inotropic support, diuretics, aggressive fluid management, and other supportive measures. Disseminated intravascular coagulation should be treated with blood products and other supportive measures as indicated. There is no evidence that steroids are of benefit. IVIG has been reported anecdotally to treat neonatal enteroviral infections with varying degrees of success (Kimura et al, 1999; Valduss et al, 1993). Only one randomized trial has systematically studied its use in 16 neonates with severe enteroviral infection; nine of these infants were randomized to receive IVIG, at a dose of 750 mg/kg. Decreased viremia and viruria along with faster resolution of irritability, jaundice, and diarrhea was demonstrated in patients administered IVIG with high titers of neutralizing enteroviral-specific antibodies. However, there were no significant differences in other major clinical outcomes, such as duration of hospitalization, fever, and symptoms of acute illness between treatment and control groups (Abzug et al, 1995). To attempt to augment the enterovirus type–specific antibody level in the setting of symptomatic neonatal disease, maternal plasma transfusion has been attempted (Jantausch et al, 1995; Rentz et al, 2006), based on the rationale that neonates typically acquire infection from their mothers in the peripartum period. Although reports of success using IVIG for treatment of neonatal enterovirus infections is largely based on anecdotal reports in limited numbers of patients, its use should be considered for severely symptomatic infections with life-threatening end-organ disease. The antiviral drug pleconaril has been developed specifically to treat picornavirus infections (enteroviruses and rhinoviruses). A small case series of infants with severe enteroviral hepatitis suggested a beneficial effect (Aradottir et al, 2001). Another small case series indicated 497 that five of six neonates with severe enteroviral infection who were treated with pleconaril survived, with minimal or no sequelae (Rotbart and Webster, 2001). However, a multicenter study of pleconaril treatment for enteroviral meningitis in children younger than 12 months conducted by the CASG and sponsored by the National Institutes of Health demonstrated no significant differences in duration of positivity by culture or PCR, hospitalization, or symptoms comparing the treatment and placebo groups (Abzug et al, 2003). Pleconaril did appear to have an effect on enteroviral meningitis in adults, promoting more rapid resolution of illness (Desmond et al, 2006), but the role of this drug in infants and children remains undefined. The drug continues to be investigational and is only available through the CASG, which is currently comparing pleconaril with placebo in severe symptomatic neonatal enteroviral infections (www.clini caltrials.gov). SHORT- AND LONG-TERM PROGNOSIS Prognostic factors for severe neonatal disease include peripartum maternal illness, earlier age of onset of neonatal disease, absence of serotype-specific antibody, and absence of fever and irritability (Abzug et al, 1993). All these risk factors are most consistent with vertical intrauterine enteroviral infection rather than postnatally acquired infection. The highest mortality rates are associated with the combination of severe hepatitis, coagulopathy, and myocarditis. Severe hepatitis caused by enteroviral infection is associated with mortality rates ranging from 30% to 80% (Abzug, 2001; Modlin, 1986). By the time disseminated intravascular coagulation has developed, the prognosis is grave. Prothrombin time longer than 30 seconds was a risk factor for death in one retrospective case review (Abzug, 2001). Few long-term follow-up studies have been published, but the available information suggests that infants who survive severe enteroviral neonatal disease have a complete recovery in most instances. Outcomes of 6 of 11 survivors with follow-up ranging from 9 to 48 months reported normal growth and no residual medical problems or liver dysfunction (Abzug, 2001). The long-term prognosis following CNS infection is unclear. A number of early studies of infants younger than 3 months with aseptic meningitis suggested that there may be some impairment of intellectual development in comparison with carefully selected control groups (Farmer et al, 1975; Sells et al, 1975). However, in a series of nine children with enteroviral meningitis and nine matched controls evaluated for sequelae at approximately 4 years of age, no differences in mean intelligence quotient, head circumference, detectable sensorineural hearing loss, or intellectual functioning were detected. Receptive language functioning of the meningitis group was significantly less than that in control subjects (Wilfert et al, 1981). A similar case-control follow-up study of 33 subjects, who were compared with siblings used as controls, reported no neurodevelopmental sequelae (Bergman et al, 1987). Older children with enterovirus 71 infection involving the CNS are at risk for significant neurodevelopmental sequelae (Chang et al, 2007). 498 PART IX Immunology and Infections PREVENTION Anecdotal reports of the use of IVIG in nursery outbreaks to prevent further horizontal transmission of enteroviral infection produce conflicting results (Carolane et al, 1985; Kinney et al, 1986; Nagington et al, 1983). Because there are multiple non–polio enteroviral serotypes that cause clinical disease, development of anti-enteroviral immunization is conceptually difficult, although vaccines are in development for enterovirus 71 (Zhang et al, 2010). Standard contact precautions should be used for the treatment of hospitalized infants with known or suspected enteroviral infections. HUMAN PARECHOVIRUS Recently two viruses formerly classified with the echoviruses, echovirus 22 and 23, were shown to comprise a separate genus within the Picornaviridae family, the genus Parechovirus; these viruses are now referred to as human parechoviruses types 1 and 2. A total of 10 distinct parechoviruses have now been recognized (Drexler et al, 2009; Harvala et al, 2010; Pajkrt et al, 2009). These viruses have been recognized recently as significant neonatal pathogens (Harvala et al, 2010). The primary site of parechovirus replication is believed to be the respiratory and gastrointestinal tract. Replication in the gastrointestinal tract is associated with prolonged shedding of infectious virus in the feces. As a result, fecal-oral transmission, as with enteroviruses, appears to be the predominant route of infection. Parechovirus infections may also be acquired via a respiratory route, with subsequent virus shedding detectable in respiratory secretions. Viremia leads to secondary seeding of other organs and systemic symptoms. Neonates with parechovirus infection, particularly parechovirus 3, have a clinical presentation similar to infants with severe enteroviral disease, with a sepsislike illness in severe cases (VerboonMaciolek et al, 2008a; Wolthers et al, 2008). Hepatitis is also a prominent feature of neonatal infection. The most frequent signs are fever, seizures, irritability, rash, and feeding problems. It has been postulated that parechovirus 3 is a newly emerging infection, and the relative lack of maternal antibody may predispose the neonate to more severe disease (Harvala et al, 2010). Parechovirus can also cause aseptic meningitis and encephalitis in the neonate, with the predominate site of injury being the periventricular white matter (Gupta et al, 2010; Levorson et al, 2009; Verboon-Maciolek et al, 2008b). Parechovirus RNA has also been found at autopsy in children younger than 2 years (Sedmak et al, 2010), including from infants with otherwise unexplained deaths. Management is similar to that for neonatal enterovirus disease—namely, supportive care. The other virus classified in the Parechovirus genus is Ljungan virus (Tolf et al, 2009). This virus was first described in 1998 and was found to cause type 1 diabetes–like symptoms and myocarditis in bank voles; it also appears to be endemic in other wild rodent populations. The virus does not appear to have a role in the development of diabetes in humans (Tapia et al, 2010). This zoonotic virus has recently emerged as a recognized cause of fetal infection (Niklasson et al, 2007, 2009b). Ljungan virus infection appears in particular to be associated with severe CNS abnormalities, including hydrancephaly and hydrocephalus (Niklasson et al, 2009a). A study from Sweden identified a strong epidemiologic association between small rodent abundance and the incidence of intrauterine fetal death in humans. Ljungan virus antigen was detected in this study in half of the intrauterine fetal death cases tested (Niklasson et al, 2009). HEPATITIS VIRUSES There are six distinct viruses known to cause viral hepatitis—hepatitis A, B, C, D (delta agent), E, and G. Hepatitis G is of importance for the pregnant patient, although only B and C are of major importance in the newborn. The main features of each virus type are listed in Table 37-3 (Jonas, 2000; Koff, 2007; Krugman, 1992). Hepatitis A virus (HAV) is passed by fecal-oral transmission and is a rare cause of neonatal disease, although it has been nosocomially transmitted in the setting of a neonatal intensive care unit. Hepatitis E virus is similar to hepatitis A in its mode of transmission and clinical manifestations, except for an increased mortality in pregnant women infected with hepatitis E (Aggarwal et al, 2009; Teshale et al, 2010). There are no data regarding perinatal transmission of hepatitis E. Hepatitis D virus may cause only coinfection or super-infection with hepatitis B virus. Its only clinical significance is that hepatitis B infection may become more severe when hepatitis D virus is present. Perinatal transmission has been described (Ramia and Bahakim, 1998), but is uncommon. Hepatitis G virus (HGV), also known as GB virus type C, has been associated with acute and chronic hepatitis and usually is noted as a coinfection with hepatitis B or C. This virus has a tropism for lymphocytes and may influence the course and prognosis in HIV-seropositive patients (Reshetnyak et al, 2008). Perinatal transmission can occur in 60% to 80% of infants born to HGV-viremic mothers, but there has not been any report of clinical hepatitis attributed to the HGV infection in this setting (Ohto et al, 2000; Wejstal et al, 1999; Zanetti et al, 1998; Zuin et al, 1999). The clinical significance and effects of HGV infection are still poorly understood. Hepatitis B and C viruses are both transmitted vertically and among the viral hepatitis are of the greatest important to the care of newborns. HEPATITIS B The issues that are most important to neonatologists are the frequency with which hepatitis B virus (HBV) is transmitted to infants at the time of birth, the short-term and long-term consequences of these infections, the importance of greater surveillance for maternal carriage of HBV, the role of antiviral therapy to prevent transmission and disease progression, and the availability of effective HBV immunoprophylaxis (Arfaoui et al, 2010; Krugman, 1988). Incidence In certain parts of the world and among certain ethnic groups, as many as 7% to 10% of all infants acquire hepatitis B infections at the time of birth, and almost all of these infections become chronic. In the United States, it has been estimated that approximately 20,000 infants are CHAPTER 37 Viral Infections of the Fetus and Newborn and Human Immunodeficiency Virus Infection during Pregnancy 499 TABLE 37-3 Viral Hepatitis Types A, B, C, D, E, and G: Comparison of Clinical, Epidemiologic, Immunologic, and Therapeutic Features Feature Hepatitis A Hepatitis B Hepatitis C Hepatitis D (Delta) Hepatitis E Virus Hepatitis A virus (HAV) Hepatitis B virus (HBV) Hepatitis C virus (HCV) Hepatitis D virus (HDV) Hepatitis E virus (HEV) Hepatitis G virus (HGC), (GB virus type C [GBV-C]) Family Picornavirus Hepadnavirus Flavivirus Satellite Calicivirus Flavivirus Genome RNA DNA RNA RNA RNA RNA Incubation period 15-40 days 50-180 days 1-5 months 2-8 weeks 2-9 weeks Unknown Usual Rare Rare Food- or ­water-borne No Usual Yes Sexual contact No Usual Yes Sexual contact less common No Usual Only with HBV Sexual contact less common Usual No Unknown Water-borne ­transmission in developing ­countries No Usual Yes Sexual contact; probably less common No No cases reported Yes Yes Yes Yes Yes Yes No No cases reported Yes Yes; controversial Yes Yes No Yes Yes Yes No No Yes No No No No No No No No No Mode of transmission Oral (fecal) Parenteral Perinatal Other Sequelae Carrier state Chronic disease Interventions Immunoglobulin Vaccine Antiviral therapy Hepatitis G/GBVC Modified from Krugman S: Viral hepatitis: A, B, C, D, and E: prevention, Pediatr Rev 13:245-247, 1992. born annually to mothers who are chronic HBV carriers (Mast et al, 1998). Since 1990 the incidence of acute hepatitis B in the United States has declined dramatically, with the largest declines occurring in children younger than 15 years (98%) and in young adults 15 through 24 years of age (90%). The frequency of transmission depends primarily on the prevalence of the hepatitis B carrier state among women of childbearing age. Hepatitis B is a carcinogenic virus; perinatal acquired infection can lead to chronic liver failure and hepatic carcinoma in adult life (Balistreri, 1988; Beasley and Hwang, 1984; Chan and Sung, 2006). The incidence of neonatal hepatitis B infection depends on the timing of infection and the overall prevalence of the disease in the population under study. Women with acute hepatitis B infection during the first or second trimester rarely transmit the virus to their infants (Krugman, 1988; Stevens, 1994). Transplacental transfer has been described, but appears to be rare. The carriage rate for hepatitis B surface antigen (HBsAg) varies from 0.1% in the United States and Europe to 15% in Taiwan and parts of Africa, with intermediate rates in Japan, South America, and Southeast Asia. Transmission rates among immigrant women in Western countries appear to parallel the rates in their countries of origin (Krugman, 1988). The most important route of transmission is transmission that occurs during labor and delivery. The likelihood of transmission is great if symptomatic acute disease is present (60% to 70% transmission; Gerety and Schweitzer, 1977). Infants of hepatitis B e antigen (HBeAg)–positive mothers have an 80% to 90% chance of becoming HBsAg carriers (Lee et al, 1978; Okada et al, 1976). The risk of an infant acquiring hepatitis B if born to an HBsAg-positive but HBeAg-negative mother is 5% to 20%. Chronic neonatal infection occurs in less than 10% of infants of HBeAg antigen–negative mothers (Krugman, 1988). Although HBsAg has been found in breast milk, breastfeeding does not appear to have any influence on the rate of transmission (Beasley et al, 1975). The World Health Organization currently recommends that all mothers who are hepatitis B positive breastfeed their infants, and that their infants be immunized at birth. Etiology and Pathogenesis HBV is a DNA virus that localizes primarily in hepatic parenchymal cells but circulates in the bloodstream, along with several subviral antigens, for periods ranging from a few days to many years. Several distinct genotypes have been identified, and these subtypes show biologic variability in transmission and disease progression (Magnius and Norder, 1995; Schaefer et al, 2009). Transplacental leakage of HBeAgpositive maternal blood is a potential source of intrauterine infection (Lin et al, 1987). During either acute or persistent viremia in the mother, the virus itself or viral antigens may rarely cross the placenta and cause intrauterine infection, but more commonly infection occurs perinatally during labor or delivery (Chisari and Ferrari, 1995; Xu et al, 2001). The finding of hepatitis antigens in the newborn might not indicate the presence of infection, but rather the passive transfer of antigen only; therefore antigen tests should be interpreted cautiously in this setting. Hepatitis B infection of placental trophoblast has been documented and is compatible with a transplacental route of infection in some circumstances (Bai et al, 2007). Most infants born to mothers infected with HBV have a negative test result for HBsAg at birth, but in the absence of prophylaxis are at risk for becoming 500 PART IX Immunology and Infections TABLE 37-4 Licensed Monovalent and Combination Hepatitis B Vaccines Monovalent Vaccines Clinical Scenario Recombivax Hepatitis B Dose, μg (mL) Enginerix-B Dose, μg (mL) Combination Vaccines Twinrix* Comvax† Pediatrix‡ Newborns born to HBsAg-negative mothers 5 (0.5) 10 (0.5) NA NA NA Newborns born to HBsAg-positive mothers§ 5 (0.5) 10 (0.5) NA NA NA Infants, children, and adolescents <20 yr old 5 (0.5) 10 (0.5) NA 5 (0.5) 10 (0.5) Adults >20 yr old 10 (1.0) 20 (1.0) 20 (1.0) NA NA *Twinrix is a combination of Engerix-B (20 mg) and hepatitis A vaccine, licensed for use in people 18 years of age and older in a three-dose series at a 0-, 1-, and 6-month schedule. †Comvax is a combination of Recombivax HB (5 μg) and Haemophilus influenzae type b (PRP-OMP) recommended for use at 2, 4, and 12 to 15 months of age. This vaccine should not be administered at birth. ‡Pediatrix is a combination of diptheria and tetanus toxoids and acellular pertussis (i.e., DTaP), inactivated poliovirus, and hepatitis B (Engerix B, 10 mg). It is recommended for use at 2, 4, and 6 months of age, but should not be administered at birth, before 6 weeks of age, or after 7 years of age. §Hepatitis B immunoglobulin (0.5 mL) should be adminstered simultaneously with vaccination. HBsAg-positive during the first 3 months of life, suggesting that transmission is primarily peripartum (Krugman, 1988, 1992; Mulligan and Stiehm, 1994; Shapiro, 1993). Infants with hepatitis B infection do not show clinical or chemical signs of disease at birth. Without immunoprophylaxis, the usual pattern is the development of chronic antigenemia with mild and often persistent enzyme elevations, beginning at 2 to 6 months of age (Mulligan and Stiehm, 1994). Less commonly the infection becomes clinically manifest, with jaundice, fever, hepatomegaly, and anorexia, followed by either recovery or chronic active hepatitis. Rarely, fulminant hepatitis is seen and can be fatal (Delaplane et al, 1983). Laboratory tests are essential in the diagnosis of hepatitis B infection. Evaluations of serum enzymes and of bilirubin reflect the extent of liver damage. Several helpful serologic tests identify the virus involved (Krugman, 1988). HBsAg appears early, usually before liver disease is found; it persists in those who become chronic carriers or disappears in the 5% of infants who resolve the infection. HBeAg and anti-HBe testing can be used to assess infectivity; although they are not as frequently used for serodiagnosis, they are useful markers of the likelihood of perinatal transmission. PCR and other nucleic acid detection techniques are important in confirming diagnosis, assessing viral load, and monitoring the response to therapy. 3. Routine immunization of children and adolescents who have not been immunized previously 4. Immunization of nonimmunized adults at increased risk of infection For infants, the three-dose vaccination schedule should be initiated in the neonatal period or by 2 months of age (Table 37-4). Four doses can be given if a birth dose is given and a combination vaccine is used to complete the series. Vaccination can be delayed until just before hospital discharge in preterm (birthweight less than 2000 g) infants born to HBsAg-negative mothers. All infants born to HBsAg-positive women should receive both active and passive immunization within 12 hours of birth. The doses and recommended options for administration of the hepatitis B vaccines that are currently licensed in the United States are provided in Table 37-4. If maternal HBsAg status is unknown, it should be checked immediately, and the infant should receive the first dose of HBV vaccine immediately. Infants born to women who test positive, or to women in whom the HBsAg is unknown, should be given hepatitis B immunoglobulin as soon as possible within 1 week after birth. The highest immunization failure rates have been observed in infants of HBeAgpositive women and those infected in utero (Farmer et al, 1987; Tang et al, 1998). It is recommended that infants born to HbsAg-positive mothers undergo postimmunization testing for anti-HBS and HbsAg 3 to 9 months after completion of the series, for identification of breakthrough carriers and of infants who may benefit from revaccination (AAP Committee on Infectious Diseases, 2010a). Prevention Treatment The primary goal of prevention strategies for hepatitis B is to prevent chronic infection and chronic liver disease. In the United States over the past two decades, a comprehensive immunization strategy has been implemented consisting of four components: 1. Universal immunization of all infants beginning at birth 2. Prevention of perinatal infection through routine screening of all pregnant women and appropriate immunoprophylaxis of infants born to HBsAg-positive women (or women whose HBsAg status is unknown) There is no therapy for acute hepatitis B infection. Interferon (IFN) α and antiretroviral drugs such as lamivudine, an antiretroviral drug that blocks HBV polymerase, and adefovir, a reverse transcriptase inhibitor, have been approved for treatment of chronic hepatitis B in children, but success is widely variable (Giacchino and Cappelli, 2010). Three therapeutic agents for chronic HBV infection in children have been approved in the United States, including standard IFN-α, lamivudine, and adefovir (Kurbegov and Sokol, 2009). These antiviral therapies are summarized in Table 37-2. IFN-α appears to be the Clinical Spectrum and Laboratory Evaluation CHAPTER 37 Viral Infections of the Fetus and Newborn and Human Immunodeficiency Virus Infection during Pregnancy most effective (approximately 30% HBeAg seroconversion; 10% HBsAg seroconversion), although benefits are primarily observed in children with alanine aminotransferase levels more than twofold the upper limit of normal. The virologic response rates for lamivudine mirror those of IFN-α (23% to 31% HBeAg seroconversion) with easier administration and a more favorable safety profile, but lower HBsAg seroconversion (2% to 3%) and high rates of drug resistance. Adefovir demonstrates a favorable safety profile and is less likely to select for resistance than lamivudine, but virologic response was limited to adolescent patients and was lower than that of lamivudine (16% HBeAg seroconversion; <1% HBsAg seroconversion). Entecavir and tenofovir, both approved therapies for adults with chronic HBV infection, are in trials for use in children. Some experts recommend “watchful waiting” of children, because current therapies are only 30% effective at best. Young children are often believed to be immune tolerant of hepatitis B infection (i.e., they have viral DNA present in serum, but normal transaminase levels and no evidence of active hepatitis). These children should have transaminases and viral load monitored, but are not typically considered to be candidates for antiviral therapy. Prognosis Most long-term follow-up studies have shown that children vaccinated at birth have high levels of protection until at least 5 years of age. Approximately 5% to 10% of infants born to HbeAg-positive mothers become chronic HBV carriers despite combined active and passive immunoprophylaxis with hepatitis B immunoglobulin and HBV vaccine (Kato et al, 1999). Failure of immunoprophylaxis may be associated with the level of maternal viremia and specific HBV genetic variants (Ngui et al, 1998). Infants who become infected with HBV perinatally have a 90% risk of chronic infection, and 15% to 25% of those with chronic infection die of HBV-related liver disease (primarily hepatocellular carcinoma) as adults. There is some evidence that the risk of carcinoma correlates with specific hepatitis B genotypes (Sherman, 2010). HEPATITIS C In 1989, hepatitis C virus (HCV) was found to be the main cause of non-A, non-B, parenterally transmitted hepatitis; subsequently, HCV has been found to account for a significant portion of the cases of sporadic acute and chronic hepatitis (Choo et al, 1989; Reyes et al, 1990; Weiss and Persing, 1995). Hepatitis C virus is a small, single-stranded RNA virus that is a member of the family Flavivirus; this family consists primarily of vector-borne infections, such as St. Louis encephalitis virus and West Nile virus. Hepatitis C virus is an exception in this family because it is not transmitted by insect vectors. Seven genotypes are described, with significant biologic differences in regard to disease progression and responsiveness to therapy (Klenerman et al, 2009). Vertically transmitted hepatitis C infection in infants is associated with a higher rate of chronic hepatitis, but less liver injury compared with adult HCV infections (Tovo et al, 2000). 501 Incidence The seroprevalence for anti-HCV antibody in pregnant women ranges from 0.7% to 4.4% worldwide (Conte et al, 2000). In the United States, seroprevalence of hepatitis C decreased during the 1990s and has remained low and stable since then. The overall seroprevalence is estimated to be 1.3%, with a seroprevalence in pregnant women ranging from 1% to 2% (AAP Committee on Infectious Diseases, 2010a). Forty percent to 50% of women with HCV have no identified known risk factors for infection (Bortolotti et al, 1998). Estimates from the 1990s suggested that the vertical transmission rate of HCV is approximately 5% to 11% in HIV-negative mothers, and ranges from 10% to 20% from mothers coinfected with HIV (Hillemanns et al, 2000; Palomba et al, 1996; Polywka et al, 1997; Tajiri et al, 2001). More recent studies, however, suggest a lower rate of vertical transmission, ranging from 2.7% to 3.6% (Ferrero et al, 2003; Syriopoulou et al, 2005). One study has shown, in a multivariate analysis, that higher risk of vertical transmission is related more to maternal use of injection drugs than to HIV infection itself, although the mechanism for this finding is still unclear (Resti et al, 2002). The risk of transmission correlates with maternal viremia, and transmission appears to be rare in the absence of viremia. The risk of HCV infection from transfused blood after the advent of HCV screening is estimated to be less than 1 in 1 million units transfused. Etiology and Pathogenesis HCV is transmitted less efficiently than HBV by sexual contact. Risk factors for HCV infection include transfusion, intravenous drug use, frequent occupational exposure to blood products, and household or sexual contact with an infected person (Weiss and Persing, 1995). In children, perinatal transmission is the most common route of infection (Mohan et al, 2010). Preparations of IVIG contaminated with HCV were reported between April 1993 and February 1994, but since that time, routine screening for HCV with PCR and application of a viral inactivation process during manufacturing have been implemented to reduce the risk of transmission (Schiff, 1994). Perinatal transmission is the leading cause of childhood HCV infection (Mohan et al, 2010). Transmission of HCV from mother to child is thought to occur either in utero or at the time of delivery. Viral genotypes and infection and replication in maternal peripheral blood monocytes can also affect the ability of the virus to infect the fetus or newborn (Azzari et al, 2000; Zuccotti et al, 1995). Perinatal transmission is confined almost always to women with detectable HCV ribonucleic acid (RNA) in the peripheral blood by the PCR, but all children born to women with anti-HCV antibodies should be tested for HCV. Data regarding the correlation of HCV RNA titer in the mother with the risk of vertical transmission are conflicting, although most studies have reported an association (Conte et al, 2000; Lynch-Salamon and Combs, 1992; Ohto et al, 1994; Resti et al, 1998; Tajiri et al, 2001). Maternal peripheral blood mononuclear cell infection by HCV, membrane rupture of longer than 6 hours before delivery, and procedures exposing the infant to maternal 502 PART IX Immunology and Infections blood infected with HCV during vaginal delivery are associated with an increased risk of transmission. Internal fetal monitoring is also a risk factor for transmission (Mast et al, 2005). Maternal coinfection with HCV and human immunodeficiency virus, maternal history of intravenous drug use and of HCV infection of the sexual partner of the mother are also risk factors for transmission (Fiore and Savasi, 2009; Indolfi and Resti, 2009). Transmission rates as high as 25% have been reported in HCV-infected, HIV-positive women. The effect of vaginal delivery versus cesarean section on transmission rates is unclear (Gibb et al, 2000; Paccagnini et al, 1995); current recommendations do not support the practice of cesarean delivery in women with HCV infection. Transmission of HCV via breast milk has not been demonstrated conclusively, although HCV can be found in breast milk and colostrum (Gurakan et al, 1994; Zimmermann et al, 1995). Studies comparing breastfed and bottlefed infants born to HCV-infected mothers have not shown a statistically significant difference in vertical transmission (Lin et al, 1995; Resti et al, 1998). Accordingly, maternal HCV infection is not a contraindication to breastfeeding (Mast, 2004); however, it may be prudent for mothers who are HCV-infected and who choose to breastfeed to consider abstaining from breastfeeding if their nipples are cracked and bleeding. Clinical Spectrum and Laboratory Evaluation Acquired HCV infection typically causes jaundice in one third of cases and significant increases in alanine aminotransferase (ALT) in almost all persons infected. Most neonates perinatally infected with HCV demonstrate little in the way of clinical symptomatology; they may have elevated liver ALT either transiently or intermittently. The increases in ALT values, when tested, are most commonly noted between 3 and 6 months of age. HCV-RNA PCR results may be negative initially at birth or within the first few days of life, but typically become positive by 1 to 2 weeks of age and remain so until at least 5 years of age. The highest sensitivity of PCR is reported after 1 month of age (Polywka et al, 2006; Thomas et al, 1997). Confirmatory anti-HCV IgG serology should be delayed until exposed infants are at least 15 to 18 months old, when at least 99% will have cleared maternal antibody (Dunn et al, 2001). Before 18 months of age, only a positive HCV PCR result can confirm diagnosis of neonatal infection; positive anti-HCV IgG results simply reflect passively acquired maternal antibody. IgM assays are not available or reliable for perinatal diagnosis of HCV. Liver ultrasonographic findings are usually normal or may consist of a mild diffuse increase in echogenicity. Liver biopsies, when performed, typically demonstrate mild to moderate chronic persistent hepatitis (Palomba et al, 1996; Tovo et al, 2000). Treatment Interferon and ribavirin are both approved by the FDA to treat adults and children with chronic hepatitis C. Antiviral agents are summarized in Table 37-2. Only 10% to 25% of adults treated with interferon have a sustained remission of disease (Bonkovsky and Woolley, 1999). Treatment with a combination of interferon and ribavirin achieves remission in closer to half of treated adults (Cornberg et al, 2002). Randomized controlled trials indicate that patients treated with pegylated interferons (so called because they are formulated and stabilized with polyethylene glycol), both as dual therapy with ribavirin and as monotherapy, experience higher sustained viral response rates than do those treated with nonpegylated interferon (Shepherd et al, 2004). Data on the use of these agents in infants and children are limited, although the use of IFN-α2b in combination with ribavirin was recently approved by the FDA (Palumbo, 2009). In a case series of four pediatric patients treated with interferon for 1 year, viremia decreased to undetectable levels during treatment in all four patients, but two patients became viremic again once the treatment was stopped (Tovo et al, 2000). There are significant genotype-dependent differences in responsiveness to antiviral therapy; patients with genotype 1 had the lowest levels of sustained virologic response, and patients with genotype 2 or 3 had the highest (Palumbo, 2009; Shepherd et al, 2004). The use of IFN-α2b in combination with ribavirin provides a much more favorable sustained virologic response in children with HCV genotype 2/3 (84%) than in those with HCV genotype 1 (36%; González-Peralta et al, 2005). For genotype 1 hepatitis C treated with pegylated interferons combined with ribavirin, it has been shown that genetic polymorphisms near the human IL28B gene, encoding interferon lambda 3, are associated with significant differences in response to the treatment (Ge et al, 2009; Thomas et al, 2009). Interferon therapies are available only in parenteral formulations and are associated with significant side effects; therefore further research into its utility in pediatric HCV infection is needed. Infants and children with persistent elevations in liver transaminases should be referred to a pediatric gastroenterologist for evaluation and management. The necessity for, as well as the frequency of, screening tests of liver function has not been established. Prognosis In 60% to 80% of HCV-infected adults, chronic hepatitis occurs, and in one third of HCV-infected adults, chronic HCV infection leads to cirrhosis or liver failure within 20 to 30 years after infection (Iwarson et al, 1995; Leone and Rizzetto, 2010). In patients with cirrhosis, the incidence of hepatocellular carcinoma is 2% to 5% per year. There are limited data regarding long-term follow-up in HCVinfected infants, but existing knowledge indicates that most children remain viremic until at least 5 to 6 years of age, and most develop chronic, persistent hepatitis. Studies monitoring infected children beyond 6 years of age have not been completed, but they are at risk for cirrhosis and hepatocellular carcinoma. Transient hepatitis C viremia with subsequent resolution has been reported (Padula et al, 1999; Ruiz-Extremera et al, 2000; Zanetti et al, 1995). All the infants who have been followed for an average of 3 to 4 years have been reported to have normal growth and development (Tovo et al, 2000). CHAPTER 37 Viral Infections of the Fetus and Newborn and Human Immunodeficiency Virus Infection during Pregnancy 503 Prevention Although development of an HCV vaccine is a major public health priority (Houghton and Abrignani, 2005), there is currently no vaccine available. Like all other infants, infants with HCV should receive routine hepatitis B immunization. In addition, they should receive hepatitis A vaccination at 2 years of age. Parents should be advised to avoid unnecessary administration of medicines known to be hepatotoxic. Standard precautions are recommended for the hospitalized infant. ADENOVIRUS Adenoviruses are so named because they were originally recovered from human adenoidal tissue; they are mediumsized (90 to 100 nm), nonenveloped icosahedral viruses composed of a nucleocapsid and a double-stranded linear DNA genome. They are the largest of the nonenveloped viruses. There are 53 described adenovirus serotypes in humans. EPIDEMIOLOGY AND CLINICAL MANIFESTATIONS Adenoviruses are responsible for a wide variety of clinical syndromes, including conjunctivitis, respiratory track disease, and gastroenteritis. Adenovirus causes 5% to 10% of upper respiratory infections in children, and many infections in adults as well. No clear seasonality has been described for adenovirus infections. ADENOVIRUS INFECTIONS IN NEWBORNS There are a limited number of reports of adenovirus infection in young infants, but published case series indicate that adenovirus can cause serious, life-threatening disease in the neonate. A review of neonatal adenovirus infection (Abzug and Levin, 1991) identified several characteristic historical features, including prolonged rupture of membranes, history of maternal illness, vaginal mode of delivery, and onset of illness within the first 10 days of life. Serotypes 2, 3, 7, 11, 13, 19, 21, 30, and 35 have been implicated (Abzug and Levin, 1991; Andiman et al, 1977; Matsuoka et al, 1990; Osamura et al, 1993; Pinto et al, 1992; Sun and Duara, 1985). Clinical findings in various case reports and case series have included lethargy, fever or hypothermia, anorexia, apnea, hepatomegaly, bleeding, and progressive pneumonia (Figure 37-5). Laboratory abnormalities include thrombocytopenia, coagulopathy, and hepatitis. Acquisition of infection from the mother via vaginal delivery is the presumed mode of transmission in most cases, although transplacental spread has also been implicated. Although most neonatal adenovirus infection is believe to be acquired in the birth canal, congenital infections have been described, presumably because of transplacental transmission. A wide range of fetal pathologies including pleural effusion, hepatitis, myocarditis, meningitis, and CNS abnormalities have been described (Baschat et al, 2003; Meyer et al, 1985; Rieger-Fackeldey et al, 2000). Adenovirus has been implicated as a potential cause FIGURE 37-5 (See also Color Plate 19.) Postmortem histologic analysis from an infant who died from disseminated adenovirus infection at 2 weeks of age. Hematoxylin and eosin stain of lung demonstrating inflammatory infiltrates (arrow) and intranuclear inclusions (arrowhead). This infant had a viral sepsis syndrome characterized by hepatic failure, disseminated intravascular coagulation, and pneumonitis from adenovirus infection presumed to have been acquired intrapartum. of chorioamnionitis and premature birth (Tsekoura et al, 2010; Van den Veyver et al, 1998). PREVENTION AND INTERVENTION Intravenous ribavirin has been administered to a neonate with disseminated adenovirus infection undergoing extracorporeal membrane oxygenation, with evidence of viral clearance within 48 hours of initiating therapy (Aebi et al, 1997). Adenovirus vaccines were at one time available for use in military personnel (Gaydos and Gaydos, 1995; Tucker et al, 2008), but are not currently in production. Adenoviruses are hardy and resistant to inactivation by physical and chemical methods that kill most viruses, adding to the challenge of hospital infection control during outbreaks of infection. RESPIRATORY VIRUSES Although relatively uncommon, any one of the respiratory viruses can cause symptomatic respiratory disease in newborn infants. The association has been described for rhinoviruses, adenoviruses, parainfluenza viruses, influenza virus, and respiratory syncytial virus. Adenovirus, rhinovirus, and parainfluenza virus infections are generally characterized by mild rhinorrhea in neonates. All of these viruses, however, can cause clinical symptoms indistinguishable from those of bacterial infection, leading to increased diagnostic testing and empiric antibiotic treatment. Influenza virus infections are usually mild, but in the absence of maternally transmitted antibody, they can be life threatening, with severe pneumonia, hypoxia, and a 504 PART IX Immunology and Infections prolonged course. During the H1N1 influenza pandemic of 2009 to 2010, pregnant women were at uniquely high risk for severe influenza (Jamieson et al, 2009), and infections were described in neonates (Sert et al, 2010). The most extensive nursery outbreaks, however, have been caused by respiratory syncytial virus (RSV; Hall et al, 1979; Wilson et al, 1989). Because of the importance that RSV plays in the newborn nursery, this virus is considered in greater detail in the following section. RESPIRATORY SYNCYTIAL VIRUS RSV is the major cause of viral pneumonia and bronchiolitis in infants and children. In temperate climates, it causes large annual epidemics during the winter and early spring months (typically ranging from November through April). During these months, RSV appears to be responsible for up to 20% of all pediatric hospital admissions (Hall et al, 2009). Nosocomial infections are frequent during these times, and illness among hospital staff members is a major factor in its spread from infant to infant. Several nursery outbreaks have been described. In one of these outbreaks, cultures were obtained prospectively so that a full picture of the virus’s pathogenicity and epidemiology could be drawn (Hall et al, 1979). Twenty-three of 66 infants hospitalized for 6 days or more were infected. Virtually all infants were symptomatic. Clinical manifestations included pneumonia, upper respiratory infection, apneic spells, and nonspecific signs. Pneumonia and apnea were seen almost exclusively in infants older than 3 weeks, and nonspecific signs were most commonly observed in younger infants. Four (17%) infants died, two unexpectedly, during the course of infection. The spread of infection in the unit was difficult to interrupt; infants in isolettes did not seem to be protected against acquisition of the infection. Eighteen of the 53 nursery personnel were infected during the outbreak; 83% of the infected nursery providers were symptomatic. Of particular importance is the observation that RSV infection, both in term and in preterm infants, is commonly associated with a new onset of apnea (Bruhn et al, 1977). Boys are at greater risk than girls for serious RSV disease. RSV-associated apnea has been the probable explanation for some case reports of deaths attributed to SIDS (Eisenhut, 2006). RSV lower respiratory track disease can be slow to resolve. After discharge from initial hospitalization, risk factors for infant rehospitalization secondary to RSV infection include premature gestational age, chronic lung disease, siblings in daycare or school, chronologic age of less than 3 months, and exposure to tobacco smoke (Carbonell-Estrany and Quero, 2001). Diagnosis can be confirmed by DFA staining of nasopharyngeal or tracheal aspirate, nasopharyngeal swab, or other respiratory secretions. Culture of RSV can require 3 to 5 days and can be used if DFA staining is not available. PCR is also available for the rapid diagnosis of infection. Treatment and prevention of RSV infection in infants attracted considerable attention during the 1990s because of the clinical and economic impacts of these infections (Groothuis, 1994; Kinney et al, 1995; Levin, 1994; Meissner, 1994). Considerable debate has ensued concerning the efficacy, safety, and potential effect on health care workers of ribavirin therapy, so its use remains controversial BOX 37-4 G  uidelines for Prophylaxis in ­Preterm Infants at Start of RSV Season PROPHYLAXIS RECOMMENDED Infants born at ≤28 weeks, 6 days of gestation during first RSV season* ll Infants born at 29 weeks, 0 days through 31 weeks, 6 days of gestation if less than 6 months of age at beginning of RSV season ll Infants and children less that 24 months of age with chronic lung disease requiring medical therapy (e.g., supplmental oxygen, bronchodilators, diuretics, corticosteroids) within the 6 months before RSV season ll Infants born at 32 weeks, 0 days through 34 weeks, 6 days of gestation if one of the following two risk factors are present: ll Attending child care ll Sibling younger than 5 years of age ll PROPHYLAXIS CONSIDERED Infants with congenital anomalies of the airways ll Infants younger than 24 months of age with hemodynamically significant congenital heart disease, as manifest by congestive heart failure, pulmonary hypertension, and cyanosis ll Infants with severe neuromuscular disease ll Infants with immune deficiencies ll Infants with cystic fibrosis ll RSV, Respiratory syncytial virus. *Earliest date to consider initiation of prophylaxis depends upon geographic location: July 1 for southeast Florida, September 15 for north central and southwest Florida, November 1 for most other areas of the United States. (Meissner, 2001; Ventre and Randolph, 2004; Wald and Dashefsky, 1994). Treatment consists of nebulization of ribavirin by a small-particle aerosol generator supplied by the manufacturer into an oxygen hood, tent, or mask from a solution containing 20 mg of ribavirin per milliliter of water (see Table 37-2). The aerosol has been administered on various schedules for 3 to 5 days (e.g., 12 to 20 h/day). The efficacy of ribavirin in this setting remains unclear. Some long-term benefit on recurrent wheezing may be realized (Ventre and Randolph, 2004). Corticosteroids are not effective in the treatment of RSV. Prevention efforts have focused on passive and active immunization. Standard immunoglobulin has not been shown to be efficacious for prevention of RSV infection in high-risk infants (Meissner et al, 1993). The efficacy of monthly prophylactic administration of RSV-specific immunoglobulin (750 mg/kg or 150 mg/kg) in 249 infants with cardiac disease or bronchopulmonary dysplasia was examined in a multicenter trial (Groothuis, 1994). In the high-dose (750 mg/kg) group, there were fewer lower respiratory tract infections, hospitalizations, days in hospital, and days in the intensive care unit as well as less use of ribavirin. Subsequently a mouse monoclonal antibody was developed, which provides similar protection against RSV (i.e., palivizumab). Unlike anti-RSV immunoglobulin, which is a pooled human blood product, the monoclonal antibody does not confer the theoretical risk of acquiring bloodborne pathogens. Palivizumab is also substantially easier to administer because it is an intramuscular injection. Recommendations by the AAP for the use of palivizumab are summarized in Box 37-4. Additional preventive measures for high-risk infants include eliminating or minimizing exposure to tobacco smoke, avoiding crowds and situations CHAPTER 37 Viral Infections of the Fetus and Newborn and Human Immunodeficiency Virus Infection during Pregnancy GASTROINTESTINAL VIRUSES The most important of the viruses that cause diarrhea from the perspective of the neonatal nursery are the rotaviruses. This important group of viruses, with at least four serotypes, is responsible for a large proportion of significant and sometimes severe diarrhea in infants 6 to 24 months of age (Cohen, 1991; Greenberg et al, 1994; Haffejee, 1991; Taylor and Echeverria, 1993). Recent evidence suggests that rotavirus extracts an unexpectedly significant burden of disease on infants as young as 2 to 3 months of age (Clark et al, 2010). Nursery-acquired infections are common; surprisingly, such infections appear to be benign in most infants. Two studies performed in nurseries in Sydney, Australia, and in London found that 30% to 50% of 5-day-old babies shed the virus (Chrystie et al, 1978; Murphy et al, 1977). However, more than 90% of the infected infants were asymptomatic. The remaining symptomatic infants had loose stools and vomiting, but this figure was only slightly higher than the proportion of uninfected infants with the same symptoms. The full scope of rotavirus disease in the newborn and in the premature infant remains to be fully defined. The recent advent and licensure of several rotavirus vaccines should have a significant effect on both nosocomial and community-acquired rotavirus disease (Bernstein, 2009). Human Immunodeficiency Virus Infection During Pregnancy CHANGING EPIDEMIOLOGY OF THE HIV/AIDS PANDEMIC From 1980, when previously healthy, young, gay men first began to have complaints of fever, malaise, weight loss, lymphadenopathy, and malaise, to the identification in 1983 of 36 cases in Kinshasa, Zaire (Congo) of acquired immunodeficiency syndrome (AIDS) by a team of representatives from the CDC and National Institutes of Health, the scope of HIV infection has evolved into a global pandemic. The Joint United Nations Program on 140 Rate (per 100,000 population) in which exposure to infected individuals cannot be controlled, careful hand hygiene education of parents, vaccinating against influenza beginning at 6 months of age, and restricting participation in child care during the RSV season whenever feasible. The prospects for infant or maternal immunization with live attenuated and protein subunit vaccines are under active investigation and development (Crowe, 2001; Fretzayas and Moustaki, 2010; Piedra, 2000). Nosocomial spread of RSV and other respiratory viruses can be minimized by emphasis on hand washing by care providers between contacts with patients. Without additional special precautions, an attack rate of approximately 26% has been observed (Madge et al, 1992). Use of standard contact precautions, such as cohort nursing and the use of gowns and gloves for all contacts with RSV-infected children, can reduce the risk of nosocomial RSV infection to 9.5% (Madge et al, 1992). 120 100 80 505 American Indian/Alaska Native Asian Black/African American Hispanic/Latino* Native Hawaiian/Other Pacific Islander White Mutiple races 60 40 20 0 Males Females FIGURE 37-6 Estimated rates of diagnoses of HIV infection among adults and adolescents, by sex and race/ethnicity, in 2008 for 37 states with confidential name-based HIV infection reporting. Estimated rates resulted from statistical adjustment that accounted for reporting delays, but not for incomplete reporting. (From Centers for Disease Control and Prevention: HIV Surveillance Report 20, 200, June 2010. Available at www.cdc.gov/hiv/topics/surveillance/resources/reports. Accessed November 24, 2010.) HIV/AIDS estimated that, as of the end of 2008, up to 35 million individuals were living with HIV/AIDS globally. Relevant to this chapter, women accounted for 60% of all adults living with HIV (Christie et al, 2008). Currently, national and international surveillance for HIV infection and the resulting illness of AIDS is routine. In the United States, CDC surveillance data for 2005 to 2008 revealed an 8% increase in the absolute number of newly diagnosed HIV infections, whereas the overall rate of new infections was stable. Women currently account for 25% of the HIV-infected U.S. population, with African American women having the highest rate of diagnosed infection among women (Figure 37-6). In Florida, an HIV prevalence study among pregnant women from 1998 to 2007 found that black women were 11-fold more likely than white women to have HIV infection (Salihu et al, 2010). The incidence of HIV infection is challenging to measure and has not been measured directly. Based on a relatively new antibody test that can distinguish recent (median, 156 days) seroconversion from longstanding infection, it is estimated that there were 13,000 to 15,000 HIV infections among U.S. women in 2006 (Hall et al, 2008). Estimates of HIV prevalence in pregnant women worldwide are challenging to obtain and may be underestimated, because those who consent to testing and counseling are more likely to be HIV negative. For example, in an urban antenatal clinic in Zambia, 25% of women accepting HIV testing were found to be HIV positive, whereas anonymous sampling of infants’ cord blood demonstrated a prevalence of 29% among women who refused testing (p <0.0001 for a difference in HIV prevalence between the two groups; Stringer et al, 2005). In Zimbabwe, HIV prevalence among antenatal clinic attendees has been declining, although rates remain high: 32.1% in 2000 and 23.9% in 2004 (Mahomva et al, 2006). Infection rates in antenatal clinics in KwaZulu-Natal, South Africa, reached 37.5% (Ramogale et al, 2007). In the country of South Africa alone, there are an estimated 220,000 pregnant women living with HIV infection. Most African cities exhibit an HIV prevalence rate in the 20% to 40% range, whereas 506 PART IX Immunology and Infections estimates in India are markedly lower—closer to 1% to 2% (Dandona et al, 2008; Lionel et al, 2008). A markedly disproportionate burden of disease is spread among a few countries as evidenced by this summary statement in a 2008 World Health Organization report: “Close to 90% of all pregnant women living with HIV in low and middle-income countries live in 20 countries and 75% are concentrated in 12 countries” (World Health Organization, 2008). The high prevalence of HIV infection among women living in settings where antiretroviral medications are not universally available results in many cases of pediatric HIV infection. Mother-to-child transmission accounts for 90% of HIV infections in children. A recent estimate is that a staggering 2.1 million children younger than 15 years are infected with HIV (World Health Organization, 2008). In addition to the expansive societal and economic cost directly resulting from maternal HIV infection, there is a ripple effect of HIV infection on prevalence and morbidity of other infections. Tuberculosis and malaria are now common opportunistic infections of HIV in low-income countries. Herpes and HIV are considered “overlapping epidemics” (Corey et al, 2004), and the interplay is complex. Reactivation of HSV increases HIV RNA levels in plasma, and mucocutaneous shedding of HIV-1 is greater during mucocutaneous replication of HSV-2. Less publicized but more morbid effects are seen on the risk for puerperal sepsis, which is increased in HIV-infected women. A study in Zimbabwe demonstrated an 11-fold increase in risk for puerperal sepsis among HIV-infected women compared with HIV-negative women (Zvandasara et al, 2006). The same trend is exhibited in European settings; the European Collaborative Study found a fourfold increased risk for puerperal sepsis among HIV-infected women in 14 HIV reference centers across five countries (Fiore et al, 2004). Clearly the effect of HIV is multidimensional, and interdisciplinary cooperation is essential to effect change. PATHOGENESIS OF HIV DISEASE Transmission of the Virus Transmission of HIV infection requires the exchange of bodily fluids. Numerous studies have demonstrated that casual household or community contacts are not associated with transmission or acquisition of HIV/AIDS. The three general mechanisms by which the virus can be passed from one individual to another are sexual, parenteral, and perinatal. Different types of exposure to infection are associated with different risks of infection (Royce et al, 1997). One sexual transmission occurs for every 100 to 1000 exposures. Male-to-female transmission risk is about 10-times greater than female-to-male transmission risk. Needlestick injury from an infected source carries a transmission risk of approximately 0.3% (Panlilio et al, 2005), with the risk from needle sharing being significantly greater. The risk of transmission through blood transfusion is greater than 90%, although this rarely occurs in the United States because of current screening methods used by blood banks. Mother-to-Child Transmission Transfer of virus from maternal to fetal or neonatal tissues is thought to occur during one of three discrete periods— antepartum (before the onset of labor), peripartum (during labor and delivery), and during breastfeeding. Studies of early viral dynamics in infected neonates suggest that the peripartum period is the time of highest risk (Mofenson, 1997). The risk of vertical transmission is heightened by premature rupture of membranes, high maternal viral load, and low maternal CD4 counts (Ioannidis and Contopoulos-Ioannidis, 1999; Mofenson et al, 1999). Without antiretroviral treatment, the risk of vertical transmission is estimated to be 15% to 20% in Europe, 15% to 30% in the United States, and 25% to 35% in Africa (Volmink et al, 2007). With highly active antiretroviral therapy, the risk of vertical transmission in the United States has been reduced to 1% to 5% (Coovadia, 2009). Transmission through breast milk carries a 5% to 20% risk of transmission, largely dependent on the viral load and CD4 count of the mother (World Health Organization, 2008). A Kenyan study that randomized infants of HIV-infected mothers to breast versus formula feeding found that breastfeeding increased the risk of HIV transmission by 16% (Nduati et al, 2000b). Prevention of Mother-to-Child Transmission The cornerstone of preventing mother-to-child transmission of HIV is knowledge of the mother’s HIV status. Updated recommendations from the CDC in 2006 (Branson et al, 2006) reinforced the importance of antenatal testing by stating that: 1. HIV screening should be included in the routine panel of prenatal screening tests for all pregnant women. 2. HIV screening is recommended after the patient is notified that testing will be performed, unless the patient declines (i.e., opt-out screening). 3. Separate written consent for HIV testing should not be required; general consent for medical care should be considered sufficient to encompass consent for HIV testing. 4. Repeated screening in the third trimester is recommended in certain jurisdictions with elevated rates of HIV infection among pregnant women (described as one diagnosed HIV case per 1000 pregnant women per year). Research has shown that HIV testing rates are highest when included in standard testing for pregnant women both in high-income and low-income countries. A study done in Birmingham, Alabama, found that HIV testing rates increased from 75% to 88% (p <0.001) subsequent to the 1999 Institute of Medicine recommendation to make HIV testing a routine part of antenatal care (Stringer et al, 2001). Similarly, in Zimbabwe, testing rates increased from 65% to 99.9% (p <0.001) after implementation of an opt-out HIV testing approach in antenatal clinics (Chandisarewa et al, 2007). HIV testing rates in most low-income countries are lower than those seen in Zimbabwe, but modest gains are being made. Globally, HIV testing in low- and middleincome countries rose from 10% of pregnant women attending antenatal clinics in 2004 to 18% in 2007 (World CHAPTER 37 Viral Infections of the Fetus and Newborn and Human Immunodeficiency Virus Infection during Pregnancy Health Organization, 2008). As antiretroviral therapy becomes more readily available and the effects of this therapy on vertical transmission become more widely known, the incentive to perform HIV testing among pregnant women should increase. ll Maternal Therapy ll The introduction of potent antiretroviral therapies in the middle 1990s markedly reduced deaths caused by AIDS in most regions of the developed world. Antiretroviral therapies have not only reversed AIDS mortality trends in the developed world; they have also dramatically altered patterns of HIV transmission from mothers to children. The results of the Pediatric AIDS Clinical Trial Group Study 076 lead to a profound change in the approach to medical care for HIV-positive pregnant women (Connor et al, 1994). This prospective, placebo-controlled, randomized trial demonstrated that zidovudine prophylaxis during pregnancy, delivery, and early infancy reduced the rate of vertical HIV transmission from 25% to 8%. Less than 2 years later, additional antiretroviral agents became available for treatment of HIV infection. It also became possible to monitor viral load via plasma HIV-1 RNA assays. These monitoring and treatment tools created the potential for the well-controlled management of HIV infection in pregnant women seen today, where it is common to see viral loads of less than 1000 copies per milliliter and vertical transmission risk as low as 1%. Numerous studies have been conducted in low- and middle-income countries as well, demonstrating the benefit of antiretroviral therapy during pregnancy for prevention of vertical transmission of HIV (Dorenbaum et al, 2002; Ioannidis et al, 2001; Namukwaya et al, 2010). Current guidelines exist for antiretroviral therapy in pregnant women (Panel on Treatment of HIV-Infected Pregnant Women and Prevention of Perinatal Transmission, 2010; available at http://aidsinfo.nih.gov/Content Files/PerinatalGL.pdf). Antiretroviral therapy is recommended for all HIV positive pregnant women during the latter trimesters of their pregnancy, with reassessment of the need to continue antiretroviral therapy after birth of the infant. If a woman is already receiving antiretroviral therapy that successfully suppresses viremia, it is recommended to continue the same regimen but find substitutes for any teratogenic drug (e.g., efavirenz). If the woman has a new HIV diagnosis, is antiretroviral therapy–naïve, and does not require antiretroviral therapy for her own health, therapy should be initiated for prophylaxis against vertical transmission based on the parameters in Box 37-5. If the mother is symptomatic from her infection or has a history of antiretroviral therapy use, the Perinatal Guidelines should be referenced for additional recommendations (Panel on Treatment of HIV-Infected Pregnant Women and Prevention of Perinatal Transmission, 2010). Internationally, women in countries with the highest HIV prevalence often have less 25% intrapartum antiretroviral therapy coverage (Figure 37-7). United Nations goals target an 80% coverage rate of antiretroviral therapy in pregnant women, but drug availability still varies widely between countries. Rapid increases in antiretroviral availability in some African countries during 2004 to 2007 prove that such services can be provided quickly and 507 BOX 37-5 T  reatment Recommendations for HIV-Infected, Antiretroviral-Naive Asymptomatic Pregnant Women ll ll ll Perform HIV antiretroviral drug resistance testing before initiating combination antiretroviral drug therapy and repeat after initiation of therapy if viral suppression is suboptimal. Prescribe a combination antiretroviral drug prophylaxis regimen (i.e., at least three drugs) for prophylaxis of perinatal transmission. ll Consider delaying initiation of antiretroviral prophylaxis until after first trimester is completed. ll Avoid use of efavirenz or other potentially teratogenic drugs in the first trimester and drugs with known adverse potential for the mother (e.g., combination stavudine/didanosine). ll Use zidovudine as a component of the antiretroviral regimen when feasible. ll If the woman has CD4 count >250 cells/mm3, use nevirapine as a component of therapy only if the benefit clearly outweighs the risk, because of an increased risk of severe hepatic toxicity. Although the use of zidovudine prophylaxis alone is controversial, consider it if the woman has plasma HIV RNA level <1000 copies/mL on no therapy. Continue antiretroviral prophylaxis regimen during intrapartum period (zidovudine given as continuous infusion during labor while other antiretroviral agents are continued orally). Evaluate the need for continuing the combination regimen postpartum; discontinue the combination regimen unless the woman has indications for continued therapy. If regimen includes a drug with a long half-life, such as NNRTI, consider stopping NRTIs at least 7 days after stopping NNRTI. Adapted from Panel on Treatment of HIV-Infected Pregnant Women and Prevention of Perinatal Transmission: [Perinatal Guidelines] Recommendations for use of antiretroviral drugs in pregnant HIV-1-infected women for maternal health and interventions to reduce perinatal HIV transmission in the United States, May 24, 2010, pp 1-117. http://aidsinfo.nih.gov/ContentFiles/ PerinatalGL.pdf. Accessed November 22, 2010. effectively in low-income countries given adequate political and financial support (Figure 37-8) (World Health Organization, 2008). Antiretroviral Drug Resistance Resistance to available antiretroviral drug therapy has become an increasingly important problem requiring continual reevaluation of antiretroviral therapy regimens for women and prophylactic regimens against perinatal transmission in both mother and infant. Drug resistance in the context of longitudinal maternal antiretroviral therapy is beyond the scope of this chapter, but current recommendations for appropriate drug regimens can be accessed in the Perinatal Guidelines (Panel on Treatment of HIV-Infected Pregnant Women and Prevention of Perinatal Transmission, 2010). It is important to conduct resistance testing before antiretroviral therapy initiation in women or infants, with the exception of the recommended 6 weeks of postnatal zidovudine for infants described in the following section). Selection of an appropriate combination drug regimen is best undertaken with the advice of a clinician specializing in the care of HIV-infected pediatric or adult patients. In low- and middle-income countries, a single dose of nevirapine is often used to prevent vertical transmission of HIV (Guay et al, 1999). Whereas single-dose nevirapine is the most widespread approach to prophylaxis against perinatal transmission of HIV because of cost, limited availability of other drugs, and ease of administration, issues of 508 PART IX Immunology and Infections 80% to 100% 50% to 80% 25% to 50% 10% to 25% Less than 10% Data not available High income areas FIGURE 37-7 Coverage of antiretroviral agents to prevent mother-to-child transmission of HIV in low- and middle-income countries: 2007. (From World Health Organization: Towards universal access: scaling up priority HIV/AIDS interventions in the health sector. Progress Report 2008. Available at www.who.int/hiv/pub/2008progressreport/en. Accessed November 22, 2010.) Estimated number of pregnant women living with HIV % of pregnant women living with HIV receiving antiretrovirals to reduce the risk of mother-to-child transmissjion of HIV, 2007 100% 250,000 UNGASS target for 2010 Number of women 200,000 90% 80% 70% 60% 150,000 50% 100,000 40% 30% 50,000 20% 10% a In di i pi a hi o Et al aw M bi a Za m Ke ny a da ga n U U ni te d R ep ub lic So ut h N Af ig e ric ria of Ta nz an ia M oz am bi qu e 0% a 0 UNGASS: United Nations General Assembly Special Session on HIV/AIDS in 2001 The bar indicates the uncertainty range around the estimate. FIGURE 37-8 Percentage of pregnant women living with HIV receiving antiretroviral agents for preventing mother-to-child transmission of HIV in the 10 countries with the highest estimated number of pregnant women living with HIV: 2007. (From World Health Organization: Towards ­universal access: scaling up priority HIV/AIDS interventions in the health sector. Progress Report 2008. Available at www.who.int/hiv/ pub/2008progressreport/en. Accessed November 22, 2010.) CHAPTER 37 Viral Infections of the Fetus and Newborn and Human Immunodeficiency Virus Infection during Pregnancy nevirapine resistance are paramount. One study published in 2010 was terminated early because of the significant difference in nevirapine resistance and associated treatment failure in infants exposed to single-dose perinatal nevirapine prophylaxis (Palumbo et al, 2010). Another study from Uganda found 100% resistance to nevirapine in 6-monthold HIV-infected infants who had received 6 weeks of daily nevirapine in addition to single-dose nevirapine prophylaxis perinatally. Resistance patterns were less prevalent in infants who had received only the single-dose perinatal prophylaxis (16% at 6 months; p=0.005; Church et al, 2008). Appropriate drug regimens for prophylaxis against perinatal transmission in the face of viral mutation and resistance patterns will likely remain a moving target in high-, middle- and low-income countries. Delivery Method In 1999, two studies (one randomized clinical trial and one metaanalysis) showed a decreased risk of vertical transmission of HIV with cesarean section before rupture of membranes or onset of labor (European Mode of Delivery Collaboration, 1999; International Perinatal HIV Group, 1999). This evidence was compelling enough for the ACOG to issue new guidelines in August 1999 recommending that all HIV positive pregnant women be offered elective cesarean at 38 completed weeks of gestation. The current ACOG recommendations, updated in 2000 and again in December 2010, recommend that all HIV-positive pregnant women with plasma viral loads greater than 1000 copies per milliliter be “counseled regarding the benefits of an elective cesarean delivery” in addition to a zidovudine infusion 3 hours before the operation (Jamieson et al, 2007). Currently, no study has answered whether elective cesarean delivery further decreases the risk of transmission in women with undetectable viral loads or in those receiving antiretroviral therapy. Infant Feeding In the United States, where replacement feeding (i.e., infant formula) is affordable and safe, it is recommended that HIV-positive mothers avoid breastfeeding entirely to minimize the estimated risk of 9% to 14% of postnatal HIV transmission through breast milk. Cell-associated HIV has been detected in human breast milk even among women receiving highly active antiretroviral therapy. Globally, there is less clarity on best practices for infant feeding. When formula feeding is not affordable, feasible, acceptable, sustainable, and safe, then breastfeeding appears to be best for all infants of HIV-infected mothers. Current World Health Organization guidelines updated in 2010 recommend that mothers known to be HIV-infected either breastfeed and receive antiretroviral interventions or avoid all breastfeeding. The guidelines further clarify that feeding recommendations should be formulated at a national or subnational level based on local epidemiology and socioeconomic and infrastructure realities rather than trying to modify feeding recommendations for each individual mother (World Health Organization, 2010). Data from a randomized controlled trial in Zambia showed that abrupt weaning at 4 months after birth versus 509 gradual weaning at the mother’s time of choice (median, 16 months) yielded no difference in HIV-free survival at 24 months of age (Kuhn et al, 2008). Among children who became infected, early weaning increased mortality. This finding contrasts with data from Kenya that showed an increased risk of HIV transmission in breastfed infants at 24 months of age (37% risk of HIV infection in breastfed versus 21% risk in formula-fed infants; Nduati et al, 2000a). Most vertical transmission occurred during the first 6 months of breastfeeding, and the overall risk of transmission from breastfeeding was 16%. The 2-year all-cause mortality was similar between the two groups in this study. Safety of breastfeeding in low-resource settings can be enhanced further by antiretroviral therapy for mothers, breastfeeding infants, or both. A synthesis of related studies was reviewed by Coovadia (2009). The mild-borne transmission of HIV-1C (MASHI) study in Botswana (Thior et al, 2006), the Six Week Extended-Dose Nevirapine study in India, Ethiopia, and Uganda (Six Week Extended-Dose Nevirapine Study Team et al, 2008), the post-exposure prophylaxis of infants (PEPI) study in Malawi (Kumwenda et al, 2008), and the Mitra study in Tanzania (Kilewo et al, 2008) have all assessed the utility of prophylactic antiretroviral agents in breastfeeding infants. There was variation in drug choice and duration, but most of these studies showed decreased vertical transmission in breastfeeding infants on some form of antiretroviral drug. A larger effect on vertical transmission, however, is observed from continuation of maternal antiretroviral therapy throughout the breastfeeding period. Maternal antiretroviral therapy in low-resource settings has decreased the vertical transmission rates during breastfeeding to 1% to 5% (Coovadia, 2009). Evaluation and Treatment of HIV-Exposed Infants Infant Testing Infants born to HIV-infected mothers should have an initial serum evaluation by HIV-DNA or RNA PCR assay. This test is most sensitive if performed initially at 2 weeks of age; however, some physicians also do this testing in the first few days after birth to rule out in utero infection, particularly when the situation is high risk. High-risk situations would include unknown maternal HIV status with concerning history or known maternal HIV infection without prophylactic antiretroviral treatment. A negative test result in the first few days after birth requires confirmatory testing at a later time. Table 37-5 shows the recommended testing and evaluation schedule for an HIV-exposed infant (Havens et al, 2009). If a mother’s HIV status is unknown at the time of delivery, a rapid antibody test should be performed on mother or baby. Test results should be available within a few hours. If this test result is positive, prophylactic antiretroviral therapy should be initiated in the infant within 12 hours of birth, even in the absence of a confirmatory test. Presumptive HIV-negative status of the infant is achieved with two negative tests results, one at 2 weeks old or later and the other at 4 weeks old or later. These are nucleic acid amplification tests (NAATs) and can be either DNA or RNA PCR. A definitive HIV-negative status is achieved 510 PART IX Immunology and Infections TABLE 37-5 Evaluation and Treatment of the Infant Exposed to HIV-1 (Birth to 18 Months of Age), in Addition to Routine Pediatric Care and Immunization Infant Age Action* Birth History and physical examination† X Assess risk of other infections X 14 days 4 wk 6 wk 8 wk 4 mo 12-18 mo X X X X Antiretroviral prophylaxis‡ X X X X Recommend against breastfeeding X X X X Hemoglobin or complete blood cell count X HIV-1 DNA PCR or RNA assay¶ ‖ X** X§ X§ X †† X Initiate PCP prophylaxis‡‡ Enzyme immunoassay for antibody to HIV-1§§ X From Havens PL et al: Evaluation and management of the infant exposed to HIV-1 in the United States, Pediatrics 123:175, 2009. PCR, polymerase chain reaction; PCP, Pneumocystis jirovechii pneumonia; *See text (in Havens et al, 2009) for detailed discussion of each action. If the infant has a diagnosis of HIV-1 infection during this period, laboratory monitoring and immunizations should follow guidelines for treatment of pediatric HIV infection. †Review maternal health information for possible exposure to coinfections. Frequency of examinations is determined, in part, by frequency of visits for immunizations in infancy. ‡Antiretroviral prophylaxis is initiated as soon as possible after birth, but certainly within 12 hours. Zidovudine prophylaxis is continued for 6 weeks. §Checked at 4 weeks by some experts and rechecked at 8 weeks if the week 4 hemoglobin level is significantly low. ¶All HIV-1–exposed infants should undergo virologic testing for HIV-1 with HIV-1 DNA PCR or RNA assays at 14 to 21 days of age and, if results are negative, repeated at 1 to 2 and 4 to 6 months of age to identify or exclude HIV-1 infection as early as possible. Any positive test result at any age is promptly repeated to confirm the diagnosis of HIV-1 infection. ‖HIV-1 DNA PCR or RNA assay testing in the first few days of life allows identification of in utero infection and might be considered if maternal antiretroviral agents were not administered during pregnancy or in other high-risk situations. A negative test result at this age requires repeated testing to exclude HIV-1 infection. **If HIV-1 RNA or DNA testing of the newborn infant was not performed shortly after birth, or if such test results were negative, diagnostic testing with HIV-1 NAAT is delayed until 14 to 21 days of age, because the diagnostic sensitivity of virologic assays increases rapidly by the age of 2 weeks. A negative test result at this age requires repeated testing to exclude HIV-1 infection. Presumptive uninfected indicates negative nucleic acid amplification test (NAAT) result at ≥14 days and ≥4 weeks (1 month) of age; definitive uninfected indicates negative NAAT result at ≥1 and ≥4 months of age. ††No NAAT is needed at 8 weeks of age if previous test results at 2 and 4 weeks of age were negative. A single negative NAAT result at 8 weeks identifies a presumptively uninfected infant. ‡‡Infants with HIV-1 infection should be given PCP prophylaxis until 1 year of age, at which time infants are reassessed on the basis of age-specific CD4+ T-lymphocyte count and percentage thresholds. Infants with indeterminate HIV-1 infection status should receive prophylaxis starting at 4 to 6 weeks of age until they are deemed to be presumptively or definitively uninfected with HIV-1. Prophylaxis is not recommended for infants who meet criteria for presumptive or definitive lack of HIV-1 infection; therefore an NAAT at 2 and 4 weeks of age allows avoidance of PCP prophylaxis if both results are negative. §§Many experts confirm the absence of HIV-1 infection with a negative HIV-1 antibody assay result at 12 to 18 months of age. when a negative NAAT test obtained at 4 weeks old or later is added to a negative NAAT results from 4 months old or later (AAP Committee on Infectious Diseases, 2010b). Many clinicians will also obtain an antibody test at 12 to 18 months after birth. By this time, an HIV-negative infant should have cleared any passively acquired maternal antibodies and have a negative antibody test result, a process referred to as seroreversion. If the infant’s antibody test is positive at this time, an NAAT test should be repeated in addition to close clinical evaluation for signs of infection. Infant Treatment Infants exposed to HIV should receive 6 weeks of prophylactic zidovudine (Table 37-6). It is important that families be given the entire quantity of medication at the time of hospital discharge to avoid challenges for compliance resulting from difficulty in filling a prescription (e.g., limited availability of the proper drug formulation, insurance coverage issues). If the mother has a positive antibody test and received no intrapartum prophylactic zidovudine, some experts would consider early initiation of additional infant antiretroviral medication. However, a careful analysis of the pros and cons of potential toxicity versus decreased transmission risk is imperative. As stated by the Committee on Pediatric AIDS, “If infant prophylaxis with antiretroviral drugs in addition to ZDV is being considered, decisions and choice of antiretroviral drugs should be determined in consultation with a practitioner who is experienced in care of infants with HIV infection,” (American Academy of Pediatrics Committee on Pediatric AIDS, 2008). If the infant has a positive NAAT test result before 6 weeks of age, antiretroviral treatment should be initiated immediately. As stated in Red Book Online (AAP Committee on Infectious Diseases, 2010b), “Initiation of antiretroviral therapy is recommended for all HIV-1-infected infants as soon as infection is confirmed, regardless of clinical, immunologic, or virologic parameters.” Before initiation, however, it is recommended that infants have resistance testing, because infants can acquire resistant virus from their mothers. Extensive guidelines exist for selecting an appropriate combination drug regimen (available at: http://aidsinfo.nih.gov/ContentFiles/PediatricG uidelines.pdf). A specialist in the treatment of pediatric HIV infection should also be consulted. Although it was once thought that treatment with antiretroviral therapy should be delayed until an infant had signs of illness, it is now known that infected infants have a decreased mortality when treatment is started early. Several studies in South Africa have found that early antiretroviral therapy slowed disease progression, and one study found a 76% reduction in early infant mortality and a 75% reduction in HIV disease progression when antiretroviral therapy was initiated early (Violari et al, 2008). CHAPTER 37 Viral Infections of the Fetus and Newborn and Human Immunodeficiency Virus Infection during Pregnancy 511 TABLE 37-6 Intrapartum Maternal and Neonatal Zidovudine Dosing for Prevention of Mother-to-Child Transmission of HIV Dosing Duration Maternal Intrapartum 2 mg/kg IV over 1 hour, followed by continuous infusion of 1 mg/kg/h Onset of labor until delivery Neonatal >35 weeks’ gestation: 2 mg/kg by mouth*,† (or 1.5 mg/kg IV) started within 6-12 h of delivery, then every 6 h Birth to 6 wk 30-35 weeks’ gestation: 2 mg/kg by mouth (or 1.5 mg/kg IV) started within 6-12 h of delivery, then every 12 h, advanced to every 8 h at 2 weeks of age Birth to 6 wk <30 weeks’ gestation: 2 mg/kg by mouth (or 1.5 mg/kg IV) started within 6-12 h of delivery, then every 12 h, advanced to every 8 h at 4 weeks of age Birth to 6 wk From Panel on Treatment of HIV-Infected Pregnant Women and Prevention of Perinatal Transmission: [Perinatal Guidelines] Recommendations for use of antiretroviral drugs in pregnant HIV-1-infected women for maternal health and interventions to reduce perinatal HIV transmission in the United States, May 24, 2010; pp 1-117. http://aidsinfo.nih.gov/ ContentFiles/PerinatalGL.pdf. Accessed November 22, 2010. IV, Intravenous. *Zidovudine dosing of 4 mg/kg given every 12 h has been used for infant prophylaxis in some international perinatal studies. Although there are no definitive data to show equivalent pharmacokinetic parameters or efficacy in preventing transmission, a regimen of zidovudine 4 mg/kg by mouth twice daily instead of 2 mg/kg by mouth four times daily may be considered when there are concerns about adherence to drug administration to the infant. †A simplified zidovudine dosing regimen has been developed for use in low resource settings. This regimen consists of 10 mg by mouth twice daily for infants weighing less than 2.5 kg at birth and 15 mg twice daily for infants weighing more than 2.5 kg at birth. This regimen could be considered for infants in higher resource settings born after 35 weeks’ gestation if simplicity in zidovudine dosing and administration is of prime importance. Prophylaxis with trimethoprim-sulfamethoxazole against opportunistic infections should also be initiated in any infant with possible HIV infection at 6 weeks of age and continued until at least 1 year of age. With a presumptive HIV-negative diagnosis, prophylaxis can be deferred. Although the occurrence of opportunistic infections has declined with more widespread access to testing and antiretroviral therapy, Pneumocystis jirovechii pneumonia remains a common serious opportunistic infection in HIV-1–infected infants, responsible for one third of pediatric AIDS diagnoses (AAP Committee on Infectious Diseases, 2010b). A Cochrane review that evaluated a study in HIV-positive Zambian children found a 33% reduction in mortality among children receiving cotrimoxazole prophylaxis versus placebo (Grimwade and Swingler, 2006). Side Effects of Antiretroviral Therapy Exposure in Infants Antiretroviral therapy has been overwhelmingly successful in the prevention of vertical transmission of HIV, but few human studies have assessed drug safety and toxicity in infants (Thorne and Newell, 2007). Most antiretroviral drugs are pregnancy category B or C; however, efavirenz was changed to a D classification in 2005 when an association with infants with neural tube defects was seen (De Santis et al, 2002; Fundarò et al, 2002). A few case reports of mitochondrial toxicity (hyperlactatemia, some with additional motor or cognitive symptoms) in HIV-uninfected, antiretroviral therapy–exposed infants have come from Europe (Blanche et al, 1999; Noguera et al, 2004; Tovo et al, 2005). Some of the symptoms resolved or improved with time. However, large cohort studies in the United States have not shown excess death owing to mitochondrial dysfunction, although they might be underpowered to detect mild dysfunction. One study done in Malawi showed a hypersensitivity reaction, primarily consisting of a rash, in 16 of 852 (1.9%) infants treated with nevirapine for 28 weeks. This reaction was not seen in the control group, which received only 1 week of zidovudine and lamivudine (Chasela et al, 2010). Studies are inconsistent regarding the risk of prematurity in antiretroviral therapy–exposed infants. European studies have shown an increased risk of prematurity in infants born to mothers receiving antiretroviral therapy regimens containing a protease-inhibitor drug (European Collaborative Study and Swiss Mother and Child HIV Cohort Study, 2000; Thorne et al, 2004). Some U.S. studies have not confirmed this association; however, a U.S. study confirmed a significantly increased risk of very low birthweight in infants born to women receiving proteaseinhibitor–containing antiretroviral therapy versus other combination antiretroviral therapy (adjusted odds ratio, 3.56) (Tuomala et al, 2002). Several studies have identified anemia in infants exposed to prophylactic zidovudine; typically it is mild and transient, but occasionally severe anemia develops (Connor et al, 1994; Taha et al, 2002; Wiktor et al, 1999). Slightly longer term effects were seen in a French study in which decreased levels of platelets, lymphocytes, and neutrophils were persistent in the treatment group after adjusting for several factors, including maternal CD4+ count and prematurity (Le Chenadec et al, 2003). Two European studies have found neutropenia in children exposed to antiretroviral therapy up to 8 years of age (Bunders et al, 2005; European Collaborative Study, 2004). These effects are of uncertain clinical significance, but as prenatal and postnatal antiretroviral therapy exposure increases in duration and number of drugs, studies to assess both subtle and significant sequelae in exposed infants will be crucial. Special Case of HIV-Exposed Uninfected Infants An increasing area of research is focused on HIV-exposed, uninfected (HIV-EU) infants. HIV-EU infants may be at increased risk of death and morbidity. One study demonstrated a fivefold increased risk of death to the infant 512 PART IX Immunology and Infections after the death of an HIV-infected mother (Newell et al, 2004). Other studies have found no increase in morbidity among HIV-EU infants when compared with unexposed infants (Taha et al, 2000). Whereas these findings may simply reflect different levels of background disease burden among vulnerable children when their mother is ill or dead, there may be an additional immunologic effect on the infant from the mother’s HIV infection. In rural Kenya, a study of tetanus antibody levels found a 22% reduction in cord:maternal ratio of tetanus antibodies after adjusting for maternal vaccination and other factors in infants born to HIV-positive mothers. An even greater reduction in infant antibody titers was seen with maternal HIV and malaria coinfection (Cumberland et al, 2007). A large study in Zambia found that HIV-EU infants had twice the risk of death or severe morbidity when the mother’s CD4+ T-cell counts were low even when controlling for maternal mortality, separation due to maternal hospitalization, lower birthweight, and other factors (Kuhn et al, 2005). Another study documented an increase in mortality and morbidity (e.g., respiratory infections) in HIV-EU infants over community baseline in Latin America and the Carribean (Mussi-Pinhata et al, 2007). All these studies are, however, limited by lack of a control group of unexposed infants in the same study setting. Early evidence supporting the possibility of impaired neonatal immunity in HIVEU infants needs further investigation. CONCLUSION Pediatric HIV infection has become a disease of the developing world, where more than 90% of HIV-exposed and more than 98% of HIV-infected children live. Without therapy, morbidity and mortality are high for all HIVinfected individuals and are worse for children. Exposed, uninfected children face the virtual certainty of becoming orphans by their tenth birthdays and a 50% chance of some day becoming infected. HIV transmission is preventable, whether sexual, vertical, or parenteral. Progression to AIDS and death is also preventable. Although the arrest of all HIV transmission remains an unfulfilled goal, potent antiretroviral combination therapy has dramatically decreased AIDS deaths and virtually eliminated vertical transmission of the virus in the United States and other high-income countries. Vigilance in researching sustainable prevention measures and continued advocacy at national and international levels can arrest HIV transmission worldwide. SUGGESTED READINGS Branson BM, Handsfield HH, Lampe MA, et al: Revised recommendations for HIV testing of adults, adolescents, and pregnant women in health-care settings, MMWR Recomm Rep 55:1, 2006. Centers for Disease Control and Prevention: HIV Surveillance Report 20, 2008. Available at www.cdc.gov/hiv/topics/surveillance/resources/reports/. Accessed November 24, 2010. Cheeran MC, Lokensgard JR, Schleiss MR: Neuropathogenesis of congenital cytomegalovirus infection: disease mechanisms and prospects for intervention, Clin Microbiol Rev 22:99-126, 2009. Coovadia H: Current issues in prevention of mother-to-child transmission of HIV1, Curr Opin HIV AIDS 4:319, 2009. Corey L, Wald A: Maternal and neonatal HSV infections, N Engl J Med 361:13761385, 2009. Corey L, Wald A, Celum CL, et al: The effects of herpes simplex virus-2 on HIV-1 acquisition and transmission: a review of two overlapping epidemics, J Acquir Immune Defic Syndr 35:435, 2004. Ergaz Z, Ornoy A: Parovivus B19 in pregnancy, Reprod Toxicol 21:421-435, 2006. Fiore S, Savasi V: Treatment of viral hepatitis in pregnancy, Expert Opin Pharmacother 10:2801-2809, 2009. Hall CB, Weinberg GA, Iwane MK, et al: The burden of respiratory syncytial virus infection in young children, N Engl J Med 360:588-598, 2009. Hall CB, Caserta MT, Schnabel KC, et al: Transplacental congenital human herpesvirus 6 infection caused by maternal chromosomally integrated virus, J Infect Dis 201:505-507, 2010. Hawkes MT, Vaudry W: Nonpolio enterovirus infection in the neonate and young infant, Paediatr Child Health 10:383-388, 2005. Kimberlin DW: Advances in the treatment of neonatal herpes simplex infections, Rev Med Virol 11:157-163, 2001. Lee JY, Bowden DS: Rubella virus replication and links to teratogenicity, Clin Microbiol Rev 13:571-587, 2000. Nassetta L, Kimberlin D, Whitley R: Treatment of congenital cytomegalovirus infection: implications for future therapeutic strategies, J Antimicrob Chemother 63:862-867, 2009. Palumbo P, Lindsey JC, Hughes MD, et al: Antiretroviral treatment for children with peripartum nevirapine exposure, N Engl J Med 363:1510, 2010. Panel on Antiretroviral Therapy and Medical Management of HIV-Infected Children: Guidelines for the use of antiretroviral agents in pediatric HIV infection, 2010, pp 1-219. Available at http://aidsinfo.nih.gov/ContentFiles/PediatricGui delines.pdf. Accessed November 22, 2010. Panel on Treatment of HIV-Infected Pregnant Women and Prevention of Perinatal Transmission: Recommendations for use of antiretroviral drugs in pregnant HIV-1-infected women for maternal health and interventions to reduce perinatal HIV transmission in the United States, 2010, pp 1-117. Available at http:// aidsinfo.nih.gov/ContentFiles/PerinatalGL.pdf. Accessed November 22, 2010. Plotkin SA: The history of rubella and rubella vaccination leading to elimination, Clin Infect Dis 43(Suppl 3):S164-S168, 2006. Prevention of respiratory syncytial virus infections: Indications for the use of palivizumab and update on the use of RSV-IGIV. Committee on Infectious Diseases and Committee of Fetus and Newborn, Pediatrics 102:1211-1216, 1998. Smith CK, Arvin AM: Varicella in the fetus and newborn, Semin Fetal Neonatal Med 14:209-217, 2009. Stagno S: Breastfeeding and the transmission of cytomegalovirus infections, Ital J Pediatr 28:275-280, 2002. Thorne C, Newell ML: Safety of agents used to prevent mother-to-child transmission of HIV: is there any cause for concern? Drug Saf 30:203, 2007. Violari A, Coton MF, Gibb DM, et al: CHER Study Team: Early antiretroviral therapy and mortality among HIV-infected infants, N Engl J Med 359:2233, 2008. Volmink J, Siegfried NL, van der Merwe L, et al: Antiretrovirals for reducing the risk of mother-to-child transmission of HIV infection, Cochrane Database Syst Rev 1:CD003510, 2007. World Health Organization: Towards universal access: scaling up priority HIV/AIDS interventions in the health sector. Progress Report 2008. Available at www.who.int/ hiv/pub/2008progressreport/en/. Accessed November 22, 2010. World Health Organization: Guidelines on HIV and infant feeding 2010. Principles and recommendations for infant feeding in the context of HIV and a summary of evidence. 2010. Available at www.who.int/child_adolescent_health/documents/ 9789241599535/en/index.html. Complete references and supplemental color images used in this text can be found online at www.expertconsult.com C H A P T E R 38 Toxoplasmosis, Syphilis, Malaria, and Tuberculosis Pablo J. Sánchez, Janna C. Patterson, and Amina Ahmed TOXOPLASMOSIS Toxoplasmosis is the disease state that results from infection with the obligate protozoan parasite Toxoplasma gondii. T. gondii is a coccidian that is ubiquitous in nature, and the cat is the definitive host. The organism exists in the following three forms: 1. An oocyst that is shed in cat feces from sporozoites formed within the cat’s intestinal tract 2. A tachyzoite or endozoite that is the proliferative form and was formerly referred to as a trophozoite 3. A tissue cyst that has an intracystic form termed cystozoite or bradyzoite. Nonfeline mammals or birds ingest infective oocysts from contaminated soil. Tissue cysts then accumulate in the organs and skeletal muscle of these animals. The possible routes of transmission from animal to human are direct contact with cat feces, ingestion of undercooked meat containing infective cysts, and ingestion of fruits or vegetables that have been in contaminated soil. Congenital infection results from placental infection and subsequent hematogenous spread to the fetus. EPIDEMIOLOGY Toxoplasmosis is a worldwide medical problem. High prevalence of infection has been documented in Europe, Central and Latin America, and parts of Africa. However, seroprevalence rates differ considerably from one country to another, from one region of a country to another, and even from one ethnic group to another in the same region. These widely disparate seroprevalence rates among different adult populations throughout the world have been explained by differences in eating and sanitation practices that contribute to acquisition of infection. Eating undercooked or raw meat or unwashed raw fruits and vegetables, drinking unpasteurized goat’s milk, working with meat, having three or more kittens, and even certain climactic conditions have been associated with higher risks of infection (Jones et al, 2009). Among women of childbearing age in the United States, the prevalence of antibody to T. gondii varies from approximately 3% to 30%, depending on the region of the country (Boyer and McAuley, 1994; Remington et al, 2001). The lowest seroprevalence rates have been found in the Mountain and Pacific states, and the highest rates have been seen in the Northeastern and Southeastern United States. Seroprevalence rates for pregnant women seem to be decreasing, although regional and ethnic differences persist. The prevalence of congenital infection in Massachusetts and New Hampshire has been documented to be 0.08 per 1000 births through immunoglobulin (Ig) M screening of newborn blood specimens collected on filter paper (Guerina et al, 1994). This finding compares with a rate of 3 to 10 per 1000 live births in Paris and Vienna, where maternal seroprevalence rates of approximately 70% and 40%, respectively, are observed. In Massachusetts, a casecontrol study involving 14 years of newborn screening for congenital toxoplasmosis found that the mother’s birth outside the United States, particularly in Cambodia and Laos, as well as the mother’s educational level and higher gravidity were strongly predictive of congenital infection (Jara et al, 2001). With approximately 4 million live births annually in the United States, there are an estimated 400 to 4000 babies born each year with congenital toxoplasmosis (Feldman et al, 2010). NATURAL HISTORY Infection of the fetus occurs as a consequence of maternal primary infection during pregnancy or, rarely, just before conception (Villena et al, 1998b). Reactivation of latent T. gondii infection during pregnancy does not lead to fetal infection, except among immunocompromised women such as those infected with the human immunodeficiency virus (HIV) or those undergoing chemotherapy (Bachmeyer et al, 2006; Dunn et al, 1997; European Collaborative Study, 1996; Langer, 1983; Mitchell et al, 1990; O’Donohoe et al, 1991; Remington et al, 2001). Under these circumstances, however, the risk is low. In addition, maternal reinfection can result in congenital toxoplasmosis (Gavinet et al, 1997; Hennequin et al, 1997). Acute maternal infection, which is usually acquired early in pregnancy, can lead to fulminant fetal infection resulting in stillbirth, nonimmune fetal hydrops, preterm birth, and perinatal death (Wong and Remington, 1994); however, chronic T. gondii infection rarely has been associated with sporadic abortion (Remington et al, 1964). Infection of the fetus occurs transplacentally during maternal parasitemia. Placental infection is an important intermediary step, and up to 16 weeks may elapse between placental infection and subsequent infection of the fetus. This time delay has been termed the prenatal incubation period (Remington et al, 2001). Congenital toxoplasmosis has occurred in twins and triplets (Couvreur et al, 1991; Sibalic et al, 1986; Wiswell et al, 1984). The clinical manifestations are usually similar in monozygotic twins, whereas discrepancies in clinical findings are common in dizygotic twins. Approximately 40% of infants born to mothers who acquired toxoplasmosis during pregnancy are infected with T. gondii. The rate of vertical transmission varies according to the trimester in which the mother became infected, with fetal infection rates increasing as pregnancy advances (Dunn et al, 1999; Remington et al, 2001; Wong and Remington, 1994). Specifically, when maternal infection 513 514 PART IX Immunology and Infections occurs in the first trimester, 15% of infants are infected, whereas maternal infection in the second and third trimesters results in transmission rates of 30% and 60%, respectively. The severity of clinical manifestations is greatest, however, when maternal infection is acquired early in pregnancy. Maternal infection in the first trimester results in severe disease in as many as 40% of infected fetuses, and in still birth or perinatal death in an additional 35% of infants. Approximately 15% of newborns have subclinical disease; however, maternal infection in the third trimester is rarely if ever associated with severe fetal disease or still birth, and approximately 90% of infants in such situations have subclinical infection. Postnatally, transmission of T. gondii can occur from transfusion of blood or blood products or from transplantation of organ or bone marrow from a seropositive donor with latent infection. Although the organism has been detected in human milk, transmission by breastfeeding has not been documented. The majority of newborns with congenital toxoplasmosis lack clinical signs of infection, although thorough evaluation may demonstrate eye or neurologic abnormalities in approximately 20% of cases. Clinically apparent disease is present in approximately 10% to 25% of infected infants (Alford et al, 1969, 1974; Guerina et al, 1994). The clinical manifestations of toxoplasmosis are often indistinguishable from those seen with other congenital infections, such as cytomegalic inclusion disease and congenital syphilis. Approximately one third of infants have a generalized form of the disease that principally involves organs of the reticuloendothelial system. The abnormalities include temperature instability, hepatosplenomegaly, jaundice, pneumonitis, generalized lymphadenopathy, rash, chorioretinitis, anemia, thrombocytopenia, eosinophilia, and abnormal cerebrospinal fluid (CSF) indices (Table 38-1; Boyer and McAuley, 1994; Eichenwald, 1960). The other two thirds of infected infants principally manifest neurologic disease. Central nervous system involvement is the hallmark of congenital T. gondii infection (Diebler et al, 1985; McAuley et al, 1994; Remington et al, 2001). Chorioretinitis, intracranial calcifications, and hydrocephalus are the most characteristic findings, occurring in approximately 86%, 37%, and 20% of symptomatic infants, respectively (see Table 38-1) (Boyer and McAuley, 1994; Eichenwald, 1960; Remington et al, 2001). This constellation of findings has been referred to as the classic triad of congenital toxoplasmosis; its presence should alert the clinician to the diagnosis. Intracranial calcifications may be single or multiple, but typically are generalized and located in the caudate nucleus, choroid plexus, meninges, and subependyma (Müssbichler, 1968); they also may occur periventricularly, as in cytomegalovirus infection. They are visualized best by computed tomography (CT), but are often detected on ultrasonography as well. Intracranial calcifications may resolve with appropriate antimicrobial therapy (McAuley et al, 1994). Hydrocephalus may be the only manifestation of disease; it results from the extensive periaqueductal and periventricular vasculitis with necrosis that causes obstruction of the ventricular system. Ventriculoperitoneal shunting is often required (Martinovic et al, 1982; McAuley et al, 1994). Abnormalities of the CSF are common; characteristically, they consist of lymphocytic pleocytosis TABLE 38-1 Clinical Findings Among Infants With Congenital Toxoplasmosis Infants with Findings (%) Neurologic Disease* (108 Cases) Generalized Disease† (44 Cases) Chorioretinitis 94 66 Abnormal cerebrospinal fluid 55 84 Anemia 51 77 Convulsions 50 18 Intracranial calcification 50 4 Jaundice 29 80 Hydrocephalus 28 0 Fever 25 77 Splenomegaly 21 90 Lymphadenopathy 17 68 Hepatomegaly 17 77 Vomiting 16 48 Microcephaly 13 0 Diarrhea 6 25 Cataracts 5 0 Eosinophilia 4 18 Abnormal bleeding 3 18 Hypothermia 2 20 Glaucoma 2 0 Optic atrophy 2 0 Microphthalmia 2 0 Rash 1 25 Pneumonitis 0 41 Finding Adapted from Remington JS, McLeod R, Thulliez P, Desmonts G: Toxoplasmosis. In Remington JS, Klein JO, editors: Infectious diseases of the fetus and newborn infant, ed 5, Philadelphia, 2001, WB Saunders, p 246. *Infants with otherwise undiagnosed central nervous system diseases in the first year of life. †Infants with otherwise undiagnosed non-neurologic diseases during the first 2 months of life. and a markedly elevated protein content. Microcephaly, when present, indicates severe brain injury. Hypothermia and hyperthermia may occur secondary to hypothalamic involvement. T. gondii has been detected in the inner ear and mastoid, with the associated inflammation resulting in deafness. An ascending flaccid paralysis with myelitis has also been reported (Campbell et al, 2001). Chorioretinitis secondary to congenital toxoplasmosis can manifest at any age. It usually manifests as strabismus in infants. Defects in visual acuity are more common in older children. Typically the eye lesion consists of a focal necrotizing retinitis that is often bilateral with involvement of the macula and even the optic nerve. Complications include blindness, iridocyclitis, and cataracts (Arun et al, 2007; Phan et al, 2008). Other less common manifestations of congenital toxoplasmosis are nonimmune hydrops fetalis, myocarditis, nephrotic syndrome, and immunoglobulin abnormalities with both hypergammaglobulinemia and hypogammaglobulinemia described. Bony abnormalities consisting of metaphyseal lucencies similar to those seen in congenital CHAPTER 38 Toxoplasmosis, Syphilis, Malaria, and Tuberculosis syphilis have also been reported (Milgram, 1974). A variety of endocrine abnormalities may occur, including hypothyroidism, diabetes insipidus (Oygur et al, 1998; Yamakawa et al, 1996), precocious puberty, and growth hormone deficiency. DIAGNOSIS Isolation of T. gondii from body fluids and tissues provides definitive evidence of infection. The organism can be isolated from placenta, amniotic fluid, fetal blood obtained by cordocentesis, umbilical cord blood, infant peripheral blood, and CSF by means of intraperitoneal and subcutaneous inoculation into laboratory mice (Foulon et al, 1999a; Remington et al, 2001; Wong and Remington, 1994). Mouse inoculation may require as long as 4 to 6 weeks for demonstration of the parasite. Although it is not a practical method, isolation of the organism should be attempted whenever possible. It is available at the Toxoplasma Serology Laboratory, Palo Alto Medical Foundation (860 Bryant Street, Palo Alto, California 94301; telephone: 415-326-8120). In addition, tissue culture has been used to isolate T. gondii from amniotic fluid. Histopathologic examination of the placenta and tissues obtained at postmortem examination or by biopsy from stillborns or infants should be performed because the specimens may demonstrate the presence of tachyzoites. In addition, tachyzoites have been demonstrated in CSF, ventricular fluid, and aqueous humor by specialized staining techniques. Polymerase chain reaction (PCR) analysis has been used successfully to detect T. gondii DNA in amniotic fluid, placenta, CSF, brain, urine, and fetal and infant blood (­Foulon et al, 1999a; Fricker-Hidalgo et al, 1998; Grover et al, 1990; Guy et al, 1996; Hohlfeld et al, 1994; Jenum et al, 1998; Romand et al, 2001). PCR performed on amniotic fluid obtained by amniocentesis has become the preferred method of confirming in utero infection (Kasper et al, 2009). Falsenegative results have been reported, however, and interlaboratory variability in performance of PCR assays has been documented (Guy et al, 1996; Romand et al, 2001). PCR performed on neonatal CSF is recommended for the evaluation of possible central nervous system involvement. Serologic assays for measurement of antibodies to T. gondii in serum and body fluids are the most widely used methods of diagnosing congenital toxoplasmosis (Boyer, 2001; Boyer and McAuley, 1994; Dannemann et al, 1990; Foudrinier et al, 1995; Guerina et al, 1994; ­Lappalainen et al, 1993; Madi et al, 2010; Naessens et al, 1999; Naot et al, 1991; Pinon et al, 2001; Rabilloud et al, 2010; ­Remington et al, 1985; Robert-Gangneux, 2001; Robert-­ Gangneux et al, 1999a, 1999b; Villena et al, 1999; Wong and ­Remington, 1994). These tests are commercially available at the Toxoplasma Serology Laboratory. The more commonly used tests that detect T. gondii-specific IgG antibodies are the Sabin-Feldman dye test, which is considered the gold standard but requires live organisms, indirect immunofluorescent antibody test, IgG enzyme-linked immunosorbent assay (ELISA), direct agglutination, and IgG avidity test. In addition, a differential agglutination test has been developed as a confirmatory test to differentiate acute from chronic maternal infection. This test compares the 515 IgG serologic titer obtained with the use of formalinfixed tachyzoites (HS antigen) with those obtained with acetone- or methanol-fixed tachyzoites (AC antigen). The latter preparation contains stage-specific T. gondii antigens that are recognized by IgG antibodies only during early infection. An additional assay to assist in ruling out maternal infection acquired in the first 3 months of pregnancy is the IgG avidity test performed by the ELISA technique. This test is based on the principle that, although the antibody-binding avidity or affinity for an antigen is initially low after primary antigenic stimulation, IgG antibodies that are present from previous antigenic stimulation are usually of high avidity. Therefore a high-avidity result in the first trimester would exclude an infection acquired in the previous 12 weeks. Finally, an enzyme-linked immunofiltration assay has been developed that allows discrimination between IgG antibodies of maternal origin and IgG synthesized by the fetus as well as identification of antibody subtypes in infected neonates (Zufferey et al, 1999). Tests that detect T. gondii-specific IgM are (1) the double-sandwich IgM ELISA, which has a sensitivity of 75% to 80% and a specificity of 100% (Guerina et al, 1994); (2) the IgM immunosorbent agglutination assay, which is the most sensitive test but should not be performed on umbilical cord blood, because even small quantities of maternal IgM antibodies contaminating the specimen will yield a false-positive result (Boyer and McAuley, 1994); and (3) the IgM immunofluorescent antibody test. The last test is not recommended because it has a much lower sensitivity than either the IgM ELISA or IgM immunosorbent agglutination assay and it has poor specificity secondary to rheumatoid factors and antinuclear antibodies, contributing to false-positive results. Other tests that are still being investigated include a T. gondii-specific IgA ELISA and IgA immunofiltration assay; a T. gondii-specific IgE immunofiltration assay; and IgG, IgM, and IgA immunoblotting tests. Because the majority of adults with acquired T. gondii infection are asymptomatic, evaluation of the pregnant woman and fetus is usually prompted by either seroconversion or an elevated maternal Toxoplasma spp. IgG titer (Couvreur et al, 1988; Daffos et al, 1988; Montoya et al, 2008). The latter may reflect chronic or past infection; therefore the acuity of the maternal infection is determined serologically with the HS-AC differential agglutination test where agglutination titers to formalin-fixed tachyzoites (HS antigen) are compared with titers against acetoneor methanol-fixed tachyzoites (AC antigen). In general, an acute pattern demonstrates high AC and HS titers, while a nonacute pattern demonstrates high AC titers and low HS titers. This method can differentiate an acute from a remote infection in pregnant women, whereas IgM and IgA antibodies detectable by ELISA or ISAGA are elevated for prolonged periods. If recent maternal infection is documented by an acute pattern on the HS-AC test, seroconversion, or rising IgG antibody titers, the fetus should be evaluated by ultrasonography, and amniotic fluid should be tested for specific Toxoplasma spp. DNA with PCR. PCR has supplanted the need for cordocentesis, and a positive result confirms fetal infection (Hohlfeld et al, 1994). Postnatally, serologic testing of paired maternal and infant sera should be performed at a reliable laboratory that will include assays for Toxoplasma spp. IgG and IgM antibodies. 516 PART IX Immunology and Infections Neonatal serum for Toxoplasma spp. IgA determination by ELISA also should be considered, because it may yield the only positive result in some infants. Subinoculation of placental tissue, amniotic fluid, and umbilical cord blood into mice should be considered. If results of these tests suggest possible infection, the newborn should be evaluated fully with complete blood cell count and platelet determination, liver function tests, CSF evaluation (including tests for IgG and IgM antibodies and PCR;) (Wallon et al, 1998), cranial ultrasound or CT imaging of the head, ophthalmologic examination, and hearing evaluation. The presence of neonatal IgM antibody in serum or CSF, or a positive PCR result for blood or CSF, indicates congenital infection. In addition, at-risk infants should undergo serologic follow-up to detect rising serum IgG titers during the 1st year of age or persistence of IgG antibody beyond 12 to 15 months of age, when maternal IgG antibody has disappeared ­(Robert-Gangneux et al, 1999a, 1999b). Uninfected infants show a continuous decline in T. gondii IgG titer with no detectable IgM or IgA antibodies. Low IgG titers and an HS-AC differential agglutination test that indicate remote maternal infection do not require further evaluation of the mother or infant unless the mother is infected with HIV. Because fetal infection has occurred during chronic T. gondii infection in HIVinfected pregnant women, their infants should be evaluated serologically at birth for evidence of congenital infection. It has been suggested that HIV-infected pregnant women who have low CD4+ T lymphocyte counts and who are seropositive for T. gondii antibody receive prophylaxis to prevent fetal infection (Beaman et al, 1992; Wong and Remington, 1994). However, insufficient data currently are available to recommend that such therapy be given routinely for this indication. Nevertheless, if such women previously have had toxoplasmic encephalitis, prophylaxis with pyrimethamine, sulfadiazine, and leucovorin (folinic acid) should be considered (Masur et al, 2002). THERAPY It is currently recommended that fetuses and infants younger than 1 year who are infected with T. gondii receive specific therapy effective against this congenital pathogen, even if they have no clinical signs of disease (­Couvreur et al, 1988; Daffos et al, 1988; Foulon et al, 1999b; ­Friedman et al, 1999; Gilbert et al, 2001; Hohlfeld et al, 1989; Koppe et al, 1986; McAuley et al, 1994; McGee et al, 1992; McLeod et al, 2009; Peyron and Wallon, 2001; Roizen et al, 1995; Vergani et al, 1998; Wallon et al, 1999; Wong and Remington, 1994). On the basis of comparison with untreated historical controls, outcome is improved substantially by neonatal treatment. The effectiveness of maternal and fetal treatment is less clear. Spiramycin has been used in pregnant women with acute toxoplasmosis to reduce transplacental transmission of T. gondii. If fetal infection is confirmed after the 17th week of pregnancy, however, treatment with pyrimethamine, sulfadiazine, and folinic acid is recommended. Prenatal treatment of congenital toxoplasmosis is believed to reduce the clinical severity of infection in the newborn while shifting the disease to a more subclinical form. This effect in turn may ameliorate the long-term neurologic complications that are commonly seen among infants who have clinical manifestations in the neonatal period. A recent metaanalysis of the effectiveness of prenatal treatment of toxoplasmosis infection found no evidence that such treatment significantly decreased clinical manifestations of disease in infected infants (Thiebaut et al, 2007). Neonatal treatment has also resulted in reductions in sensorineural hearing loss and neurodevelopmental and visual handicaps. Table 38-2 shows the recommended guidelines for the treatment of congenital toxoplasmosis. In infants with congenital toxoplasmosis, the treatment consists of pyrimethamine, sulfadiazine, and folinic acids (Boyer and McAuley, 1994; McAuley et al, 1994; McLeod et al, 1992; Remington et al, 2001). The actual duration of therapy is not known, although prolonged courses of at least 1 year are preferred. Currently most experts recommend combined treatment until the patient is 1 year old (Remington et al, 2001; Villena et al, 1998a, 1998b). Complete blood cell counts and platelet determination must be monitored closely while the patient is receiving therapy, because granulocytopenia, thrombocytopenia, and megaloblastic anemia can occur. These parameters usually improve once a higher dose of folinic acid is administered or pyrimethamine and sulfadiazine are discontinued temporarily. The indications for adjunctive therapy with corticosteroids such as prednisone (0.5 mg/kg twice per day) are CSF protein concentration 1 g/dL or higher and chorioretinitis that threatens vision; corticosteroid treatment is continued until either condition resolves. Current therapies are not effective against encysted bradyzoites and therefore might not prevent reactivation of chorioretinitis and neurologic disease. PROGNOSIS Maternal toxoplasmosis acquired during the first and second trimesters has been associated with still birth and perinatal death secondary to severe fetal infection in approximately 35% and 7% of cases, respectively. Among infants born with congenital toxoplasmosis, the mortality rate has been reported to be as high as 12%. In addition, infants with congenital toxoplasmosis are at high risk for ophthalmologic, neurodevelopmental, and audiologic impairments, including mental retardation (87%), seizures (82%), spasticity and palsies (71%), and deafness (15%) (Eichenwald, 1960; Hohlfeld et al, 1989; Koppe et al, 1986; McAuley et al, 1994). Of neonates with subclinical infection, long-term follow-up reveals eye or neurologic disease in as many as 80% to 90% by the time they reach adulthood (Couvreur and Desmonts, 1962; Couvreur et al, 1984; McLeod et al, 2000; Saxon et al, 1973; Wilson et al, 1980). Data from the United States National Collaborative Treatment Trial show that treatment of neonates with congenital toxoplasmosis early and for 1 year resulted in more favorable outcomes than were reported for untreated infants or infants who were treated for only 1 month. PREVENTION Pregnant women whose serologic status for T. gondii is negative or unknown, as well as women who are attempting to conceive, should be educated on the prevention CHAPTER 38 Toxoplasmosis, Syphilis, Malaria, and Tuberculosis 517 TABLE 38-2 Treatment Guidelines for Toxoplasmosis Condition Therapy Dose (Oral Unless Specified) Duration Pregnant woman with acute toxoplasmosis Spiramycin for first 21 wk of gestation or until term if fetus not infected* 1 g every 8 h without food Until fetal infection documented or excluded at 21 wks; if fetal infection documented, replaced with pyrimethamine, leucovorin, and sulfadiazine (see below) Pyrimethamine (if fetal infection confirmed after 18th week of gestation or if infection acquired in last few weeks of gestation) and Loading dose: 100 mg/day in two divided doses for 2 days followed by 50 mg/day Until delivery Sulfadiazine* and Loading dose: 75 mg/kg/day in two divided doses (maximum, 4 g/d) for 2 days; then 100 mg/kg/day in two divided doses (maximum, 4 g/day) Until delivery Leucovorin† 10-20 mg/d Until delivery Pyrimethamine and Loading dose: 2 mg/kg/day for 2 days; then 1 mg/kg/day for 2 or 6 months; then 1 mg/kg/day on Mon, Wed, and Fri each week ≥1 yr Congenital Toxoplasma gondii infection in infant ≥1 yr Sulfadiazine* and 100 mg/kg/day in 2 daily divided doses Leucovorin (folinic acid)† 10 mg 3 times weekly Corticosteroids (prednisone)‡ ≥1 yr 1 mg/kg/daily in 2 daily divided doses Until resolution of elevated (≥1 g/dL) CSF protein or active chorioretinitis that threatens vision Data from Boyer KM, McAuley B: Congenital toxoplasmosis, Semin Pediatr Dis 5:42, 1994; and Remington JS, McLeod R, Thulliez P, Desmonts G: Toxoplasmosis. In Remington JS, Klein JO, editors: Infectious diseases of the fetus and newborn infant, ed 5, Philadelphia, 2001, WB Saunders, p 293. CSF, Cerebrospinal fluid. *Available only on request from the U.S. Food and Drug Administration (301-827-2127; fax, 301-927-2475). †Monitor blood and platelet counts weekly; adjust dosage for megaloblastic anemia, granulocytopenia, or thrombocytopenia. ‡When signs of inflammation or active chorioretinitis have subsided, dose can be tapered and eventually discontinued; use only in conjunction with pyrimethamine, sulfadiazine, and leucovorin. of congenital toxoplasmosis through avoidance of at-risk behaviors that may expose them to cat feces or encysted bradyzoites in raw meat (Centers for Disease Control and Prevention, 2000; Eskild et al, 1996; Foulon et al, 2000; Jones et al, 2001; Wilson and Remington, 1980). Such women should be taught to wear gloves when changing cat litter boxes or gardening and to wash hands after such activities. Daily changing of cat litter will also decrease the chance of infection, because oocysts are not infective during the first 1 to 2 days after passage. In addition, feeding cats commercially prepared foods rather than undercooked meats or wild rodents reduces the likelihood of their becoming infected and capable of transmitting the infection to a pregnant woman. Oral ingestion of T. gondii can be prevented by either cooking meat to well done, smoking it, or curing it in brine, and by washing kitchen surfaces that come into contact with raw meat. Vegetables and fruits should be washed, and hands and kitchen surfaces should be cleaned after handling fruits, vegetables, and raw meat. Flies and cockroaches may serve as transport hosts for T. gondii, so their access to food must be prevented. Routine serologic screening of women during pregnancy has been an effective means of prevention in such countries as France and Austria, where the incidence of congenital toxoplasmosis is high. No such screening is currently recommended in the United States. However, high-risk women, including those who are immunocompromised, should be screened early in pregnancy. Neonatal screening for IgM antibody has also been advocated so that asymptomatic infants can be detected and treated before neurologic symptoms develop (Peterson and Eaton, 1999). This strategy, however, has been hampered by the lack of readily available and reliable IgM test kits. Moreover, such screening will not detect the approximately 25% of infected infants who lack anti-Toxoplasma spp. IgM antibody. Further study involving cost analyses is needed to define the best preventive strategy for congenital toxoplasmosis in specific populations, regions, and countries. SYPHILIS Syphilis is caused by infection with the spirochete Treponema pallidum. In adults, this spirochete is transmitted through sexual contact, but infants acquire the infection from their mothers, either in utero or during delivery. The history of syphilis epidemics is one of intermittent peaks. During the 1930s and 1940s in the congenital syphilis clinic of the Harriet Lane Home (Baltimore, Md.), 60 to 80 infants and children attended each week for arsenic therapy. Many more were lost to follow-up before completing their 2- to 3-year course of treatment. It was unusual if fewer than three or four new examples were discovered in the general outpatient department in the course of 1 week. Then, for several decades, the frequency of new cases of congenital syphilis declined. 518 PART IX Immunology and Infections INCIDENCE AND EPIDEMIOLOGY After years of declining incidence, congenital syphilis is again on the rise. Recent Centers for Disease Control and Prevention (CDC) surveillance data show a 23% increase in congenital syphilis cases from 2005 to 2008 as shown in Figure 38-1 (Centers for Disease Control and Prevention, 2010a). This increase is directly linked to a rise in primary and secondary syphilis in women from 2004 to 2007. In 2008, there was a total of 431 congenital syphilis cases in the U.S. Half of these cases were in infants born to black mothers, primarily in the South. Approximately 30% of the mothers of these infected infants did not receive prenatal care. When mothers did receive prenatal care and their infants still became infected, 27% were screened <30 days before delivery (likely resulting in a false negative test) and 24% screened positive but were not treated (Centers for Disease Control and Prevention, 2010a). An outbreak was also reported among Pima Indians in Arizona in 2007-2009 in which a total of 106 cases were identified, including six congenital cases, two of which resulted in stillbirth (Centers for Disease Control and Prevention, 2010b). A similar epidemic was seen in Alabama with a peak of 238 cases in 2006 with the largest increase seen in heterosexual women (Centers for Disease Control and Prevention, 2009). Worldwide more newborns are affected by congenital syphilis than by any other neonatal infection (Schmid, 2004). Countries such as Ethiopia, Swaziland, and Mozambique have prevalence rates of maternal syphilis as high as 12% to 13% (World Health Organization, 2007). There are an estimated 2 million pregnancies affected annually, and most women infected with syphilis within 1 year of their pregnancy will transmit the infection to their infant. Adverse outcomes in these pregnancies are severe: 17% to 40% result in stillbirth, 10% to 23% in neonatal death, and 10% to 30% result in congenital syphilis infection (World Health Organization, 2007). The global problem of congenital syphilis is further confounded by the high prevalence of HIV, which reaches 25% to 40% in some African cities. Mothers who are coinfected with syphilis and HIV may be less likely to respond to penicillin treatment, which increases the risk of congenital syphilis in the fetus (Lukehart et al, 1988). In addition, HIV-infected mothers with untreated syphilis may be at increased risk of transmitting HIV to their fetus secondary to a placentitis that allows the virus to pass from maternal to fetal circulation (Pollack et al, 1990). In one study of HIV-infected women from Tanzania, coinfection with syphilis doubled the risk of still birth (Kupka et al, 2009). The extent of negative synergy between these two infections will become increasingly clear as research progresses. PREVENTION Screening for syphilis in pregnancy is the primary method of preventing congenital syphilis. The American Academy of Pediatrics and the American Congress of Obstetricians and Gynecologists in Guidelines for Perinatal Care (2002) recommend screening all pregnant women at the first prenatal visit, after exposure to an infected partner, and at delivery. Reinforcing this recommendation is the Reaffirmation Recommendation Statement by the U.S. Preventive Services Task Force (2009) that “Grade A evidence” is present for screening all pregnant women for syphilis infection. As summarized by Miller and Karras (2010) in reference to the latest CDC surveillance data, “the increase in the 50 8 40 6 30 4 20 2 10 0 P&S rate (per 100,000 females) CS rate (per 100,000 live births) CS rate P&S syphilis rate among females 0 1995 1997 1999 2001 2003 Year 2005 2007 * CS rates from 1995 to 2006 were calculated using yearly live birth data as denominators. Rates for 2007 and 2008 were calculated using live birth data for 2006. Available at http:// www.cdc.gov/nchs/births.htm. † P&S syphilis rates were calculated using bridged race population estimates for 2000-2007 based on 2000 U.S. Census counts. Available at hhtp://wonder.cdc.gov/wonder/help/ bridged-race.html. FIGURE 38-1 Congenital syphilis (CS) rate among infants less than 1 year old and the rate of primary and secondary syphilis (P&S) among females 10 years or older. National Electronic Telecommunication System for Surveillance, United States, 1995-2008. (Data from Centers for Disease Control and Prevention: Congenital syphilis: United States, 2003-2008, MMWR Morb Mortal Wkly Rep 59:413, 2010.) 519 CHAPTER 38 Toxoplasmosis, Syphilis, Malaria, and Tuberculosis congenital syphilis rate, the substantial burden of primary and secondary syphilis among black women in the south, and the high case/fatality ratio associated with congenital syphilis require that congenital syphilis prevention be given high priority in areas with high syphilis morbidity and evidence of heterosexual syphilis transmission.” ETIOLOGY AND PATHOGENESIS The organism responsible for syphilis is Treponema pallidum. This delicate, corkscrew-shaped, flagellated, highly motile spirochete is almost identical in appearance to Treponema pertenue, which causes yaws. Because Treponema spp. enter the fetal bloodstream directly, the primary stage of infection is completely bypassed. There is no chancre and no local lymphadenopathy. Instead, the liver, the immediate target of the invasion, is flooded with organisms that then penetrate all the other organs and tissues of the body to a lesser degree. Other sites of invasion include skin, mucous membranes of the lips and anus, bones, and the central nervous system. If fetal invasion has taken place early, the lungs may be heavily involved in a characteristic pneumonia alba, but this condition is usually life threatening. Treponema spp. may be found in almost any other organ or tissue of the body, but seldom cause inflammatory and destructive changes in loci other than the ones named previously. Under the microscope, the tissue alterations consist of nonspecific interstitial fibrosis with or without evidence of low-grade inflammatory response in the form of round cell inflammation. Necrosis follows fairly regularly in bone, but only rarely in other tissues. Localization and gumma formation are not common in the neonate; however, extramedullary hematopoiesis in the liver, spleen, kidneys, and other organs can be seen. Syphilis in infants is acquired primarily by transplacental transmission, which can occur at any time during pregnancy but ordinarily occurs during 16 to 28 weeks’ gestation. Fetuses infected early may die in utero or are at high risk for significant neurodevelopmental morbidity. The usual outcome of a third-trimester infection is the birth of an apparently normal infant who becomes ill within the first few weeks of life. Virtually all infants born to women with primary or secondary infection have congenital infection, but approximately half are clinically symptomatic. As hospitals and insurers strive to cut costs, early discharge of new mothers has hindered identification of congenitally infected, asymptomatic infants whose mother’s infections occurred late in the third trimester and whose syphilis serologic test results are not yet positive at the time of delivery (Dorfman and Glaser, 1990). It is critical that at-risk infants have a source of primary health care capable of tracking both maternal and infant syphilis status (Chhabra et al, 1993; Zenker and Berman, 1991). Early latent infection results in a 40% infant infection rate, and late latent infection results in a 6% to 14% infant infection rate (Wendel, 1988). CLINICAL PRESENTATION Congenital syphilis is challenging to diagnose in neonates because 60% of infected infants are asymptomatic. Early presentation is typically within the first 4 weeks after birth and is characterized by the signs and symptoms summarized in Table 38-3. Persistent rhinitis (“snuffles”) was estimated to occur in two thirds of patients in the early literature, but is now less prevalent (Ingall et al, 2006). Prematurity and low birthweight is seen in 10% to 40% of infants (Saloojee et al, 2004). Additional diagnoses associated with congenital syphilis include nonimmune hydrops, nephrotic syndrome, and myocarditis. Cutaneous lesions can appear at any time from the 2nd week after birth and onward. These copper-colored lesions can be either sparse or numerous; round, oval, or iris-shaped; and circinate or desquamative (Figure 38-2). Even more characteristic than their appearance is their distribution, which most frequently includes the perioral, perinasal, and diaper regions. Palms and soles are also involved, but the rash is soon replaced there by diffuse reddening, thickening, and wrinkling. In heavily infected infants, the rash may become generalized. Mucocutaneous junctions become involved in typical fashion. The lips become thickened and roughened and tend to weep. TABLE 38-3 Clinical Features of Congenital Syphilis in the Neonatal Period Feature Prevalence (%) Hepatomegaly with or without splenomegaly 33-100 Radiographic bone changes 75-100 Lymphadenopathy 50 Blistering skin rash 40 Respiratory distress 34 Jaundice 33 Pseudoparalysis of Parrot 12 Fever 16 Bleeding 10 Adapted from Remington JS, Klein JO, Wilson CB, et al, editors: Infectious diseases of the fetus and newborn infant, ed 6, Philadelphia, 2006, Saunders, p 572 and Saloojee H, Velaphi S, Goga Y, et al: The prevention and management of congenital syphilis: an overview and recommendations, Bulletin of the World Health Organization 82, 2004. FIGURE 38-2 Congenital syphilis with desquamation over the hand. (From American Academy of Pediatrics Committee on Infectious Diseases: Syphilis: clinical manifestations images. Red Book Online Visual Library, 2009, American Academy of Pediatrics. Available at http:// aapredbook.aappublications.org/cgi/content/figsonly/2009/1/3.129. Accessed November 19, 2010.) 520 PART IX Immunology and Infections Radial cracks appear that traverse the vermillion zone up to and a bit beyond the mucocutaneous margins of the lips. These cracks are the beginnings of the radiating scars that may persist for many years as rhagades. Similar mucocutaneous lesions involve the anus and vulva, but in these locations, the white, flat, moist, raised plaques known as condylomata are also encountered, although less frequently. Radiographs of the bones show characteristic osteochondritis and periostitis in 80% to 90% of infants with symptomatic congenital syphilis (Figure 38-3). In most cases, the bone lesions are asymptomatic, but in a few they are severe enough to lead to subepiphyseal fracture and epiphyseal dislocation with an extremely painful pseudoparalysis of one or more extremities. Approximately 20% of asymptomatic, congenitally infected infants have metaphyseal changes consistent with congenital syphilis. Radiographic alterations include an unusually dense band at the epiphyseal ends, below which is a band of translucency whose margins are at first sharp but that later become serrated, jagged, and irregular. The shafts become generally more opaque, but spotty areas of translucency throughout may give them a moth-eaten look. The periosteum of the long bones becomes more thickened. Epiphyses separate because the dense end plate breaks away from the shaft by fracture through the subepiphyseal zone of decalcification. Pseudoparalysis of Parrot can occur because of bone pain from these changes. Signs of visceral involvement include hepatomegaly, splenomegaly, and general glandular enlargement. ­Palpable epitrochlear nodes are not pathognomonic, but are highly suggestive of congenital syphilis. The liver may be greatly enlarged, firm, and nontender. Associated with this finding may be jaundice, which appears in the 2nd or 3rd weeks of life, is seldom intense, and does not persist for many days. Anemia, probably indicative of bone marrow infection and hematopoietic suppression, may become severe. Lesions in the gastrointestinal tract and pancreas can occur and produce distention and delay in the passage of meconium. Clinical signs of central nervous system involvement seldom appear in the newborn infant, although one third to FIGURE 38-3 The radiograph displays the characteristic “celery stalking” and widening of the metaphases in long bones found in untreated congenital syphilis. (From American Academy of Pediatrics Committee on Infectious Diseases: Syphilis: clinical manifestations images. Red Book Online Visual Library, 2009, American Academy of Pediatrics. Available at http://aapredbook.aappublications.org/cgi/content/figsonly/ 2009/1/3.129.) half of those infected suffer such involvement. One study by Beeram et al (1996) found no difference in leukocyte values or protein between the CSF of symptom-free infants born to syphilis-seropositive mothers (positive rapid plasmin reagin [RPR] test result confirmed by positive fluorescent treponemal antibody-absorbent [FTA-ABS]) and symptom-free control infants for whom sepsis for associated sepsis risk factors had been ruled out. In addition, only 2 of the 329 infants born to syphilis-infected mothers with no or inadequate treatment had a positive venereal disease research laboratory ( VDRL) test result on their CSF specimen. Another study of neonatal CNS invasion by T. pallidum found the sensitivity and specificity of CSF VDRL to be 71% and 92%, respectively (Sanchez et al, 1993). Overall, the added value of CSF evaluation in asymptomatic infants appears marginal, because few cases of congenital syphilis are diagnosed based on CSF abnormalities in the absence of other laboratory or clinical findings. However, the CDC and the European International Union against Sexually Transmitted Infections guidelines still recommend that infants with confirmed disease (e.g., those with increasing quantitative titers) or infants who have a normal physical examination result but with an inadequately treated, syphilis-infected mother have a CSF evaluation for VDRL, cell count, and protein (Centers for Disease Control and Prevention, 2006; French et al, 2009). Further research may inform future guideline development. DIAGNOSIS The suggestion of congenital syphilis infection should be spurred by a positive maternal antibody test (RPR or VDRL). If the mother’s treponemal test (treponema pallidum particle agglutination assay [TP-PA] or FTA-ABS) is also positive, a RPR or VDRL test should be performed on the infant’s serum per current Red Book recommendations by the American Academy of Pediatrics (2009). Cord blood is not a reliable testing source for neonatal infection as false-positive RPR rates of 10% and false-negative RPR rates of 5% have been reported (Rawstron and Bromberg, 1991). However, a serum-positive antibody test alone does not confer a diagnosis of congenital syphilis, because of transplacental transfer of maternal nontreponemal and treponemal IgG antibodies. Because IgG is transferred across the placenta, its finding in the baby’s serum simply means that the mother either currently has or has had syphilis. She may have been successfully treated during pregnancy and yet still has antibodies in her blood, or she may not have received treatment at all and still has not passed the disease on to her fetus. Conversely, if an antibody test is done on the infant’s serum because of clinical suspicion in the absence of a positive maternal antibody test, the infant’s antibody test can also be falsely negative if the mother acquired infection late in pregnancy, because she may not have had time to form antibodies herself or has not passed them to her fetus. IgM does not cross the placenta, so the presence of specific IgM antibodies in the infant is generally diagnostic. Whereas the IgM immunoblot appears to be the best ­available method to detect T. pallidum-specific IgM in neonates, the CDC (2006) does not identify this as a commercially available and recommended test (Herremans CHAPTER 38 Toxoplasmosis, Syphilis, Malaria, and Tuberculosis et al, 2010). Treponemal tests on infant serum are not necessary. Comparison of maternal and infant nontreponemal serologic titers at delivery and followed over time are most informative for infant diagnosis. These titers should be accessed via the same test and in the same laboratory to ensure valid comparisons. If the infant titer is fourfold higher than the mother’s titer, it is considered highly suggestive of congenital infection. Additional criteria for treatment in the neonate include inadequate maternal treatment in a mother with syphilis and the presence of clinical, laboratory, or radiographic evidence in the infant. 521 Evaluation of an infant with a concerning physical examination result and maternal profile should include darkfield or fluorescent antibody test of skin lesions or mucous discharge. In addition, a complete blood cell count with differential should be obtained, and CSF studies for cell count, protein and VDRL are recommended. If clinically indicated, long-bone radiographs can be diagnostic. Clearly the evaluation for congenital syphilis is complex, and serial investigations may be necessary for definitive diagnosis. An algorithm for evaluation is shown in Figure 38-4. Reactive maternal RPR/VDRL Nonreactive maternal TP-PA or FTA-ABS False-positive reaction: no further evaluation Reactive maternal TP-PA or FTA-ABSa Infant RPR/VDRL nonreactive; Treatment (1) Infant RPR/VDRL less than 4 times maternal RPR/VDRL Infant physical examination normal a b c d Maternal treatment before prenancyb Maternal penicillin treatment during pregnancy and more than 4 weeks before delivery No evaluation or treatment No evaluation; Treatment (2) Infant RPR/VDRL reactive Infant RPR/VDRL at least 4 times maternal RPR/VDRL Infant physical examination abnormal Maternal treatment: none; or undocumented; or 4 wk or less before delivery; or nonpenicillin drug; or maternal evidence of reinfection or relapse (fourfold or greater increase in titers) Evaluatec Treatment (3) Evaluated Normal evaluation Abnormal, not done, or incomplete evaluation Treatment (2) Treatment (3) RPR indicates rapid plasma reagin (test); VDRL, Venereal Disease Research Laboratory (test); TP-PA, Treponema pallidum particle agglutination (test); FTA-ABS, fluorescent treponema antibody absorption (test). Test for human immunodeficiency virus (HIV) antibody. Infants of HIV-infected mothers do not require different evaluation or treatment. Women who maintain a VDRL titer 1:2 or less (RPR 1:4 or less) beyond 1 year after successful treatment are considered serofast. Evaluation consists of complete blood cell (CBC) and platelet count; cerebrospinal fluid (CSF) examination for cell count, protein, and quantitative VDRL. Other tests as clinically indicated: long-bone and chest radiographs, neuroimaging, auditory brainstem response, eye examination, liver function tests. CBC, platelet count; CSF examination for cell count, protein, and quantitative VDRL; long-bone radiography. TREATMENT: (1) If the mother has had no treatment, undocumented treatment, treatment 4 weeks or less before delivery or evidence of reinfection or relapse (fourfold or greater increase in titers) AND the infant’s physical examination is normal, THE treat infant with a single intramuscular (IM) injection of benzathine penicilin (50 000 U/kg). If these criteria are not met, no treatment is required. In both scenarios, no additional evaluation is needed. (2) Benzathine penicillin G, 50 000 U/kg, IM,  1 dose. (3) Aqueous penicillin G, 50 000 U/kg, IV, every 12 hours (1 week of age or younger), every 8 hours (older than 1 week), or procaine penicillin G, 50 000 U/kg, IM, single daily dose,  10 days. FIGURE 38-4 Algorithm for evaluation and treatment of infants born to mothers with reactive serologic tests for syphilis. (From American Academy of Pediatrics Committee on Infectious Diseases: Syphilis, Red Book Online, 2009, American Academy of Pediatrics. Available at http://aapredbook.aappublication s.org/cgi/content/full/2009/1/3.129. Accessed November 19, 2010.) 522 PART IX Immunology and Infections TABLE 38-4 Recommended Treatment of Pregnant Patients With Syphilis Stages of Syphilis Drug (Penicillin) Route Dose (Units) Early (<1 yr duration) Primary, secondary, or early latent HIV antibody-negative Recommended: benzathine IM 2.4 million single dose; possibly repeat in 1 wk HIV antibody-positive* Recommended: benzathine IM 2.4 million single dose; possibly repeated weekly for 3 wk IM 2.4 million weekly for 3 wk IV 3-4 million every 4 hr for 10-14 days IM 2.4 million daily for 10-14 days Alternative: penicillin desensitization Latent (>1 yr duration)† Recommended: benzathine Alternative: penicillin desensitization Neurosyphilis Recommended: aqueous Alternative: procaine‡ From Remington JS, Klein JO, Wilson CB, et al, editors: Infectious diseases of the fetus and newborn infant, ed 6, Philadelphia, 2006, WB Saunders, p 567. IM, Intramuscular; IV, intravenous. *With normal cerebrospinal fluid findings, if performed. †Lumbar puncture to exclude neurosyphilis is recommended for HIV-antibody–positive patients. ‡Probenecid, 500 mg orally, four times per day for 10 to 14 days, should also be prescribed. TREATMENT AND LONG-TERM OUTCOMES Treatment of all pregnant women infected with syphilis is recommended. Guidelines for maternal treatment are found in Table 38-4 and are tailored to the stage of disease (e.g., primary, secondary, latent). Treatment of an infected infant requires 10 days of intravenous or intramuscular penicillin. Specific dosing and treatment regimens are outlined in Table 38-5, followed by long-term follow-up recommendations in Table 38-6. Notably, if a single dose is missed during the treatment of an infected infant, restarting the entire series is recommended. Late manifestations of infection in untreated infants, even if initially asymptomatic, can occur years after birth. Pathognomonic presentations include the Hutchinson triad (interstitial keratitis, eight cranial nerve deafness and Hutchinson’s [peg-shaped] teeth), mulberry molars (first lower molar with many small cusps), and Clutton’s joints (synovitis with hydrarthrosis, tenderness, and limited range of motion) (Fiumara et al, 1970). Mental retardation can also be a feature of late, untreated congenital syphilis (Brosco et al, 2006). CONGENITAL MALARIA Malaria is a parasitic disease of epidemic proportion. An estimated 200 to 500 million cases occur worldwide each year, resulting in 2 to 3 million deaths. The greatest burden of disease occurs in Sub-Saharan Africa, although malaria is increasingly recognized as a significant public health problem in Asia and Oceania. In areas of high transmission, mortality is concentrated largely among young children and pregnant women. Malaria is caused by four Plasmodium species: P. falciparum, P. vivax, P. ovale, and P. malariae. Of these, P. falciparum is the major cause of morbidity and mortality. Humans typically acquire infection through the bite of the Anopheles spp. mosquito. Transmission may also occur through blood transfusion or vertically from the mother to fetus, resulting in congenital malaria. Despite the prevalence of malaria in the developing world, it has historically been believed that congenital malaria occurs infrequently in areas of high transmission (Bruce-Chwatt, 1952; Covell, 1950; McGregor, 1984). As of 1995, only 300 cases of congenital malaria had been reported in the literature (Balatbat et al, 1995), largely from outside of malarious areas. The rarity of vertical transmission has been attributed to the effectiveness of the placenta as a barrier against passage of maternal parasitized red blood cells as well as the protective effect of maternally derived antibodies in the fetus and newborn. Recent reports, however, suggest that congenital malaria is no longer a rare occurrence in endemic areas and has likely been underrecognized and underreported (Akindele et al, 1993; Falade et al, 2007; Fischer, 1997, Ibhanesebhor, 1995; Menendez and Mayor, 2007; Uneke, 2007a, 2007b). Although endemic malaria has been eliminated from the United States, approximately 1200 cases of malaria are reported to the CDC annually, almost exclusively in travelers and immigrants from endemic countries. From 1966 to 2005, 81 cases of congenital malaria were reported to the National Malaria Surveillance System of the CDC (Lesko et al, 2007). The majority of infants were born to foreign-born women. As an increasing number of people travel to and emigrate from malarious areas, the number of cases of malaria (and consequently congenital malaria) will continue to increase. The lack of familiarity with this disease in the United States renders it a diagnostic and therapeutic challenge for clinicians, with delays in diagnosis potentially leading to significant morbidity and mortality (Griffith et al, 2007). It is therefore critical to maintain a high index of suspicion for congenital malaria in the evaluation of infants born to women from endemic countries. DEFINITION Congenital malaria is defined as malaria acquired by the fetus or newborn from the mother, either in utero or at parturition, but there exists no consensus on the application of this definition. Most commonly, congenital malaria is defined as the presence of Plasmodium spp. parasites CHAPTER 38 Toxoplasmosis, Syphilis, Malaria, and Tuberculosis TABLE 38-5 Recommended Treatment of the Newborn With Syphilis Penicillin Dosage Aqueous penicillin G 50,000 units/kg every 12 hours during first 7 days of life and every 8 hours thereafter for a total of 10 days or Penicillin G, procaine 50,000 units/kg/day, intramuscular, in a single dose every day for 10 days From Centers for Disease Controls and Prevention: Sexually transmitted diseases treatment guidelines, 2006. Available at www.cdc.gov/std/treatment/2006/congenital-syphilis.htm. Accessed November 17, 2010. TABLE 38-6 Follow-up after Treatment or Prophylaxis for ­Congenital Syphilis Patient Category Follow-up Procedures Infants with diagnosis of congenital syphilis RPR testing every 2-3 mo until negative or decreased fourfold; if RPR titer is stable or increasing after 6-12 mo after treatment, reevaluate and treat again; perform treponemal antibody test after age of 15 mo; if initial CSF analysis is abnormal or infant shows signs of CNS disease, repeat CSF evaluation every 6 mo until normal; with abnormal CSF not due to intercurrent illness on retesting, treat again; careful developmental evaluation, vision testing, and hearing testing are indicated Infants who received treatment in utero or at birth because of maternal syphilis RPR testing at birth and then every 3 mo until result is negative; treponemal antibody test after age of 15 mo Women who received treatment for syphilis during pregnancy RPR testing as often monthly until delivery, then every 6 mo until negative result obtained or titer decreased fourfold; repeated treatment any time there is a fourfold rise in RPR titer From Remington JS, Klein JC, Wilson CB, et al, editors: Infectious diseases of the fetus and newborn infant, ed 6, Philadelphia, 2006, Saunders, 2006, p 572. CSF, Cerebrospinal fluid; RPR, rapid plasmin reagin. in the peripheral blood in the first 7 days of life (Covell, 1950; Menendez and Mayor, 2007; Moran and Couper, 1999; Sotimehin et al, 2008; Uneke, 2007a). This definition is applicable in areas of high malaria transmission where, among older infants, it would be difficult to distinguish congenitally acquired from mosquito-acquired disease. Outside endemic areas, where postnatal transmission can be reasonably excluded, clinical onset of disease often does not occur until after the 1st week of life, and age-specific criteria are not useful for the diagnosis of congenital malaria. It is likely that because of the delay in clinical presentation, many cases of congenital malaria in endemic areas are misclassified as being acquired from mosquitoes. Alternate applications of the definition of congenital malaria include the detection of parasites in peripheral blood or in umbilical cord blood (Fischer, 1997) within the first 24 hours of life (Larkin and Thuma, 1991). It has been argued that congenital malaria should be distinguished 523 from cord blood parasitemia, which is not uncommon in some endemic areas and frequently clears without involving the peripheral circulation. Reinhardt (1978) found parasites in thick smears of umbilical cord blood in 22% of 19 infants born to women in the Ivory Coast, but the peripheral blood smears were negative for all of the infants. Similarly, 4% of 1009 infants born to Tanzanian women had parasites in cord blood, but parasitemia was detected on peripheral smear in only 2 of 11 infants (McGregor, 1984). Parasitemia detected shortly after birth may also resolve without evolving into clinically symptomatic disease. In a recent multicenter study, the correlation between maternal, placental, umbilical cord, and peripheral parasitemia was evaluated in 1875 mother-baby pairs in Nigeria. Thick and thin blood smears were obtained within 4 hours of birth, and smear-positive (cord or peripheral) neonates were retested on days 2, 3, and 7 of life, with treatment for those with symptoms or persistent parasitemia. The overall prevalence of congenital malaria was 5.1%, with parasitemia detected in 19% of infants born to mothers with peripheral parasitemia, 21% of those born to mothers with placental malaria, and 45% of those with positive cord smears. Spontaneous clearance of parasitemia occurred in 62% of infants before day 2, whereas 33% were symptomatic within 3 days of birth (Falade et al, 2007). It has now been well established that (1) maternal and placental parasitemia are important risk factors for congenital malaria, (2) a correlation between umbilical cord parasitemia and neonatal parasitemia exists, and (3) congenital malaria can occur in the absence of symptoms (Mukhtar et al, 2006; Sotimehin et al, 2008; Uneke, 2007a). However, the definition of congenital malaria has yet to be standardized to facilitate a better understanding of the epidemiology of this disease and the outcomes of prevention measures. EPIDEMIOLOGY Congenital malaria has traditionally been considered a rare consequence of malaria in pregnant women living in endemic areas. Covell (1950) published a review of cases of congenital malaria reported in areas with rates of placental parasitemia ranging from 5% to 74%, and rates of maternal malaria ranging from 1% o 68%. The prevalence of congenital malaria—defined as parasitemia detected in the first 7 days of life—was estimated to be 0.3% (16 of 5324 births) among immune mothers. Subsequent reports supported the observed low frequency of congenital malaria—defined as umbilical cord parasitemia or parasitemia in the first 24 hours of life—among indigenous populations (Bruce-Chwatt, 1952; Cannon, 1958; MacGregor, 1984; Williams and McFarlane, 1970). These observations have been cited repeatedly in the literature to support the notion that congenital malaria is an uncommon occurrence in endemic areas despite the high prevalence of maternal and placental malaria. More recent reports, however, suggest that the prevalence of congenital malaria was underestimated, with rates from endemic and nonendemic areas ranging from 8% to 33% (Desai et al, 2007; Menendez and Mayor, 2007). In Zambia during a season of heavy malaria transmission, incidence rates for congenital malaria ranged from 4% to 15% (Nyirjesy et al, 1993). A survey of seven sites in 524 PART IX Immunology and Infections Sub-Saharan Africa indicated a prevalence of congenital malaria—defined as umbilical cord blood parasitemia— ranging from 0% to 23% (Fischer, 1995). Congenital malaria, defined as neonatal parasitemia, was detected in 15.3% and 17.4% of neonates born in two sites in Nigeria (Mukhtar et al, 2006; Runsewe-Abiodun et al, 2006). The apparent increase in the frequency of congenital malaria has been attributed to increasing resistance of P. falciparum to antimalarial drugs resulting in increased maternal parasitemia, increased virulence of the parasite, and reduced transmission of antibody from mother to newborn because of malaria chemoprophylaxis administered to pregnant women. It is equally likely that congenital malaria was previously underreported because of the difficulty in differentiating congenital malaria from postnatally acquired malaria in neonates, or it was underdetected because of low parasite densities in newborns. It is evident that congenital malaria is not an uncommon occurrence in areas of high transmission. The variability in prevalence of disease may be attributed to differences in the definition of congenital malaria or the methods used for parasite detection. The inconsistency may also represent true environmental differences with differences in levels of maternal immunity. In contrast to the rarity of congenital malaria in indigenous populations, Covell (1950) found the prevalence of congenital malaria among nonimmune populations (i.e., Europeans residing in or visiting endemic areas) to be approximately 7%. It was postulated that congenital malaria is more common in infants born to these women because of lower levels of malaria-specific maternal antibodies being transmitted to the fetus. Most published reports of congenital malaria describe infants born in nonendemic countries whose mothers emigrated from malarious areas to areas free of malaria. These women presumably developed a recrudescence of disease because of waning immunity from lack of continued exposure In the United States, the occurrence of congenital malaria is well documented because the country has been free of indigenous disease since the 1950s. From 1950 to 1991, 49 cases of congenital malaria were reported in the literature (Hulbert, 1992), and additional cases were reported during the next 15 years (Balatbat et al, 1995; Baspinar et al, 2006; D’Avanzo et al, 2002; Gereige and Cimino, 1995; Starr and Wheeler, 1998; Viraraghavan and Jantausch, 2000). In an updated review, Lesko et al (2007) tabulated 81 cases reported to the CDC between 1966 and 2005. Almost all the cases were among infants whose mothers were foreign born, suggesting that congenital malaria is primarily a health problem of recent immigrants rather than of U.S.-born travelers to malaria-endemic countries. Forty-four women (54%) had emigrated from Asia, 27 (33%) from South or Central America, and 7 (9%) from Africa. Until 1979, one to two cases were reported annually (Malviya and Shurin, 1984). An abrupt rise to 16 cases around 1981 (Figure 38-5) correlated with an increase in the total number cases of malaria that occurred as a result of a large influx of refugees and immigrants from Southeast Asia, with 15 of the 16 infants being born to mothers from that region (Quinn et al, 1982). Congenital malaria occurs with all Plasmodium species. Among the 107 cases of congenital malaria reported by Covell (1950), mostly from Africa, 64% were infected with P. falciparum, 32% by P. vivax, and 2% by P. malariae. Although P. falciparum remains the predominant pathogen in Sub-Saharan Africa, P. vivax may account for a larger proportion of cases in Asia. Among 27 cases reported in Thailand between 1981 and 2005, 82% of were cause by P. vivax (Wiwanitkit, 2006). P. malariae is less frequently a causative agent, with fewer than 10 cases reported worldwide since 1950 (de Pontual et al, 2006). Concurrent infection with P. malariae and P. vivax has been documented (MacLeod et al, 1982). In the United States, the predominant Plasmodium species causing congenital malaria 12 Number of cases 10 8 6 4 2 19 66 19 68 19 70 19 72 19 74 19 76 19 78 19 80 19 82 19 84 19 86 19 88 19 90 19 92 19 94 19 96 19 98 20 00 20 02 20 04 0 Year FIGURE 38-5 Number of cases of congenital malaria reported to the National Malaria Surveillance System per year, 1966-2005. (From Lesko CR, Arguin PM, Newman RD: Congenital malaria in the United States: a review of cases from 1966 to 2005, Arch Pediatr Adolesc Med 161:1062-1067, 2007.) CHAPTER 38 Toxoplasmosis, Syphilis, Malaria, and Tuberculosis reflects the countries of origin of the mothers, as noted previously. In Hulbert’s (1992) review of 49 cases, 82% of infections were caused by P. vivax. In the updated review by Lesko et al (2007), the predominant infecting species remained P. vivax (81%), although all four species were represented. PREGNANCY AND MALARIA Malaria in pregnancy is an immense public health problem in the developing world, with substantial effects on maternal, fetal, and infant health. It is well established that both the frequency of disease and density of parasitemia are higher in pregnant women compared with nonpregnant women (Coll et al, 2008; Desai et al, 2007; Rogerson et al, 2007). In Sub-Saharan Africa, 25 million pregnant women are at risk for P. falciparum infection every year (Desai et al, 2007). Among 20 studies conducted between 1985 and 2000, the median prevalence of maternal malaria infection (defined as peripheral or placental infection) in all was 28% (Steketee et al, 2001). Thus, one in four pregnant women in areas of stable transmission in Africa have evidence of malaria infection at the time of delivery. The increased susceptibility likely represents a combination of immunologic and hormonal changes combined with the unique ability of infected erythrocytes to sequester in the placenta (Rogerson et al, 2007). The risk is indisputably greater for primigravidae, with primigravidae having a twofold to fourfold increased risk of placental malaria compared with multigravidae (Desai et al, 2007; McGregor, 1984; Uneke, 2008). The clinical features of P. falciparum malaria in a pregnant woman depend to a large degree on her immune status, which in turn is determined by her prior exposure to malaria. In pregnant women with little or no preexisting immunity, such as women from nonendemic countries or travelers to malarious areas, infection is associated with high risks of severe disease with significant maternal and perinatal mortality. In contrast, women residing in areas of stable malaria transmission usually have a high level of immunity to malaria. Infection may be frequently asymptomatic and therefore unsuspected or undetected, but it is associated with placental parasitization, with consequent effects on maternal and fetal outcomes. The most significant consequence of pregnancy-associated malaria is maternal anemia. It is estimated that in Sub-Saharan Africa between 200,000 and 500,000 pregnant women develop anemia as a result of malaria, and that up to 10,000 maternal anemia-related deaths are a consequence of P. falciparum parasitemia (Uneke, 2008). Malaria in pregnancy also has potentially devastating effects on the fetus and newborn, including spontaneous abortion, still birth, premature delivery, congenital infection, and neonatal death (Coll et al, 2008; Fischer, 2003). The most notable consequence is low birthweight (LBW), likely because of fetal growth restriction. In areas of high transmission in Africa, the risk of LBW approximately doubles if women have placental malaria, with the greatest effect in primigravidae (Desai et al, 2007). Pregnancy-associated malaria is believed to be responsible for 30% to 35% of LBW infants and for 75,000 to 200,000 infant deaths each year (Coll et al, 2008). 525 TRANSMISSION The timing and mechanism of transmission of Plasmodium spp. parasites from the mother to the fetus is not well understood. Postulated mechanisms include maternal transfusion into fetal circulation either during pregnancy or at delivery, or direct penetration of parasitized red blood cells through the chorionic villi or through premature separation of the placenta. In utero transmission is supported by the finding of malarial parasites in fetal tissues at autopsy (Mertz et al, 1981), by umbilical cord blood parasitemia (McGregor, 1984) and the onset of clinical signs of malaria within hours of birth (Brandenburg and Kenny, 1982; Covell, 1950; Gereige and Cimino, 1995). Antenatal transmission is also suggested by findings from Malawi, where approximately 50% of newborns with cord blood parasitemia were infected with parasites of a different genotype than their mothers at the time of delivery (Fischer, 2003). Alternatively, clinical findings in infants with congenital malaria may be delayed for several weeks after birth, suggesting infection at parturition. It is generally understood that the placenta acts as an effective barrier to prevent the transfer of malaria parasites from maternal into fetal circulation, supporting transmission at parturition as the most likely mechanism. Vertical transmission of malaria probably does not occur as a result of transplacental passage of exoerythrocytic parasites. More likely, transmission occurs by transfusion of parasitized maternal erythrocytes through a breach in the placental barrier that may occur either prematurely during pregnancy or during labor. Transmission of malaria by breastfeeding is not known to occur. The fate of the Plasmodium spp. parasite is unclear after it is transmitted to the fetus. As discussed previously, parasites detected in umbilical cord blood or shortly after birth may be cleared spontaneously, resulting in no disease manifestation. Alternatively, parasitemia may be maintained and proliferate until multiplication permits the development of clinical disease. CLINICAL PRESENTATION The clinical picture of overt congenital malaria is detailed in cases reported outside of endemic areas (Harvey et al, 1969; Hindi and Azimi, 1980; Hulbert, 1992; Lesko et al, 2007). The manifestation of disease, although occasionally noted within hours of birth (Brandenburg and Kenny, 1982; Gereige and Cimino, 1995), is typically delayed until the infant is several weeks old. In the classic review of 49 infants with congenital malaria reported in the United States between 1950 and1992, the mean age at onset of symptoms was 5.5 weeks, with 96% of infants presenting between 2 and 8 weeks of age (Hulbert, 1992). Among cases reported to the CDC from 1966 to 2005 (Lesko et al, 2007), the median age of symptom onset for 81 infants was 21.5 days for all species combined (Figure 38-6). Infants infected with P. malariae were significantly older at symptom onset (mean, 53 days) compared with those infected with P. vivax or P. falciparum. The prolonged interval between birth and onset of clinical manifestations may be explained by transmission late in pregnancy or at delivery, such that multiple erythrocytic 526 PART IX Immunology and Infections 18 16 14 Number of cases 12 10 8 6 4 2 0 <1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 Age at symptom onset, wk FIGURE 38-6 Age in weeks at symptom onset of infants with reported congenital malaria, United States, 1966-2005. (From Lesko CR, Arguin PM, Newman RD: Congenital malaria in the United States: a review of cases from 1966 to 2005, Arch Pediatr Adolesc Med 161:1062-1067, 2007.) life cycles are required to produce clinically evident disease. Alternatively, the delay may be attributed to the presence of transplacentally acquired maternal antimalarial antibodies. When such antibodies are present in sufficient concentrations, as in infants born to immune mothers, parasitic replication can be prevented or attenuated, and clinical signs can be mild, delayed, or even absent. The presence of a high concentration of fetal hemoglobin in newborns may also promote resistance to multiplication of parasites. Among infants born to mothers with low or nonexistent immunity, parasitic replication is more likely uninhibited, and clinical signs of malaria may supervene. Preterm infants, who do not benefit from passive immunity, can manifest clinical signs earlier than full-term infants. In a review of premature neonates with congenital malaria, 4 of 5 infants received a diagnosis in the 1st week of life (Ahmed et al, 1998), although the prompt medical evaluation afforded these infants may have facilitated earlier detection. The clinical features of congenital malaria are nonspecific and often resemble those of bacterial or viral sepsis and other congenital infections. Fever is almost uniformly present, although without the classic paroxysmal pattern described for malaria beyond the neonatal period. Hulbert (1992) noted fever in all 44 infants for whom clinical information was available. In the cases reported from 1966 to 2005, fever was reported in 70 of 81 cases (86%) (Lesko et al, 2007). Hepatomegaly and splenomegaly suggestive of a transplacentally acquired infection are found in a substantial portion of infants (Table 38-7). Anemia (often hemolytic), thrombocytopenia, and hyperbilirubinemia are the most commonly reported laboratory findings. Additional signs, symptoms, and laboratory findings are listed in Table 38-7. TABLE 38-7 Frequency of Symptoms, Signs, and Laboratory Findings Among 81 Infants With a Diagnosis of Congenital Malaria (United States, 1966-2005) Symptoms, Signs, and Laboratory Findings Fever Anemia Splenomegaly Hepatomegaly Thrombocytopenia Jaundice Irritability Anorexia Vomiting Cough Diarrhea Lethargy Hemolysis Pallor Hyperbilirubinemia Failure to thrive Seizures Dyspnea Purpura Tachycardia Monocytosis Infants, No. (%)* 70 (86) 28 (36) 25 (31) 16 (20) 12 (15) 11 (14) 8 (10) 8 (10) 8 (10) 6 (7) 3 (4) 3 (4) 3 (4) 3 (4) 2 (3) 2 (3) 2 (3) 1 (1) 1 (1) 1 (1) 1 (1) From Lesko CR, Arguin PM, Newman RD: Congenital malaria in the United States: a review of cases from 1966 to 2005, Arch Pediatr Adolesc Med 161:1062-1067, 2007. *Percentages do not total 100% because each case can have more than one symptom, sign, or laboratory finding. In endemic areas, the traditional belief has been that congenital malaria is rare and that when it occurs the infant is typically asymptomatic and develops no clinical features. The lack of symptoms has been attributed to transplacentally acquired antibodies from the mother as CHAPTER 38 Toxoplasmosis, Syphilis, Malaria, and Tuberculosis well as the protective effects of high levels of fetal hemoglobin. Depending on the region, spontaneous clearance of peripheral parasitemia has been documented in 87% to 100% of neonates (Lesko et al, 2007; Mukhtar et al, 2006). Larkin and Thuma (1991) found peripheral parasitemia within 24 hours of age in 19 of 51 newborns (65%), but only 7 had clinical signs of disease. Because all 19 newborns received antimalarial therapy, it is unknown how many would have manifested disease if untreated. More recently, Falade et al (2007) noted spontaneous clearance of parasitemia in 62% of 95 neonates before day 2 of life. Of the remaining infants, 34% were symptomatic within 3 days of birth, with fever and refusal to eat being the most common signs of disease. When active surveillance for malaria was conducted in newborns being evaluated for possible bacterial sepsis in Nigeria, 16 of 203 (8%) neonates had parasitemia, and 10 (5%) met the definition of congenital malaria (Ibhanesebhor, 1995). Predominant features of disease included fever, respiratory distress, anemia, and hepatomegaly. In another area in Nigeria, of 202 neonates less younger than 1 week who were admitted for evaluation of sepsis, 71 (35%) were diagnosed with congenital malaria (Ekanem et al, 2008). Fever was the most common symptom and was present in 93% of infants. Refusal to feed and jaundice were reported in approximately 33%. These observations suggest that, as in infants diagnosed outside endemic areas, the clinical presentation of congenital malaria in endemic areas does not differ significantly from bacterial sepsis. Because the clinical symptoms of congenital malaria may be indistinguishable from that of neonatal sepsis, it is suggested that screening for malaria be included as part of routine investigation of newborns with fever in areas of high malaria transmission (Ekanem et al, 2008; Runsewe-Abiodun et al, 2006). DIAGNOSIS Diagnostic tests for malaria include blood smears, rapid antigen detection tests (RDTs) and PCR. Definitive diagnosis of congenital malaria is based on the microscopic demonstration of parasites on stained thick and thin blood films. Thick blood smears test for the presence of parasites by concentration of red blood cells, whereas thin blood smears allow species identification and quantification of parasitemia. In cases of suggested congenital malaria, specimens for smears should be obtained from both the infant and the mother. If test results from the initial set of smears are negative, additional sets should be obtained every 12 to 24 hours; three sets are generally considered sufficient for diagnostic evaluation. Response to therapy may also be measured by clearance of parasitemia on blood films. RDTs are based on the immunochromatographic detection of parasite-specific antigens circulating in the bloodstream. Many RDTs are commercially available outside the United States, and, depending on the antigens targeted, the tests may detect only P. falciparum or all Plasmodium species. RDTs are simple to use, do not require specialized training or facilities, and offer a useful alternative to microscopy in situations where reliable microscopic diagnostics are not readily available. However, the tests have 527 demonstrated mixed results in multiple trials, and sensitivity remains a problem, especially at low parasite densities. Information regarding the sensitivity of these tests is limited for neonatal or congenital malaria. In the evaluation of an RDT for the diagnosis of congenital malaria in Nigeria, parasitemia was detected by microscopy in 21 of 192 newborns (10.9%) at 0 to 3 days of age, whereas the RDT result (OptiMAL [Flow Inc, Portland, Oregon]) was negative in all infants, including those diagnosed by microscopy. Whether the poor performance of the test was due to low-level parasitemia or suboptimal state of the parasites, the RDT was determined not to be useful for the diagnosis of congenital malaria (Sotimehin et al, 2007). It is recommended that microscopy be conducted in parallel with RDTs because low level parasitemia may not be detected or, if parasitemia is detected, for species identification and determination of parasite density. The major advantage of PCR for the diagnosis of malaria is its ability to detect low level parasitemia and identify parasites to a species level. Currently PCR is used mainly to confirm positive blood smears, particularly when the results of the smear are not definitive or there is a mixed species infection. PCR may detect DNA from circulating nonviable parasites after treatment, resulting in difficulty differentiating an active infection from a recently cleared infection. Although PCR is a highly sensitive alternative to microscopy, the infrastructure and expertise required preclude its use in malaria-endemic areas and many health care settings in the United States. As with malaria in general, the diagnosis of congenital malaria outside of endemic areas is often delayed because of nonspecific features and lack of clinical suspicion. Among the 81 cases reviewed by Lesko et al (2007), a median length of delay of 8.5 days was noted for 15% of the infants. Occasionally the diagnosis is made incidentally. In all four cases of congenital malaria reported by Quinn et al (1982), Plasmodium spp. parasites were noted by hematology technicians on routine smears performed for blood cell counts. Maternal history of recent travel to or emigration from an endemic area may suggest the diagnosis, but is often obscured by the lack of clinical or laboratory findings in the mother. Lesko et al (2007) found that, of the mothers for whom a history was available, 67% reported having fever during pregnancy, and 26% reported a diagnosis of malaria during pregnancy. Maternal blood films were performed after either symptomatic illness or malaria diagnosis in the infant. Overall only 42% of women had parasitemia detected, although it is not clear whether an adequate number of smears were conducted for each patient. As a result, lack of peripheral parasitemia in the mother of an infant with suspected congenital malaria does not exclude the diagnosis. Further confounding the early recognition of disease in the infant is the potentially prolonged lapse between malaria exposure in the mother and transmission of infection to the infant. P. vivax and P. ovale may remain dormant in the liver, especially if the infected individual did not receive therapy for the exoerythrocytic stage, which can cause a delayed relapse of malaria in travelers or immigrants. P. malariae can persist for 20 to 40 years before clinical symptoms or demonstrable parasitemia appear (D’Avanzo et al, 2002). Congenital malaria has been 528 PART IX Immunology and Infections reported in an infant whose mother lived in the United States for 5 years before delivery and had no signs or symptoms or malaria for more than 20 years (Harvey et al, 1969). In North Carolina, congenital P. malariae infection was reported in a 10-week-old infant who was born to a mother who had emigrated from the Democratic Republic of Congo 4 years before delivery (D’Avanzo et al, 2002). In a review by Lesko et al (2007), the median duration from the mother’s last exposure to delivery was 9.5 months. The time elapsed since exposure was longest for those with P. malariae infection, ranging from 2 to 12 years. Recognizing that congenital malaria is an exceptional occurrence in the United States, it is still important to include malaria in the differential diagnosis of fever in infants born to mothers who have been exposed to malaria, even if the exposure is remote and even if the woman is asymptomatic. Of 11 infants with congenital malaria in the United States born to women known to have parasitemia at or shortly after delivery, only five underwent testing by blood smears, and all five had negative test results at the time of delivery (Lesko et al, 2007). Data are insufficient to determine the overall risk of an infant developing congenital malaria when born to a woman at risk for parasitemia or identified with parasitemia at birth. Consequently, the evaluation of infants born outside endemic areas to women with epidemiologic risk factors for parasitemia should be individualized. In malaria endemic areas, and as a public health measure, it has been recommended that blood smears should be checked as part of the evaluation of neonates with fever born to mothers who have had fever within a few weeks of delivery (Uneke, 2007a). TREATMENT The management of malaria consists of supportive care and antimalarial therapy. Information regarding treatment of congenital malaria is limited, and recommended chemotherapy is similar to that of noncongenital infections. The treatment regimen is based on the infecting species, the possibility of drug resistance, and the severity of disease. For mild infections caused by P. vivax, P. ovale, and P. malariae or chloroquine-sensitive P. falciparum, chloroquine orally (10 mg base/kg initially followed by 5 mg base/kg 6, 24, and 48 hours later) is recommended. Treatment with primaquine is not necessary for congenitally acquired P. vivax or P. ovale infection because, like transfusion-associated malaria, congenital infection does not involve the exoerythrocytic phase. The treatment of congenital malaria due to chloroquine-resistant P. falciparum is poorly defined. In older children, three treatment options currently recommended are: (1) oral quinine plus either tetracycline, doxycycline, or clindamycin; (2) atovaquone-proguanil; or (3) mefloquine. For the treatment of congenital malaria, oral quinine sulfate and trimethoprim-sulfamethoxazole for 5 days was recommended by Quinn et al (1982), who used the regimen to treat a 1-month-old infant. Ahmed et al (1998) used a similar regimen for the treatment of an infant born at 28 weeks’ gestation to a mother from Zaire. Other regimens used successfully in neonates include oral quinine sulfate and pyrimethamine-sulfadoxine (Gereige and Cimino, 1995) and intravenous quinine hydrochloride followed by oral quinine (Airede, 1991). Intravenous quinine is no longer available in the United States. Because of the rarity of congenital malaria in the United States, the changing pattern of resistance, and the potential toxicity associated with drugs used for therapy, current treatment recommendations should be sought from the Malaria Branch of the CDC (www.cdc.gov/malaria). For health care professionals, assistance with management of malaria is also available 24 hours a day through the CDC Malaria Hotline (770-448-7788). Severe malaria occurs most commonly with P. falciparum infection and is characterized by one or more of the following: (1) parasitemia greater than 5% of red blood cells, (2) central nervous system or other end-organ involvement, (3) shock, (4) acidosis, (5) severe anemia, or (6) hypoglycemia. Management of severe malaria involves parenteral treatment in an intensive care setting. Until recently, the only parenteral therapy available in the United States was quinidine gluconate. Quinidine is more cardiotoxic than quinine and should be administered with continuous cardiac monitoring. Exchange transfusion may be warranted when parasitemia exceeds 10% or if there are complications at lower parasite densities. The efficacy of treatment should be monitored by examining blood smears (i.e., malaria smears) every 12 hours until negative for malaria parasites. Response to therapy with chloroquine for non-P. falciparum malaria is usually favorable (Brandenburg and Kenny, 1982; Dowell and Musher, 1991; Hindi and Azimi, 1980). It has been suggested that infants born to mothers with parasitemia at delivery should be treated presumptively for congenital malaria (Lesko et al, 2007). Data are insufficient to determine the risk of an infant developing congenital malaria when born to a mother with parasitemia. Although there is evidence from endemic areas that parasitemia detected at or shortly after delivery may clear spontaneously, the clinical relevance of this observation in nonendemic areas is unclear. It is recommended that physicians judge each case individually, considering factors such as access to medical care and reliability of follow up in deciding whether to treat infants presumptively. PROGNOSIS Malaria during pregnancy is likely an underappreciated risk factor for increased infant morbidity and mortality in endemic areas. In a review of studies published between 1985 and 2000, Steketee et al (2001) determined population-attributable risks for maternal malaria of 3% to 8% for infant mortality. It was estimated that 75,000 to 200,000 infant deaths annually are associated with malaria during pregnancy, although what proportion of these are related to congenital malaria is unknown. Outside endemic areas, the short-term outcome of congenital malaria has been favorable. Most infants respond rapidly to therapy with clearance of parasitemia. There were no reports of death or adverse outcomes in the 49 cases reported from 1950 to 1992 or in the 81 cases reported to the CDC from 1966 to 2005 (Hulbert, 1992; Lesko et al, 2007). It is unclear, however, whether the outcomes are due to an overall favorable prognosis or reporting bias. CHAPTER 38 Toxoplasmosis, Syphilis, Malaria, and Tuberculosis PREVENTION The prevention of congenital malaria is based on a pregnant woman’s avoidance of exposure and use of chemoprophylaxis. Malaria infection in pregnant women is more common and more severe than in nonpregnant women. Malaria also increases the risk for adverse pregnancy outcomes, including prematurity, abortion, and still birth. Nonimmune pregnant women are at the highest risk for these adverse outcomes. Based on these observations, the CDC advises women who are pregnant or likely to become pregnant to avoid travel to areas with malaria transmission. If such travel is unavoidable, consultation with an infectious disease or malaria expert is advised. The use of mosquito netting, mesh screens on windows, insecticides, and mosquito repellents can decrease potential exposure to malaria parasites. For pregnant women traveling to areas where there is no chloroquine-resistant P. falciparum malaria, prophylaxis with chloroquine is advised. The safety of chloroquine for the fetus when used at the recommended doses for malaria prophylaxis is well established (MacLeod et al, 1982). For travel to areas where chloroquine resistance has been reported, mefloquine is the only medication that is currently recommended for prophylaxis during pregnancy. Atovaquone-proguanil is not recommended during pregnancy because of insufficient information on potential adverse effects. Doxycycline is contraindicated because of adverse effects on the fetus caused by a related drug, tetracycline, which include dysplasia and discoloration of teeth and inhibition of bone growth. Health care professionals caring for women who cannot take the recommended antimalarial agent should contact the CDC Malaria Hotline (770-488-7788). Pregnant women originally from areas where malaria is endemic but who are now living in nonendemic areas may be only partially immune. When traveling to their countries of origin, they should be considered nonimmune and thus should receive the same recommendations as nonimmune women. The burden of malaria among pregnant women in endemic areas is well recognized, and prevention and control strategies for areas of high P. falciparum transmission are aimed at reducing maternal and infant mortality. The World Health Organization has proposed a three-pronged approach: (1) intermittent preventative treatment with an effective antimalarial agent at scheduled antenatal visits, (2) insecticide-treated nets, and (3) effective case management of clinical infection. A metaanalysis of more recent intervention trials suggests that successful prevention of these infections reduces the risk of severe maternal anemia by 38%, LBW by 43%, and perinatal mortality by 27% among paucigravidae women (Desai et al, 2007). Unfortunately, the full implementation of antenatal malaria prevention efforts is burdened by the challenges associated with health care delivery in the developing world. CONGENITAL TUBERCULOSIS Tuberculosis remains one of the deadliest communicable diseases worldwide. More than 2 billion people, equivalent to one third of the world’s population, are infected with Mycobacterium tuberculosis. Each year, an estimated 529 9 million new tuberculosis cases are identified, and 2 million people die from the disease. The greatest burden of disease is in developing countries, where tuberculosis remains a major public health threat. While case rates have declined steadily in the United States and Europe, the corresponding numbers have increased dramatically in the former Soviet Union and Sub-Saharan Africa, in part fueled by the epidemic of human immunodeficiency virus (HIV). Despite the prevalence of tuberculosis worldwide, congenital tuberculosis occurs rarely, with fewer than 400 cases reported in the English-language literature. The majority of reports describe infants born in low-burden countries to mothers who have emigrated from high-burden countries. Although some reports originate from countries where tuberculosis is endemic, it is likely that congenital tuberculosis is underrecognized and underreported in these areas, largely because of the nonspecific clinical features of disease and the limited diagnostic capability. In the United States, 10,000 to 15,000 cases of tuberculosis are reported annually. Although the overall rate of tuberculosis infection has been declining steadily since 1992, the proportion of cases in foreign-born populations has increased. More than 50% of cases reported in 2008 occurring among the foreign born. With increased global mobility and the epidemiologic trends of tuberculosis, it is likely that tuberculosis and congenital tuberculosis will continue to be observed in developed countries. The nonspecific features of congenital tuberculosis and the mortality associated with untreated disease underscore the importance of maintaining a high index of suspicion for tuberculosis in pregnant women and young infants. DEFINITION Transmission of M. tuberculosis from the mother to the neonate can occur in utero, intrapartum, or postpartum. Although congenital infection is classically considered the result of in utero infection of the fetus, the term congenital tuberculosis has historically referred to infection acquired either in utero or intrapartum. The infection can be transmitted by direct spread to the fetus from the placenta via the umbilical vein or by aspiration or ingestion of infected amniotic fluid, either in utero or intrapartum. Beitzke (1935) proposed diagnostic criteria to distinguish congenital tuberculosis from postnatally acquired tuberculosis. The criteria required that the infant have proven tuberculosis lesions and one of the following: (1) a primary hepatic complex as evidence of dissemination of the tubercle bacilli via the umbilical vein or (2) in the absence of a primary complex, the presence of tuberculous lesions in the first few days of life or the exclusion of postnatal infection by separation of the infant at birth from the mother and other potential sources of infection. These criteria were developed before the introduction of chemotherapy, when infant mortality with congenital tuberculosis was high and diagnosis was largely based on autopsy findings. It is difficult and no longer practical to apply Beitzke’s criteria. The demonstration of a primary hepatic complex with liver and regional node involvement requires an open surgical procedure; a percutaneous liver biopsy may demonstrate caseating granulomas, but the 530 PART IX Immunology and Infections primary complex will seldom be identified. Cantwell et al (1994) proposed revised criteria that are more applicable to current practice and improve diagnostic sensitivity. To meet the criteria, the infant must have proven tuberculous lesions and at least one of the following: (1) lesions in the first week of age, (2) a primary hepatic complex or caseating hepatic granulomas, (3) tuberculous infection of the placenta or maternal genital tract, or (4) exclusion or postnatal transmission by thorough investigation of contacts. Whereas the distinction between congenital tuberculosis and postnatally acquired disease may be relevant for academic or epidemiologic purposes, it does not affect the management, treatment, or prognosis of disease. EPIDEMIOLOGY From 1953 through 1984, the incidence of tuberculosis in the United States declined steadily, reaching a nadir of 9.4 cases per 100,000 population. From 1985 through 1992, there was a 20% increase in the total number of cases. The resurgence of disease was attributed to multiple factors, including the HIV epidemic, increased immigration, and a decline in public health funding for tuberculosis control. With the availability of antiretroviral therapy for HIV and fortification of public health measures, the epidemiologic trend was reversed. Since 1992, the case rates for tuberculosis have decreased annually to a case rate of 4.2 per 100,000 in 2008. Despite the decrease in the total burden of disease, tuberculosis continues to disproportionately affect the foreign-born and racial and ethnic minorities. In 2008, 59% of all cases of tuberculosis in the United States occurred in foreign-born persons (CDC, 2008). The current epidemiology of tuberculosis in pregnancy is not well delineated. With the resurgence of tuberculosis in the 1980s, the largest increase in the incidence of disease occurred in the 25- to 44-year age group, and the number of cases among women of childbearing age rose by 40% (Cantwell et al, 1994). In 1991, almost 40% of tuberculosis cases in minority women occurred in those between 15 and 35 years of age (Smith and Teele, 1995). These trends persist through the present time, thus placing women of childbearing age, especially those who are foreign born, and their newborns at continued risk. The risk of congenital tuberculosis in infants born to women with tuberculosis is unknown. Blackall (1969) reported only three cases among infants born to 100 mothers with tuberculosis. Ratner et al (1951) identified no cases among infants born to 260 mothers with the disease. In a study of 1369 infants separated at birth from their tuberculous mothers and placed in foster care, only 12 became tuberculin-positive during 4 years of observation, and in all 12 cases there was a source of infection in the postnatal environment (Smith and Teele, 1995). The low incidence of congenital tuberculosis is in part attributable to the high likelihood of infertility in women who have endometrial tuberculosis (Balasubramanian et al, 1999). However, in areas with high rates of tuberculosis transmission, neonates may be undiagnosed or underreported, and the incidence of congenital infection or vertical transmission remains unknown. Fewer than 400 cases of congenital tuberculosis have been reported in the literature, with the majority being in the prechemotherapy era (Laartz et al, 2002). Hageman et al (1980) reported two cases of congenital tuberculosis and reviewed another 24 reported in the English-language literature since the introduction of isoniazid (INH) in 1952. In the subsequent 30 years, more than 30 additional cases of neonates with congenital tuberculosis have been described (Abughali et al, 1994; Cantwell et al, 1994; Chen and Shih, 2004; Doudier et al, 2008; Grover et al, 2003; Hatzistamatiou et al, 2003; Laartz et al, 2002; Manji et al, 2001; Mazade et al, 2001; Nicolaidou et al, 2005; Pejham et al, 2002; Saitoh et al, 2001). The more recent reviews are cited in Table 38-8. Not surprisingly, the majority of infants were born to foreign-born mothers living in nonendemic areas. PATHOPHYSIOLOGY In pregnant women, tuberculous bacillemia can result in dissemination of infection to the placenta, the endometrium, or the genital tract. Genital tuberculosis that occurred before pregnancy may be asymptomatic, but often results in sterility, thus likely accounting for the low overall frequency of congenital infection. Vertical transmission may occur in one of three ways: (1) hematogenous spread from the infected placenta via the umbilical vein, (2) in utero aspiration or ingestion of amniotic fluid infected from the placenta or endometrium, or (3) ingestion of infected amniotic fluid or secretions from maternal genital lesions during delivery. The hematogenous route and in TABLE 38-8 Reviews of Cases of Congenital Tuberculosis Cases Reported in the English-Language Literature in the Era of Chemotherapy Years Cases Reported No. of Cases Hageman et al, 1980 1952-1980 26 NR 2 of 14 Respiratory distress, fever, hepatomegaly 46 (12) Cantwell et al, 1994 1980-1994 31 Median 24 (range 1 to 84) 0 of 9 Hepatosplenomegaly, respiratory distress, fever 38 (22) Abughali et al, 1994 1952-1994 58 NR 1 of 19? Respiratory distress, hepatomegaly, fever 45 (14) Laartz et al, 2002 1994-2002 16 Mean 17.4 (range, 1 to 60) 1 of 4 Respiratory distress, hepatomegaly, fever 20 Reference NR, Not reported; TST, tuberculin skin test. Age at Clinical Presentation (d) No. of Infants With Reactive TST Common Symptoms Mortality (%) (With Treatment) 531 CHAPTER 38 Toxoplasmosis, Syphilis, Malaria, and Tuberculosis utero aspiration each probably account for approximately half of the cases of congenital tuberculosis. Tuberculous bacilli have been demonstrated in the decidua, amnion, and chorionic villi of the placenta. It is unlikely that the fetus can be infected directly from the mother without the presence of a caseous lesion in the placenta, although massive involvement of the placenta does not always result in congenital tuberculosis. When a tubercle ruptures into the fetal circulation, bacilli in the umbilical vein can infect the liver, forming a primary focus with involvement of periportal lymph nodes. The bacilli also may pass through the liver and right ventricle and into the lung, or they can enter the left ventricle via the foramen ovale and pass into the systemic circulation. The organisms in the lung remain dormant until after birth, when oxygenation and circulation result in their multiplication and the subsequent development of a primary pulmonary focus. Alternatively, if the caseous lesion in the placenta ruptures directly into the uterine cavity and infects the amniotic fluid, the fetus can inhale or ingest the bacilli, leading to primary foci in the lung, intestine, or middle ear. Pathologic examination of tuberculosis in the fetus and newborn usually demonstrates disseminated disease, with the liver and lungs being principally involved. In Siegel’s study (1934) of 38 postmortem cases, the lungs were involved in 97%, the liver in 82%, and the spleen in 76% of the infants. Other sites described are the gastrointestinal tract, kidneys, adrenal glands, and skin (Agrawal and Rehman, 1995; Hageman et al, 1980; Sood et al, 2000). It is not always possible to determine whether sites represent multiple primary foci or are secondary to primary lesions in the lung or liver. The only lesion in the neonate that is unquestionably associated with congenital infection is a primary complex in the liver; all others may be acquired congenitally or postnatally. M. tuberculosis infection acquired in utero or perinatally may be indistinguishable from postpartum infection. Postnatal acquisition of M. tuberculosis acquired by airborne inoculation, either from the mother or another contagious adult in the infant’s environment, is the most common route of infection of the neonate. In addition, postnatal infection can occur from ingestion of infected breast milk from a mother with a tuberculous breast abscess. In the absence of a breast abscess, transmission of tuberculosis via breast milk has not been documented. The distinction between congenital tuberculosis and postnatally acquired disease may be important for academic or epidemiologic purposes, but the management, treatment, and prognosis of the disease processes are the same. CLINICAL PRESENTATION The clinical presentation of congenital tuberculosis is neither distinctive nor specific. Manifestations of disease resemble those of neonatal sepsis or other congenital infections. The affected infant is commonly born prematurely (Amodio et al, 2005; Davis et al, 1960; Foo et al, 1993; Katumba-Luenya et al, 2005; Premkumar et al, 2008; Wanjari et al, 2008). A retrospective cohort study from Mexico of infants born to 35 mothers with pregnancies complicated by tuberculosis demonstrated an approximately twofold risk of prematurity compared with newborns of mothers without tuberculosis (Figueroa-Damian and Arredondo-Garcia, 2001). Clinical signs may be evident shortly after birth, but typically do not appear until 2 to 4 weeks of age (see Table 38-8). Among the 29 cases reviewed by Cantwell et al (1994), the median age of presentation was 24 days. In an updated review of 16 cases reported since 1994, the mean age at presentation was slightly younger at 17.4 days (Laartz et al, 2002). Before the availability of INH, congenital tuberculosis was almost uniformly fatal. Notable signs included failure to thrive, jaundice, and central nervous system involvement. In the post-INH era, the most commonly described features of disease are respiratory distress, hepatomegaly with or without splenomegaly, and fever (Abughali et al, 1994; Cantwell et al, 1994; Hageman et al, 1980). Additional findings are listed in Table 38-9. Although it is important to evaluate for meningitis in an infant with suspected congenital tuberculosis, central nervous involvement occurs in fewer than 50% of cases (Hageman et al, 1980; Starke, 1997). Otitis media with aural discharge has been described as the presenting sign of congenital tuberculosis (Ng et al, 1996; ­Senbil et al, 1997), accompanied by regional lymphadenopathy (Gordon-Nesbitt and Rajan, 1973; Hatzistamatiou et al, 2003; Figure 38-7) or facial palsy (Pejham et al, 2002). It is presumed that the infection is due to the accumulation of infected amniotic fluid in the eustachian tube, either in utero or at birth. Cutaneous manifestations of congenital tuberculosis include papular, pustular, or vesicular lesions often surrounded by erythema (Al-Katawee et al, 2007; Azimi and Grossman, 1996; Hageman et al, 1980; Loeffler et al, 1996; Sood et al, 2000). Biopsy of the lesions is often confirmatory, demonstrating granulomatous inflammation and the presence of acid fast bacilli (AFB) on tissue stain (Hageman et al, 1980; Loeffler et al, 1996). A unique case of congenital tuberculosis involving the spine was recently reported in India (Grover et al, 2003). TABLE 38-9 Clinical Signs of Congenital Tuberculosis in 58 Infants Sign No. of Patients % of Patients Respiratory distress 44 76 Hepatomegaly with or without splenomegaly 38 65 Fever 33 57 Lymphadenopathy 19 33 Poor feeding 18 31 Lethargy, irritability 16 30 Abdominal distention 15 26 Failure to thrive 9 15 Ear discharge 9 15 Rash 5 9 Abnormal funduscopic findings 4 7 Jaundice 4 7 Seizure 3 5 Bloody diarrhea 3 5 Ascites 3 5 Adapted from Abughali N, Van Der Kuyp F, Annable W, et al: Congenital tuberculosis, Pediatr Infect Dis J 13:738-741, 1994. 532 PART IX Immunology and Infections FIGURE 38-7 Cervical and suboccipital tuberculous lymphadenitis in a 6-day-old premature infant. (From Hatzistamatiou Z, Kaleyias J, Ikonomidou U, et al: Congenital tuberculous lymphadenitis in a preterm infant in Greece, Acta Paediatr 92:392-394, 2003.) DIAGNOSIS The timely diagnosis of congenital tuberculosis requires a high index of suspicion. Clinical signs of disease in the neonate are nonspecific, and disease in the mother may be unsuspected, contributing to further delay in diagnosis. The diagnosis of congenital tuberculosis should be considered in any neonate with suspected infection who is unresponsive to conventional antimicrobial therapy. Evaluation for suspected disease should include a tuberculin skin test (TST), chest radiography, lumbar puncture, and mycobacterial culture of appropriate specimens. Biopsy specimens of affected tissue, either from the infant or the mother, and the placenta have been confirmatory in several case reports (Abughali et al, 1994; Cantwell et al, 1994; Chou, 2002; Hageman et al, 1980; Laartz et al, 2002; Loeffler et al, 1996). The TST is the most commonly used diagnostic test for tuberculosis. The test uses 5 tuberculin units of purified protein derivative injected intradermally on the volar surface of the forearm. The reaction is measured 48 to 72 hours later as millimeters of induration. As many as 10% to 40% of immunocompetent children with culture-proven tuberculosis do not initially react to a TST. Host factors such as young age and immunocompromised state can also decrease the sensitivity of the TST. Specificity may be compromised by cross reactivity with bacille-CalmetteGuérin vaccine or with environmental nontuberculous mycobacteria. The TST result is usually negative in neonates with congenital or perinatal tuberculosis, either secondary to immature cell-mediated immunity or because of overwhelming disease (see Table 38-8). Hageman et al (1980) found that only 2 of 14 infants who underwent skin testing had positive test results; on repeated testing, seven infants subsequently demonstrated positive tuberculin skin tests, the earliest being at 6 weeks of age, almost 4 weeks after presentation with clinical signs. Similarly, results of TSTs performed in 9 of 29 patients described by Cantwell et al (1994) were all negative, with results of subsequent testing being positive in 2 of the 9 infants. Among the 16 infants with congenital tuberculosis recently reviewed by Laartz (2002), three of four infants tested had nonreactive TST results. Recent advances in diagnostic tools for tuberculosis include whole-blood interferon γ release assays (IGRAs), which are immunologically based tests that measure interferon γ production from lymphocytes in response to antigens that are fairly specific to M. tuberculosis. The two types of assays currently available include the Quantiferon Gold (Cellestis, Valencia, California) and the enzymelinked immunosorbent spot assay. Advantages of these tests include lack of cross-reactivity with bacille-CalmetteGuérin vaccine and most nontuberculous mycobacteria. The correlation between IGRAs and TSTs is variable, and negative results do not definitively exclude tuberculosis. Published experience with the use of IGRAs in children is limited, and the negative predictive value of these tests in this population is unclear. Although IGRAs are endorsed by the CDC for use in circumstances in which a TST is indicated (Mazurek et al, 2010), the tests are not recommend for use in children younger than 5 years of age (American Academy of Pediatrics, 2009). Data on the use of IGRAs in newborns is limited to case reports, and these assays should not be substituted for TSTs in the evaluation of congenital tuberculosis. Given the frequency of respiratory distress in infants with congenital tuberculosis, it is not surprising that chest radiograph findings are frequently abnormal at first examination. Typically, a nonspecific parenchymal infiltrate is noted, although a miliary pattern representative of disseminated disease is occasionally observed (Agrawal and Rehman, 1995; Airede, 1990; Nemir and O’Hare, 1985; Pal and Ghosh, 2008; Polansky et al, 1978; Singh et al, 2006; Figure 38-8). Sixteen of 26 patients (62%) reviewed by Hageman et al (1980) had abnormal radiographic findings on presentation; seven had a miliary pattern and nine had nonspecific changes. Radiographic abnormalities developed subsequently in four additional infants. Among the 29 cases reviewed by Cantwell et al (1994), 23 infants (79%) had chest radiographic abnormalities, the majority being nonspecific infiltrates. Cavitation secondary to progressive pulmonary involvement has been reported (Cunningham et al, 1982). CT imaging of the chest may demonstrate adenopathy suggestive of tuberculosis or confirm miliary disease (Das et al, 2008; Singh et al, 2006). An ultrasound or CT image of the abdomen may reveal enlargement of the liver, spleen, or both, possibly with areas of abscesses (Amodio et al, 2005; Berk and Sylvester, 2004; Grover et al, 2003; Senbil et al, 1997). Congenital tuberculosis involving the spine was identified by radiographs and confirmed by magnetic resonance imaging in India (Grover et al, 2003) (Figure 38-9). Microbiologic confirmation of disease in the neonate should be sought using specimens from multiple sites. For infants and children unable to expectorate sputum, gastric aspirates are considered the specimens of choice. Additional sources for culture include endotracheal aspirate, bronchial washing, middle-ear discharge, and lymph node tissue. CSF should be analyzed and cultured, although isolation of M. tuberculosis from CSF is uncommon (Abughali et al, 1994; Hageman et al, 1980). Traditionally the detection of mycobacterial organisms by smear or culture has been considered difficult, because children have paucibacillary disease relative to adults. With three morning gastric aspirates collected appropriately in hospitalized CHAPTER 38 Toxoplasmosis, Syphilis, Malaria, and Tuberculosis FIGURE 38-8 Miliary tuberculosis in a neonate with congenital tuberculosis. (From Singh M, Kothur K, Dayal D, et al: Perinatal tuberculosis a case series, J Trop Pediatr 53:135-138, 2006.) FIGURE 38-9 Saggital magnetic resonance image on the left demonstrates destruction of T9 to T11 vertebral bodies with the collapse of T10, leading to a kyphotic deformity causing cord compression. A large prevertebral collection is also seen. The image on the right demonstrates destruction with collapse of L5 and S1 vertebral bodies. (From Grover SB, Pati NK, Mehta R, et al: Congenital spine tuberculosis: early diagnosis by imaging studies, Am J Perinatol 20:150, 2003.) 533 children with a clinical diagnosis of tuberculosis, only 40% of children had positive cultures (Starke and TaylorWatts, 1989). In comparison, cultures of aspirates from infants evaluated at the same institution had a 75% yield (Starke and Taylor-Watts, 1989; Vallejo et al, 1994). The improved diagnostic yield in infants likely reflects more widely disseminated and progressive disease, with higher bacillary loads. Hageman et al (1980) found positive cultures of M. tuberculosis in 10 of 12 gastric aspirates, 3 of 3 liver biopsy specimens, 3 of 3 lymph node specimens, and 2 of 4 bone marrow biopsy specimens. Among the 31 cases reviewed by Cantwell et al (1994), noninvasive procedures and biopsy were useful for the diagnosis of congenital malaria in the majority infants (Table 38-10). More recent reports confirm the high yield of cultures from a variety of specimens in neonates (Berk and Sylvester, 2004; Chou, 2002; Mazade et al, 2001; Premkumar et al, 2008; Wanjari et al, 2008). Histologic examination of tissue may suggest the diagnosis before culture results are available. Biopsy of skin lesions, lymph nodes, and the liver have suggested the diagnosis in several cases by demonstration of granulomas or acid-fast bacilli on staining before culture results are available (Berk and Sylvester, 2004; Davis et al, 1960; Hageman et al, 1980). Whereas M. tuberculosis can require 7 to 21 days for growth by standard culture technique, results from techniques such as PCR may be available within 48 hours. A comparison of PCR, AFB smear, and culture with clinical diagnosis in children found a sensitivity of 60% and a specificity of 97% (Smith, 2002). Although PCR has been useful for diagnosing congenital tuberculosis in a few case reports, it is not sensitive enough to preclude obtaining specimens for culture. Isolation of M. tuberculosis is still important to determine susceptibilities and to optimize treatment. The mother of a newborn in whom congenital tuberculosis is suspected is often asymptomatic or has subclinical disease. In the series of congenitally infected infants reported by Hageman et al (1980), the majority of mothers did not have a diagnosis until after the disease became apparent in their infants. Cantwell et al (1994) found that 50% of the mothers of infected infants were not ill at the time their newborns exhibited clinical signs of disease. Evaluation of the mother should include a TST, a chest radiograph, and, if the radiograph is consistent with tuberculosis disease, collection of sputum for microbiological confirmation. It is not unusual for the mother to have extrapulmonary disease such as meningitis or peritonitis (Laartz et al, 2002; Naouri, 2005), and evaluation may need to be extended to identify such sites if pulmonary disease is not discovered. In mothers with no clinical evidence of disease, endometritis should be considered. Pathologic examination and culture of the placenta (if available) or endometrial biopsy can confirm the diagnosis of genital transmission (Asensi et al, 1990; Balasubramanian et al, 1999; Cantwell et al, 1994; Cooper et al, 1985; Niles, 1982; Surve et al, 2006). In several case reports, the diagnosis of maternal tuberculosis was solely and ultimately made by endometrial biopsy and culture (Pejham et al, 2002). In addition, culture of amniotic fluid should be performed. All mothers with tuberculosis should be tested for HIV infection and, if the mother is seropositive, the infant should be evaluated for perinatally acquired HIV infection. 534 PART IX Immunology and Infections TREATMENT AND MANAGEMENT The successful management of congenital tuberculosis depends on early recognition and treatment of disease. In suspected cases, treatment should not be delayed until results of cultures or other diagnostic tests are available. TABLE 38-10 Results of Diagnostic Procedures Performed on 29 Infants With Congenital Tuberculosis Reported from 1980 to 1994 Type of Specimen Acid-Fast Smear* Mycobacterial Culture Smear or Culture Gastric aspirate 8/9 8/9 9/11 Endotracheal aspirate 7/7 7/7 7/7 Ear discharge 2/2 1/1 2/2 Cerebrospinal fluid 1/2 1/2 1/2 Urine 0/2 0/2 0/2 Peritoneal fluid 1/1 1/1 1/1 Bronchoscopic specimen 1/1 1/1 1/1 14/19 11/12 16/21 Lymph node 7/8 6/6 7/8† Liver 4/6 1/2 4/6† Skin 1/3 1/1 1/3 Lung 1/1 1/1 2/2 Bone marrow — 1/1 1/1 Ear 1/1 1/1 1/1 Biopsy specimen Adapted from Cantwell MF, Shehab ZM, Costello AM, et al: Brief report: congenital tuberculosis, N Engl J Med 330:1051, 1994. *Results expressed as number of positive results per number of patients tested. †All biopsy specimens of lymph node and liver that tested negative on smear and culture showed histopathologic changes consistent with tuberculosis (i.e., giant cell transformation of granulomas, with or without caseation). Multiple drug therapy for an extended duration has long been recognized as the standard of care for tuberculosis. Because of the rarity of the condition, clinical trials have not been conducted to establish the optimal treatment regimen for congenital tuberculosis. It is assumed that the regimens used for older infants and children are safe and effective for the treatment of neonates with congenital tuberculosis. Consultation with a pediatric infectious disease specialist or tuberculosis expert is advised. Until susceptibility results are known, infants with proven or suspected tuberculosis should be treated with a four-drug regimen consisting of INH rifampin (RIF), pyrazinamide (PZA), and ethambutol (Table 38-11). Some experts would recommend administration of three drugs (INH, RIF, PZA) if antimicrobial resistance is not suspected in the mother, because either she or the source case are known to have a susceptible strain or she has no risk factors for resistant M. tuberculosis. The adjunctive use of corticosteroids is recommended for the treatment of tuberculosis meningitis based on decreased mortality and morbidity demonstrated in adults and children (Girgis et al, 1991). Supplementation with pyridoxine, although not routinely recommended for otherwise healthy older children, should be provided to breastfeeding infants receiving INH. If the M. tuberculosis isolate is determined to be susceptible, the regimen can be narrowed to three drugs (INH, RIF, PZA) for the first 2 months of initial treatment, and subsequently to 2 drugs (INH, RIF) to complete the continuation phase of treatment. Once the infant is discharged to home, directly observed therapy is recommended to ensure adherence and to prevent relapse. The optimal duration of treatment for infants with congenital tuberculosis is unknown. The typical duration of treatment for susceptible M. tuberculosis is 6 months for pulmonary disease, pulmonary disease with hilar adenopathy, or hilar adenopathy alone (American Academy of TABLE 38-11 Commonly Used Drugs for Treatment of Tuberculosis in Infants, Children, and Adolescents Daily Dose (mg/kg) Twice per Week Dose (mg/kg) Tablets (100,400 mg) 20-25 50 Scored tablets (100, 300 mg) Syrup 10 mg/mL 10-15† 20-30 Pyrazinamide Scored tablets (500 mg) 30-40 50 Rifampin Capsules (150, 300 mg) Syrup formulated capsules 10-20 10-20 Drugs Dose Forms Ethambutol Isoniazid Maximum Dose Adverse Reactions 2.5 g Optic neuritis (usually reversible), decreased red-green color discrimination, gastrointestinal tract disturbances, hypersensitivity 300 mg daily Mild hepatic enzyme elevation, hepatitis,† peripheral neuritis, hypersensitivity Diarrhea and gastric irritation caused by vehicle in the syrup 900 mg twice per week 2g Hepatotoxic effects, hyperuricemia, arthralgia, gastrointestinal tract upset 600 mg Orange discoloration of secretions or urine, staining of contact lenses, vomiting, hepatitis, ­influenza-like reaction, thrombocytopenia, pruritus; oral contraceptives may be ineffective From American Academy of Pediatrics Committee on Infectious Diseases: Section 3. Summaries of Infectious Diseases. Tuberculosis. In Pickering LK, editor: 2009 red book: report of the committee on infectious diseases, ed 28, Elk Grove Village, Ill, 2009, American Academy of Pediatrics, p 688. †When isoniazid in a dose exceeding 10 mg/kg/d is used in combination with rifampin, the incidence of hepatotoxic effects may be increased. CHAPTER 38 Toxoplasmosis, Syphilis, Malaria, and Tuberculosis Pediatrics, 2009). For extrapulmonary disease, the duration is extended to 9 to 12 months, and for drug-resistant M. tuberculosis the duration may be extended even further to prevent failure or relapse. Most experts would treat infants with congenital tuberculosis for 9 to 12 months because of the decreased immunocompetence of neonates (Starke, 1997). Although there is a fair amount of data to support the safety of INH, data on the safety and pharmacokinetics of other agents are limited. Careful monitoring for signs and symptoms of hepatitis and other adverse effects of drug therapy is recommended. Routine determination of serum transaminases in children is indicated with severe tuberculosis (e.g., military or meningitis), for those with concurrent liver or biliary disease, or for those receiving other potentially hepatotoxic drugs. Routine monitoring of liver function also should be considered for neonates with congenital tuberculosis given the paucity of data on adverse effects of anti-tuberculosis agents in this age group (Patel et al, 2008). The risks of optic neuritis with ethambutol should be considered when this agent is used, and vision should be monitored periodically. The management of infants born to mothers who have latent tuberculosis infection (LTBI) or tuberculosis disease is outlined in Figure 38-10. Recommendations are based on the categorization of infection in the mother and the potential risk of transmission of tuberculosis to the infant (American Academy of Pediatrics, 2009). Infants born to mothers with potentially contagious tuberculosis should be evaluated for congenital tuberculosis. Separation of the infant and mother is necessary only in cases in which the mother is highly infectious at the time of delivery. The mother with latent tuberculosis infection is not contagious. To prevent reactivation disease in the mother and subsequent exposure of the infant, the mother should receive treatment with INH for LTBI. In addition, latent tuberculosis infection in the mother may be a marker for contagious tuberculosis within the household, and it is recommended that all household members and close contacts of the mother be evaluated for tuberculosis. Breastfeeding is not contraindicated in women with LTBI. The breast milk of a woman with tuberculosis does not contain tubercle bacilli. For women with tuberculosis who are potentially infectious and separated from the newborn, breast milk may be manually expressed and fed to the infant. Once the mother is noninfectious or the infant is receiving therapy, breastfeeding can be resumed. The exception, however, is the mother with an active tuberculous breast lesion. In this situation the breast milk may be pumped and discarded until resolution of the lesion (Efferen, 2007). PROGNOSIS The prognosis for congenital tuberculosis was dismal in the prechemotherapy era, the diagnosis often being only made at autopsy. Although the survival rate subsequently improved, mortality remained approximately 50% secondary to delayed diagnosis. In a review of 26 cases reported between 1952 and 1980, 12 (46%) patients died, 9 of whom were untreated but with a diagnosis made at autopsy. The subsequent reviews demonstrate a decrease in case fatality 535 (see Table 38-8), with earlier diagnosis and treatment. Timely diagnosis and initiation of anti-tuberculosis therapy are critical for a favorable outcome. PREVENTION Prevention of congenital tuberculosis requires the treatment and prevention of disease in women of childbearing age. Risk factors for acquiring tuberculosis infection or progressing to disease should be assessed at prenatal visits, and women identified as being high risk for LTBI or progression to disease should undergo tuberculin skin testing as recommended by the CDC (Centers for Disease Control, 2000). Women who are TST positive should undergo evaluation for active disease. The treatment of active tuberculosis during pregnancy is considered standard, and early treatment has been shown to improve maternal and neonatal outcome (Figueroa-Damian and ArredonondoGarcia, 1998). While not without potential complications, treatment during pregnancy is less of a hazard to a pregnant woman and her fetus than tuberculosis itself. The treatment of LTBI during pregnancy is somewhat more controversial. Some experts advocate treatment during pregnancy, whereas others support a delay in therapy until weeks to months after delivery. Although there is no demonstrated teratogenic potential for the use of INH, there is concern that pregnant and postpartum women are more vulnerable to INH-related hepatotoxicity (Franks et al, 1989). Most children with pulmonary tuberculosis, especially those younger than 10 years, have paucibacillary disease and often have little to no cough. Isolation of the hospitalized pediatric patient is directed at accompanying adult contacts who may be source cases and potentially contagious. Visitation of the hospitalized pediatric patient should be restricted to adults in whom contagious tuberculosis has been excluded. Hospitalized children with negative sputum AFB smears (if obtained) require standard precautions, assuming that contagious tuberculosis has been excluded in the visitors. Airborne precautions are recommended for the following pediatric patients: (1) children and adolescents with adult-type cavitary disease, (2) extensive pulmonary infection, (3) those with smears positive for AFB, and (4) congenitally infected neonates undergoing endotracheal intubation (American Academy of Pediatrics, 2009). Compared with older children, neonates likely have a higher concentration of bacilli in their sputum. As noted previously, AFB smears on tracheal aspirates and other specimens are frequently positive in this population compared with older children. Transmission of tuberculosis from congenitally infected neonates to health care workers and other hospitalized infants has been reported and is likely related to aerosolization of bacilli during respiratory manipulation (Crockett et al, 2004; Laartz et al, 2002; Lee et al, 1998; Mouchet et al, 2004). Neonates suspected of having congenital tuberculosis should be placed in respiratory isolation if intubated or if undergoing any procedure with the potential for aerosolization of infected sputum. Exposed infants, visitors, and health care workers should undergo evaluation for tuberculosis infection or disease. 536 PART IX Immunology and Infections Mother: Tuberculin skin test positive* Symptomatic infant † Asymptomatic infant Maternal chest x-ray normal; no active disease Maternal chest x-ray abnormal † Mother with clinical or radiographic evidence of contagious TB † Maternal evaluation consistent with TB? • TST, chest x-ray • Lumbar puncture • AFB culture: gastric aspirate, endotracheal aspirate, CSF • Head CT scan or MRI • Multi-drug treatment (see text) • Airborne isolation • Consult infectious disease specialist Yes No ‡ Maternal treatment for TB 2 wk, sputum AFB negative, and strain not multi-drug resistant Maternal treatment none or 2 wk, sputum AFB positive, or strain is multi-drug resistant No infant evaluation or therapy required • Separate infant from mother until she is non-contagious (treatment 2wk) or infant on appropriate TB drug(s) • Evaluate infant for congenital tuberculosis: • TST; chest x-ray • AFB culture: gastric aspirate, endotracheal aspirate • Consider lumbar puncture and AFB culture of CSF • Consult infectious disease specialist; notify local health department Evaluation consistent with congenital tuberculosis? No • INH for 3–4 months • If maternal TB strain is multi-drug resistant, consider BCG vaccine • Follow-up tuberculin skin test at 3–4 months: • Negative test: stop INH • Positive test: • Reassess for TB as in symptomatic infant • If no other evidence of TB, INH for 9 months Yes • Lumbar puncture if not previously done • Head CT scan or MRI • Multi-drug treatment (see text) FIGURE 38-10 Management of infants born to mothers with a positive tuberculin skin test (TST) result. BCG, Bacille Calmette-Guérin; CSF, cerebrospinal fluid; CT, computed tomography; INH, isoniazid; MRI, magnetic resonance imaging. (Adapted from American Academy of Pediatrics: Tuberculosis. In Pickering LK, editor: 2003 Red book: report of the committee on infectious diseases, ed 26, Elk Grove Village, Ill, 2003, American Academy of Pediatrics, p 642.) *Household contacts should have a TST and further evaluation for contagious tuberculosis (TB). Consult local health department. The mother should receive treatment for latent tuberculosis infections. All persons with TB should be tested for human immunodeficiency virus (HIV) infection. †Acid-fast bacillus (AFB) culture of amniotic flid and placenta, if available; placenta for histopathologic examination. ‡Includes mother with chest radiographic findings consistent with old, healed TB. CHAPTER 38 Toxoplasmosis, Syphilis, Malaria, and Tuberculosis SUGGESTED READINGS Toxoplasmosis Beaman MH, Luft BJ, Remington JS: Prophylaxis for toxoplasmosis in AIDS, Ann Intern Med 117:163-164, 1992. Boyer K: Diagnostic testing for congenital toxoplasmosis, Pediatr Infect Dis J 20:59-60, 2001. Centers for Disease Control and Prevention: CDC recommendations regarding selected conditions affecting women’s health: Preventing congenital toxoplasmosis, MMWR Morb Mortal Wkly Rep 49(RR-2):59-68, 2000. Feldman DM, Timms D, Borgida AF: Toxoplasmosis, parvovirus and cytomegalovirus in pregnancy, Clin Lab Med 30:709-720, 2010. Jones JL, Dargelas V, Roberts J, et al: Risk factors for Toxoplasma gondii infection in the United States, Clin Infect Dis 49:878-884, 2009. Kasper DC, Sadeghi K, Prusa AR, et al: Quantitative real-time polymerase chain reaction for the accurate detection of Toxoplasma gondii in amniotic fluid, Diagn Microbiol Infect Dis 63:10-15, 2009. Koppe JG, Loewer-Sieger DH, DeRoever-Bonnet H: Result of 20‑year follow-up of congenital toxoplasmosis, Lancet 1:254-256, 1986. Masur H, Kaplan JE, Holmes KK, et al: Guidelines for preventing opportunistic infections among HIV-infected persons: 2002 recommendations of the U.S. Public Health Service and the Infectious Diseases Society of America, MMWR Morb Mortal Wkly Rep 51(RR-8):5, 2002. McLeod R, Kieffer F, Sautter M, et al: Why prevent, diagnose and treat congenital toxoplasmosis?, Mem Inst Oswaldo Cruz 104:320-344, 2009. Rabilloud M, Wallon M, Peyron F: In utero and at birth diagnosis of congenital toxoplasmosis: use of likelihood ratios for clinical management, Pediatr Infect Dis J 29:421-425, 2010. Syphilis American Academy of Pediatrics Committee on Infectious Diseases: Syphilis, Red Book Online: 2009. Available at http://aapredbook.aappublications.org/cgi/ content/full/2009/1/3.129. Accessed November 17, 2010. Beeram MR, Chopde N, Dawood Y, et al: Lumbar puncture in the evaluation of possible asymptomatic congenital syphilis in neonates, J Pediatr 128:125-129, 1996. Centers for Disease Control and Prevention: Sexually Transmitted Diseases Treatment Guidelines, 2010. MMWR 59:1-110, 2010. Accessed 537 Centers for Disease Control and Prevention: Congenital syphilis: United States, 2003-2008, MMWR 59:413-417, 2010. Herremans T, Kortbeek L, Notermans DW: A review of diagnostic tests for congenital syphilis in newborns, Eur J Clin Microbiol Infect Dis 29:495-501, 2010. Ingall J, Sanchez P, Baker C: Syphilis, In Remington JS, Klein JO, Baker C, et al: Infectious diseases of the fetus and newborn infant, ed 6, Philadelphia, 2006, WB Saunders, pp 545-580. World Health Organization Department of Reproductive Health and Research: The global elimination of congenital syphilis: rationale and strategy for action: 2007. Available at www.who.int/reproductivehealth/publications/rtis/ 9789241595858/en/index.html. Accessed November 17, 2010. Malaria Desai M, ter Kuile FO, Nosten F, et al: Epidemiology and burden of malaria in pregnancy, Lancet Infect Dis 7:93-104, 2007. Hulbert TV: Congenital malaria in the United States: report of a case and review, Clin Infect Dis 14:922-926, 1992. Lesko CR, Arguin PM, Neman RD: Congenital malaria in the United States: a review of cases from 1966 to 2005, Arch Pediatr Adolesc Med 161:1062-1067, 2007. Menendez C, Mayor A: Congenital malaria: the least known consequence of malaria in pregnancy, Semin Fetal Neonatal Med 12:207-213, 2007. Tuberculosis Abughali N, Van Der Kuyp F, Annable W: Kumar ML: Congenital tuberculosis, Pediatr Infect Dis J 13:738-741, 1994. Cantwell MF, Shehab ZM, Costello AM, et al: Brief report: congenital tuberculosis, N Engl J Med 330:1051-1054, 1994. Hageman J, Shulman S, Schreiber M, et al: Congenital tuberculosis: critical reappraisal of clinical findings and diagnostic procedures, Pediatrics 66:980-984, 1980. Laartz BW, Narvarte HJ, Hold D, et al: Congenital tuberculosis and management of exposures in a neonatal intensive care unit, Infect Control Hosp Epidemiol 23:573-579, 2002. Complete references and supplemental color images used in this text can be found online at www.expertconsult.com C H A P T E R 39 Neonatal Bacterial Sepsis Patricia Ferrieri and Linda D. Wallen NEONATAL BACTERIAL SEPSIS Throughout pregnancy, the fetus is protected from bacterial and viral infections by the chorioamniotic membranes, the placenta, and various antibacterial factors, that are poorly described in amniotic fluid. It is thought that some subclinical infections of the fetus, amniotic fluid, membranes, or placenta may contribute to the onset of preterm labor and the delivery of preterm infants. There are several mechanisms by which bacteria can reach the fetus or newborn and initiate infection. Maternal blood stream infections, caused by bacteria such as Listeria monocytogenes and Mycobacterium tuberculosis, can reach the fetus and cause infection. Bacteria such as group B streptococcus (GBS) can be acquired from the vagina, cervix, or fecal contamination of the birth canal through either ruptured or intact membranes, leading to amnionitis, intrauterine pneumonitis, and premature delivery (Ancona et al, 1980; Ferrieri, 1990; Larsen and Sever, 2008; Payne et al, 1988). Finally, infection can occur via aspiration of birth canal contents or colonization of mucosal surfaces during passage through the birth canal, leading to pneumonia, followed by bacteremia and sepsis after 1 day or later; this may be the mode by which Neisseria gonorrhoeae, Escherichia coli, and GBS are acquired. Early-onset bacterial sepsis remains a major cause of neonatal morbidity and mortality, although the sepsisassociated death rates per 100,000 live births have declined significantly from 2001-2011. Much of this decline in mortality is because of the introduction of intrapartum antibiotic prophylaxis in pregnant women during labor and delivery (Centers for Disease Control and Prevention, 2007, 2009; Schrag et al, 2002; Schrag and Stoll, 2006). Mortality rates in infected premature infants and very immature infants are significantly higher than in term infants. Major improvements in neonatal intensive care and early identification and recognition of infected infants have all contributed to reduced mortality rates in the newborn period. PATHOGENESIS OF EARLY-ONSET NEONATAL BACTERIAL INFECTIONS There are multiple portals through which bacteria can enter and infect the newborn. The primary portals of entry appear to be the respiratory tract, as suggested by the high frequency of acute respiratory distress and pneumonia, which occurs in infants with early-onset disease. Acquisition via the placenta is suggested in some instances by the presentation of high-grade bacteremia and severe sepsis clinically apparent at the time of birth in the presence of intact membranes in infants born via cesarean section. The primary maternal event in this sequence, leading to infection of the fetus and newborn infant, is colonization of the maternal genital tract with organisms such as the GBS. 538 Bacteria that reside in the cervix, vagina, or rectum can ascend into the amniotic cavity through intact or ruptured membranes and lead to chorioamnionitis. Bacteria can initially spread into the choriodecidual space and can occasionally cross intact chorioamniotic membranes. Although organisms recovered from the amniotic sac in the mother are usually polymicrobial and include such organisms as the GBS, group D enterococcus, aerobic gram-negative bacteria, and anaerobes such as Bacteroides spp. (Gibbs et al, 1980), a single organism causing bacterial sepsis is the rule in de novo sepsis of the newborn infant. Genital microplasmas are at times recovered from women as well as Chlamydia spp., but the precise pathogenic role is unclear (Krohn et al, 1995; Pankuch et al, 1984). Ureaplasma spp., and Chlamydia spp. can be isolated from infants’ respiratory tract after birth; these two plus Mycoplasma hominis can be recovered from the respiratory tract after birth, but they are not associated with sepsis syndrome. Although many microorganisms recovered from the amniotic cavity are thought to induce spontaneous preterm labor, and possibly premature rupture of membranes, the exact mechanisms by which this may occur are debatable. Clinical or subclinical chorioamnionitis can incite a marked inflammatory response with the release of cytokines that can contribute to the onset of preterm labor and premature rupture of membranes. Other risk factors for clinical intraamniotic infection include young maternal age, prolonged labor, prolonged rupture of membranes (≥18 hours), internal scalp fetal monitoring, the presence of urinary tract infections, and a history of bacterial vaginosis (Newton et al, 1989; Soper et al, 1989). Despite inherent antibacterial properties in amniotic fluid, these may not be sufficient to overcome a large bacterial inoculum, because of rapid multiplication of bacteria during a prolonged labor or the absence of type-specific maternal antibodies for various pathogens (Ferrieri, 1990). Infants who immediately display signs of respiratory distress and after birth undoubtedly have onset of infection before or during labor and delivery. Particularly with hypoxia in utero, the infant may gasp and inhale contaminated amniotic fluid, leading to pneumonia, blood stream infection, sepsis, and a severe systemic response syndrome. Infants who display such signs at birth or within a short time after birth have the highest mortality rates. Infants who have an initial asymptomatic period after birth may display symptoms gradually as the organisms multiply in the lung and in the blood. An example of another invasive site of entry is the scalp lesion created by a monitoring device, which becomes contaminated in the setting of amniotic fluid infected with the GBS. An overarching mechanism for continued bacteremia is the absence of sufficient local and systemic host defenses, such as adequate complement levels or type-specific immunity against the invading microorganism (Ferrieri, 1990). CHAPTER 39 Neonatal Bacterial Sepsis The inflammatory cascade is initiated by activation of macrophages by bacterial cell wall constituents, toxins, or enzymes. A number of proinflammatory cytokines can be released, such as interleukin (IL) 6, IL-8, and tumor necrosis factor-α (TNF-α). These cytokines can alter vascular permeability and vascular tone, decrease myocardial contractility, activate clotting systems, increase pulmonary vascular resistance, and activate other phagocytic cells, such as polymorphonuclear leukocytes (PMNs). Ideally, proinflammatory and antiinflammatory cytokines would be balanced; however, this is usually not the case and the bacteria persist with subsequent consequences. It is common in newborn infants and, particularly in preterm infants, to have dissemination of bacteria to other organs such as the meninges, kidneys, and bone. EPIDEMIOLOGY OF EARLY-ONSET BACTERIAL INFECTIONS Before the availability and use of antibiotics in the late 1940s and 1950s, there were few survivors of neonatal bacterial sepsis, contributing to the high perinatal mortality rate of that period. There have been changes in the types of bacteria responsible for neonatal infection over the years. In the 1930s and 1940s, the group A streptococcus was a prominent cause of neonatal sepsis; this organism is now rather rare (Bizzarro et al, 2005). In the 1950s, nursery outbreaks of Staphylococcus aureus infections appeared across North America and Europe, prompting changes in techniques of hygiene and encouraging the development and use of penicillinase-resistant antibiotics (Bizzarro et al, 2005). In the 1960s, Escherichia coli became the most common cause of bacterial sepsis, followed in the 1970s by the GBS. Even in an era of intrapartum antibiotic prophylaxis of GBS colonized mothers, the GBS remains the most common bacterial pathogen in neonatal centers of North American and Europe, followed by E. coli (Schrag et al, 2006). An update on neonatal sepsis at Yale, between 1989 to 2003, revealed an overall decrease in sepsis caused by both GBS and E. coli (Bizzarro et al, 2005). Regional differences exist, however, and must be considered before attempting to apply epidemiologic data to individual perinatal units. For example, Listeria monocytogenes is a frequent isolate in some western European countries, and S. aureus is found commonly in Germany and Scandinavia (PosfayBarbe and Wald, 2009). The incidence of early-onset bacterial infection is variable and ranges from one to five per 1,000 live births; however, it is clear that the incidence has declined as a result of intrapartum antibiotic therapy (Centers for Disease Control and Prevention, 2007, 2009). Recent data from the CDC revealed a downward trend from 2000 to 2003 (from 0.52 to 0.31 case per 1000 live births), followed by an increase from 2003 to 2006 (from 0.31 to 0.40 case per 1000 live births) (Centers for Disease Control and Prevention, 2009). Stratified by race, the incidence increased significantly among black infants from 2003 to 2006 (from 0.53 to 0.86 case per 1000 live births), whereas the incidence among white infants did not change significantly (from 0.26 to 0.29 case per 1000 live births). When stratified by gestational age, the average incidence of early-onset GBS disease among preterm infants during 2003 to 2006 539 was 2.8-fold higher among black infants compared with white infants (1.79 case versus 0.67 case per 1000 live births). It is of interest that both preterm black and white infants had increases in early-onset disease from 2003 to 2006 that were not statistically significant. Early-onset disease among full-term white infants was stable during 2003 to 2006, whereas term black infants had a significant increase of the incidence during this period, from 0.33 to 0.7 case per 1000 live births. The overall rates of late-onset GBS disease remained stable from 2000 to 2006 (0.36 case versus 0.30 case per 1000 live births) (Centers for Disease Control and Prevention, 2009). No overall incidence trend was observed from 2003 to 2006. When stratified by race, late-onset disease incidence among black infants decreased significantly by 42% from 2005 to 2006 (0.95 case versus 0.55 case per 1000 live births). Between 2003 and 2006 there were no significant trends among black or white infants. Infants described with early-onset sepsis frequently have one or more identifiable risk factors (Dutta et al, 2010). Prematurity is considered the single greatest risk factor for early-onset bacterial infections. Because it is accepted that ELBW infants have impairment of host defenses, and since preterm birth may be associated with low-grade chorioamnionitis, it is not surprising that the attack rates for infection by pathogens such as GBS are 26- to 30-fold higher in preterm infants, compared to term newborn infants, with an associated high mortality. Other risk factors for early-onset sepsis are maternal age, health and nutrition, colonization with well-known pathogens (e.g., GBS), and rupture of membranes for longer than 18 hours (Schuchat et al, 2000). Neonatal susceptibility to GBS infection is increased with deficiencies in circulating levels of GBS type-specific antibody and complement, which is further heightened by any element of neutrophil dysfunction, as may be seen in the more premature infants (Ferrieri, 1990; Foxman, 2007; Makhoul et al, 2009; Nandyal, 2008). BACTERIAL PATHOGENS IN EARLYONSET INFECTIONS GROUP B STREPTOCOCCAL INFECTIONS Since the early 1930s when Rebecca Lancefield reported her grouping system for hemolytic streptococci, group A streptococcus (Streptococcus pyogenes) was widely acknowledged as the major pathogen associated with puerperal sepsis. GBS was initially thought to be a commensal until 1938, when Frye reported seven cases of GBS-associated puerperal fever with three deaths (Eickhoff et al, 1964). Before the 1960s, GBS was not recognized frequently as a cause of human disease. However, in the late 1960s GBS emerged as the leading cause of neonatal sepsis in newborn infants (Bizzarro et al, 2005; Zaleznik et al, 2000). Before the era of maternal intrapartum prophylaxis, GBS had a reported national incidence of approximately two per 1000 live births and was associated with approximately 50% mortality in the newborn infant. As mentioned previously, over the past decade with the introduction of antibiotic maternal prophylaxis, there has been a significant decrease in the incidence of GBS to its current rate of approximately 0.32 per 1000 live births for early-onset disease. 540 PART IX Immunology and Infections Transmission of GBS from Mothers to Infants In the United States, approximately 20% to 35% of pregnant women are asymptomatic carriers of GBS in the genital and gastrointestinal tract during pregnancy and at the time of delivery (Ferrieri, 1990; Ferrieri et al, 2004b; Zaleznik et al, 2000). The prevalence of GBS colonization during pregnancy varies. Among women who were positive for GBS between 26 and 28 weeks’ gestation, only 65% remain colonized at term, whereas 8% of those with negative prenatal cultures were positive for GBS at term (Ancona et al, 1980; Zaleznik et al, 2000). Treatment of GBS-colonized women during pregnancy only temporarily eradicates the organism, and most women are recolonized within several weeks. At birth, 50% to 65% of infants who were born to GBS-colonized mothers have positive GBS cultures from mucous membranes and skin (external ear canal, throat, umbilicus, and anal or rectal sites) (Shet and Ferrieri, 2004). Before the introduction of intrapartum antibiotic prophylaxis, approximately 1% to 2% of colonized infants developed GBS infection, and the overall incidence of neonatal GBS infection was approximately two per 1000 live births in the United States. With intrapartum prophylaxis, approximately 60% to 80% of GBS cases occur in infants born to women with negative antenatal GBS screens (Verani and Schrag, 2010). A small number of GBS-infected infants acquired their bacteremia because of hematogenous transmission through the placenta. In these situations, the mother commonly displays signs and symptoms of chorioamnionitis, although it may occur in the absence of maternal symptoms (Baker and Edwards, 1995). Detection of GBS colonization has been emphasized since approximately 1996; studies to determine the optimal sites of sampling have been key to the effectiveness of intrapartum prophylaxis. GBS resides in the genitourinary and gastrointestinal tracts, where large numbers of gram-negative bacteria are also present. The majority of colonization studies have revealed high rates of both rectal and vaginal colonization with GBS (Ancona et al, 1980; Hickman et al, 1999; Zaleznik et al, 2000). The use of the selective broth enrichment medium that inhibits the growth of gram-negative enteric bacilli and other normal flora can increase culture sensitivity for GBS to greater than 90%. The most widely used selective medium is Todd-Hewitt broth with either gentamicin or colistin and nalidixic acid. As recommended in a 2002 publication from the Centers for Disease Control and Prevention, the optimal time for performing antenatal cultures is between 35 and 37 weeks’ gestation, and the highest culture yield is obtained when both the lower vaginal area and anal or rectal sites are sampled (Schrag et al, 2002). Epidemiologic studies of GBS-colonized women have shown that those with heavy colonization (3+ to 4+) are more likely to transmit GBS to their infants (Ancona et al, 1980). The colonization of GBS in pregnant women may be long standing, intermittent, or transient. There is a definite association between GBS colonization and other risk factors for neonatal sepsis; these include preterm labor, preterm delivery, premature rupture of membranes, prolonged rupture of membranes, and maternal fever. In the past few years, rapid diagnostic tests to detect GBS colonization in pregnant women have included realtime polymerase chain reaction (PCR); compared with broth enrichment cultures, there is an approximately 10% to 15% increased sensitivity (P. Ferrieri, unpublished data). The advantages of PCR detection of maternal vaginal or rectal colonization are the rapid turnaround time, as well as the increased sensitivity. Although more expensive than traditional culture-based detection assays, the results are available 1 to 2 days sooner. The argument that this does not provide semiquantitative data on the degree of GBS colonization in the mother is moot, because even low grade colonization in pregnant women is a risk factor for neonatal GBS sepsis. However, women with heavy (3+ to 4+) colonization, determined by semiquantitative assessment of vaginal or rectal cultures, are more likely to pass the microorganism to their infants (Ancona et al, 1980). Chemoprophylaxis and Intrapartum Antibiotics Prevention is of key importance in decreasing invasive GBS disease. The challenge has been to widely promulgate screening cultures in pregnant women. Revised guidelines from the CDC were published in 2002 and presented only the culture screening based approach for prevention and chemoprophylaxis, rather than the two preventive approaches published in 1996: a culture screening-based and a risk-based approach (Schrag et al, 2002). The riskbased approach involved the use of antibiotics based solely on the presence of antenatal or intrapartum risk factors such as maternal fever, preterm labor or premature rupture of membranes (<37 weeks’ gestation); prolonged rupture of membranes (≥18 hours); history of a previous newborn infant with GBS disease; and GBS bacteruria during pregnancy. Challenges to the implementation of the CDC guidelines, such as failure to seek prenatal care and the use of suboptimal laboratory culture techniques continue in certain populations. Data for the United States as a whole show a decrease in the incidence of early-onset GBS concurrent with the implementation of maternal GBS screen and intrapartum antibiotic prophylaxis guidelines. The current estimate for the overall United States population for early-onset GBS disease is 0.32 per 1000 live births (Van Dyke et al, 2009). Intrapartum Antibiotic Prophylaxis GBS is sensitive to penicillin, which is the drug of choice because of its narrow spectrum; the alternative is ampicillin. If a mother is allergic to penicillin but not at high risk for anaphylaxis, the use of cefazolin has been proposed (Schrag et al, 2002). When patients are at high risk for anaphylaxis in which the GBS is known to be susceptible to clindamycin and erythromycin, they can receive either of these drugs intravenously (IV) until delivery. When GBS is resistant to clindamycin or erythromycin or the antibiotic susceptibility is unknown, vancomycin given every 12 hours IV until delivery is the current recommendation (Schrag et al, 2002). In the United States, GBS exhibits CHAPTER 39 Neonatal Bacterial Sepsis considerable resistance against erythromycin (5% to 32%) and clindamycin (3% to 21%) (Castor et al, 2008). It is therefore important to have antibiotic testing done on the group B streptococcal isolates from pregnant women. For laboratories performing PCR on maternal vaginal or rectal cultures, it is recommended that the swabs be placed in a selective enrichment broth containing inhibitory antibiotics (either colistin and nalidixic acid or gentamicin and nalidixic acid) against gram-negative bacteria. If the PCR result is positive, the broth culture can be subcultured and antibiotic testing can be pursued. Because of the higher sensitivity of PCR, compared with the selective brothenrichment culture, the organism will not grow in 10% to 15% of occasions. Group B Streptococcal Sepsis in Neonates The majority of infections in newborn infants occur within the first week of life and are designated as early-onset disease (Table 39-1). Late-onset infections occur in infants 7 days or older, with the majority of these infections ­appearing in the first 3 months of life. Although chemoprophylaxis has led to a significant decrease in the incidence of early-onset GBS disease, there is no evidence that chemoprophylaxis prevents late-onset disease (CohenWolkowiez et al, 2009; Hamada et al, 2008; Jordan et al, 2008). Young infants with early-onset invasive GBS disease usually have pneumonia, sepsis, often; less often they have meningitis, osteomyelitis, or septic arthritis (Koenig and Keenan, 2009). The frequency of meningitis, osteomyelitis, or septic arthritis is higher among infants with late-onset disease. There are nine antigenically distinct GBS serotypes, based on their capsular polysaccharide analysis (types Ia, Ib, II to VIII) and a proposed new type, IX (Diedrick et al, 2010; Henrichsen et al, 1984; Slotved et al, 2007). In the United States and Western Europe, types Ia, II, and III account for the majority of isolates from infants TABLE 39-1 Manifestations of Early-Onset and Late-Onset Group B Streptococcal Disease Early-Onset ­Disease Late-Onset ­Disease Age at onset Birth through day 6 of life Day 7 to 3 months Symptoms Respiratory ­distress, apnea Irritability, fever, poor feeding Findings Pneumonia, sepsis Sepsis, meningitis, osteoarthritis Maternal ­obstetrical ­complications Frequent Uncommon Mode of transmission Vertical, in utero, or intrapartum Nosocomial, horizontal Predominant serotypes Ia, III, V* III, Ia, V* Effect of intrapartum antibiotic prophylaxis recommended by the Centers for Disease Control and Prevention Reduces incidence by 85%-90% No effect Characteristic *In decreasing order of frequency 541 with early-onset disease (Diedrick et al, 2010; Zaleznik et al, 2000). However, recent studies in the United States have demonstrated that serotypes Ia, III, and V, the latter emerging in recent years (Diedrick et al, 2010; Elliott et al, 1998; Harrison et al, 1998), account for the majority (70-75%) of early-onset invasive disease in newborn infants and parturient women (Diedrick et al, 2010; Zaleznik et al, 2000). Late-onset GBS disease in infants is dominated by serotype III followed by serotypes Ia and V (Shet and Ferrieri, 2004). A polysaccharide capsule is considered the most important virulence factor (Cieslewicz et al, 2005; Kasper et al, 1996; Paoletti et al, 1997, 2001); however, the role of surface localized GBS proteins (Ferrieri et al, 2004a; Johnson and Ferrieri, 1984; Lindahl et al, 2005; Smith et al, 2004) in pathogenesis and immune protection has gained favor (Maione et al, 2005; Tettelin et al, 2005). The remaining GBS isolates from invasive disease consist primarily of types Ib and II, but types IV, VI, VII, and VIII compose a small fraction. Type IV GBS represented between 0.4% and 0.6% of colonizing GBS isolates (Diedrick et al, 2010), but it was relatively uncommon for type IV isolates to be found in invasive GBS (Ferrieri et al, 2008; Puopolo and Madoff, 2007). Recent studies in the United Arab Emirates, Turkey, and Zimbabwe showed large proportions of type IV GBS among their isolates (Amin et al, 2002; Diedrick et al, 2010; Ekin and Gurturk, 2006; Moyo et al, 2002). In Zimbabwe it was the fourth most common serotype, comprising 4.6% of the colonizing and invasive isolates (Moyo et al, 2002). Serotypes VII and VIII are uncommon in Western countries. Infected infants have low levels of type-specific antibody to the infecting GBS serotype (Baker and Edwards, 2003). Vaccines against the common GBS serotypes have been shown to elicit a specific antibody response in humans (Baker and Edwards, 2003; Kasper et al, 1996; Paoletti et al, 1997, 2001). Sera from these vaccinated individuals protected against GBS challenge in neonatal mice, thereby showing the potential of vaccines to prevent invasive neonatal GBS disease in infants (Paoletti et al, 1997). The prospect of a multivalent GBS vaccine, with or without conjugated GBS surface localized proteins, makes the study of the common GBS serotypes important because of the possibility of serotype replacement or capsular switch (Cieslewicz et al, 2005; Lipsitch, 1999; Maione et al, 2005; Tettelin et al, 2005). ESCHERICHIA COLI INFECTIONS Historically, E. coli has been the second-most common pathogen causing sepsis and meningitis in newborn infants. The antigenic structure of E. coli is complex, composed of approximately 150 somatic or cell wall O antigens, 50 flagellar H antigens, and approximately 80 capsular K antigens. However, a limited number of K antigen E. coli strains cause meningitis, and approximately 80% of the strains causing meningitis and 40% of the strains causing bacteremia or sepsis express K1 (Mulder et al, 1984). The capsular K1 polysaccharide antigen is highly homologous to the capsular antigen of group B Neisseria meningitidis. Because a high percentage of women may have bacteriuria with strains of E. coli that express the K1 antigen or are 542 PART IX Immunology and Infections colonized with it at the time of delivery, it is surprising that E. coli sepsis or meningitis is not more common. It has been estimated that disease occurs in 1 in 100 to 200 infants colonized by K1 E. coli. Surveillance data from the National Institute of Child Health and Human Development Neonatal Research Network, a consortium of 16 U.S. academic neonatal centers, revealed that in the era of widespread implementation of antibiotic prophylaxis, E. coli sepsis increased from 3.2 to 6.8 cases per 1000 live births. This increase was observed in the 1998-2000 era and persisted in 2002 to 2003. Approximately 85% of E. coli infections in very low-birthweight (VLBW) infants were ampicillin resistant (Stoll et al, 2002). However, most evidence suggests that intrapartum antibiotic prophylaxis has not been associated with a concomitant adverse impact of increasing E. coli or other non-GBS bacterial causes. Among preterm infants, however, the incidence of E. coli and ampicillin-resistant E. coli infections increased significantly (Bizzarro et al, 2005). A recent retrospective case-control study between 1997 and 2001 concluded that exposure to intrapartum antibiotic prophylaxis therapy in mothers did not increase the odds of invasive, early-onset E. coli infection. In fact, among full term infants, exposure to 4 hours or greater of intrapartum antibiotic therapy was associated with decreased odds of early-onset E. coli infection (Schrag et al, 2006). LISTERIA MONOCYTOGENES INFECTIONS Listeria monocytogenes is a small, facultatively anaerobic, gram-positive motile bacillus that produces a narrow zone of beta hemolysis on blood agar plates, and can be confused with GBS unless a careful Gram stain, a catalase reaction, and other tests are performed. Most disease is due to three primary serotypes: 1a, 1b, and 4b. The last serotype has been described in most outbreaks of listeriosis (Posfay-Barbe and Wald, 2009). Most cases of listeriosis appear to be food borne, including those acquired by pregnant women. Foods that can be contaminated by L. monocytogenes include raw vegetables such as cabbage, raw milk, fish, poultry, processed chicken, beef, and hot dogs (Schlech, 2000). Transmission to the fetus occurs through either a hematogenous (transplacental) route or via an ascending infection through the birth canal. Frequently, infections with Listeria spp. early in gestation result in abortion; later in pregnancy, infection with Listeria spp. can result in premature delivery of a stillborn or infected newborn. Approximately 70% of Listeria-infected women deliver before 35 weeks’ gestation. Illness in the mother may be undetected because of vague influenza-like illnesses that may not come to medical attention. In approximately half of perinatal cases, illness in the mother has preceded delivery by 2 days to 2 weeks. At autopsy of stillborn infants or of those who die in the perinatal period, granulomas may be found throughout such organs as the liver and lungs, and infection is widely disseminated, including involvement of the meninges. Treatment of Listeria spp. infection or bacteremia during pregnancy can prevent infection in the fetus (Kalstone, 1991). Like GBS infection, Listeria spp. infection may have either an early-onset or late-onset presentation. Epidemics of neonatal Listeria spp. infection have been described after ingestion of contaminated foods such as cheese or coleslaw. The first clearly documented food borne (coleslaw) outbreak of listeriosis was in 1981 from the Eastern Maritimes in Canada (Schlech, 2000); it was associated with a fatality rate of 27%. There are reports of repeated abortions in women with colonization in the gastrointestinal tract, and cold-enrichment cultures can be performed to try to detect fecal carriage in such women. However, cold enrichment cultures are inferior to selective media for Listeria spp. in isolating the organism from various foods or stool specimens. Rapid antigen tests based on nucleic acid amplification are not in common use in clinical diagnostic laboratories. There is no current vaccine for Listeria spp. infection, but preventive measures have included the surveillance programs from the U.S. Department of Agriculture, prohibiting the sale of contaminated meats. Between 1996 and 2006, the incidence of Listeria spp. infections declined by 36%; however, an outbreak of disease in 2002 related to contaminated turkey meat led to 54 illnesses, eight deaths, and three fetal deaths in nine states (Posfay-Barbe and Wald, 2009). MISCELLANEOUS BACTERIAL PATHOGENS Bacteria responsible for early-onset neonatal sepsis have changed dramatically over time. There are regional differences in the organisms commonly responsible for early-onset sepsis. In addition to the organisms mentioned previously, other bacterial pathogens associated with earlyonset bacteremia or sepsis in newborn infants are Enterococcus spp., viridans group Streptococcus spp., Klebsiella spp., Enterobacter spp., Haemophilus influenzae (typeable and nontypeable), S. aureus, Streptococcus pneumoniae, group A streptococcus and other beta-hemolytic streptococci, and coagulase-negative staphylococci. CLINICAL SIGNS OF BACTERIAL SEPSIS There is great variability in the clinical presentations of infants with early-onset bacterial sepsis (Box 39-1). Most infants exhibit respiratory distress in the first 12 hours of life, frequently immediately after birth. In these infants, the progression may be rapid with cardiovascular instability, shock, and death. Presentation within the first 12 hours of life suggests that the infection with pneumonia and bacteremia occurred at or near the time of birth BOX 39-1 C  ommon Clinical Signs of ­Neonatal Sepsis ll ll ll ll ll ll ll ll ll ll Abnormal neurologic status: irritability, lethargy, poor feeding Abnormal temperature: hyperthermia or hypothermia Apnea Bleeding problems: petechiae, purpura, oozing Cardiovascular compromise: tachycardia, hypotension, poor perfusion Cyanosis Gastrointestinal symptoms: abdominal distention, emesis, diarrhea Jaundice Respiratory distress: tachypnea, increased work of breathing, hypoxemia Seizures CHAPTER 39 Neonatal Bacterial Sepsis 543 Bacterial Infection Present YES NO POSITIVE True Positive TP False Positive FP Positive Predictive Value  (TP)/(TPFP) NEGATIVE False Negative FN True Negative TN Negative Predictive Value  (TN)/(FNTN) Sensitivity  (TP)/(TPFN) Specificity  (TN)/(FPTN) Laboratory Test Result FIGURE 39-1 Diagram of the relationships among sensitivity, specificity, positive predictive value, and negative predictive value. FN, Number with infection incorrectly diagnosed as healthy by the test. FP, number of healthy infants incorrectly diagnosed as infected by the test; TN, number of healthy infants correctly diagnosed as not infected by the test; TP, number of infants with infection correctly diagnosed by the test. or during the immediate postnatal period. Infants with hypoxia in utero may gasp, inhaling contaminated amniotic fluid and setting the stage for early-onset pneumonia, bacteremia, and sepsis. The signs of early-onset infection may be subtle, with tachypnea suggesting “wet lung disease” or may be more overt with grunting, flaring, and subcostal and intercostal retractions. Because the signs of sepsis can be relatively nonspecific, such as poor feeding and increased sleepiness, they can be overlooked. In newborn or intermediate or intensive care nurseries, one must be attuned to subtle abnormal findings in newborn infants. The clinical signs of neonatal sepsis include hyperthermia or hypothermia, respiratory distress, apnea, cyanosis, jaundice, hepatomegaly, abdominal distention, feeding abnormalities, and neurologic abnormalities. Autopsy findings in preterm infants with fatal early-onset GBS infection suggested that surfactant deficiency respiratory distress syndrome was common (Payne et al, 1988). LABORATORY TESTING There are many laboratory tests that have been evaluated for infants with possible sepsis, and the results must be interpreted with caution, assessing the sensitivity and specificity of a particular test as well its positive and negative predictive accuracy (Figure 39-1). The sensitivity of a test is defined as the proportion of individuals with proven or probable sepsis in whom the result is abnormal; the specificity is the proportion of healthy or noninfected infants in whom the result is normal. Ideally, a test would have a high sensitivity and a high specificity, but this is rarely achievable. High sensitivity is the most desirable characteristic when dealing with serious and treatable diseases such as neonatal sepsis. Because sepsis is generally treated with antibiotic agents that have a low toxicity, diagnostic tests do not need to have a high specificity, but should have a high sensitivity, which will allow sepsis to be excluded. A positive predictive accuracy is the probability that an infant with an abnormal laboratory result is infected; a negative predictive value is the probability that infection with a normal or negative result is free of infection. The more sensitive the test, the greater its negative predictive value; the more specific a test, the higher its positive predictive value. MICROBIOLOGIC CULTURES Previously, cultures of superficial body sites in newborn infants (external auditory canal, gastric aspirate, umbilicus, and nasopharynx) were used to identify bacterial pathogens when more specific blood culture results were negative. Most centers no longer use surface cultures to make clinical decisions regarding either the institution or the discontinuation of antibiotic therapy, because surface cultures are of limited value in predicting the etiology of bacterial sepsis in newborn infants (Evans et al, 1988; Shenoy et al, 2000). Examination of gastric aspirates by Gram stain as a screening mechanism has also lost favor, as has examination for PMNs. Although the presence of increased numbers of PMNs may represent amnionitis and an inflammatory response, these cells may reflect a maternal origin and have no specificity in predicting bacterial sepsis for the infants (Vasan et al, 1977). Blood Cultures The gold standard for detection of bacteremia in newborn infants with suspected sepsis is a positive blood culture. With the introduction of newer blood culture detection instruments that are semiautomated and examined for the presence of growth by CO2 production by the internal computer of the instrument every minute, the sensitivity of detecting positive blood cultures has increased. Another variable that influences the sensitivity of detection of bacteremia is the volume of blood obtained and placed in the culture bottles. Ideally 1 to 3 mL of blood from infants should be obtained, but this not always possible in very small infants. Most positive blood cultures are detected within 24 to 48 hours using the new technology (Garcia-Prats et al, 2000). However, the use of intrapartum antibiotics for prophylaxis in mothers with either GBS colonization or suspected amnionitis on account of any cause can reduce the ability to detect bacteremia in newborn infants. In a term infant who was asymptomatic at the initiation of antibiotic therapy, it may be reasonable to stop antibiotic administration if the blood cultures remain negative after 48 hours. However, the decision to discontinue treatment with antibiotics should include the assessment of results of other laboratory tests used for sepsis screens, and should not solely rely on a negative blood culture. When 544 PART IX Immunology and Infections TABLE 39-2 Predictive Values of Components of the White Blood Cell Count and of C-Reactive Protein Predictive Value (%) Laboratory Result Sensitivity (%) Specificity (%) Positive Negative ANC <1750 cells/mm3 38-96 61-92 20-77 96-99 ANC <10% (≈5580 cells/mm3 for term) 48 73 4 98 I/T ratio ≥ 0.2 90-100 30-78 11-51 98 I/T ratio ≥ 0.3 35 89 7 98 CRP > 1mg/dL 70-93 78-94 27 100 Adapted from Gerdes JS: Diagnosis and management of bacterial infections in the neonate, Pediatr Clin North Am 51:939-959, 2004. ANC, Absolute neutrophil count; CRP, C-reactive protein; I/T, immature to total neutrophils (ratio), calculated by the number of immature neutrophils (bands, metamyelocytes, and myelocytes) divided by the total number of neutrophils (immature plus mature). the suspicion of sepsis is high, clinicians should consider continuing antibiotic therapy for a complete course despite negative blood cultures (Ottolini et al, 2003). Urine Cultures The frequency of positive urine cultures in infants with early-onset sepsis is relatively low, and it is rare to find bacteriuria in infants with negative blood culture results (DiGeronimo, 1992). Infants with late-onset sepsis tend to have a higher rate of positive urine cultures (Visser and Hall, 1979). In the era of widespread intrapartum antibiotic prophylaxis in the mother, positive urine cultures may be obscured because of excretion of antibiotics in the urine of newborn infants. When pyelonephritis is found in newborn infants, it likely represents metastatic seeding of the kidney during a bout of bacteremia. In the first 72 hours of life, because the yield from urine cultures is low, it is not generally recommended to obtain these specimens. However, in the older newborn infant, a urine sample collected by an aseptic technique (urinary catheter or suprapubic bladder aspiration) is an important part of the sepsis workup. Cerebrospinal Fluid Lumbar punctures are deferred in infants with any instability or uncorrected bleeding disorders. The details of examination of cerebrospinal fluid and the diagnostic approach for examining cerebrospinal fluid will be discussed under the section on bacterial meningitis. WHITE BLOOD CELL COUNT AND NEUTROPHIL INDICES Normal white blood cell (WBC) counts range from 9000 to 30,000 cells/mm3 at the time of birth, and differences in the site of sampling can affect these values. The absolute neutrophil count (ANC), the absolute band count of immature neutrophils, and the ratio of immature neutrophils to total neutrophils (I/T) are regarded as more useful than total leukocyte counts in the diagnosis of neonatal sepsis. The lower limit or total neutrophil count rises to 7200 cells/mm3 by 12 hours of age, and then declines to approximately 1720 cells/mm3 by 72 hours of age (Manroe et al, 1979; Schmutz et al, 2008). Postnatally the absolute band count also undergoes similar changes, with peak values of 1400 cells/mm3 at 12 hours of age, and then declines. In contrast, the I/T ratio is maximum at birth and then declines to a value of 0.12 beyond 72 hours of age (Manroe et al, 1979; Schelonka et al, 1994). However, in VLBW infants there is a greater reference range for the total ­neutrophil counts (Manroe et al, 1979; Mouzinho et al, 1994). There are no significant differences in I/T ratio or absolute ­immature neutrophil counts in VLBW infants. There is considerable overlap of the cutoff values of the ANC, absolute band count, and I/T ratio between healthy and infected newborns. There are a number of clinical conditions that affect the total neutrophil count. Prolonged crying, meconium aspiration syndrome, maternal fever, and asphyxia are all associated with an increase in the total neutrophil count, and there may be an increase in the total immature neutrophil forms, as well as an increased I/T ratio. Maternal hypertension is associated with a decrease in total neutrophils. At high altitudes, a higher upper limit of neutrophil values occurs (Schmutz et al, 2008). In approximately two thirds of infants with sepsis, the total neutrophil count is abnormal. Neutropenia is the best predictor of sepsis, whereas neutrophilia does not correlate well. The absolute neutrophil band count is not a sensitive marker of sepsis, but has a relatively good predictive value and specificity. The I/T ratio is considered to have the best sensitivity of all of the neutrophil indices (Table 39-2) (Gerdes, 2004). PLATELET COUNTS Approximately 25% to 30% of infants exhibit thrombocytopenia at the time of diagnosis of sepsis, and this frequency increases during the course of infection. Accelerated platelet destruction and possibly depressed production caused by bacterial products on the bone marrow are the underlying mechanisms for thrombocytopenia in infected infants. Disseminated intravascular coagulation may be seen in some infants with severe sepsis. ACUTE PHASE REACTANTS AND ERYTHROCYTE SEDIMENTATION RATE A number of acute phase reactants have been studied to help identify infants with likely sepsis. Most biochemical markers currently in use are derived from components of CHAPTER 39 Neonatal Bacterial Sepsis the complex inflammatory response to an invading pathogen. These markers have been studied during the past 20 years and continue to be under investigation. Of issue is the availability of tests for these inflammatory mediators when investigating an infant with possible sepsis, and the value of these tests in assisting with the early diagnosis of sepsis. It is known that some proinflammatory cytokines peak rapidly, within 1 to 4 hours after a sepsis stimulus (Lam and Ng, 2008). C-reactive protein (CRP) rises to a maximum at 12 to 24 hours, and procalcitonin (PCT) rises at 4 hours, peaks at 6 hours, and plateaus 8 to 24 hours after a stimulus. CRP is induced by proinflammatory cytokines. Among the early markers are the proinflammatory mediators, such as IL-1β, IL-6, the chemokine IL-8, TNF-α, and interferon gamma; these activate host defenses against bacterial and other infecting agents, whereas antiinflammatory mediators such as IL-4, IL-10, and transforming growth factor β1 (TGF-β1) are important in regulating and limiting the inflammatory response, thus preventing an excessive reaction that could lead to organ damage and tissue cell death (Mehr and Doyle, 2000). Survival should be considered a fine balance between the proinflammatory cytokines and the antiinflammatory mediators. Because of the limitations of other assessment markers of potentially infected neonates, such as WBC counts and band counts, the advantages of having a surrogate marker readily available are apparent. CRP is probably the most studied acute phase reactant in neonatal sepsis (Benitz et al, 1998; Jaye and Waites, 1997; Lam and Ng, 2008; Weitkamp and Aschner, 2005). Monitoring of CRP levels has been widely promulgated as a way to diagnose neonatal infection and to adjust the duration of antibiotic therapy in infants with suspected versus proven sepsis (Benitz et al, 1998; Ehl et al, 1997; Gerdes, 2004; Philip and Mills, 2000). Depending on the laboratory, a CRP value of 1 to 8 mg/dL is considered the upper limit of normal; it is important to know the different cutoff values in the laboratory supporting one’s ­neonatal units. Theoretically, results can be available within 30 minutes. CRP is produced by the liver in response to stimulation by the proinflammatory cytokine IL-6, which is produced by both T and B cells (Lam and Ng, 2008; Weitkamp and Aschner, 2005). Because exposure of the host to bacterial products results in a substantial and rapid increase in IL-6 concentrations, it appears that IL-6 is potentially a more useful marker than CRP during the early phase of an infection. In one study, the IL-6 concentration had a sensitivity of 89% versus 60% for CRP at the onset of clinical suspicion of a neonatal infection (Lam and Ng, 2008). The negative predictive values of IL-6 are much higher than those for CRP (Lam and Ng, 2008), and this was also found in cord blood IL-6 levels, where the sensitivity of detection was high. This finding can be useful in deciding which infants do not require antibiotic therapy and those in whom antibiotics can be discontinued after a relatively short course. In a recent study, the high-sensitivity CRP (hsCRP) measurement has been shown to provide increased sensitivity for detecting neonatal infection (Edgar et al, 2010). Not all diagnostic laboratories can provide hsCRP in a timely fashion. In addition, the optimum diagnostic cutoff levels for CRP and hsCRP are debatable. 545 A recent study examined the combination of hsCRP, serum soluble intercellular adhesion molecule 1 (slSAM1), soluble E-selectin (sE-selectin), and serum amyloid A, individually and in combination, for the diagnosis of sepsis in a neonatal intensive care unit (Edgar et al, 2010). In this study, all four measurements had some diagnostic value for neonatal infection; however, s1SAM-1, hsCRP, and sE-selectin demonstrated the highest negative predictive value individually (sISAM, 84%; hsCRP, 79%; and sE-selectin, 74%) (Edgar et al, 2010). Use of a combination of these measurements enhanced the diagnostic value, with sensitivities of 90.3% and a negative predictive value of 91.3% (Edgar et al, 2010). However, the application of this set of diagnostic markers is not available for most facilities, and more investigative work is needed to confirm their role in excluding early-onset infection. In another study, early markers (IL-6, TNF-α, IL-8, interferon gamma, CRP, IL-18, the antiinflammatory cytokine IL-10, the I/T ratio, and PCT, a later marker of infection) were studied for use in detection of early-onset sepsis in 123 newborn infants (Bender et al, 2008). This study concluded that IL-6 combined with PCT was a fair measure of evaluating early-onset infection, and that the traditional I/T ratio was almost as efficient as IL-6. Combining an early marker such as IL-6 and an I/T ratio may reduce the number of diagnostic tests and nonconclusive values. The combined value of IL-6 and PCT at the first blood drawing had a sensitivity of 71% and specificity of 88% (Bender et al, 2008). There are many studies of procalcitonin in the literature, and most have concluded that procalcitonin is superior to CRP levels in the early diagnosis of neonatal sepsis (Lam and Ng, 2008). PCT is the precursor of calcitonin, normally synthesized in the C-cells of the thyroid gland. PCT is induced by systemic inflammation and bacterial sepsis and is produced by such cells as hepatocytes, nephrons, and monocytes. The physiologic function of PCT is unknown. In bacterial infections, plasma PCT concentrations increase from 0.001 to 0.01 ng/mL (baseline range) to values ranging from 1 to 1000 ng/mL. PCT concentrations rise much faster than CRP (6 to 8 hours versus 48 hours for maximum levels). In healthy newborn infants, plasma PCT concentrations increase gradually after birth, reaching peak levels at approximately 24 hours of age (range, 0.1 to 20 ng/mL) and then decrease to normal values less than 0.5 ng/mL by 48 to 72 hours of age (Stocker et al, 2010). Various studies on the use of PCT as a marker of neonatal sepsis have yielded contradictory results regarding its application to clinical decision making for both diagnosis and adjustment of the length of antibiotic therapy. In a recent study of 121 newborn infants with suspected early-onset sepsis, serial PCT determinations allowed shortening of the duration of antibiotic therapy (Stocker et al, 2010). However, there are currently few institutions or intensive care units that are applying PCT as a regular measurement for diagnosis or for adjustment of length of antibiotic therapy. Larger studies are needed to determine the true value of PCT in diagnosis and therapy. To date, studies on the role of various cytokine determinations in assisting with diagnosis and treatment of early-onset sepsis are intriguing, but have not translated to widespread use. Other molecular technologies are also 546 PART IX Immunology and Infections being studied to diagnose neonatal sepsis by the rapid identification and differentiation of gram-negative and gram-positive bacterial bloodstream infections. PCR, using universal bacterial primers, targets conserved regions of the 16SrRNA gene common to all bacteria, but not found in other organisms (Dutta et al, 2009). In one study, ­universal primer PCR was performed in newborn infants with clinically suspected sepsis. PCR was performed before starting antibiotic therapy, and repeated at 12, 24, and 48 hours after starting drug therapy (Dutta et al, 2009). The sensitivity, specificity, and positive and negative ­predictive ­values of universal primer PCR were 96.2%, 96.3%, 87.7%, and 98.8% respectively. Results of testing in two patients were blood culture positive, but 0-hour PCR negative, and results in 7 patients were 0-hour PCR positive, but blood culture result was negative. Of the patients with a result of 0-hour PCR positive, 7 remained positive at 12 hours but none remained positive at 24 and 48 hours after starting antibiotic therapy. Although universal bacterial primer PCR may be a useful test for diagnosing an early episode of culture-proven sepsis, it cannot be used for diagnosis if the patient has been exposed to 12 hours or more of antibiotic therapy (Dutta et al, 2009). Much larger studies are certainly required before this assay can be recommended for routine clinical use in newborn infants suspected of sepsis. PREVENTION INTRAPARTUM ANTIBIOTIC PROPHYLAXIS There were initial concerns about adverse effects of implementing the CDC consensus strategies, but these have proved unwarranted. The primary risks considered were maternal anaphylaxis from administered antibiotics, and these were unfounded. The possibility of the emergence of infections in mothers and infants caused by antibiotic resistant organisms (e.g., E. coli) was addressed in part by studies conducted by the Neonatal Network of the National Institute of Child Health and Human Development. In 2002, they reported a change in the pathogens causing early-onset sepsis, with an increase in sepsis caused by E. coli from 3.2 to 6.8 per 1000 live births (Stoll et al, 2002). Most of the E. coli isolates were resistant to ampicillin. Mothers of infants with ampicillin-resistant E. coli infections were more likely to have received intrapartum antibiotic prophylaxis than were those with ampicillin-sensitive strains (Stoll et al, 2002). In another study comparing a period of no prophylaxis to periods of risk-based and universal screening-based prophylaxis, no change in the incidence of infection with ampicillin-resistant organisms was observed overall or among VLBW infants (Puopolo and Eichenwald, 2010). However, an increased proportion of infections was caused by ampicillin-resistant organisms. Mothers of infants with ampicillin-resistant infections were also more likely to have been treated with ampicillin (Chen et al, 2005). A 10-year study of the effect of intrapartum antibiotic prophylaxis on GBS and E. coli sepsis in Australasia revealed a steady decline in early-onset GBS infection, but also a trend to decreasing early-onset E. coli sepsis in all infants and a stable rate for this infection in VLBW infants (Daley et al, 2004). INTRAPARTUM MANAGEMENT OF PARTURIENTS The universal GBS screening strategy recommended by the CDC is done at 35 to 37 weeks’ gestation. Antepartum antibiotic therapy is recommended for women with a previous infant with invasive GBS disease, a positive GBS screening culture result during pregnancy, unknown GBS status (culture either not performed or incomplete or results unknown), and any of the following features: delivery before 37 weeks’ gestation, rupture of chorioamniotic membranes for greater than 18 hours, intrapartum temperature of 100.4° F (38° C) or higher, and GBS bacteriuria during pregnancy. Women who are not allergic to penicillin can be given penicillin G (5,000,000 units as a loading IV dose followed by 2.5 million units every 4 hours until delivery) or ampicillin (2g IV as a loading dose followed by 1g every 4 hours until delivery). For women who are allergic to penicillin, the recommendation is to determine, ideally during prenatal care, whether the patient is at high risk for anaphylaxis (i.e., history of immediate hypersensitivity reactions). Women who are not at high risk for anaphylaxis can receive cefazolin (2g IV and 1g IV every 8 hours until delivery). Women who are at high risk for anaphylaxis whose GBS isolate is not resistant to clindamycin or erythromycin can receive clindamycin (900 mg IV every 8 hours until delivery) or erythromycin (500 mg IV every 6 hours until delivery) (Schrag et al, 2002). Women with clindamycin-erythromycin–resistant GBS isolates can receive vancomycin (1g IV every 12 hours until delivery). INTRAVENOUS IMMUNE GLOBULIN FOR PREVENTION OF EARLY-ONSET SEPSIS Premature infants are susceptible to early-onset sepsis because of diminished transplacental transfer of immunoglobulins and decreased synthesis of immunoglobulin (Ig) G. An intravenous infusion of immunoglobulin (IVIg) will increase the typically low levels of serum immunoglobulin in preterm infants, and it has been proposed to improve immune function that may lead to improved clinical outcome (Wynn et al, 2009). Clinical trials examining the possible benefit of IV-Ig began over 15 years ago (Chirico et al, 1987; Conway et al, 1990; Haque et al, 1986). The very premature infants are the group most likely to receive and potentially benefit from these adjuvant treatments and preventative interventions. A Cochrane metaanalysis showed a reduction in mortality of newborn infants treated with IV-Ig for what were subsequently proved to be bacterial infections (relative risk, 0.55; 95% confidence interval, 0.31 to 0.98) (Ohlsson and Lacy, 2010). However, a review of the studies included in the metaanalysis revealed the inclusion of few very premature infants. A large international, placebo-controlled, double-blind, randomized clinical trial has recently been undertaken in ­approximately 5000 newborn infants to examine the benefit of polyclonal IV-Ig treatment for proven or suspected sepsis (Wynn et al, 2010). However, because randomization will allow the inclusion of neonates of all gestational and postnatal ages with suspected or proven infection, the subgroup sample sizes may leave the issue of clinical efficacy for CHAPTER 39 Neonatal Bacterial Sepsis early- and late-onset sepsis in the different gestational age groups unanswered. A metaanalysis that included almost 4000 preterm (<37 weeks’ gestation) or low birthweight (<2500 g) infants showed a modest reduction in the development of sepsis after IV-Ig prophylaxis in preterm neonates (Wynn et al, 2010). Currently available data do not support the suggested immunologic enhancements expected for IV-Ig in very premature neonates. There is continued interest in the potential of IV-Ig as treatment or prophylaxis for infection in very premature infants, but large trials are needed to provide convincing data. Other immunomodulating factors such as granulocyte colony-stimulating factor (G-CSF) have been studied to both prevent and treat neonatal sepsis. An earlier metaanalysis showed that G-CSF administration improved mortality in all neonates and in the subgroups of neutropenic and preterm infants (Bernstein et al, 2001). A recent Cochrane metaanalysis concluded that prophylaxis with G-CSF does not significantly reduce mortality in all infants, although in premature infants with neutropenia (ANC < 1750) mortality may be improved (Carr et al, 2009). It is possible that combination therapies using IV-Ig and other immunomodulating factors, such as granulocyte-macrophage colony-stimulating factor (GM-CSF), G-CSF, and complement-containing blood products, may lead to improved immune function in these highly susceptible infants; however, these studies are not currently being conducted. DIAGNOSTIC APPROACH TO NEONATES WITH SUSPECTED SEPSIS Obviously, all symptomatic newborn infants must be carefully evaluated for the possibility of bacterial sepsis and treated with antibiotics, if necessary. Although the presence of various risk factors should increase the suspicion of sepsis, the absence of risk factors in a symptomatic infant cannot be dismissed. In adjusting to postnatal life, some infants exhibit abnormal signs transiently, such as tachypnea, before becoming asymptomatic. However, any infant who has other findings or is still symptomatic with only one finding by 6 hours of life should have a diagnostic evaluation performed with a complete blood cell count (CBC) and differential, a blood culture, and as appropriate, a lumbar puncture and a chest radiograph. Antibiotic therapy can be stopped when the physical findings are normal, the clinical suspicion of sepsis is low, and the screening results for sepsis, including the blood culture, remain negative. If the blood culture result is positive, the lumbar puncture findings abnormal, or there are clinical signs of sepsis, then the infant should be treated with an appropriate course of antibiotics (Figure 39-2). Management of the asymptomatic infant with risk factors for sepsis is more controversial. Escobar et al (2000) retrospectively reviewed 2785 newborns with birthweight greater than 2000 g and who were evaluated for sepsis after birth. Asymptomatic status was associated with a significantly decreased odds ratio for infection. Low absolute neutrophil count and meconium-stained amniotic fluid were associated with an increased risk of infection. The highest infection rates occurred in infants whose mother had a clinical diagnosis of “definite” chorioamnionitis, but chorioamnionitis was significantly associated with an 547 increased risk of neonatal infection only in infants whose mothers had not received intrapartum antibiotics (Escobar et al, 2000). Current practice management for the infant with no signs or symptoms of sepsis, but with risk factors such as maternal chorioamnionitis or incomplete GBS maternal prophylaxis, includes performing a limited sepsis evaluation (CBC with differential and blood culture). If the WBC count, absolute neutrophil count, and the I/T ratio are all normal, then the infant is usually observed in the hospital and discharged at approximately 48 hours. If the screening tests are abnormal, the infant undergoes a complete sepsis evaluation with subsequent antibiotic therapy initiated. Recent CDC and American Academy of Pediatrics guidelines for the prevention of early-onset neonatal GBS sepsis have a slightly different approach to management of the at-risk asymptomatic infant (see Figure 39-2; Verani and Schrag, 2010). If there are signs of sepsis in the newborn, then a full sepsis workup and antibiotic therapy are recommended. If there is maternal chorioamnionitis, a limited sepsis evaluation and antibiotic are recommended. For infants whose mother received greater than 4 hours of antibiotic therapy, it is recommended that the infant be observed in the hospital for approximately 48 hours. For infants whose mother did not receive more than 4 hours of antibiotic prophylaxis, a 48 hour in-hospital observation is recommended. If the infant is less than 37 weeks’ gestation, or the duration of rupture of membranes is greater than or equal to 18 hours, then a limited evaluation and observation in the hospital for approximately 48 hours are recommended. TREATMENT ANTIMICROBIAL THERAPY The choice of antibiotic administration for an infant with suspected early-onset sepsis depends on the predominant bacterial pathogens and the antibiotic susceptibility profiles for the microorganisms causing early-onset disease in a particular geographic region. Any decision to discontinue antimicrobial therapy should be based on the level of suspicion for sepsis at the time treatment was begun, the culture results, laboratory test results, and the clinical behavior and course of the infant. If sepsis is highly suspected in an infant, antibiotics should be considered for a full course even if the culture results are negative. Empiric therapy for early-onset sepsis generally consists of combinations of antibiotics effective against grampositive (e.g., GBS, L. monocytogenes) and gram-negative pathogens (e.g., E. coli). The two most commonly used combinations are (1) ampicillin with an aminoglycoside, usually gentamicin, and (2) ampicillin with a third-­ generation cephalosporin, usually cefotaxime. Cefotaxime has minimal toxicity and is well tolerated by newborn infants. However, the third-generation cephalosporins, as well as vancomycin, a glycopeptide antibiotic, have been associated with the development of vancomycin-resistant enterococci and, in the case of cephalosporins, with the induction of various β-lactamase producing gram-negative bacteria, including extended-spectrum β-lactamase producing organisms (Bryan et al, 1985). These latter bacteria 548 PART IX Immunology and Infections Signs of neonatal sepsis? yes Full diagnostic evaluation* Antibiotic therapy† yes Limited evaluation¶ Antibiotic therapy†¶ no Routine clinical care†† yes Observation for 48 hours††§§ yes Observation for 48 hours†† yes Limited evaluation¶ Observation for 48 hours¶¶ no Maternal chorioamnionitis?§ no GBS prophylaxis indicated for mother?** yes Mother received 4 hours of penicillin, ampicillin or cefazolin IV? no 37 weeks AND duration of membrane rupture <18 hours? no Either <37 weeks OR duration of membrane rupture 18 hours? * Includes CBC with differential, platelets, blood culture, chest radiograph (if respiratory abnormalities are present), and LP (if patient stable enough to tolerate procedure and sepsis is suspected). † Antibiotic therapy should be directed towards the most common causes of neonatal sepsis including GBS and other organisms (including gram negative pathgoens), and should take into account local antibiotic resistance patterns. § Consultation with obstetric providers is important to determine the level of clinical suspicion for chorloamnionitis. Chorioamnionitis is diagnosed clinically and some of the signs are non-spedific. ¶ Includes blood culture (at birth), and CBC with differential and platelets. Some experts recommend a CBC with differential and platelets at 6-12 hours of age. ** GBS prophylaxis indicated if one or more of the following: (1) mother GBS positive at 35-37 weeks’ gestation, (2) GBS status unknown with one or more intrapartum risk factors including <37 weeks’ gestation, ROM 18 hours or T 100.4F (38.0C), (3) GBS bacteriuria during current pregnancy, (4) history of a previous infant with GBS disease. †† If signs of sepsis develop, a full diagnostic evaluation should be done and antibiotic therapy initiated. §§ If 37 weeks’ gestation, observation may occur at home after 24 hours if there is a knowledgeable observer and ready access to medical care. ¶¶ Some experts recommend a CBC with differential and platelets at 6-12 hours of age. FIGURE 39-2 Algorithm for recommended management of newborns at risk for group B streptococcal disease. (Adapted from Verani JR, Schrag SS: Group B streptococcal disease in infants: progress in prevention and continued challenges, Clin Perinatol 37:375-392, 2010.) are resistant to all β-lactam antibiotics and frequently are resistant to other antibiotics, but not to meropenem. Another disadvantage of the cephalosporin antibiotics is the lack of effectiveness against enterococci or L. monocytogenes. L. monocytogenes is usually treated with ampicillin and an aminoglycoside until the blood culture result is negative and the infant has shown an improved outcome. In infants with bacteremia and sepsis caused by GBS, gentamicin is frequently combined with ampicillin or penicillin, although there are no data to suggest that the addition of the aminoglycoside improves outcome. However, it is common practice to use the combination of these two drugs during the first few days of therapy and then to continue the full course of therapy with ampicillin or penicillin alone. When the likelihood of infection is very low, the antibiotic therapy should be stopped. In most hospitals using modern blood culture instrumentation, 48 hours should be considered sufficient to determine whether a blood culture result is negative, assuming that no antibiotics were being given when the culture was obtained (Garcia-Prats et al, 2000). Blood culture bottles with antimicrobial binding resins are in common use in microbiology laboratories, enhancing the ability to demonstrate positive blood cultures. Infants with proven bacteremia, but without meningitis, are commonly treated for 7 to 10 days. The use of antibiotics with nephrotoxicity (i.e., aminoglycosides) should be monitored using appropriate drug levels. IMMUNOLOGICAL THERAPIES FOR TREATMENT OF EARLY-ONSET SEPSIS Various adjunctive therapies have been proposed to improve the immune status of the infant in an attempt to mitigate the mortality of neonatal sepsis. These included IV-Ig, granulocyte transfusions and G-CSF or GM-CSF treatment. Overall, there are insufficient data to recommend the routine use of any of these therapies for the treatment of sepsis in newborn infants. However, mortality in cases of subsequently proven infection was reduced in the IV-Ig–treated infants, and a large randomized controlled trial of IV-Ig use was recently completed (Ohlsson and Lacy, 2010). Clinical studies also did not show that granulocyte transfusions provided a significant benefit to neonates with culture-proven early-onset infections (Vamvakas and Pineda, 1996). However, in infants with CHAPTER 39 Neonatal Bacterial Sepsis overwhelming sepsis, severe neutropenia, and depletion of bone marrow neutrophil stores, treatment with granulocyte transfusions improved survival (Cairo et al, 1992). Using GM-CSF and G-CSF to enhance the quantity and quality of neutrophils has also been studied in human neonates (Carr et al, 2009). Although these agents have induced circulating numbers of neutrophils and appeared to be relatively safe, they have not significantly reduced neonatal sepsis mortality (Wynn et al, 2009). In two metaanalyses, subgroup analysis showed a significant reduction in mortality in the group of infants with systemic infection and neutropenia (Bernstein et al, 2001; Carr et al, 2009). Future studies may prove that selective use of these therapies is beneficial in specific subgroups of septic infants. NEONATAL BACTERIAL MENINGITIS Neonatal bacterial meningitis is ominous because of its mortality and morbidity, and it is associated with the same pathogens that cause bacterial sepsis, with GBS and E. coli accounting for approximately 70% of all cases, and L. monocytogenes accounting for an additional 5% in the first week of life. On occasion, it is possible to isolate ­Streptococcus pneumoniae and Haemophilus influenzae, and in infants who are older than 1 week residing in neonatal intensive care units, coagulase-negative staphylococci are the most common isolates. The underlying pathogenesis of bacterial meningitis is a seeding of the meninges during a bacteremic phase in the infant. Studies in neonatal rats have shown that high-grade bacteremia is more likely to lead to bacterial meningitis than lower grade bacteremia. GBS meningitis (with a mortality approaching 30% and a morbidity of 50%) usually presents as late-onset disease, and the most common GBS serotype identified is III (Levent et al, 2010). In a recent paper by Ansong et al (2009), GBS meningitis complicated 22 in 145 (15%) episodes of early-onset GBS sepsis and 13 in 23 (57%) of episodes of late-onset GBS sepsis. GBS meningitis can occur in the presence of negative blood cultures, and 20% of infants in this study had negative blood cultures (Ansong et al, 2009). Approximately 80% of all serotypes of E. coli that cause meningitis in newborn infants possess the K1 or capsular antigen. The K1 capsular polysaccharide antigen is considered one of the primary virulence factors of this capsular type of E. coli, because antibody against K1 antigen has been shown to be protective in neonatal rat models of infection. Mortality rates for neonatal E. coli meningitis vary from 20% to 30% in some centers and 50% to 60% in others (Dodge, 1994). 549 instability of temperature regulation, vomiting, respiratory distress, and apnea. A bulging fontanel may be seen, but this is usually a late manifestation. Seizures are frequently observed and can be caused by either direct central nervous system inflammation or by metabolic abnormalities such as hypoglycemia or hyponatremia. DIAGNOSIS The gold standard for diagnosis of meningitis is the analysis of the cerebrospinal fluid, including the WBC count, glucose and protein levels, Gram stain, and culture. The interpretation of cerebrospinal fluid cell counts in newborn infants may be difficult (Garges et al, 2006; Greenberg et al, 2008; Polk and Steele, 1987; Unhanand et al, 1993). During the first week of life, the cerebrospinal fluid WBC count slowly decreases in full-term infants, but may remain high or even increase in premature infants. There is no change in cerebrospinal fluid WBC counts or protein content with gestational age, but there is a significant decrease with postnatal age (Mhanna et al, 2008). The cerebrospinal fluid cell counts, protein, and glucose concentrations from healthy infants may overlap with those from infants with meningitis, and from 1% to 10% of infants with proven meningitis have normal results on a cerebrospinal fluid analysis (Hristeva et al, 1993; Garges et al, 2006; Greenberg et al, 2008). Finally, approximately 30% of all infants with positive results of cerebrospinal fluid cultures for bacteria have negative blood culture results (Garges et al, 2006; Wiswell et al, 1995). A Gram stain of cerebrospinal fluid must be examined carefully for every infant with suspected meningitis. The stains for approximately 20% of newborns with proven meningitis are reported as showing “no bacteria seen.” Although an increase is expected in neutrophils with bacterial meningitis, one may see a predominance of lymphocytes within a conversion to PMNs. With L. monocytogenes, a mononuclear cellular response is found in examination of the cerebrospinal fluid. In clinical care units, it is routine to repeat the cerebrospinal fluid examination and culture 2 to 3 days after the initiation of antibiotic therapy. This examination is especially important if the patient has not responded clinically and is experiencing seizures or ­continued fever. At times it is difficult to eradicate the organism from the cerebrospinal fluid, and consideration can be given to examining the inhibitory and bactericidal concentrations in cerebrospinal fluid. It is especially important to repeat the cerebrospinal fluid examination before stopping antibiotics in patients with more complicated courses, and for enteric gram-negative bacterial meningitis. PATHOLOGY AND CLINICAL MANIFESTATIONS THERAPY At autopsy, infants who die of meningitis have purulent exudates of the meninges and the surfaces of the ventricles associated with inflammation. Historically, hydrocephalus and a noninfectious encephalopathy were demonstrated in approximately 50% of infants who died of bacterial meningitis. The signs and symptoms of neonatal meningitis are not easy to distinguish from those of sepsis. The most common presenting symptoms are lethargy, feeding problems, Infants with bacterial meningitis are frequently ill and should be monitored in intensive care units when the critical needs can be met with aggressive management. These patients may require mechanical ventilation, complex fluid management to attenuate the effects of cerebral edema and effects of secretion of inappropriate anti-diuretic hormone (ADH), seizure control, vasopressor support, and cardiopulmonary monitoring. The choice of appropriate antibiotic therapy is based on several factors, including 550 PART IX Immunology and Infections the achievable cerebrospinal fluid levels of drugs that have in vitro efficacy against the microorganism. In the case of gram-positive bacteria, the use of penicillin and ampicillin will achieve 10- to 100-fold higher concentrations than the minimal inhibitory concentrations needed to inhibit the bacteria, and there is rapid sterilization of the cerebrospinal fluid. In contrast, aminoglycosides, such as gentamicin and tobramycin, achieve only 40% of peak serum levels and may not achieve minimal inhibitory concentrations more than those equal to or slightly greater than found in vitro for gram-negative bacteria. In many intensive care units, ampicillin and gentamicin are recommended for the initial therapy for neonatal meningitis. An alternative regimen of ampicillin and cefotaxime can also be used, recognizing the potential for the introduction of cephalosporin-resistant gram-negative isolates in the unit. Although used in the past, neither intrathecal nor intraventricular administration of antibiotics has been found to improve the morbidity or mortality of gram-negative meningitis (Shah et al, 2008). Once the microorganism has been identified and the antibiotic susceptibility results are available, either a single drug or a combination of drugs found to be effective in vitro should be used. Usually penicillin or ampicillin is used for GBS meningitis; in the first few days of therapy, dual therapy with an aminoglycoside may be used. For L. monocytogenes, it is common to treat with ampicillin with or without gentamicin. Ampicillin or cefotaxime with or without an aminoglycoside can be used for infection with gramnegative enteric bacteria. The precise length of antibiotic therapy depends on the rapidity of response and sterilization of cerebrospinal fluid. In general, continue therapy for approximately 2 weeks after sterilization of the cerebrospinal fluid, or a minimum of 2 weeks for gram-positive meningitis and a minimum of 3 weeks for gram-negative meningitis. In difficult situations, therapy may be required for as long as 4 to 6 weeks. It may be prudent to repeat a cerebrospinal fluid examination and culture after initiation of therapy, especially if the clinical response is less than satisfactory. If organisms are seen on gram-stained smears of the fluid, modification of the therapeutic regimen should be considered. In general, approximately 3 days or more are required for an antibiotic regimen to sterilize the cerebrospinal fluid in infants with gram-negative meningitis. In infants with gram-positive meningitis, sterilization is usually seen within 36 to 48 hours. Neuroimaging should be considered to exclude parameningeal foci and abscess formation and to assist in assessing the infant’s prognosis. PROGNOSIS Complications from neonatal meningitis include brain abscess, communicating or noncommunicating hydrocephalus, subdural effusions, ventriculitis, deafness, and blindness. Generally the severity of complications is related to severity of the disease during the early neonatal period. It is imperative to follow hearing competency and to examine these infants for prolonged periods after recovery. The infant who has experienced meningitis may appear relatively healthy at the time of discharge, and only after careful follow-up do perceptual difficulties, reading problems, or signs of brain damage become apparent. Approximately 40% to 50% of survivors have some evidence of neurologic damage, with severe damage being obvious in 11%. Infants who survive neonatal meningitis should have regular audiology, language, and neurologic evaluations until they enter school (Edwards et al, 1985; Stevens et al, 2003). SUGGESTED READINGS Ansong AK, Smith PB, Benjamin DK, et al: Group B streptococcal meningitis: cerebrospinal fluid parameters in the era of intrapartum antibiotic prophylaxis, Early Hum Dev 85:S5-S7, 2009. Bender L, Thaarup J, Varming K, et al: Early and late markers for the detection of early-onset neonatal sepsis, Dan Med Bull 55:219-223, 2008. Castor ML, Whitney CG, Como-Sabetti K, et al: Antibiotic resistance patterns in invasive group B streptococcal isolates, Infect Dis Obstet Gynecol 727505:2008, 2008. Centers for Disease Control and Prevention: Trends in perinatal group B streptococcal disease: United States, 2000-2006, MMWR Morb Mort Wkly Rep 58:109-112, 2009. Diedrick MJ, Flores AE, Hillier SL, et al: Clonal analysis of colonizing group B Streptococcus, serotype IV, an emerging pathogen in the United States, J Clin Microbiol 48:3100-3104, 2010. Edgar JD, Gabriel V, Gallimore JR, et al: A prospective study of the sensitivity, specificity and diagnostic performance of soluble intercellular adhesion molecule 1, highly sensitive C-reactive protein, soluble E-selectin and serum amyloid A in the diagnosis of neonatal infection, BMC Pediatr 10:22, 2010. Ferrieri P, Baker SJ, Hillier SL, et al: Diversity of surface protein expression in group B streptococcal colonizing and invasive isolates, Indian J Med Res 119:191-196, 2004a. Garges HP, Mood MA, Cotton CM, et al: Neonatal meningitis: what is the correlation among cerebrospinal fluid cultures, blood cultures, and cerebrospinal fluid parameters? Pediatrics 117:1094-1100, 2006. Jordan HT, Farley MM, Craig A, et al: Revisiting the need for vaccine prevention of late-onset neonatal group B streptococcal disease: a multistate, population based analysis, Pediatr Infect Dis J 27:1057-1064, 2008. Koenig JM, Keenan WJ: Group B streptococcus and early-onset sepsis in the era of maternal prophylaxis, Pediatr Clin North Am 56:689-708, 2009. Lam HS, Ng PC: Biochemical markers of neonatal sepsis, Pathology 40:141-148, 2008. Levent F, Baker CJ, Rench MA, et al: Early outcomes of group B Streptococcal meningitis in the 21st century, Pediatr Infect Dis J 29:1009-1012, 2010. Puopolo KM, Eichenwald EC: No change in the incidence of ampicillin-resistant, neonatal, early-onset sepsis over 18 years, Pediatrics 125:e1031-1038, 2010. Verani JR, Schrag SJ: Group B streptococcal disease in infants: progress in prevention and continued challenges, Clin Perinatol 37:375-392, 2010. Wynn JL, Neu J, Moldawer LL, et al: Potential of immunomodulatory agents for prevention and treatment of neonatal sepsis, J Perinatol 29:79-88, 2009. Wynn JL, Seed PC, Cotton CM: Does IVIg administration yield improved immune function in very premature neonates? J Perinatol 30:635-642, 2010. Complete references used in this text can be found online at www.expertconsult.com C H A P T E R 40 Health Care–Acquired Infections in the Nursery David A. Munson and Jacquelyn R. Evans* For decades, nosocomial or health care–acquired infection (HAI) has been considered by many as an unavoidable problem associated with prolonged stays in an intensive care nursery. Immature skin and immune systems, pathologic skin flora, and necessary invasive interventions all put the premature neonate at high risk for HAI. But some centers have clearly shown that, despite these risk factors, HAIs ranging from catheter-associated infections to acquisition of respiratory viruses can largely be prevented (Andersen, 2005; Bloom et al, 2003; Kilbride et al, 2003). There is a direct financial cost to the health care system associated with infections acquired while in the hospital, but more importantly there is growing evidence that the deleterious effects of the inflammatory response to infection may adversely affect long-term outcomes, including increased rates and severity of bronchopulmonary dysplasia and neurodevelopmental impairment (Adams-Chapman and Stoll, 2001, 2006; Gonzalez et al, 1996; Hintz et al, 2005; Jobe and Ikegami, 1996; Leviton et al, 1999; Rojas et al, 1995; Shah et al, 2008; Stoll et al, 2004). The growing numbers of surviving very low-birthweight (VLBW) infants magnifies the impact of this problem. Clinicians must make efforts to minimize the exposure of newborns to known risk factors for infection. Continuous surveillance and monitoring of endemic HAI rates and patterns of responsible pathogens are necessary to establish a reference point in each nursery and facilitate early identification of epidemics. Prevention of infections in the neonatal patient requires an expanding skill set in addition to organized hospital-wide initiatives. Familiarity with quality improvement concepts, standardization of practice, and hospital programs such as antimicrobial stewardship are all essential to preventing harm in the smallest patients. DEFINITIONS The Centers for Disease Control and Prevention (CDC) provides the following surveillance definition for an HAI: illness associated with a pathogen or its toxins that is not present or incubating at the time of admission to the intensive care unit (Horan et al, 2008). As simple as this definition seems, it does leave some room for confusion for the clinician caring for the neonatal population. Infections that manifest early in the 1st week of life are typically related to perinatal risk factors and vertical transmission from the mother. HAIs are more often related to patient colonization and environmental risk factors. Unfortunately, there *Acknowledgment: In writing this chapter, the authors excerpted significant portions of the chapter Nosocomial Infections in the Nursery by Ira Adams-Chapman and Barbara Stoll from the previous edition of this book. is no specific postnatal age that distinguishes maternally transmitted infections from HAIs (Baltimore, 1998), and it may be difficult to differentiate a late-onset, perinatally acquired infection from an HAI. Most sources define HAIs as infections occurring after 3 days of age (Baltimore, 1998; Stoll et al, 2002a). The CDC defines HAI as any infection that occurs after admission to the neonatal intensive care unit (NICU) and that was not transplacentally acquired (Garner et al, 1988). Other authorities suggest that HAI be defined as any infection occurring more than 5 to 7 days after birth (Baltimore, 1998). The majority of HAIs in the neonatal population are bloodstream infections associated with an intravascular device. Criteria for central line–associated bloodstream infections (CLABSIs), as defined by the CDC, are (1) isolation of a recognized pathogen from one blood culture specimen or of a skin commensal from two blood culture specimens; (2) one or more clinical signs of infection, such as temperature instability, apnea, and bradycardia; and (3) the presence of an intravascular device at the time the culture specimen is collected (Horan and Gaynes, 2004). Ventilator-associated pneumonias are difficult to diagnose in the NICU, but can contribute to morbidity, especially in patients with evolving lung disease. Although the use of Foley catheters is low in the NICU compared with other intensive care units, the acquisition of urinary tract infections while in the NICU is still a problem. A summary of definitions of important HAIs in infants is provided in Box 40-1. The prevention of the spread of respiratory viruses is another area that demands attention, especially in a unit caring for infants with significant bronchopulmonary dysplasia. The NICU, as with other intensive care units, must remain vigilant in the control of bacteria resistant to antibiotics such as methicillinresistant Staphylococcus aureus (MRSA) and vancomycinresistant enterococcus (VRE). As community-acquired MRSA is becoming more prevalent, it is becoming more difficult to determine the origin of MRSA in a neonatal patient. INCIDENCE WELL BABY (HEALTHY NEWBORN) NURSERY Accurate rates of HAI in the well baby nursery are difficult to ascertain because there is no systematic reporting or surveillance network for this issue; however, the incidence appears to be low. Some researchers estimate rates of less than 1 per 100 patients discharged (Baltimore, 1998). Typical risk factors for acquiring a HAI, such as invasive procedures and presence of an intravascular device, are uncommon in the healthy newborn population. Discharge 551 552 PART IX Immunology and Infections from the hospital within 48 hours of birth and rooming-in practices have helped further to decrease the risk of exposure in modern mother-baby units. Early discharge may also limit surveillance efforts, because patients may be discharged home before they become symptomatic from HAIs (Goldmann, 1989). NEONATAL INTENSIVE CARE UNIT The majority of neonatal HAIs occur in term and preterm infants hospitalized in special care nurseries. It is difficult to determine and compare reports of endemic HAI rates in different NICU populations. Any comparison of surveillance data between institutions is limited by differences in patient demographics (e.g., birthweight, BOX 40-1 D  efinitions of Nosocomial ­Bacteremia and Pneumonia for Patients Younger than 12 Months, from the Centers for Disease ­Control and Prevention NOSOCOMIAL BLOODSTREAM INFECTIONS Recognized pathogen isolated from blood culture and pathogen is not related to infection at another site. ll Must have at least one of the following clinical symptoms: fever (body temperature >38°C, rectal), hypothermia (body temperature <37°C, rectal), apnea, bradycardia, and a skin commensal isolated from two blood culture specimens on separate occasions. ll PNEUMONIA Chest radiograph with new or progressive infiltrate, cavitation, consolidation, or pneumatoceles ll Worsening gas exchange and three of the following: ll Temperature instability ll White blood cell count <4000 or >15,000 with ≥10% bands ll New-onset purulent sputum, change in character of sputum, increased respiratory secretions, or increased suctioning requirements ll Physical examination findings consistent with increased work of breathing or apnea ll Wheezing, rales, or rhonchi ll Cough ll Bradycardia or tachycardia ll URINARY TRACT INFECTION Clinical symptoms (fever [body temperature >38°C, rectal], hypothermia [body temperature <37°C, rectal], apnea or bradycardia, dysuria, lethargy or vomiting) and positive urine culture (≥100,000 microorganisms/mL) with no more than two species identified ll May also qualify as a UTI if the above symptoms are present without another cause and has at least one of the following ll Positive dipstick for leukocyte esterase, nitrate, or both ll Pyuria ll Organisms seen on Gram stain of unspun urine ll At least two urine cultures with the same uropathogen with ≥100 colonies/mL ll ≤100,000 colonies/mL of a single uropathogen if the patient is receiving an effective antimicrobial agent ll Physician diagnosis of UTI ll Physician institutes appropriate therapy for a UTI ll Modified from Horan TC, Andrus M, Dudeck MA: CDC/NHSN surveillance definition of health care-associated infection and criteria for specific types of infections in the acute care setting, Am J Infect Control 36:3091-332, 2008. UTI, Urinary tract infection. gestational age distribution, underlying severity of disease, birth in the hospital versus elsewhere), back transport policies, and the use of different definitions for HAI. Among a cohort of 6215 VLBW infants (weight <1500 g) who were monitored by the National Institute of Child Health and Human Development (NICHD) Neonatal Research Network, 21% of those who survived beyond 3 days of age had at least one episode of late-onset sepsis (Stoll et al, 2002b). There was significant variability among the participating centers, with rates ranging from 10.6% to 31.7% at individual sites (Stoll et al, 2002b). Moreover, rates were inversely related to birthweight and gestational age, ranging from 43% for infants weighing 401 to 750 g at birth to 7% for infants weighing 1251 to 1500 g at birth. These data are similar to those reported by Brodie et al (2000). In their study, 19% of 1354 infants weighing less than 1500 g at birth had HAIs, and the rates were highest for infants at the lowest birthweights (39%, <750 g; 27%, 750 to 999 g; 10%, 1000 to 1499 g). A point prevalence survey conducted by the 29 level II to level IV nurseries participating in the Pediatric Prevention Network revealed a prevalence of 11.4% for HAI (Sohn, 2001). The CDC National Healthcare Safety Network (NHSN) was developed in 2005 and replaced the National Nosocomial Infections Surveillance system that reported national data before that time. Some investigators believe that reporting overall incidence rates may be misleading, because of the wide variations in practice and patient populations in different units; therefore the NHSN system monitors device-associated HAI rates by using an approach that accounts for variability in device use and length of hospital stay (Edwards et al, 2008; Emori et al, 1991; Gaynes et al, 1996). These data are also stratified by birthweight categories and expressed as incidence density per 100 or 1000 patient-days. These two adjustments modify the relative risk on the basis of the severity of the illness and the duration of exposure to the risk factor. A number of states in the United States have mandatory reporting of HAIs through the NHSN, and the number of participating medical centers is growing. The NHSN data are similar to institutional and collaborative epidemiologic data reported elsewhere. Rates of device-associated infections are presented in Table 40-1. These rates remained constant in the 1990s (Gaynes et al, 1996; National Nosocomial Infections Surveillance System Report, 2000), but appear to have dropped considerably in the 2007 report (Edwards et al, 2008), suggesting that preventive strategies for CLABSIs may be having an effect. ANATOMIC SITES OF INFECTION WELL BABY NURSERY HAIs in the well baby nursery are most commonly superficial, involving the skin, mouth, or eyes. These infections include omphalitis, pustules, abscesses, and bullous impetigo (Goldmann, 1989). Nursery epidemics of diarrhea caused by bacterial and viral enteropathogens have been reported, but they occur infrequently (Goldmann, 1989). CHAPTER 40 Health Care–Acquired Infections in the Nursery NEONATAL INTENSIVE CARE UNIT National and institutional surveillance data demonstrate that bloodstream infections are responsible for the majority of HAIs among NICU patients (Gaynes et al, 1996; National Nosocomial Infections Surveillance System Report, 2000; Sohn et al, 2001). The remaining cases involve the respiratory tract, eye, ear, nose, throat, or gastrointestinal tract (Figure 40-1). Surveillance data for meningitis are limited. Reports suggest that late-onset meningitis may be underdiagnosed in the high-risk population of VLBW infants (Stoll et al, 2002b). There are widespread differences in clinical practice regarding the inclusion of a lumbar puncture with cerebrospinal fluid TABLE 40-1 Rate of Device-Associated HAIs Birthweight (g) NHSN 2008 CLABSI per 1000 catheter days NHSN 2008 VAP per 1000 ­ventilator days <750 3.7 2.6 751-1000 3.3 2.1 1001-1500 2.6 1.5 1501-2500 2.4 1.0 >2500 2.0 0.9 Modified from Edwards JR, Peterson KD, Andrus ML, et al: National Healthcare Safety Network (NHSN) Report, data summary for 2006 through 2007, issued November 2008. National Healthcare Safety Network Facilities, Am J Infect Control 36:609-626, 2008. Level III NICU pooled data. CLABSI, Central line–associated bloodstream infection; HAI, health care–acquired infection NHSN, National Healthcare Safety Network; VAP, ventilator-associated pneumonia. 553 analysis in the evaluation of a neonate with possible sepsis. This variation in practice has been challenged by recent data suggesting that generally accepted cerebrospinal fluid parameters are not sensitive or specific for meningitis in the neonate (Smith et al, 2008). Although early onset meningitis remains a rare event, clinicians need to consider including a lumbar puncture in the initial evaluation of a neonate with signs and symptoms of infection. RISK FACTORS Some risk factors for acquiring an HAI are reflections of the patient population and are difficult or impossible to modify. Other risk factors are directly related to the level of supportive care associated with intensive care medicine and may be difficult yet possible to modify. Risk factors for HAI include lower gestational age and low birthweight; invasive procedures; invasive devices, especially intravascular catheters and endotracheal tubes; parenteral nutrition and intravenous lipids; colonization of skin, gastrointestinal tract, and airway with invasive organisms; selected drugs administered to the neonate; and issues surrounding nursery staffing, nursery crowding, and hand washing (Box 40-2) (Gaynes et al, 1996; Kawogoe et al, 2001; Perlman et al, 2007; Stoll et al, 1996, 2002a; Stover et al, 2001; Suara et al, 2000). Minimizing exposure to known risk factors for infection is important to reduce the rate of HAIs in the nursery. The risk of developing an HAI is inversely related to gestational age and birthweight (Gaynes et al, 1996, Kawogoe et al, 2001; National Nosocomial Infections Surveillance System Report, 2000; Stoll et al, 1996, 2002a; Stover 1000 g 1501–2500 g BSI (49%) BSI (32%) EENT (21%) SST (6%) EENT (8%) SSI (1%) GI (7%) PNEU (16%) Other (13%) GI (11%) SST (10%) SSI (3%) Other (10%) PNEU (13%) (N = 3,987) (N = 3,547) 1001–1500 g >2500 g BSI (45%) BSI (36%) EENT (13%) EENT (14%) SST (7%) SSI (1%) GI (10%) PNEU (12%) Other (11%) SST (9%) GI (5%) SSI (7%) Other (12%) PNEU (18%) (N = 1,881) (N = 3,764) FIGURE 40-1 Distribution of health care–acquired infections by site of infection. (Data from Gaynes RP, Edwards JR, Jarvis WR, et al: Nosocomial infections among neonates in high-risk nurseries in the United States, Pediatrics 98:357-361, 1996.) 554 PART IX Immunology and Infections BOX 40-2 R  isk Factors for Acquiring ­Health Care–Acquired Infections in the ­Neonatal Intensive Care Unit ll ll ll ll ll ll ll ll ll ll ll ll ll ll ll ll ll Prematurity Low birthweight Invasive device Intravascular device (CVC, PAL, PICC, PIVC, umbilical catheter) Mechanical ventilator Urinary catheter Ventriculoperitoneal shunt Medications Histamine2-blocking agents Steroids Others Prolonged administration of hyperalimentation Intralipid administration Delayed enteral feedings Feeding with formula rather than human milk Inadequate nursery staffing and overcrowding Poor compliance with handwashing CVC, Central versus catheter; PAL, percutaneous arterial line; PICC, peripherally inserted central catheter; PIVC, percutaneous intravenous catheter. et al, 2001; Suara et al, 2000). Previous reports estimate that there is a 3% lower risk of acquiring an HAI with each 500-g increment in birthweight (Goldmann, 1989). Infants with birthweights less than 1000 g have twice the rate of nosocomial bloodstream infections than do infants with birthweights greater than 1000 g (Brodie et al, 2000; Geffers et al, 2008; Gaynes et al, 1996; National Nosocomial Infections Surveillance System Report, 2000; Stoll et al, 2002a; Stover et al, 2001). Severity of illness scores may be more predictive than birthweight alone, because they reflect physiologic stability and the cumulative need for intervention and invasive therapies (Goldmann, 1989; Gray et al, 1995). The use of any type of invasive device increases the risk for infection. The most common invasive devices used in the nursery are intravascular catheters, mechanical ventilators, ventriculoperitoneal shunts, and urinary catheters. In general, the risk rises as the duration of exposure lengthens. Compared with adult patients, neonates are at higher risk for catheter-related bloodstream infections and at lower risk for ventilator-associated pneumonia and urinary tract infections (Langley et al, 2001; National Nosocomial Infections Surveillance System Report, 2000). These patterns correlate with the frequency with which these invasive devices are used in the neonatal patient population. Prolonged duration of mechanical ventilation is the primary risk factor for development of hospital-acquired pneumonia. Contamination of respiratory equipment— especially with gram-negative organisms that thrive in moist environments, such as Acinetobacter, Pseudomonas, and Flavobacterium spp.—frequently leads to colonization of the respiratory tract. There is some evidence that aspiration of gastric and oropharyngeal secretions around the uncuffed endotracheal tube may occur, and this could be a mechanism for ventilator-associated pneumonia (Farhath et al, 2008). The various closed-system suctioning devices may decrease the risk of iatrogenic contamination during suctioning. Defining a ventilator-associated pneumonia in the neonatal population remains a challenge. Patients in the NICU do not often meet all of the parameters defined by the CDC (see Box 40-1), but acute worsening of pulmonary function in an infant with a ventilator could represent a new-onset pneumonia. Collaboration among the largest NICUs is needed to better define this entity in the neonatal population. Intravascular devices commonly used in the neonatal population are peripheral intravenous catheters (PIVs), umbilical catheters, peripherally inserted central catheters (PICCs), surgically placed central venous catheters (CVCs) and percutaneous arterial catheters. Regardless of the type of device used, the rate of catheter-related bloodstream infections is directly related to the number of days the catheters are in place and inversely related to the gestational age and birthweight of the patients (Gaynes et al, 1996; Kawogoe et al, 2001; National Nosocomial Infections Surveillance System Report, 2000; Stoll et al, 2000a; Suara et al, 2000). Coagulase-negative Staphylococcus (CONS) remains the primary pathogen associated with catheter-related bloodstream infections (Garland et al, 2001). PIVs are the most commonly used device for vascular access in neonates. In adults, the removal of such catheters after 72 hours is recommended. Data in neonates are insufficient to recommend elective removal of PIVs after 72 hours, because studies have not shown a clear correlation between the higher colonization rate noted after 72 hours and an increased rate of catheter-related bloodstream infection (Oishi, 2001; Pearson, 1996). Further study is needed to answer this question. Data comparing the infection rates of the various types of intravascular catheters are limited. Theoretically the risk should be lower for tunneled catheters, because the Dacron cuff proximal to the exit site of a surgically placed catheter can inhibit the migration of organisms into the catheter tract (Mermel et al, 2001). Adult data suggest that tunneled catheters have lower infection rates than nontunneled catheters; however, a report of 79 surgical neonates requiring tunneled lines showed a rate of infection of 9.9 per 1000 catheter days (Klein et al, 2003), comparable to or worse than reported rates of PICC line infections in other NICU populations. This issue needs further study in the neonatal population, especially because PICCs are being used with greater frequency for long-term vascular access in neonates (Stoll et al, 2002a). Some recent trials comparing PIVs to PICC lines for the provision of parenteral nutrition have challenged the notion that the presence of a central line increases the risk of infection above the risk that exists simply with the use of peripheral catheters. It appears that the use of a PICC line decreases the risk of complications associated with PIVs without increasing the rate of infection, at least in the initial weeks of life (Ainsworth et al, 2007). There is certainly much work to be done to determine the ideal access for the different populations cared for in the NICU. The use of intravenous lipid emulsions increases the risk for infection (Freeman et al, 1990). Lipid emulsions decrease the flow rate through the intravenous catheter and potentiate the growth and proliferation of some microorganisms. Lipid emulsions can interfere with host 555 CHAPTER 40 Health Care–Acquired Infections in the Nursery defense mechanisms by impairing the function of neutrophils and reticuloendothelial macrophages (Freeman et al, 1990; Langevin et al, 1999; Nugent, 1984). Extrinsic contamination has been reported but rarely occurs in the United States (Hernandez-Ramos et al, 2000). Freeman et al (1990) reported that the use of a lipid emulsion was positively and independently predictive for the development of CONS bacteremia. Infants who demonstrated CONS sepsis were 5.8-fold more likely to have been exposed to lipid emulsions. Administration of such emulsions has also been linked to a higher risk for HAI with Candida and Malassezia spp. in neonates (Long and Keyserling, 1985; Redline et al, 1985; Saiman et al, 2000). Histamine-blocking agents and postnatally administered corticosteroids are the medications most commonly associated with an increased risk of HAIs among newborns. It is hypothesized that the reduced gastric pH associated with the use of histamine2-blocking agents promotes bacterial overgrowth and invasion of pathogenic bacteria (Beck-Sague et al, 1994; Stoll et al, 1999). Dexamethasone has been used in ventilator-dependent preterm infants to facilitate weaning from the ventilator and to minimize the risk of chronic lung disease. The use of dexamethasone has decreased in the last decade in VLBW infants secondary to concerns of spontaneous bowel perforation and adverse effects on growth and neurodevelopmental outcome (O’Shea et al, 1999; Stark et al, 2001, Vohr et al, 2000). However, there is renewed interest in a role for steroids after the first week of life, especially in infants who have demonstrated a high risk for developing chronic lung disease (Doyle et al, 2005, 2006, 2007). Because the use of dexamethasone has been associated with an increased risk of infection in VLBW infants (Stoll et al, 1999; Yeh et al, 1997), neonatologists will need to include this concern in any risk-benefit analysis of the use of steroids in their patients. Nursery design and staffing influence the risk of infection. Overcrowding and larger workloads decrease compliance with hand washing and raise the risk of HAI (Archibald et al, 1997; Fridkin et al, 1996; Harbarth et al, 1999; Robert et al, 2000; Vicca, 1999). Inadequate numbers of staff and the use of temporary or inexperienced staff members both adversely affect the rate of infection. The adverse effects of inadequate staffing and overcrowding were demonstrated by Hayley and Bregman (1982), who showed a relationship between nurse-to-patient ratio and colonization of patients with MRSA. Fridkin et al (1996) found that patient-to-nurse ratio was an independent predictor of development of a catheter-related bloodstream infection. Furthermore, strategic nursery design and improvement in nursing staffing correlate with lower rates of HAIs (Gladstone et al, 1990). DISTRIBUTION BY PATHOGEN The predominant pathogens responsible for nosocomial bloodstream infections have changed over time. Goldmann (1989) proposed that these trends are explained by changes in the neonatal intensive care patient population and advancing technology. S. aureus was the most common nosocomial pathogen in the 1950s and 1960s. In the 1960s and 1970s, gram-negative organisms emerged TABLE 40-2 Distribution of Pathogens Responsible for Bloodstream Infections Pathogen No. % Coagulase-negative staphylococci 3833 51.0 Staphylococcus aureus 563 7.5 Group B streptococci 597 7.9 Enterococcus spp. 467 6.2 Candida spp. 518 6.9 Escherichia coli 326 4.3 Other Streptococcus spp. 205 2.7 Enterobacter spp. 219 2.9 Klebsiella pneumoniae 188 2.5 Data from Gaynes RP, Edwards JR, Jarvis WR, et al: Nosocomial infections among neonates in high-risk nurseries in the United States, Pediatrics 98:357-361, 1996. as the predominant pathogens; globally, these organisms remain the most important pathogens responsible for HAIs in the nursery (Stoll, 2001). National surveillance data in the United States indicate that CONS is currently the most common nosocomial pathogen (Gaynes et al, 1996; National Nosocomial Infections Surveillance System Report, 2000) (Table 40-2). In the United States, the distribution of pathogens has not changed significantly over the past decade (Gaynes et al, 1996; Stoll et al, 1996, 2002a) (Box 40-3). Among a cohort of infants with birthweights less than 1500 g with late-onset infections, the NICHD Neonatal Research Network reported that gram-positive organisms were responsible for 70% of cases, gram-negative organisms for 18%, and fungi for 12% (Stoll et al, 2002a). These findings are similar to those of a 10-year retrospective analysis of pathogens in a single center, in which gram-positive organisms caused 57% of late-onset infections (Karlowicz et al, 2000). The distribution of infecting pathogens did not differ with birthweight or timing of infection (Stoll et al, 2002a). More recent articles evaluating interventions and mechanisms of infection confirm the continued predominance of CONS in neonatal line infections in the United States (Garland et al, 2008) as do review articles (Curtis and Shetty, 2008). For the clinician, understanding the specific colonization and resistance patterns in the individual NICU is perhaps more important than being aware of national trends. The emergence of nosocomial pathogens with antimicrobial resistance is a concern. In patients of all ages, reports estimate that 50% to 60% of HAIs are caused by resistant organisms (Jones, 2001; Weinstein, 1998). Organisms showing patterns of increasing antibiotic resistance of importance to NICU patients are VRE, MRSA, and multidrug-resistant gram-negative organisms (Bizzarro and Gallagher, 2007). GRAM-POSITIVE BACTERIA CONS is the most common endemic nosocomial pathogen in neonates (Brodie et al, 2000; Garland et al, 1996, 2008; Gray et al, 1995; National Nosocomial Infections Surveillance system report, 2000; Stoll et al, 2002a). The majority of CONS infections are bloodstream infections. Reported 556 PART IX Immunology and Infections BOX 40-3 C  ommon Pathogens Causing Health Care–Acquired Infections GRAM-POSITIVE ORGANISMS Coagulase-negative staphylococci ll Staphylococcus aureus ll Enterococcus spp. ll Group B Streptococcus ll GRAM-NEGATIVE ORGANISMS Escherichia coli ll Klebsiella spp. ll Pseudomonas spp. ll Enterobacter spp. ll Serratia spp. ll Haemophilus spp. ll Acinetobacter spp. ll Salmonella spp. ll FUNGAL ORGANISMS Candida albicans ll Candida parapsilosis ll Malassezia furfur ll VIRAL ORGANISMS Respiratory syncytial virus ll Influenza ll Varicella zoster ll Rotavirus ll Enterovirus ll Data from Gaynes RP, Edwards JR, Jarvis WR, et al: Nosocomial infections among neonates in high-risk nurseries in the United States, Pediatrics 98:357-361, 1996; and Stoll BJ, Hansen N, Fanaroff AA, et al: Late-onset sepsis in very-low-birth-weight neonates: the experience of the NICHD Neonatal Research Network, Pediatrics 110:285-291, 2002. incidence ranges from 51% to 78% among VLBW infants (Gray et al, 1995; Isaacs et al, 1996; Stoll, 1996, 2002a). Gray et al (1995) reported a cumulative incidence of 17.5 episodes of CONS sepsis per 100 patient-days and an incidence density of 6.9 episodes of CONS sepsis per 1000 patient-days. Known risk factors for CONS infection are low birthweight, lower gestational age, use of central venous catheters, prolonged administration of hyperalimentation, use of intravenous lipid emulsions, postnatal administration of corticosteroids, and prolonged hospital stay (Brodie et al, 2000; Freeman et al, 1990; Goldmann, 1989; JohnsonRobbins et al, 1996). CONS is the pathogen most commonly associated with catheter-related infections, partially because it produces a capsular polysaccharide adhesin (poly-N-succinyl glucosamine), which enhances its ability to adhere to intravascular devices. Although some studies suggest that prophylactic use of vancomycin reduces the risk of CONS catheter-related infections, this practice is not recommended because of the serious risk of encouraging antibiotic-resistant organisms, especially VRE and staphylococci. The use of a vancomycin lock, where vancomycin is instilled into the catheter and then removed after a specified dwell time, has more promise in terms of balancing the risk of systemic exposure and prophylaxis against contamination of the catheter lumen. Garland et al (2005) demonstrated a reduction in the rate of CLABSI from 17.8 to 2.4 per 1000 catheter days using this technique. Although promising, several units have achieved comparable reductions without the use of this kind of intervention. Consequently, the role of an antibiotic lock will likely find its place in certain circumstances or in populations that have demonstrated persistently elevated risks for infection. Enterococci are responsible for both endemic and epidemic HAIs in the NICU. Use of central venous catheters, prolonged hospital stay, and prior antibiotic use are recognized risk factors for colonization with these organisms. The gastrointestinal tract is often the primary source of infection; however, the pathogens are typically spread via the hands of health care workers or through environmental contamination. The widespread use of antibiotics has led to the emergence of VRE. There are published guidelines to prevent the spread of VRE, which include hand washing, isolation, barrier precautions, and cohorting of infected patients (Gross and Pujat, 2001; Hospital Infection Control Practices Advisory Committee, 1995). Educational programs to limit the indiscriminate use of antibiotics have been effective in decreasing the spread of VRE (Goldmann et al, 1996; Isaacs, 2000). S. aureus has caused epidemics in well baby nurseries and in NICUs. In the nursery, the major reservoirs for staphylococci are colonized or infected infants. These infants transmit the organism to health care workers, who subsequently infect other infants. The skin, nares, and umbilicus are the most common sites of colonization. Unfortunately, routine surveillance cultures are not useful because colonization rates correlate poorly with infection rates. Studies in adult patients have shown that many patients who demonstrate S. aureus bacteremia were colonized with the identical strain at the time of admission to the hospital, suggesting that some of the infections with S. aureus are community acquired rather than hospital acquired (von Eiff et al, 2001). Similarly, health care workers can be colonized with a community-acquired strain of S. aureus, which they then transfer to vulnerable infants. Importantly, there has been a significant increase in methicillin-resistant strains of S. aureus causing infections in the NICU. A review of the National Nosocomial Infections Surveillance data from 1995 to 2004 indicated a 308% increase in the incidence of MRSA over that 10-year time frame. As of the 2004, MRSA accounted for 34% of S. aureus infections in the reported data (Lessa et al, 2009). Consequently, when covering for a possible S. aureus infection, it is critical to select an antibiotic that is effective against methicillinresistant strains. Group B streptococcus (GBS) remains an important cause of late-onset infection in neonates, but do not have a clear role as an HAI. Intrapartum antibiotic treatment to prevent vertical transmission of GBS from a colonized mother to her infant has led to a decrease in early-onset GBS disease. In contrast, the incidence of late-onset GBS disease has remained unchanged, presumably because prophylaxis does not eradicate colonization of the genital tract or the environment. GRAM-NEGATIVE ORGANISMS Gram-negative organisms are a particularly important cause of nosocomial bloodstream infections, pneumonia, and meningitis because they generally cause severe disease. CHAPTER 40 Health Care–Acquired Infections in the Nursery Escherichia coli is the most common gram-negative pathogen. Other gram-negative organisms responsible for HAI are Klebsiella, Pseudomonas, Enterobacter, Acinetobacter, Serratia, Haemophilus, and Salmonella spp. (see Box 40-3). The attributable mortality is much higher for gram-negative infections than for gram-positive infections. Stoll et al (2002a) reported that infants with gram-negative infections had a 3.5-fold higher risk of death. Karlowicz et al (2000) found that gram-negative infections were associated with fulminant death within 48 hours of a positive blood culture result in 69% of cases. Pseudomonas spp. appear to be particularly virulent, causing death in 42% to 75% of infected neonates (Karlowicz et al, 2000; Leigh et al, 1995; Stoll et al, 2002a). A more recent study by Makhoul et al (2005) and colleagues in Israel confirms that gram-negative infections are associated with a substantially higher mortality. FUNGAL ORGANISMS Fungal infections are discussed in detail in Chapter 41. Invasive fungal infection is estimated to occur in 1% to 4% of VLBW infants and the incidence in extremely lowbirthweight (ELBW) infants has been reported to be significantly higher. The risk may even approach 20% in infants with a birthweight less than 750 g (Bartels et al, 2007; Benjamin et al, 2006; Clerihew et al, 2006; Fridkin et al, 2006; Makhoul et al, 2002, 2007; Rodriguez et al, 2006; Stoll et al, 2002a). Rates and predominant fungal species vary among clinical centers. The smallest and most premature infants appear to be at the highest risk, particularly when they are exposed to broad-spectrum antibiotics or long courses of antibiotics. Other identified risk factors are prolonged mechanical ventilation, prolonged use of central venous catheters, prior use of lipid emulsions, and the use of histamine type 2 antagonists (Benjamin et al, 2001; Long and Keyserling, 1985; Makhoul et al, 2002; Saiman et al, 2000). Efforts at preventing nosocomial fungal infection have focused on the use of prophylactic fluconazole for VLBW infants. This strategy has been shown to decrease the rate of invasive disease in randomized trials, but the baseline rates in the control arms in some studies have been high (Kaufman et al, 2001). Consequently it has been difficult to determine how generalizable the findings are. Furthermore, metaanalysis has not demonstrated an associated improvement in mortality (Clerihew et al, 2007). A number of retrospective studies describing the implementation of fluconazole prophylaxis in individual units have also suggested a reduction in invasive fungal disease (Aghai et al, 2006; Bertini et al, 2005; Dutta et al, 2005; Healy et al, 2005; Manzoni et al, 2006; Uko et al, 2006). Although no study has found an increase in fluconazole-resistant fungus, resistance remains a concern with any prophylactic strategy. Limiting a prophylactic strategy to units with a high rate of invasive candidal disease and to patients with the highest risk will likely ultimately provide the greatest benefit with the lowest risk. VIRAL ORGANISMS Viral organisms that cause HAI in the NICU include respiratory syncytial virus (RSV), rhinovirus, metapneumovirus, influenza, varicella, rotavirus, and enterovirus. 557 Isolated infections generally result from contact with infected caregivers or family members. Nursery epidemics may occur in addition to isolated individual cases. Respiratory Syncytial Virus Careful hand washing is the most effective measure to prevent RSV infection. Recommendations for preventing RSV epidemics in the inpatient setting include cohorting of patients, barrier precautions, and careful hand washing (Goldmann, 2001; Hall, 2000; Karanfil et al, 1999; Mlinarić-Galinović and Varda-Brkić, 2000; Snydman et al, 1988). Rapid testing to detect the virus in nasal washings facilitates efforts to cohort infected patients, which has been shown to be an effective control measure (Madge et al, 1992; Snydman et al, 1988). RSV is a fastidious organism capable of surviving on inanimate objects for prolonged periods; therefore some authorities advocate the use of gowns and gloves, because of the increased risk of contamination through casual contact between the patient and the environment (Goldmann, 2001; Hall, 2000; Karanfil et al, 1999; Mlinarić-Galinović and Varda-Brkić, 2000). Several case reports have described the use of palivizumab, an RSV monoclonal antibody, to control nosocomial outbreaks; however, the efficacy of administering monthly injections to hospitalized patients has not been critically or systematically evaluated (Cox et al, 2001; Macartney et al, 2000). Preterm infants born before 32 weeks’ gestation and those with chronic lung disease remain at high risk for RSV infection after hospital discharge; therefore the American Academy of Pediatrics (AAP) recommends that all highrisk infants (<32 weeks’ gestation, chronic lung disease, or asymptomatic acyanotic congenital heart disease, such as a patent ductus arteriosus or ventricular septal defect) receive up to five doses of palivizumab during the RSV season (American Academy of Pediatrics, 1998, 2009a; Meissner et al, 1999). These guidelines were updated in 2003 to also include patients with cyanotic congenital heart disease. They have been modified again in 2009 to clarify the risk factors for the 32- to 35-week gestational age group (American Academy of Pediatrics, 2009a). Influenza Influenza is spread primarily via airborne transmission. Hand washing and immunization of health care workers are the primary tools to prevent nosocomial spread of this virus (Nichol and Hauge, 1997). Most infection control guidelines recommend that every health care worker wear a mask during contact with infected patients. Although several drugs are available for the prophylaxis and treatment of influenza, they have not been studied in newborns and cannot be recommended at this time (Meissner, 2001). At-risk infants should receive the influenza vaccine during the winter months once they reach the age of 6 months. With the appearance of the novel H1N1 influenza strain, the season for possible flu exposure has increased substantially. In addition, newborns are thought to be in a high-risk population if they are born to mothers who are actively infected. Guidelines for minimizing risk to the infant have been offered by the CDC. Pregnant women 558 PART IX Immunology and Infections with symptoms that are consistent with H1N1 influenza should receive treatment as soon as possible. If tolerated, the mother should wear a mask during labor and delivery. After the delivery, the infant should be cared for away from the mother until he or she has received 48 hours of treatment and has become afebrile and able to control cough and secretions. The infant should be assisted in lactation if she intends to breastfeed, because the breast milk itself is not thought to be a means of viral transmission (CDC, 2009). Varicella Potential exposures and epidemics of nosocomial varicella have been reported, but true nosocomial infection in the NICU is rare. The index case is typically an asymptomatic infected health care worker or family member who has had contact with a susceptible infant before the onset of clinical disease in the health care worker or family member. The incubation period lasts 14 to 16 days, and infected persons are contagious 24 to 48 hours before the appearance of the rash. The relative risk of infection varies according to the intensity of the exposure and the presence of maternal antibody in the infant. The majority of transplacental antibody transfer occurs during the third trimester; therefore most extremely preterm infants are born before this process is complete. ELBW infants may be at risk even if their mothers have a documented titer of varicella antibodies. The AAP recommends administration of varicella zoster immunoglobulin (VZIG) to any newborn whose mother shows signs of varicella infection within 5 days before delivery or within 2 days after delivery. If the exposure is postnatal, whether via a family member or a health care worker, an exposed preterm infant born before 28 weeks’ gestation or weighing less than 1000 g birthweight should receive immunoprophylaxis regardless of the mother’s varicella history or serologic status. A postnatally exposed preterm infant born after 28 weeks’ gestation should receive VZIG only if the mother lacks clinical or serologic evidence of prior disease (American Academy of Pediatrics, 2009b). Airborne and contact precautions are recommended for all infants with active varicella disease for at least 5 days after vesicles appear and until all lesions have crusted. Airborne and contact precautions should be used for any exposed susceptible patient from 8 to 21 days after the exposure. Administration of VZIG potentially prolongs the incubation period; therefore all exposed infants who receive this product should be isolated for up to 28 days after the exposure. An infant born to a mother with active varicella should be isolated for 21 days, or for 28 days if the infant received VZIG. The infants should also be isolated from the mother until all of her lesions have crusted. Rotavirus Although rare, epidemics of rotavirus diarrhea may occur in the nursery (Herruzo et al, 2009; Jain and Glass, 2001; Lee et al, 2001; Widdowson et al, 2000); they are primarily caused by inadequate hand washing and cross-contamination between patients. Standard and contact precautions should be followed throughout the duration of the illness. Some patients have prolonged fecal shedding of low concentrations of the virus; therefore some infection control experts recommend contact precautions for the duration of the hospitalization of such patients. Rotavirus is an important cause of diarrhea in older infants and should be suspected with any apparent epidemic of diarrhea. There are now two live virus vaccines available for the prevention of rotavirus infection, but their use in the NICU remains controversial because of concerns about the possibility of illness occurring in immunocompromised infants via spread of the vaccine related virus. Currently the AAP recommends the administration of rotavirus vaccine at the time of discharge or after discharge (American Academy of Pediatrics, 2009c). Enterovirus There are numerous serotypes of enteroviruses, including polioviruses, Coxsackie viruses A and B, echovirus, and nonassigned subtypes (Chambon et al, 1999). Enterovirus infections have been described among neonates in the well baby nursery and the neonatal intensive care setting (Chambon et al, 1999; Isaacs et al, 1989; Sizun et al, 2000; Takami et al, 2000, Wreghitt et al, 1989). Both individual cases and epidemics can occur. The seasonal distribution of enterovirus infections in neonates mirrors what is seen in the community. The clinical presentation associated with enteroviral infection is variable. Many patients are asymptomatic; however, several case reports describe overwhelming infection with multisystem organ dysfunction and death (Jankovic et al, 1999; Keyserling, 1997). Clinical manifestations of severe disease include meningoencephalitis, hepatic dysfunction, disseminated intravascular coagulation, myocarditis, pneumonitis, gastroenteritis, and muscle weakness (Abzug et al, 1993; Isaacs et al, 1989; Keyserling, 1997; Wreghitt et al, 1989). The severity of disease and the likelihood of death are often more pronounced in perinatally acquired cases than in nosocomially acquired cases, presumably related to the lack of maternal antibody present in the neonate (Isaacs et al, 1989; Modlin et al, 1981). Infection control measures during an enteroviral epidemic include hand washing with alcohol-based preparations or antimicrobial soaps and cohorting of infected patients. Surveillance cultures of specimens from the throat and rectum may be helpful in identifying asymptomatic colonized infants. Blood and cerebrospinal fluid cultures should be obtained from any patient with clinical symptoms of disease. Polymerase chain reaction analysis is helpful in making a rapid diagnosis (Nigrovic, 2001). No antiviral agents are currently available to treat enteroviral infections in newborns (Nigrovic, 2001). Although commercially available intravenous immunoglobulin preparations have high levels of neutralizing antibodies to common enterovirus serotypes, there is no clear evidence that administration of immunoglobulin alters the process or outcome of enteroviral infection (Abzug et al, 1993; Dagan et al, 1983; Keyserling, 1997); therefore the use of these products remains controversial in patients with enteroviral infection. EPIDEMICS Numerous nursery epidemics have been reported. Common sources for infection are contaminated equipment (i.e., breast pump, thermometer, ventilator), environmental CHAPTER 40 Health Care–Acquired Infections in the Nursery reservoirs, soaps, and lapses in hand hygiene practices (Focca et al, 2000; Hervas et al, 2001; Hoque et al, 2001; Jeong et al, 2001; Jones et al, 2000; McNeil et al, 2001; Reiss et al, 2000; van den Berg et al, 2000; Wisplinghoff et al, 2000; Yu et al, 2000). Clinicians must have a high index of suspicion to detect nursery outbreaks, especially when clusters of infections with unusual pathogens occur in a nursery. Continuous surveillance and monitoring of the endemic infection rates are crucial for determining when there has been a significant change in the baseline pattern of infection for a given nursery. Modern molecular technology facilitates identification and tracking of specific strains of bacteria to help determine the common source of infection (Chambon et al, 1999; Jeong et al, 2001; Jones et al, 2000; Takami et al, 2000). Epidemics must be identified promptly, and immediate control measures must be instituted. Efforts should be made to identify the causal agent and the mode of transmission. Surveillance of staff members may be necessary, even if they are asymptomatic. Reinforcement of hand-washing policies is of utmost importance. Cohorting infected patients may be helpful in limiting the spread of organisms in the nursery. In an effort to minimize the risk of cross-contamination between patients, staff members caring for colonized or infected infants should not care for infants who are not infected or colonized. INFECTION CONTROL The CDC has developed a two-tiered approach to infection control. Standard precautions should be used with all patient contact regardless of the underlying diagnosis or infectious status. These precautions consist of universal precautions (designed to prevent blood and body fluid contamination) and body substance precautions (designed to prevent contamination with moist substances). Transmission-based precautions are necessary when a patient is infected with a known or suspected pathogen that is associated with a high risk of contamination via airborne or droplet transmission or contact with the skin or contaminated surfaces (Garner, 1996). The routine use of gowns is not an effective measure to decrease the endemic HAI rate in the nursery (Garner, 1996; Goldmann, 1989, 1991). Gowns should be used in specific circumstances in which the risk for contamination is high or the infant is being held. Gowns should be changed after each patient encounter. Visitation policies are not restrictive in most modern nurseries. Various regulatory agencies have developed guidelines for sibling visitation (Box 40-4) (American Academy of Pediatrics, 1985). Infection and colonization rates have not risen with use of the current standards. Siblings should be screened for infection or recent exposures before visiting. PREVENTION OF HEALTH CARE–ACQUIRED INFECTIONS Several intensive care nurseries have described their success in decreasing the rate of central line–associated infections (Andersen et al, 2005; Bloom et al, 2003; Kilbride et al, 2003). Until recently it was considered implausible 559 BOX 40-4 G  uidelines for Sibling Visitation in the Nursery ll ll ll ll Siblings should not have been exposed to known communicable diseases (e.g., varicella). Siblings should not have fever or symptoms of acute illness (i.e., upper respiratory infection or gastroenteritis). Children should be supervised by their parents or a responsible adult during the entire visit. Children should be prepared in advance for the visit. Adapted from the American Academy of Pediatrics Committee on the Fetus and Newborn: Postpartum (neonatal) sibling visitation, Pediatrics 76:650, 1985. by many that HAIs could be eliminated completely. But the challenge that faces neonatologists, nurses, respiratory therapists, physical therapists, clerks, environmental service staff, and anyone else who affects the care of the neonate now must be framed by the following question: can we get to zero? Much is known about the mechanisms of catheter-related infections and the factors that increase risk. Effective prevention strategies focus on modifying the known risks, such as standardization of procedures that relate to central line care, strategic nursery design, adequate staffing, hand-washing compliance, minimization of catheter days, and promotion of enteral nutrition, especially with human milk (Adams-Chapman and Stoll, 2002; Goldmann, 1989; Horbar et al, 2001). Care “bundles” incorporating a group of interventions aimed at standardizing central line care can be effective if a particular intensive care unit is committed to implementing them. One hundred percent compliance with adequate hand hygiene is required to eliminate the spread of viruses and bacteria, particularly those that are drug resistant. Monitoring, surveillance, and benchmarking of the HAI rates in the nursery are also critical components of any prevention program (Box 40-5). There is power in simply knowing how an individual unit compares to others, and there is empowerment in knowing that preventing HAIs is possible (Schulman et al, 2009). HAND HYGIENE Historic Perspective Hand washing has clearly been shown to be the most effective and least expensive means of preventing the spread of HAI; however, compliance with this simple practice is poor (Jarvis, 1994; Larson, 1999; Pittet, 2000; Pittet et al, 2000). Historically, the benefits of hand washing have been known since the early nineteenth century. Labarraque, a French pharmacist, was one of the first to demonstrate that cleansing the hands with solutions containing lime or soda could be used as disinfectants and antiseptics (Boyce et al, 2002). The issue was further elucidated by Ignaz Semmelweis in the 1850s, who noted that the mortality and puerperal infection rates were higher among women receiving care from physicians than in those receiving care from midwives. He postulated that the puerperal fever in these patients was caused by “cadaverous particles” transmitted from the autopsy suite on the hands of the students and physicians. After 560 PART IX Immunology and Infections BOX 40-5 P  rinciples for the Prevention of Health Care–Acquired Infections in the NICU ll ll ll ll ll ll ll ll Observe recommendations for standard precautions with all patient contact. Observe recommendations for transmission-based precautions as indicated. ll Gowns ll Gloves ll Masks ll Isolation Use good nursery design and engineering. ll Appropriate nurse-to-patient ratio ll Avoidance of overcrowding and excessive workload ll Readily accessible sinks, antiseptic solutions, soaps, and paper towels ll Maintain hand-washing practices. ll Improving hand-washing compliance ll Washing of hands before and after each patient encounter ll Appropriate use of soap, alcohol-based preparations, or antiseptic solutions ll Alcohol-based antiseptic solution at each patient’s bedside ll Emollients provided for nursery staff ll Education and feedback for nursery staff Minimize risk of contamination of CVCs. ll Maximal sterile barrier precautions during CVC insertion ll Local antisepsis with chlorhexidine gluconate ll Minimal entries into the line for laboratory tests ll Aseptic technique when entering the line ll Minimal CVC days ll Sterile preparation of all fluids to be administered via a CVC Provide meticulous skin care. Encourage early and appropriate advancement of enteral feedings. Provide education and feedback for nursery personnel. Perform continuous monitoring and surveillance of HAI rates in the NICU. CVC, Central venous catheter; HAI, health care–acquired infection; NICU, neonatal intensive care unit. he instructed physicians to cleanse their hands with a chlorine solution between patients, the maternal infection and mortality rates dropped dramatically (Boyce et al, 2002; Semmelweis, 1983). This intervention represents the first clinical evidence suggesting that cleansing contaminated hands with an antiseptic agent was more effective than washing them with plain soap and water and that hand antisepsis could reduce the spread of contagious disease via the hands of health care workers (Boyce et al, 2002). Guidelines for Hand Hygiene Practices in the Neonatal Intensive Care Unit Clinical Indications for Hand Hygiene Hand hygiene techniques are effective in decreasing the colonization rate of resident and transient flora and have been shown to reduce cross-contamination among patients. Several studies have shown that the hands of health care workers become contaminated during routine patient care activities, including “clean” procedures such as lifting patients and checking vital signs, as well as during contact with intact skin (Casewell and Phillips, 1977; Pittet et al, 1999). Attempts have been made to stratify the type of activity with the likelihood of contamination, but they have not been validated by quantitative bacterial contamination analysis. Direct patient contact and BOX 40-6 R  ecommendations for Hand Hygiene Practices in the NICU INDICATIONS FOR HAND HYGIENE Wash hands with a nonantimicrobial soap or an antimicrobial soap and water when hands are visibly soiled or contaminated with proteinaceous material. ll If the hands are not visibly soiled, then alcohol-based, waterless antiseptic agents are strongly preferred for routine decontamination of hands in all other clinical situations. ll Alcohol-based, waterless antiseptic agents should be available at each patient area and other convenient locations, and in individual pocket-sized containers for health care providers. ll Antimicrobial soaps can be considered in settings with few time constraints and easy access to hand hygiene facilities. ll Decontaminate hands after contact with intact patient skin (i.e., checking pulse or lifting). ll Decontaminate hands after contact with body fluids or excretions, mucous membranes, nonintact skin, or wounds. ll Decontaminate hands before applying sterile gloves or inserting a central intravascular catheter. ll Decontaminate hands before inserting indwelling urinary catheters or other invasive devices not requiring surgical procedures. ll Decontaminate hands after removing gloves. ll Decontaminate hands before caring for patients with severe neutropenia or severe immunosuppression. ll Decontaminate hands after contact with inanimate objects in the immediate vicinity of the patient. ll RECOMMENDED TECHNIQUES FOR HAND HYGIENE When using a waterless antiseptic agent, apply enough of the product to cover all surfaces of the hands and fingers, and rub hands together until they are dry. Each manufacturer has guidelines for the volume to be used; in general, enough should be applied so that it takes 15 to 25 seconds to dry. ll When using a nonantimicrobial or antimicrobial soap, wet hands, apply 3 to 5 mL of solution to the hands, and rub for at least 15 seconds. Be sure to cover all surfaces of the hands and fingers. Rinse hands with warm water, and dry thoroughly. Foot pedals, automatic faucets, or towel barriers should be used to turn off the water. ll Adapted from Boyce JM, Pittet D, et al: Guidelines for hand hygiene in health-care settings: recommendations of the Healthcare Infection Control Practices Advisory Committee and the HICPAC/SHEA/APIC/IDSA Hand Hygiene Task Force, Infect Control Hosp Epidemiol 23(Suppl 12):S33-S40, 2002. NICU, Neonatal intensive care unit. respiratory tract care seem to be particularly associated with contamination (Pittet et al, 1999). Organisms such as RSV, S. aureus, and gram-negative bacilli are able to survive on inanimate objects, so changing diapers or holding an infant infected with one of these organisms, and even touching items in the infant’s environment, may result in contamination (Boyce et al, 2002; Goldmann, 1989; Hall, 2000). Recommendations concerning indications and techniques for hand hygiene are summarized in Box 40-6 (Boyce et al, 2002; Garner and Favero, 1986; Larson, 1995). Current recommendations strongly support the use of waterless alcohol-based preparations as the primary agents for hand hygiene, except when the hands are soiled with organic material (Boyce et al, 2002). The bottom line is that hands need to be cleaned before and after every patient contact. Until there is 100% compliance with this simple intervention, HAIs will continue to be a problem. CHAPTER 40 Health Care–Acquired Infections in the Nursery 561 TABLE 40-3 Antimicrobial Spectrum and Characteristics of Hand Hygiene Antiseptic Agents Group* GramPositive Bacteria GramNegative Bacteria Mycobacteria Fungi Viruses Alcohols +++ +++ +++ +++ +++ Fast Optimum concentration 60%90%, no persistent activity Chlorhexidine (2% and 4% aqueous) +++ +++ + + +++ Intermediate Persistent activity, rare allergic reactions Iodine compounds +++ +++ +++ ++ +++ Intermediate Causes skin burns, usually too irritating for hand hygiene Iodophors +++ +++ + ++ ++ Intermediate Less irritating than iodine Phenol derivatives +++ + + + + Intermediate Activity neutralized by nonionic surfactants Triclosan +++ ++ + − +++ Intermediate Acceptability varies Quaternary ammonium compounds + ++ − − + Slow Used only in combination with alcohols, ecologic concerns Speed of Action Comments Reprinted with permission from Boyce JM, Pittet D, et al: Guidelines for hand hygiene in health-care settings: recommendations of the Healthcare Infection Control Practices Advisory Committee and the HICPAC/SHEA/APIC/IDSA Hand Hygiene Task Force, Infect Control Hosp Epidemiol 23(Suppl 12):S33-S40, 2002. +++, Excellent; ++, good, but does not include the entire bacterial spectrum; +, fair; −, no activity or not sufficient. *Hexachlorophene is not included because it is no longer an accepted ingredient of hand disinfection. Preparations for Hand Hygiene Washing the hands with soap causes suspension and mechanical removal of microorganisms and dirt from the hands. Hand disinfection refers to the same process, but with the use of an antimicrobial product to kill or inhibit microorganisms. Table 40-3 compares properties of the various hand hygiene agents. The cleansing activity of plain soap results from its detergent properties (Boyce et al, 2002). Hand washing with soap alone reduces bacterial colonization of the hands, but has no significant antimicrobial activity and is ineffective at removing pathogenic flora. Soaps containing antimicrobial agents are typically used in the NICU environment. The antimicrobial activity of alcohol stems from its ability to denature proteins (Boyce et al, 2002). Because proteins are not readily denatured in the absence of water, solutions containing 50% to 80% alcohol are most effective (Boyce, 2000; Boyce et al, 2002). Alcohol has a rapid onset of action and reduces bacterial colonization, but has no residual activity. The efficacy of alcoholbased products is influenced by the type of alcohol used, concentration, contact time, volume used, and whether the hands are wet when the product is applied (Boyce et al, 2002; Mackintosh and Hoffman, 1984). When used in adequate amounts, alcohol is usually more effective than other hand hygiene products (Boyce, 2000; Boyce et al, 2002). Chlorhexidine gluconate is a cationic bis-biguanide whose antimicrobial activity is caused by attachment and disruption of cytoplasmic membranes (Boyce et al, 2002). The onset of action is slower than the alcohol-based preparations; however, its major benefit is the persistent antimicrobial activity, which may last up to 6 hours after application (Boyce et al, 2002; Larson, 1995). Varying percentages of this product have been added to other hand hygiene preparations, especially the alcohol-based preparations, to confer greater residual activity. Hexachlorophene is a bisphenol compound with bacteriostatic properties. Its activity is due to its ability to inactivate essential enzymes systems. It has good activity against S. aureus, but weak activity against gram-negative bacteria, fungi, and Mycobacterium tuberculosis. Hexachlorophene also has residual activity. Hexachlorophene was used for routine bathing of newborn infants until 1972, when the U.S. Food and Drug Administration (FDA) warned against its use because of an increased occurrence of cystic degeneration of the cerebral white matter in infants who had been bathed in the 3% solution (Shuman et al, 1975). This product should be considered only in term infants during a severe outbreak of S. aureus. Most experts recommend diluting the material 1:4 with water to decrease the risk of systemic absorption. Iodine and iodophors penetrate the cell walls of organisms, impairing protein synthesis and altering the cellular membranes (Boyce et al, 2002). The amount of iodine present determines the level of antimicrobial activity. Combining iodine with polymers (i.e., povidone or poloxamer) increases the solubility, promotes sustained release of iodine, and decreases skin irritation. The activity of this product is affected by pH, exposure time, temperature, the presence of organic (blood or sputum) or inorganic material, and the concentration of iodine. Compliance Despite the fact that the benefits of hand washing have been reported since the nineteenth century, compliance with hand-washing protocols remains unacceptably low. The overall compliance rate is approximately 40% (Pittet, 2000). Reported barriers to compliance with hand hygiene recommendations include skin irritation, poor accessibility of sinks or cleansing agents, greater priority for patient needs, insufficient time, heavy workload and understaffing, and lack of information (Box 40-7). A common misconception is that using gloves obviates the need for adequate hand hygiene. 562 PART IX Immunology and Infections BOX 40-7 F  actors Influencing Compliance With Hand Hygiene Practices OBSERVED RISK FACTORS FOR POOR ADHERENCE TO RECOMMENDED PRACTICES ll Physician status (rather than a nurse) ll Nursing assistant status (rather than a nurse) ll Working in an intensive care unit ll Working during the week compared with the weekend ll Wearing gowns and gloves ll Automated sink faucet ll Activities with high risk of contamination ll High numbers of indications for hand hygiene per hour of patient care SELF-REPORTED FACTORS FOR POOR ADHERENCE TO RECOMMENDED PRACTICES ll Agents are irritating or drying ll Inconvenient location of sinks and supplies ll Inadequate supplies (e.g., soap, paper) ll Too busy ll Patient needs take priority ll Perceived low risk of acquiring infection ll Gloves obviate the need for hand hygiene ll Lack of knowledge of guidelines and recommendations ll Disagreement with recommendations Data from Pittet D: Improving compliance with hand hygiene in hospitals, Infect Control Hosp Epidemiol 21:381-386, 2000; and Boyce JM: Using alcohol for hand antiseptics: dispelling old myths, Infect Control Hosp Epidemiol 21:438-441, 2000. Leakage and contamination of gloves have been reported (Boyce et al, 2002; Larson, 1995). Disposable single-use gloves should be removed after each patient encounter, and hands should be washed before and after their use. Hand hygiene is extremely cost effective. The additional hospital charges associated with a single HAI almost equal the yearly hand hygiene budget. Pittet et al (2000) estimated the cost of a hand hygiene intervention program to be less than $57,000 per year. Assuming that only 25% of the observed decrease in infections was attributable to improved hand hygiene practices, they estimated a savings of $2,100 for every infection averted. PREVENTION OF CENTRAL LINE–ASSOCIATED BLOODSTREAM INFECTIONS The care and maintenance of central venous catheters may affect the risk of catheter-related bloodstream infection. Before insertion of an intravascular device, attempted sterilization of the skin insertion site is of utmost importance. The recommended technique for skin antisepsis is to use two consecutive 10-second applications or a single 30-second application of the selected antibacterial agent. Sterile water rather than alcohol should be used to remove antiseptics from the skin. Povidone-iodine (10%) solution is commonly used for skin antisepsis; however, data suggest that chlorhexidine gluconate is more effective at decreasing skin colonization and subsequent CLABSIs in adults, as well as colonization of peripheral intravenous catheters in neonates (Darmstadt and Dinulos, 2000; Garland et al, 1995). Hub colonization and repeated entry into the line for administration of medications or collection of specimens for laboratory studies are both associated with an increased risk for infection. Salzman et al (1993) conducted a prospective study in which surveillance cultures of catheter hubs were performed three times per week, and blood cultures and hub cultures were conducted for any suspected episode of sepsis. These investigators found that 54% of 28 episodes of CLABSI were preceded by or coincided with colonization of the catheter hub with the same pathogen. This mechanism has been confirmed in the neonatal population by Garland et al (2008). Mahieu et al (2001b) reported that catheter manipulation for blood sampling and fluid or tubing changes raised the risk of colonization. Despite the lack of large randomized controlled trials, the available data suggest that the use of heparin in intravenous fluids is associated with a lower risk of bacterial colonization and thrombosis (Appelgren et al, 1996; Mahieu et al, 2001b). It is hypothesized that heparin prevents bacteria from adhering to the catheter and that the preservatives in heparin preparations have some limited antibacterial properties. Various products have been designed to decrease the risk of catheter-related infection in adult patients. Although data on their use are promising, many of the currently available products are not specifically designed for the neonate, and their safety and efficacy have not been tested in this patient population. Antiseptic- or antibiotic-impregnated catheters have been shown to significantly decrease the risk of CLABSI in adults (Elliott, 2000; Marin et al, 2000; Pai et al, 2000; Tennenberg et al, 1997; Veenstra et al, 1999). Pierce et al (2000) reported that pediatric patients randomly assigned to the use of a heparin-bonded device had a lower incidence of CLABSI than controls (4% versus 33%; p < 0.005). Studies are in progress to evaluate the efficacy of antibiotic lock therapy in neonates. Johnson et al (1994) treated a small group of pediatric patients with CLABSIs with antibiotic lock therapy for 10 days. They reported that 10 of 12 catheters were salvaged. There are limited data describing the use of antibiotic lock therapy in neonates. Garland et al (2002) randomly assigned 82 VLBW infants to saline control or vancomycin antibiotic lock therapy of catheters in an effort to prevent CLABSI. These researchers found a significantly lower incidence density of catheter-related infections, no evidence of toxicity, and no cases of VRE in the neonates receiving antibiotic lock therapy. UMBILICAL CORD CARE Umbilical cord infections remain a significant cause of neonatal mortality in developing nations, although not in the United States (Zupan and Garner, 2000). In clinical practice, there are numerous approaches to umbilical cord care in the healthy term infant because the available data are insufficient to recommend the use of a specific agent or regimen (Zupan and Garner, 2000). Most authorities agree that the umbilical cord should be separated from the placenta with the use of good aseptic technique. Subsequently, some experts recommend “natural drying,” but others support the use of an antiseptic agent such as alcohol, silver sulfadiazine, chlorhexidine, triple dye, or gentian violet (Darmstadt CHAPTER 40 Health Care–Acquired Infections in the Nursery and Dinulos, 2000; Hsu et al, 1999; Pezzati et al, 2002; Zupan and Garner, 2000). Hexachlorophene and iodine are used sparingly because of concerns about their systemic absorption and toxicity. Studies have shown that antiseptic products decrease umbilical cord colonization. Unfortunately, this decrease has not clearly resulted in a lower incidence of umbilical cord infections. In a metaanalysis that included “no intervention” as an alternative, Zupan and Garner (2000) were unable to determine which regimen was superior. In general, they found that time to cord separation was prolonged when antiseptics were used, but there were no significant differences in the incidence of infection or death with the use of a particular agent for cord care. SKIN CARE The skin of VLBW preterm infants is immature and functions as an ineffective barrier to prevent transepidermal loss of water and invasion of bacteria. The stratum corneum (the outermost layer of skin) has both mechanical and chemical properties that decrease the risk of infection (Darmstadt and Dinulos, 2000). This layer of skin matures at approximately 32 weeks’ gestation. In a prematurely born neonate, the maturation process is accelerated and is usually complete by 2 to 4 weeks after birth (Darmstadt and Dinulos, 2000). Unfortunately, there is no consensus on the most effective skin care practices for VLBW infants (Baker et al, 1999; Munson et al, 1999). Neonatologists had hoped that the application of a topical emollient would protect the developing epidermal layer, reduce the risk of infection, and prevent transepidermal water loss. Several studies have documented that such a practice decreases water loss; however, there was no significant reduction in HAI in VLBW infants randomly assigned to receive routine application of one emollient product (Aquaphor) in a study performed by Edwards et al (2001). Infants randomly assigned to emollient therapy had more nosocomial bloodstream infections, particularly with CONS. This difference was most evident in infants with birthweights of 501 to 750 g. The efficacy of other topical agents warrants further study. Efforts to prevent traumatic injury to the skin of VLBW infants include the use of transparent dressing over bony prominences, semipermeable barriers between the skin and adhesive tape, and water-activated electrodes and the avoidance of agents such as tincture of benzoin, Mastisol, and adhesive removers (Adams-Chapman and Stoll, 2002; Darmstadt and Dinulos, 2000; Hoath and Narendran, 2000). MANAGEMENT OF HEALTH CARE–ACQUIRED INFECTIONS CATHETER-RELATED BLOODSTREAM INFECTIONS Management of neonates with CLABSIs is problematic because of the limited intravenous access in most patients, as well as the lack of consensus among clinicians. Current recommendations for adults regarding catheter removal 563 with associated CLABSI suggest the removal of nontunneled central venous catheters associated with bacteremia or fungemia, unless the pathogen is CONS (Mermel et al, 2001). A patient with a catheter-related infection caused by CONS should undergo catheter removal if culture results are persistently positive or if the patient’s condition is unstable (Benjamin et al, 2001; Karlowicz et al, 2000). Benjamin et al (2001) retrospectively reviewed data on infants with CLABSIs and compared outcomes in patients in whom catheters were removed at the onset of infection with those in whom catheters remained in place. Forty-six percent (59 of 128) of infants in whom catheter sterilization was attempted had complications, compared with 8% (2 of 25) of those in whom catheters were removed. In particular, infants with gram-negative infections were more likely to have complications if their catheters remained in place. A study of infants with CLABSIs caused by CONS found no difference in the complication or mortality rate in patients in whom removal of the catheter was delayed (Karlowicz et al, 2000). However, patients with a CONS infection related to their central line were more likely to have persistently positive culture results when their lines were not removed with the first positive culture (43% versus 13% for immediate catheter removal). The attempt to retain the catheter was never successful if culture results remained positive for more than 4 days. Additional prospective randomized trials are needed to validate these observational data. ANTIBIOTIC AND ADJUNCTIVE THERAPIES Antibiotic therapy that is effective against a cultureproven or suspected pathogen is the primary treatment for HAIs. As a general rule, antibiotic choice should initially cover a broad spectrum of pathogens and should then be narrowed as soon as possible to cover the specific bacteria identified once sensitivities are known, or it should be discontinued if infection is not proved and is not likely. One should consider coverage for Pseudomonas spp. or other resistant gram-negative organisms in patients with a rapid clinical deterioration (Karlowicz et al, 2000; Stoll et al, 2002a). However, empiric broadspectrum antibiotic use should be limited as much as possible to avoid complications of resistant bacterial and fungal infections. A metaanalysis of the prophylactic use of intravenous immunoglobulin (IVIG) in preterm neonates found only a 3% reduction in HAI and no reduction in mortality (Modi and Carr, 2000; Ohlsson and Lacy, 2001a). The benefit in patients with culture-proven sepsis remains unclear, and IVIG is therefore not recommended for routine use in such patients (Modi and Carr, 2000). Some authorities speculate that the benefit of IVIG would be greater if products containing high concentrations of specific antibodies against pathogens frequently responsible for neonatal infections were developed (Hill, 2000; Ohlsson and Lacy, 2001b; Weisman et al, 1994). Hemopoietic colony-stimulating factors (granulocyte and granulocytemacrophage) are effective in raising the neutrophil count, but have not consistently decreased HAI rates or mortality 564 PART IX Immunology and Infections (Carr et al, 2003; Modi and Carr, 2000). Studies are being conducted in adults to evaluate the efficacy of administering various cytokine preparations known to modulate the inflammatory response. CONCLUSION Interventions to reduce HAI are urgently needed. Infants with HAIs have significantly longer hospital stays (79 versus 60 days; p < 0.001) and higher hospital costs (Gray et al, 1995; Mahieu et al, 2001a; Pittet et al, 1994; Stoll et al, 2002a). The higher costs are primarily caused by daily hospital charges and pharmaceutical fees (Mahieu et al, 2001a). The attributable cost of a nosocomial infection has been estimated to be approximately $40,000 per adult case and $1200 per neonatal case (Mahieu et al, 2001a; Pittet et al, 1994). The magnitude of these costs is significant, especially when considering the growing number of surviving VLBW infants, who are at the highest risk for acquiring an HAI. Moreover, infants who experience an HAI are significantly more likely to die than are those who remain uninfected (Stoll et al, 1996, 2002a). A variety of new therapeutic alternatives is currently being investigated. Clinicians must continue to focus on developing effective prevention strategies, including strict hand-washing policies, minimal use of invasive devices, promotion of enteral nutrition surveillance of infection patterns, and education for the nursery staff members. Quality improvement approaches can be effective in implementing the necessary practices within a given unit. SUGGESTED READINGS Bloom BT, Craddock A, Delmore PM, et al: Reducing acquired infections in the NICU: observing and implementing meaningful differences in process between high and low acquired infection rate centers, J Perinatol 23:489-492, 2003. Clerihew L, Austin N, McGuire: Prophylactic systemic antifungal agents to prevent mortality and morbidity in very low birth weight infants, Cochrane Database Syst Rev 4:CD003850, 2007. Curtis C, Shetty N: Recent trends and prevention of infection in the neonatal intensive care unit, Curr Opin Infect Dis 21:350-356, 2008. Kilbride HW, Wirtschafter DD, Powers RJ, et al: Implementation of evidencebased potentially better practices to decrease nosocomial infections, Pediatrics 111(4 Pt 2):e519-e533, 2003. Meissner HC, Long SS: American Academy of Pediatrics Committee on Infectious Diseases and Committee on Fetus and Newborn: revised indications for the use of palivizumab and respiratory syncytial virus immune globulin intravenous for the prevention of respiratory syncytial virus infections, Pediatrics 112:14471452, 2003. Perlman SE, Saiman L, Larson EL: Risk factors for late-onset health care-associated bloodstream infections in patients in neonatal intensive care units, Am J Infect Control 35:177-182, 2007. Parellada JA, Moise AA, Hegemier S, et al: Percutaneous central catheters and peripheral intravenous catheters have similar infection rates in very low birth weight infants, J Perinatol 19:251-254, 1999. Schulman J, Wirtschafter DD, Kurtin P: Neonatal intensive care unit collaboration to decrease hospital acquired bloodstream infections: from comparative performance reports to improvement networks, Pediatr Clin N Am 56:865-892, 2009. Stoll BJ, Adams-Chapman I, et al: Abnormal neurodevelopmental outcome of ELBW infants with infection, Pediatr Res Suppl 53:2212, 2003. Stoll BJ, Hansen NI, Adams-Chapman I, et al: National Institute of Child Health and Human Development Neonatal Research Network: neurodevelopmental and growth impairment among extremely low-birth-weight infants with neonatal infection, J Am Med Assoc 292:2357-2365, 2004. Complete references used in this text can be found online at www.expertconsult.com C H A P T E R 41 Fungal Infections in the Neonatal Intensive Care Unit Margaret K. Hostetter EPIDEMIOLOGY Invasive fungal infection occurs in approximately 1% to 2% of all infants admitted to U.S. neonatal intensive care units (NICUs) (Stoll et al, 1996), and the incidence rises dramatically with decreasing gestational age. Among the fungi, Candida spp. is the dominant pathogen, with infections evenly distributed among C. albicans and non–C. albicans species (C. parapsilosis, C. orthopsilosis, C. metapsilosis, C. glabrata, C. guilliermondii, C. krusei, and C. lusitaniae). Candida spp. are the third most common cause of lateonset sepsis in the NICU, with a fatality rate more than sevenfold greater than that of Staphylococcus epidermidis, the most common pathogen (Benjamin et al, 2000; Saiman et al, 2001). Other fungi encountered include Aspergillus fumigatus, A. flavus, Malassezia furfur, M. pachydermatis. The yeast Cryptococcus neoformans and the fungi Histoplasma capsulatum, Blastomyces dermatitidis, and Coccidioides immitis are rarely if ever seen in the NICU. The most important risk factor for fungal infection is gestational age. In a survey of 2847 infants from six different nurseries, the incidence of candidemia in infants weighing less than 800 g (7.55%) was 25-fold that of the infant weighing more than 1500 g (Figure 41-1) (Saiman et al, 2000). This latter group acquires blood-borne candidal infection in association with congenital anomalies, especially those of the gastrointestinal tract (Rabalais et al, 1996). Candidemia carries a mortality rate exceeding 25% in most studies (Chapman and Faix 2000; Weese-Mayer et al, 1987). Colonization with ubiquitous fungal species occurs in at least 25% of very low-birthweight infants (Baley et al, 1986), and both the amount of Candida spp. in the gastrointestinal tract (Pappu-Katikaneni et al, 1990) and colonization at sites such as endotracheal tubes (Rowen et al, 1994) have been correlated with increased risk of invasive disease caused by Candida spp. Prospective studies correlating colonization by other fungal genera (e.g., Aspergillus, Malassezia spp.) with risk of invasive disease have not been done. Apart from colonization and gestational age, other host factors that contribute to the susceptibility of the infant in the NICU to fungal infection include 5-minute Apgar scores less than 5, and an age-dependent immunocompromised state ascribable to reduced numbers of T cells, impaired phagocyte number and function, and reduced levels of complement (Marodi et al, 1994; Rebuck et al, 1995; Witek-Janusek et al, 2002; Zach and Hostetter, 1989). Concomitants of nursery care that are thought to increase the risk of fungal infections include length of stay greater than 1 week, indwelling central venous catheters, abdominal surgery, parenteral nutrition, intralipids, H2 (histamine) receptor antagonists, endotracheal intubation, and prolonged use of broad-spectrum antimicrobials, especially third-­ generation cephalosporins (Cotten et al, 2006; Saiman et al, 2001; Saiman et al, 2000). Other centers have identified associations with systemic steroids and catecholamine infusions in retrospective studies and with topical petrolatum in a prospective case-control study (Benjamin et al, 2000; Botas et al, 1995; Campbell et al, 2000). Interestingly, a number of variables appear not to associate with candidal colonization, including the use of antibiotics in the mother, premature rupture of the membranes, the infant’s gender, the use of antimicrobial agents other than third-generation cephalosporins in the infant, surgical procedures, or frequency of intubation (Saiman et al, 2001). Although ≈5% of NICU staff members carry C. albicans on the hands and 19% carry C. parapsilosis, there is no correlation with site-specific rates of infant colonization (Saiman et al, 2001). This chapter will place major emphasis on infections caused by Candida spp. and other fungi, as well as the approach to diagnosis, treatment, and management of infants with fungal infection. INFECTIONS CAUSED BY CANDIDA SPECIES CONGENITAL CANDIDIASIS Appearing within the first 24 hours of life in both full-term and premature infants, congenital candidiasis, a very rare entity, manifests as a deeply erythematous skin rash in the setting of pronounced neutrophilia, with white blood cell counts often rising to 50,000 or more, and infrequently Candida spp. funisitis (see Local Infections, later). In the fullterm infant, there are no invasive consequences, and desquamation typically ensues within 2 to 3 days. In contrast, the condition is life threatening in the premature infant (Dvorak and Gavaller, 1966; Johnson et al, 1981) and is distinguished by a pustular rash, hazy infiltrates reminiscent of respiratory distress syndrome on chest radiograph, and frequently positive blood cultures. The premature infant is thought to acquire the organism from inhaling infected amniotic fluid. Diagnosis in both premature and full-term infants requires the visualization of the organism on Gram stain from a bullous lesion or an opened pustule. On rare occasions, the placenta has yielded the diagnosis. Treatment for the full-term infant requires only the full-body application of topical antifungal creams containing either nystatin or azoles, such as miconazole or clotrimazole. In the premature infant, initiating parenteral amphotericin B deoxycholate at a dose of 1 to 1.5 mg/kg is mandatory, but respiratory involvement typically heralds death despite anti-fungal therapy. 565 566 PART IX Immunology and Infections 8 Incidence (%) 7 6 5 4 3 2 1 0 800 801–1000 1001–1500 1500 Birthweight in grams FIGURE 41-1 Incidence of candidemia related to birthweight in grams. (Data from Saiman L et al: Risk factors for candidemia in Neonatal Intensive Care Unit patients. The National Epidemiology of Mycosis Survey study group, Pediatr Infect Dis J 19:319-324, 2000.) LOCAL INFECTIONS WITH CANDIDA SPECIES Diaper Dermatitis Diaper dermatitis manifests as an erythematous, erosive dermatitis of the perineal region, typically with pustular satellite lesions beyond the borders of the rash. Predisposing factors include systemic antibiotics, glucosuria, and wet diapers. Care must be taken to differentiate this tractable condition from invasive fungal dermatitis (see Invasive Fungal Dermatitis, later). The infection responds well to topical antifungal ointments. Funisitis Infection of the umbilical cord with Candida spp., although rare, is an indicator of chorioamniotitis and carries a poor prognosis, especially in the premature infant (Qureshi et al, 1998). A minority (16%) of infants in one study had associated congenital candidiasis. Intrauterine contraceptive devices or cervical cerclage were reported in 16% of the mothers. Urinary Tract Infection Isolation of Candida spp. from a catheterized specimen or suprapubic aspiration, as opposed to a bagged sample, is a reliable indicator of infection, although asymptomatic colonization of urinary catheters, stents, or nephrostomy tubes can be difficult to distinguish from true infection (Lundstrom and Sobel, 2001). The presence of candiduria in infants in the NICU is associated with renal candidiasis—the latter manifested by cortical abscesses or fungal mycelia in the collecting system (“fungus balls”)—almost half the time and may be a cause of frank obstruction (Bryant et al, 1999). Therefore, in contrast to older children or adults, the finding of candiduria in an infant in the NICU should at least prompt blood cultures and renal imaging. If blood cultures prove to be positive, a full evaluation for disseminated candidiasis should be undertaken (see Systemic Infection, later). Because of the high prevalence of associated upper tract disease, imaging of the kidneys by ultrasonography should occur after isolation of the organism from a sterile urine specimen. Approximately one half of the patients who eventually develop upper tract manifestations will display them on the first ultrasound (Bryant et al, 1999). Therefore follow-up imaging is recommended both to ensure the clearance of fungal mycelia, if present, and to monitor for later development of this complication. Unfortunately, no standard interval for monitoring has been proposed; however, in the infant with persistent funguria or candiduria, a single negative ultrasound should not be considered definitive. Removal of a colonized urinary catheter may suffice for treatment in a patient without pyuria or systemic symptoms. Disease confined to the lower tract is best addressed with azoles (e.g., fluconazole, 4 to 6 mg/kg/day). Upper tract disease requires parenteral amphotericin B in systemic doses (1 to 1.5 mg/kg/day). Liposomal amphotericin B may not be an acceptable alternative. The particles in at least one liposomal preparation (Abelcet) appear to be too large to penetrate the adult kidney (Hell et al, 1999), and this author has seen a premature newborn whose persistent candiduria failed to resolve until treatment was changed to amphotericin B deoxycholate. PERITONITIS Candida spp. peritonitis typically develops as a consequence of bowel perforation or rarely as a complication of peritoneal dialysis. In the former situation, multiple organisms such as gram-negative rods and enterococci may also be involved, and the neonate is at risk for sepsis with any one of them (Johnson et al, 1980). Peritonitis associated with a peritoneal dialysis catheter usually occurs as an isolated process, and the outcome is much better. Spontaneous intestinal perforation associated with Candida spp. peritonitis with or without sepsis has been described within 7 to 10 days of birth in infants weighing less than 1000 g, typically in the absence of necrotizing enterocolitis (Adderson et al, 1998; Holland et al, 2003; Meyer et al, 1991; Mintz and Applebaum, 1993). Hallmark clinical findings include bluish discoloration of the abdomen and a gasless pattern on abdominal film. A substantial proportion of these infants will have systemic candidiasis, although Staphylococcus epidermidis can also be seen. In a small study of seven patients (Holland et al, 2003), deficiency of the muscularis propria was found in six. Diagnosis requires visualization of the organism on a Gram stain of peritoneal fluid sterilely obtained or culture of the organism from the same source. Isolation of Candida spp. from the peritoneal fluid should always prompt a search for bowel perforation, by either radiology or surgical exploration, depending upon the clinical circumstances. Treatment of Candida spp. peritonitis caused by necrotizing enterocolitis or bowel perforation requires surgical evaluation, supportive therapy, and direct address of all contaminating microorganisms in the peritoneal fluid and the bloodstream. The typical regimen may include ampicillin and an aminoglycoside for enterococci and gramnegative rods, clindamycin for anaerobes, and systemic antifungal therapy, most likely with amphotericin B. The isolation of Candida spp. from peritoneal dialysate can be treated with removal of the catheter and a short course (7 to 10 days) of amphotericin B therapy in a dose of CHAPTER 41 Fungal Infections in the Neonatal Intensive Care Unit 0.3 to 0.5 mg/kg/day. The catheter can typically be reinserted within 24 to 48 hours, once the Gram stain is free of yeast cells. SYSTEMIC INFECTION CANDIDEMIA ASSOCIATED WITH CENTRAL VENOUS CATHETERS The association between prematurity and blood-borne candidal infections has been recognized for 25 years (Baley et al, 1984; Johnson et al, 1984). Over this same period of time, the incidence of candidemia has escalated from 25 to 123 cases per 10,000 NICU admissions (Kossoff et al, 1998; Saiman et al, 2000). The median time of onset is approximately 30 days of age (Baley et al, 1984). In a large multicenter study, colonization of the gastrointestinal tract preceded candidemia in 43% of cases (Saiman et al, 2000). A variety of nonspecific clinical findings may be associated with this presentation of candidal disease, including respiratory decompensation, feeding intolerance, temperature instability, or mild thrombocytopenia. It is unclear whether the latter manifestation relates more to the use of heparin in vascular catheters or to the presence of Candida spp. in the blood stream. Isolation of Candida spp. from a blood culture should never be regarded as a contaminant and should prompt an immediate search for evidence of dissemination, which occurs in approximately 10% of premature newborns with candidemia (Noyola et al, 2001; Patriquin et al, 1980). A thorough evaluation includes ophthalmologic examination and ultrasonography of the heart, venous system, and abdomen. When lumbar puncture is performed, 10% to 50% of infants with candidemia may have associated meningitis (Benjamin et al, 2006; Faix, 1984); in one prospective study, almost 50% of extremely low-birthweight (ELBW) infants with Candida spp. meningitis (13/27) had negative blood cultures (Benjamin et al, 2006). Numerous studies have shown that central venous catheters should be removed within 24 hours after identification of yeasts in the blood culture (Karlowicz et al, 2000); in particular, removal of the central venous catheter within 3 days is associated with a significantly shorter median duration of candidemia (3 versus 6 days) and a reduced mortality rate (0% versus 39%). In at least one study of candidemia, delayed removal of central venous catheters was associated with neurodevelopmental impairment at 18 to 22 months (Benjamin et al, 2006). Many experts recommend routine echocardiograms for patients with catheterassociated candidemia to look for thrombi before removal of the catheter. However, even with the prompt removal of the catheter and the institution of appropriate antifungal therapy, a substantial proportion of infants may exhibit prolonged candidemia lasting 1 to 3 weeks (Chapman and Faix, 2000). DISSEMINATED CANDIDIASIS Mortality rates approach 30% for this dreaded complication of candidal infection. C. albicans is the leading pathogen. Organ involvement is most common in the vascular tree at catheter sites (15.2%), followed by the 567 kidneys (7.7%) (Noyola et al, 2001; Patriquin et al, 1980). Eye involvement occurs in approximately 6% of infants. Thrombi within the vascular bed may be particularly difficult to eradicate with antifungal therapy; infants with right atrial thrombi may benefit from atriotomy (Foker et al, 1984). Other sites less frequently involved include the liver, spleen, and skeletal system. In infection of the bones and joints in premature newborns, Candida spp. are typically the second most likely pathogen, preceded by Staphylococcus aureus (Ho et al, 1989). ANTI-FUNGAL THERAPY FOR SYSTEMIC INFECTION As is true of most medications used in the neonatal ICU, dosing recommendations for antifungal therapies have not undergone rigorous testing in this patient population. With that caveat in mind, it is important to discuss practice guidelines for this difficult clinical problem. Amphotericin B is the standard antifungal therapy for treatment of systemic neonatal fungal infection. The drug binds to ergosterol in the membrane of fungi, facilitating membrane leakage. Rapid institution of parenteral amphotericin B deoxycholate in doses of 1.0 to 1.5 mg/kg/day, given by intravenous infusion over 2 to 6 hours, is the therapy of choice for systemic infection, including catheter-associated candidemia and disseminated candidiasis. No more than 24 hours should elapse before the infant is receiving a dose of 0.7 to 1.0 mg/kg/day. Some experts recommend instituting empiric antifungal therapy in acutely thrombocytopenic ELBW infants (Benjamin et al, 2003). Dose adjustment for renal dysfunction is necessary only if serum creatinine increases significantly during therapy. Amphotericin B has poor cerebrospinal fluid penetration; therefore accompanying meningitis should prompt the addition of 5-flucytosine in doses of approximately 12.5 to 37.5 mg/kg per dose given by mouth every 6 hours (in patients with normal renal function). Peak serum concentrations of this drug should be kept between 40 and 60 μg/mL to avoid bone marrow suppression or hepatotoxicity. Fluconazole may also be added to amphotericin B for meningeal penetration, if 5-flucytosine is unavailable or cannot be used. Infants being treated with amphotericin B who experience a twofold rise in creatinine, which is evidence of renal tubular compromise or renal tubular acidosis, may benefit from use treatment with one of the liposomal amphotericin preparations. AmBisome has been used in neonates at well-tolerated doses of 5 to 7 mg/kg per dose every 24 hours intravenously, infused over 2 hours (JusterReicher et al, 2003; Weitkamp et al, 1998), although the failure of many liposomal preparations to penetrate renal parenchyma can militate against fungal clearance from this organ. With prompt removal of an offending central venous catheter, and no evidence of dissemination, the duration of therapy for catheter-associated candidemia is typically 10 to 14 days after the blood culture becomes negative (Donowitz and Hendley, 1995). Disseminated candidiasis, including Candida spp. meningitis, requires at least 3 weeks or more of parenteral therapy; the course is typically completed when all foci have been eradicated. Most infectious disease experts will use fungicidal doses of parenteral 568 PART IX Immunology and Infections amphotericin B or a liposomal preparation for the entire course. Azoles (such as fluconazole at 6 mg/kg intravenously at varying dosing intervals) are antifungal agents that interfere with ergosterol synthesis by inhibiting C-14 alpha demethylase, a cytochrome P450 enzyme. Use of an azole to complete the latter part of an antifungal course must account for the inability of the infant’s immune system to compensate for the fungistatic activity of the azoles. Azoles such as itraconazole and posaconazole are preferable to fluconazole for aspergillus and zygomycetes, but no studies have been done to recommend neonatal dosing guidelines. Although not a first-line antifungal medication, echinocandins such as caspofungin, which interrupt biosynthesis of β-(1,3)-D-glucan, an integral part of the fungal cell wall, have also been used in doses of 1 to 2 mg/kg/day to treat invasive candidal disease in the newborn in several case reports; this drug is particularly helpful for species such as Candida glabrata, Candida krusei or Candida lusitaniae, which may have decreased susceptibility or de novo resistance to amphotericin B (Odio et al, 2004; Saez-Llorens et al, 2009). ANTI-FUNGAL PROPHYLAXIS Five randomized, controlled trials comparing intravenous fluconazole (3 mg/kg/day) with placebo or no treatment in very low-birthweight (VLBW) or ELBW infants for 4 to 6 weeks (Cabrera et al, 2002; Kaufman et al, 2001; Kicklighter et al, 2001; Manzoni et al, 2007; Parikh et al, 2007) met criteria for analysis in Cochrane reviews (Clerihew et al, 2007). Kaufman and Manzoni reported significantly lower incidence of invasive fungal infection, whereas there was no difference in treated versus untreated infants in other studies (Cabrera et al, 2002; Kicklighter et al, 2001; Parikh et al, 2007). The only study to evaluate neurologic outcomes found no difference in neurologic impairment at 16 months (Kaufman et al, 2001). Fluconazole prophylaxis did not create a significant difference in the risk of death before discharge in any of the five studies or in a metaanalysis (Clerihew et al, 2007). No study documented clinically significant adverse effects of fluconazole or the emergence of fluconazole resistance. One study from a single center compared nonrandomized fluconazole prophylaxis in 2002 to 2006 with an untreated, retrospective cohort (2000 to 2001) and reported that invasive candidiasis decreased from 0.6% to 0.3%, although the proportion of invasive disease caused by non–C. albicans species increased from 26% to 41% after the introduction of fluconazole prophylaxis (Healy et al, 2008). Interestingly, fluconazole prophylaxis in this study was extended to several infants with birthweights greater than 1000 g, if risk factors (e.g., maternal HIV infection, intestinal abnormalities) were present. The finding that prophylactic fluconazole reduces the incidence of invasive fungal infection must be interpreted with caution (Clerihew et al, 2007, 2008): 1. The incidence of invasive fungal infection in the placebo groups (Kaufman et al, 2001; Manzoni et al, 2007; Parikh et al, 2007) was significantly higher (13% to 16%) than in other large cohort studies of VLBW or ELBW infants in the United States (6% to 7%) or the United Kingdom (1% to 2%). 2. Fluconazole prophylaxis may have impaired the microbiologic isolation of some fungal species and led to underdiagnosis of infection in the treatment group. 3. Six years after the introduction of fluconazole prophylaxis, one study reported that non–C. albicans species with relatively reduced susceptibility to the azoles were the most common causes of invasive fungal infection (Parikh et al, 2007). This study did not detect a significant effect of fluconazole prophylaxis in reducing invasive candidal disease. Based on these results and cautionary notes, infants weighing less than 1000 g who receive third-generation cephalosporins and central venous catheters may be the best group to be evaluated for a statistically significant benefit of fluconazole prophylaxis in preventing invasive fungal infection, neurologic complications, and death before hospital discharge. Study durations less than 6 years may be insufficient to detect the emergence of fluconazole resistance. INFECTIONS ASCRIBABLE TO OTHER FUNGI INVASIVE FUNGAL DERMATITIS Invasive fungal dermatitis typically manifests in the infant weighing less than 1000 g who displays macerated or bruised lesions that are contaminated with fungal species. In the initial report, three of seven confirmed cases had C. albicans, C. parapsilosis, C. tropicalis, Trichosporon beigelii, Curvularia spp., or Aspergillus niger and A. fumigatus were cultured from the remainder (Rowen et al, 1995). Among cases considered “probable,” seven of eight had C. albicans. Systemic complications including fungemia, meningitis, or infection of the urinary tract occurred in four of seven confirmed cases and seven of eight probable cases. More cases than controls had postnatal steroids and prolonged hyperglycemia. Disseminated infection occurred in 69%, all ascribable to Candida spp. Diagnosis requires a skin biopsy specimen demonstrating fungal invasion beyond the stratum corneum or a positive potassium hydroxide preparation of skin scrapings; growth of the identical organism from an otherwise sterile site (blood, cerebrospinal fluid, or urine obtained via supra pubic aspiration) is confirmatory. Treatment requires systemic doses of amphotericin B in the range of 0.7 to 1.0 mg/kg/day; in infants who do not develop systemic infection, oral therapy with fluconazole or topical antifungal creams may suffice. Oral therapy is not advisable for pathogens like Aspergillus spp., and repeated skin biopsies may be necessary to define duration of therapy. LINE INFECTIONS CAUSED BY LIPOPHILIC ORGANISMS The species Malassezia furfur and Malassezia pachydermatis are lipophilic organisms commonly carried on the skin, even in patients without tinea versicolor (Marcon and CHAPTER 41 Fungal Infections in the Neonatal Intensive Care Unit Powell, 1992). Cutaneous colonization can infect hyperalimentation fluids or parenteral lipid formulations. Infants typically exhibit mild but nonspecific signs: respiratory decompensation, glucose intolerance, or thrombocytopenia (Dankner et al, 1987; Stuart and Lane, 1992). Diagnosis requires isolation of the organism from blood by growth on fungal medium overlaid with olive oil, because Malassezia spp. will not grow in the absence of lipids (Marcon et al, 1986). Removal of the intravascular catheter usually suffices for therapy, although some experts recommend the addition of amphotericin B in dosages of 0.5 mg/ kg/day for 7 days. MISCELLANEOUS FUNGAL INFECTIONS Aspergillus species Although rarely seen in neonates, systemic infection with Aspergillus spp. suggests severe immune system compromise, such as DiGeorge Syndrome or myeloperoxidase deficiency (Chiang et al, 2000; Marcinkowski et al, 2000). Disseminated disease has occurred in premature newborns without additional immunologic abnormalities (Rowen et al, 1992). Diagnosis requires isolation of the fungus from a normally sterile tissue site or visualization by Gomori-methenamine silver stain on a biopsy specimen of infected tissue. Of note, a commercially available enzyme-linked immunosorbent assay for diagnosis of aspergillosis on serum specimens had an 83% rate of false positive results in premature newborns (Siemann et al, 1998). Treatment requires systemic amphotericin B in doses of 1.0 to 1.5 mg/kg/day. Fungistatic therapies such as the triazoles are not recommended for aspergillosis. 569 Trichosporon beigelii In a cluster of five neonatal cases of infection caused by T. beigelii, a yeast found ubiquitously in soil, no common source was identified (Fisher et al, 1993). Two of three premature infants infected with this organism died. Resistance to achievable concentrations of amphotericin B complicates therapy. SUGGESTED READINGS Benjamin DK Jr, Stoll BJ, Fanaroff AA, et al: Neonatal candidiasis among extremely low birth weight infants: risk factors, mortality rates, and neurodevelopmental outcomes at 18 to 22 months, Pediatrics 117:84-92, 2006. Clerihew L, Austin N, McGuire W: Prophylactic systemic antifungal agents to prevent mortality and morbidity in very low birth weight infants, Cochrane Database Syst Rev 4:CD003850, 2007. Cotten CM, McDonald S, Stoll B, et al: The association of third-generation cephalosporin use and invasive candidiasis in extremely low birth-weight infants, Pediatrics 118:717-722, 2006. Faix RG: Systemic Candida infections in infants in intensive care nurseries: high incidence of central nervous system involvement, J Pediatr 105:616-622, 1984. Juster-Reicher A, Flidel-Rimon O, Amitay M, et al: High-dose liposomal amphotericin B in the therapy of systemic candidiasis in neonates, Eur J Clin Microbiol Infect Dis 22:603-607, 2003. Karlowicz MG, Hashimoto LN, Kelly RE Jr, et al: Should central venous catheters be removed as soon as candidemia is detected in neonates? Pediatrics 106:E63, 2000. Noyola DE, Fernandez M, Moylett EH, et al: Ophthalmologic, visceral, and cardiac involvement in neonates with candidemia, Clin Infect Dis 32:1018-1023, 2001. Odio CM, Araya R, Pinto LE, et al: Caspofungin therapy of neonates with invasive candidiasis, Pediatr Infect Dis J 23:1093-1097, 2004. Rowen JL, Atkins JT, Levy ML, et al: Invasive fungal dermatitis in the < or = 1000gram neonate, Pediatrics 95:682-687, 1995. Saiman L, Ludington E, Pfaller M, et al: Risk factors for candidemia in Neonatal Intensive Care Unit patients. The National Epidemiology of Mycosis Survey study group, Pediatr Infect Dis J 19:319-324, 2000. Saiman L, Ludington E, Dawson JD, et al: Risk factors for Candida species colonization of neonatal intensive care unit patients, Pediatr Infect Dis J 20: 1119-1124, 2001. Complete references and supplemental color images used in this text can be found online at www.expertconsult.com P A R T X  Respiratory System C H A P T E R 42 Lung Development: Embryology, Growth, Maturation, and Developmental Biology Maria Victoria Fraga and Susan Guttentag The primary function of the lung is to accomplish exchange of oxygen and carbon dioxide to accommodate the needs of aerobic cellular respiration. The oxygen consumption of the adult human ranges from 250 mL/min at rest to 5500 mL/min at peak exercise (Warburton et al, 2000). To accommodate these metabolic needs, a large surface area, and a thin alveolar-capillary membrane are required to enable efficient diffusion of oxygen more so than carbon dioxide. Ultimately the zone of gas exchange will attain a surface area of 50 to 100 m2 and a volume of 2.5 to 3.0 L in the adult human; therefore a primary goal of lung organogenesis is to expand the lung surface area to meet these needs. A second goal of lung organogenesis is to minimize the diffusing distance from alveolus to red blood cell, coordinating the development of an extensive capillary network with a thin, expansive alveolar epithelial surface. A third goal of lung development is production of a protective aqueous barrier overlying the delicate alveolar epithelium while mitigating the effects of the surface tension generated by this barrier, specifically alveolar collapse, through the production of a surface active agent or surfactant. The trachea, airways, and alveoli are in constant contact with the external environment. Consequently with every inhalation, epithelial surfaces encounter large numbers of microorganisms and potentially toxic particles and gases. Lung organogenesis must also incorporate mechanisms for clearance of microorganisms and allergens that may result in epithelial infection or injury. Similarly the lung must defend against nonparticulate gases that are potentially harmful. Oxygen, although critical to cellular function, can be the source of harmful reactive oxygen species and inhaled pollutants similarly require detoxification. The appropriate development and maintenance of these lung functions are critical to the health and survival of newborn infants. This chapter focuses on developmental aspects of each function that place the premature neonate at increased risk for lung injury and disease. KEY EVENTS IN LUNG DEVELOPMENT Lung formation begins early in human gestation (by day 25), and growth extends well into childhood (Burri, 2006; Kumar et al, 2005). Lung development can be organized into stages (embryonic, pseudoglandular, canalicular, saccular, and alveolar), although the timing of these stages is somewhat imprecise and considerable overlap can occur. Figure 42-1 shows a timeline of fetal and postnatal lung development. DEVELOPMENT OF AIRWAYS AND GAS EXCHANGE SURFACES The initial phase of lung development, the embryonic phase, is marked by the formation of the lung bud and initial branching of presumptive airways. The lung bud is first recognizable as a laryngotracheal groove of the ventral foregut at 25 days’ gestation. Within a few days the groove closes so that the only remaining lumenal attachment to the foregut is in the region of the developing hypopharynx and larynx. The lung bud, consisting of epithelium and surrounding mesenchyme, then begins the first of a series of dichotomous divisions that give rise to the conducting airways and five primordial lung lobes (two left and three right). Tracheoesophageal fistulas, tracheal atresia, and tracheal stenosis result from errors in separation of the laryngotracheal groove, whereas failure of partitioning of the lung bud can result in pulmonary agenesis, most typically of the right lung. Branching continues into the pseudoglandular stage of lung development. By 7 weeks’ gestation, the trachea and the segmental and subsegmental bronchi are evident. By the end of 16 weeks’ gestation, all bronchial divisions are completed. It is important to remember that, although the conducting airways will certainly enlarge as the fetus and newborn grow (airway diameter and length increase twofold to threefold between birth and adulthood), large airway branching ceases after 16 weeks’ gestation. Closure of the pleuroperitoneal folds is a critical event of the pseudoglandular phase, reaching completion by 7 weeks’ gestation and resulting in separation of the thoracic cavity from the peritoneal cavity. Failure of pleuroperitoneal closure results in a diaphragmatic defect and continuity between these cavities. At 10 weeks’ gestation, when the midgut returns to the peritoneal cavity from the umbilical cord, abdominal contents are free to pass into the thoracic cavity and restrict the space into which the lung grows. The resulting congenital diaphragmatic hernia leads to pulmonary hypoplasia of the lung ipsilateral to the diaphragmatic defect as bowel and solid viscera migrate into the thorax. Pulmonary hypoplasia can also extend to the contralateral lung as the mediastinum shifts because of accumulating abdominal viscera in the thorax. 571 572 PART X Respiratory System Lung growth Late alveolarization Microvascular maturation Bulk alveolarization Saccular Canalicular Pseudoglandular Embryonic 0 5 10 15 20 25 30 35 40 wk 1 yr 2 yr 3 yr FIGURE 42-1 Human lung development. Timeline of major events in human fetal and postnatal lung development. The canalicular phase is marked by completion of the conducting airways through the level of the terminal bronchioles, and the development of the rudimentary gas exchange units that are no longer invested with cartilaginous support. The acinus is the gas exchange unit of the lung and encompasses a respiratory bronchiole and all of its associated alveolar ducts and alveoli. A terminal bronchiole with all its associated acinar structures constitutes a lobule. Branching of these distal airspaces continues on a more limited basis during the canalicular phase, finally achieving a total of 23 airway subdivisions. Evolution of the relationships between the airspaces, capillaries, and mesenchyme acquires greater significance during the saccular phase of lung development (24 to 38 weeks’ gestation), enabling an alveolocapillary membrane sufficient to participate in gas exchange (0.6 μm) by approximately 24 weeks. Beyond this point, the efficiency of gas exchange is determined by the available surface area. Lengthening and widening of the terminal sacs expands the gas exchange surface area. Each saccule consists of smooth-walled airspaces with thickened interstitial spaces containing a double capillary network. These will give rise to two to three alveolar ducts, further expanding the available surface area. Expansion of these rudimentary gas exchange units continues well into the 3rd trimester of human gestation; therefore the human lung is not fully mature structurally, even at term delivery. Postnatal lung development can be subdivided into additional stages (Burri, 2006). True alveoli become evident as early as 36 weeks’ gestation, initiating the alveolar phase of lung development. The development of primary alveoli is followed by a further expansion of the gas-exchange surface area through the formation of septae or secondary crests (see Alveolarization, later). Postnatal alveolarization extends from term through 1 to 2 years of age. An initial phase of bulk alveolarization occurs within the first 6 months postnatally, with a more modest addition of secondary alveoli through the remainder of this period. The alveoli of the infant lungs are different from adult alveoli. These immature secondary alveoli contain a double capillary bed, whereas adult alveoli are invested by a single capillary bed. Microvascular maturation, the next phase of postnatal lung development, occurs between the first few postnatal months of life through 3 years of age (see Development of the Pulmonary Vasculature, later). There is considerable controversy regarding when the lung ceases to add alveoli. Estimates have ranged from as early as 2 years to as late as 20 years old in humans; this is further complicated by the observation that alveolar expansion can occur in response to pneumonectomy in adult animals and humans. The acquisition of alveoli after the maturation of the microvasculature has been termed late alveolarization. This activity has been most often demonstrated in subpleural regions of the lung and likely invokes mechanisms similar to secondary crest formation. The addition of alveoli is not the only means for expanding the surface area of the lung. While alveolarization wanes over the first 3 years of life in the human, growth of the lung continues to expand the gas exchange surface. Between 2 years of age and adulthood, lung tissue expands with lung volume roughly proportionately to the increase in bodyweight of the child. Thus, because of the combined processes of prenatal lung development, postnatal lung development, and lung growth, there is tremendous potential for expansion of the gas exchange surface area that is developmentally programmed into the fetal lung to account for the growing needs of the infant, child, and adult for aerobic cellular respiration. The extent to which these developmental mechanisms can be harnessed after premature birth, with or without superimposed lung injury, is a topic of active investigation. COMPOSITION OF AIRWAYS AND ALVEOLI As branching morphogenesis proceeds, the epithelium lining the successive generations of airways and alveoli gives rise to specialized cells that participate in gas exchange, surfactant production, mucociliary clearance, detoxification, and host defense. Differentiation proceeds in a centrifugal fashion from proximal to distal airspaces, lagging behind branching. Temporal and contextual signals foster the regionalization of epithelial cell types. Proximal Airways The airway epithelium is tall and columnar, decreasing to a more cuboidal appearance more distally (Jeffrey, 1998; Snyder et al, 2009). The endodermal epithelial lining cells of the trachea and bronchi partition into four cell types: undifferentiated columnar, ciliated, secretory-goblet, and basal cells. Ciliated cells critical to the process of mucus clearance are first apparent between 11 and 16 weeks’ gestation and become less prevalent in more distal airways. Three types of secretory cells—those with largely mucous CHAPTER 42 Lung Development: Embryology, Growth, Maturation, and Developmental Biology granules, those with serous granules, and some with both types of granules—can be seen as early as 13 weeks’ gestation. The number of mucin-producing goblet cells in airways peaks at mid-gestation in the fetus and declines into adulthood. Finally, immature basal cells expressing epidermal keratin have been noted as early as 12 weeks’ gestation. Basal cells have a critical role in regenerating injured large airway epithelium (see Stem and Progenitor Cells in the Lung, later). Cartilaginous support of the tracheobronchial tree begins and proceeds in a centrifugal fashion beginning in the primitive trachea at 4 weeks, reaching the main bronchi by 10 weeks, and proceeding to the most distal terminal bronchioles by approximately 25 weeks’ gestation. Cartilaginous investment of airways is complete by the 2nd month postnatally. Submucosal glands are found in the interstitium between the cartilaginous tissue and surface epithelium, and they have a major role in airway host defense. Submucosal gland development can be characterized by five stages: (1) epithelial budding and invasion of the lamina propria, (2) development of a lumen, (3) initiation of tube branching, (4) dichotomous branching, and (5) repeated dichotomous branching. By comparison, the airways of infants and children contain relatively more submucous glands than do adults. The glands are lined by mucous cells proximally and serous cells more distally, the latter comprising 60% of the total epithelial cell content of the glands. Serous cells secrete water, electrolytes, and proteins with antimicrobial, antiinflammatory, and antioxidant properties, whereas the mucous cells produce primarily mucins. In addition to this host defense role, submucosal glands also contain a population of basal cells that respond to injury of the airway by replenishing the airway epithelium. Muscular investment of the airways begins as early as 6 to 8 weeks gestation as smooth muscle cells are identifiable around the trachea and large airways. Fetal airway smooth muscle is innervated and able to contract during the first trimester. It is also responsive to methacholine challenge that is reversible with β-adrenergic agonists. Muscularization increases through fetal life and childhood such that there is an increase in the amount of smooth muscle relative to airway size compared with adults. Furthermore, there is a rapid increase in bronchial smooth muscle immediately after birth, whether born at term or prematurely. An additional airway cell deserves mention because of its role in a wide variety of pediatric diseases. Pulmonary neuroendocrine cells (PNECs) are found throughout the airways, often in innervated clusters known as pulmonary neuroendocrine bodies (NEBs) located at branch points in the bronchial tree (Cutz et al, 2007). Although they arise from foregut endoderm, the cell of origin is distinct from other epithelial components of the lung. Solitary PNECs are sensitive to stretch- and hypoxia-mediated secretion, producing both amines (i.e., serotonin) and peptides (i.e., bombesin) that are important in regulating bronchial tone. Pathologic conditions recently associated with PNECs and NEBs, most often characterized by hyperplasia, include bronchopulmonary dysplasia, disorders of respiratory control (congenital central hypoventilation syndrome and sudden infant death syndrome), cystic fibrosis, and pulmonary hypertension. Neuroendocrine hyperplasia of infancy is a rare form of interstitial lung disease of infancy 573 associated with expansion of the number of PNECs and NEBs, yet little is known about the mechanism of disease. Distal Airways The bronchiolar epithelium differs from the more proximal airway epithelium. In addition to being more cuboid in appearance, the epithelium contains progressively fewer ciliated cells and goblet cells, which are ultimately absent from the terminal bronchioles. Instead, the nonciliated, secretory Clara cell is found in increasing numbers and density down the conducting airways, such that the Clara cell is the most abundant cell of the terminal bronchiole (Jeffrey, 1998). Clara cells are first evident by 16 to 17 weeks’ gestation, initially exhibiting large glycogen stores that are replaced by secretory granules. Between 23 and 34 weeks’ gestation, there is a dramatic increase in Clara cell numbers in distal bronchioles. Clara cells are critical to the host defense and detoxification functions of the lung. This specialized cell produces the highest levels of cytochrome P-450 and flavin monooxygenases in the lung. While critically important in detoxification, these enzymes participate in the bioactivation of procarcinogens as well, placing the Clara cell in a precarious position as a primary target of toxic metabolites. The Clara cell also has an important role in immunoregulation in the distal airways. Important host defense products of the Clara cell include Clara cell secretory protein (CCSP or CC10), surfactant proteins A, and D, leukocyte protease inhibitor, and a trypsinlike protease. The function of Clara cell SP-B in airways is less certain, especially because Clara cells do not produce a mature 8-kDa SP-B protein, but may also contribute to host defense. The secretion of antiproteases from Clara cells suggests that they modulate the protease-antiprotease balance in the distal lung. Alveolar Epithelium During the 4th through 6th months of gestation, the epithelial cells lining the acini begin to differentiate further (Mallampalli et al, 1997). The cuboidal epithelial cells accumulate large glycogen stores and develop small vesicles containing loose lamellae. The large glycogen pools provide a ready source of substrate required for the production of increasing amounts of surfactant phospholipids, and they decrease in size as surfactant production advances in the fetal lung. In cells destined to become type 2 cells, lamellar bodies become larger, more numerous, and more densely packed with surfactant phospholipids and proteins, whereas those cells destined to become type 1 cells, upon losing their relationship to mesenchymal fibroblasts, lose the prelamellar vesicles and become progressively thinner, thereby adopting a phenotype more suitable for gas exchange. Alveolar type 1 and 2 cells are readily identified early in the saccular stage of fetal lung development. There remains considerable controversy regarding the origin of type 1 cells. In culture, these cells demonstrate slow turnover, with a doubling time estimated to be between 40 and 120 days, suggesting that they are functionally terminally differentiated in vivo. Furthermore, in response to epithelial denudation occurring with lung injury, type 2 cells proliferate to reestablish epithelial continuity, and then lose 574 PART X Respiratory System Lamellar body Alveolus Lung Monolayer–multilayer film Tubular myelin Adsorption Nucleus Lipid vesicle Turnover Clearance Secretion Recycling A Lamellar body Synthetic SP-B, SP-C Alveolar macrophage Alveolar type II cell Multivesicular body B 2 microns FIGURE 42-2 Surfactant life cycle. A, Schematic diagram depicting the life cycle of surfactant. B, Electron micrograph of a type II alveolar epithelial cell showing the prominent lamellar bodies near the apical surface. (A, Reprinted with permission from Whitsett J, Weaver T: Hydrophobic surfactant proteins in lung function and disease, N Engl J Med 347:2141–2148, 2002. Copyright 2002 Massachusetts Medical Society. All rights reserved.) phenotypic features such as lamellar bodies and acquire markers of type 1 cells, suggesting that rapid repopulation of type 1 cells requires a type 2 cell intermediary. There is increasing appreciation for the alveolar type 1 cell as more than a passive membrane for gas exchange (Williams, 2003). The large surface area and small cytoplasm/nucleus ratio provides for a thin alveolocapillary membrane to facilitate gas exchange. However, this large surface area also provides a large absorptive surface in the lung. The presence of water and ion channels, some distinct from those in type 2 cells, facilitates the maintenance of a relatively dry alveolus. Type 1 cells may also regulate cell proliferation locally, signal macrophage accumulation, and modulate the functions of local peptides, proteases and growth factors. Although most notable for its role in surfactant production, the alveolar type 2 cell provides additional important functions in the alveolus (Fehrenbach, 2001). Alveolar type 2 cells are local progenitor cells, as mentioned previously. Like type 1 cells, alveolar type 2 cells contain specialized ion and water channels as well as ion pumps in both the apical and basal membranes that contribute to the movement of water and ions across the epithelium (see Static Stretch: Fetal Lung Fluid Protection, later). Type 2 cells contain and secrete important antioxidants (superoxide dismutases 1, 2, and 3, glutathione) and molecules of innate host defense (SP-A, SP-D, lysozyme) to participate in detoxification and sterilization of the alveolar microenvironment. More recently, it is becoming clear that alveolar type 2 cells may also play a part in exacerbating alveolar pathology. The type 2 cell participates in the coagulation-fibrinolysis cascade through the production of fibrinogen, urokinasetype plasminogen activator, and tissue factor, especially under pathologic circumstances. Type 2 cells are increasingly recognized as a source of cytokine and chemokine production in the lung, as well as growth factors that can promote fibrosis. Finally, cross-talk between epithelial cells, cell matrix, interstitial cells, and local inflammatory cells can foster the resolution of injury and inflammation or prolong lung remodeling after injury with detrimental effects, such as lung destruction and fibrosis. Although previously heralded as the defender of the alveolus, the alveolar type 2 cell have a much more complex role in alveolar health and disease. SURFACTANT Pulmonary surfactant is essential to alveolar health. A thin layer of liquid is constantly secreted into the alveolar lumen to protect the delicate alveolar epithelium. The surface tension generated by this aqueous layer opposes alveolar inflation and promotes alveolar collapse at the end of expiration, because of Laplace’s law, which states that the collapsing pressure on the alveolus is directly proportional to the surface tension while inversely proportional to the radius of the alveolus. The film of pulmonary surfactant at the air-liquid interface lowers surface tension as alveolar surface area decreases, thereby preventing end-expiratory atelectasis, maintaining functional residual capacity, and lowering the force required for subsequent alveolar inflations. Pulmonary surfactant is a complex mixture of phospholipids, neutral lipids, and proteins that is synthesized, packaged, and secreted by alveolar type 2 cells (Zuo et al, 2008). The life cycle of surfactant is depicted in Figure 42-2. Storage of surfactant occurs in the lamellar body, a lysosome-derived membrane-bound organelle that undergoes regulated secretion in response to a variety of stimuli, including stretching. In the alveolus, surfactant phospholipids transition through an extracellular storage form— tubular myelin. Phospholipid and protein components are recycled out of the surfactant monolayer at the air-liquid interface and taken back into the alveolar type 2 cell, where they can be repackaged into lamellar bodies. Alternatively, alveolar macrophages are able to engulf and degrade surfactant components. The predominant surfactant phospholipid is saturated dipalmitoylphosphatidylcholine, with the remaining phospholipids consisting of monounsaturated phosphatidylcholine, phosphatidylglycerol and other phospholipids (Table 42-1). Dipalmitoylphosphatidylcholine is the only surface active component of lung surfactant capable of lowering surface pressure to nearly zero. The presence of unsaturated phospholipids and other lipid components like CHAPTER 42 Lung Development: Embryology, Growth, Maturation, and Developmental Biology TABLE 42-1 Composition of Pulmonary Surfactant Component Lipid Percentage (by Weight) 90 Saturated phosphatidylcholine 45 Unsaturated phosphatidylcholine 25 Phosphatidylglycerol Other phospholipids Neutral lipids Protein 5 5 10 10 Surfactant proteins 5 Serum proteins 5 cholesterol enables the monolayer to remain fluid at body temperature during the respiratory cycle. Phospholipid content in the fetal lung increases with advancing gestation because of increased activity of enzymes responsible for phospholipid synthesis within alveolar type 2 cells. The expression and activity of enzymes of the choline incorporation pathway, the predominant pathway for surfactant phospholipid synthesis, are developmentally regulated and induced by hormones. The inductive hormones that have direct clinical relevance are glucocorticoids and agents that increase intracellular cyclic adenosine monophosphate (cAMP) such as the β-adrenergic agonist (and tocolytic) terbutaline. Surfactant contains a group of specific proteins with importance to surfactant function and host defense. The four surfactant proteins SP-A, -B, -C, and -D are subdivided based on their physical characteristics into either hydrophobic (SP-B and -C) or hydrophilic (SP-A and -D) proteins. The hydrophobic surfactant proteins play a major role in the surface-active properties of surfactant, whereas the primary roles of the hydrophilic surfactant proteins are in host defense, immunomodulation, and surfactant clearance and metabolism (see Surfactant, earlier). Together, the hydrophobic proteins facilitate the mobilization of surfactant phospholipid from tubular myelin to the surface monolayer, promote spreading of phospholipids in the surfactant film, and assist in film stability at end-expiration (Zuo, 2008). SP-B plays a central role in alveolar health because of its critical function in surfactant homeostasis. It is a secretory protein that exhibits strong association with membranes, unlike SP-C, which contains a membrane-spanning domain and covalently attached fatty acids (palmitate) that render it integral to phospholipid membranes (Conkright et al, 2001). Both SP-B and SP-C are synthesized as large precursor proproteins that undergo extensive posttranslational processing as they pass through the secretory pathway, ultimately reaching the lamellar body. SP-B is essential for the process of lamellar body formation, and the alveolar type 2 cells of infants with inherited deficiency of SP-B are devoid of lamellar bodies. Because the lamellar body is where SP-C processing is completed, infants with inherited deficiency of SP-B are also deficient in mature SP-C, instead accumulating a larger, nonfunctional precursor of SP-C. Therefore patients with inherited deficiency of SP-B, despite having relatively normal surfactant phospholipid profiles, make a 575 pulmonary surfactant with poor surface tension properties because of the combined defects in SP-B and SP-C. Conversely, because SP-C does not have either a direct nor indirect role in SP-B protein processing, animals with SP-C deficiency have normal SP-B, normal lamellar bodies, and relatively normal surfactant function, and they exhibit no perinatal lethality because of surfactant dysfunction. Like the enzymes of surfactant phospholipid production, SP-B and SP-C exhibit developmental and hormonal regulation of expression (Mendelson, 2000). In human fetuses, SP-C mRNA is detected as early as 12 weeks’ gestation and SP-B mRNA by 14 weeks’ gestation, but the mature proteins are not detectable in fetal lung tissue until after 24 weeks’ gestation. SP-B protein is not detectable in amniotic fluid until after 30 weeks’ gestation, increasing toward term (Pryhuber et al, 1991), because of developmental regulation of posttranslational events in the proteolytic processing of proSP-B and proSP-C (Guttentag, 2008). Consequently, infants delivered prematurely have reduced levels of both surface active components of surfactant, phospholipid, and hydrophobic surfactant proteins, because of the developmental regulation of surfactant proteins and the enzymes of phospholipid production in alveolar type 2 cells. The rate of type 2 cell differentiation, and secondarily surfactant production by the fetal lung, is modulated by levels of endogenous corticosteroids and is accelerated by administering antenatal glucocorticoid to women in preterm labor. The response of the surfactant system to glucocorticoid involves all the lipid and protein components, and it occurs primarily through increased gene expression, thus representing precocious maturation mimicking the normal developmental pattern. Endogenous thyroid hormones, prostaglandins, and catecholamines also have stimulatory effects on type 2 cell maturation and clearance of lung fluid at birth. Certain proinflammatory cytokines (e.g., tumor necrosis factor [TNF]-α and transforming growth factor [TGF-β]) inhibit surfactant production in experimental systems and may downregulate surfactant in conditions such as sepsis and inflammation. A partial list of hormones capable of inducing or inhibiting lung maturation is presented in Table 42-2. DEVELOPMENT OF THE PULMONARY VASCULATURE The pulmonary vasculature consists of the vascular supply to the acini and the bronchial circulation (Hislop, 2005). During early fetal life, the airways act as a template for pulmonary blood vessel development. The earliest pulmonary vessels form de novo in the tissue that surrounds the lung bud by differentiation of mesenchymal cells into endothelial cells and then capillaries, a process known as vasculogenesis. Mesodermal cells within the mesenchyme investing the developing lung tube differentiate into endothelial cells, proliferate, organize into chords, and develop a central lumen. As each new airway buds into the ­mesenchyme, a new plexus forms and adds to the pulmonary circulation, thereby extending the arteries and veins. By 5 weeks’ gestation, a capillary network surrounds each bronchus and circulation of blood between the right ventricle and the left atrium via this network is evident. 576 PART X Respiratory System TABLE 42-2 Hormonal Regulators of Lung Maturation Inducers Inhibitors Glucocorticoids (cortisol) Major endogenous ­modulator of ­alveolar development and ­surfactant production β-Adrenergic ­agonists ­(epinephrine), cAMP Increase surfactant ­production and secretion, especially during labor and delivery Thyroid hormones (T3, T4) Enhance glucocorticoid effects on lipid synthesis Retinoic acid May interact with ­glucocorticoids to regulate surfactant phospholipid and protein production Bombesin-related peptides, ­parathyroid hormone–related protein May contribute to surfactant lipid synthesis Protein kinase C activators (­proinflammatory cytokines) Inhibit surfactant protein gene transcription during infection, inflammation TGF-β family Inhibit type 2 cell ­maturation during early gestation and with ­inflammation TNF-α Inhibit SP-B and SP-C gene transcription during infection Insulin Inhibit surfactant ­protein gene transcription in infants of poorly ­controlled diabetic women Dihydrotestosterone Delayed type 2 cell ­maturation in males By the canalicular stage of lung development, continued branching of the airways is accompanied by thinning of the epithelium. At this stage, new blood vessels form from preexisting vessels, a process known as angiogenesis. By comparison, angiogenesis is initiated by endothelial cell proliferation and sprouting from established vessels, resulting in the extension of a vascular network into undervascularized regions. Vasculogenesis is the primary mode of pulmonary vascular development until 17 weeks’ gestation, when all preacinar airways and their accompanying vessels are present, whereas angiogenesis becomes the predominant mode of vascular development in the later stages of lung development. Although originally thought to be sequential processes, it is generally accepted that vasculogenesis occurring in the periphery and more central angiogenesis occur concurrently during lung development (deMello et al, 1997). Interconnections between vascular networks arising from both angiogenesis and vasculogenesis increase in the saccular phase of lung development. In the human lung, a second circulatory system, the bronchial circulation, arises from the dorsal aorta supplying systemic blood. The bronchial vasculature develops after the pulmonary circulation, with bronchial vessels first apparent by 8 weeks’ gestation. The network of bronchial vessels is extensive, with bronchial arteries demonstrated as distal as the alveolar ducts in the adult respiratory tree. The inappropriate branching of bronchial vessels from the dorsal aorta is implicated in the formation of bronchopulmonary sequestration, a space-occupying lung malformation that can result in hypoplasia of the ipsilateral lung. Vasculogenesis and angiogenesis are the primary mechanisms of vascular development throughout intrauterine life. The human lung at term contains only a small portion of the adult number of alveoli, and the airspaces walls are represented by a thick primary septum consisting of a central layer of connective tissue surrounded by two capillary beds, each of them facing one alveolar surface (Burri, 2006). As alveolar architecture changes with the appearance of secondary septa, or secondary crests, folding of one of the two capillary layers occurs within the secondary septa. This double capillary network is not present in the adult lung. Microvascular maturation involves fusion of the double capillary network into a single capillary system. The expansion of surface area and lumenal volume compresses the interstitium, bringing the capillary networks in close proximity to potential air spaces and thereby promoting both alveolar surface area expansion and capillary bed fusion. Interestingly, by the third postnatal week during which lung volume increases by 25%, there is a concomitant 27% decrease of the interstitial tissue volume that is believed to promote focal microvascular fusions. Subsequently, there is preferential growth of fused areas that continues until 3 years of age. Lastly, it is well known that lung volume increases approximately 23-fold between birth and young adulthood, and capillary volume expands 35-fold. It has been recently shown that this increase in capillary volume occurs by insertion of new capillary meshes in the absence of capillary sprouting. This new concept in capillary network growth has been named intussusceptive microvascular growth, and involves the formation of transluminal tissue pillars that then expand, resulting in increased capillary surface area (Burri, 2006). Muscularization can be detected early in development of the pulmonary arteries (Hislop, 2005). Initially the muscular investment of the vasculature is derived from the migration of bronchial smooth muscle cells from adjacent airways. Muscularization of preacinar and resistance arteries of the pulmonary vasculature begins in the canalicular stage and continues through the remainder of gestation. This second phase of smooth muscle cells investing pulmonary vessels develops from the surrounding mesenchyme. Fibroblasts in close proximity to developing arteries alter their cellular shape and begin to express α-smooth muscle actin, a marker of smooth muscle cells. A third phase of vascular muscularization has been described, largely in the distal lung, in which capillary endothelial cells undergo a process of endothelial-mesenchymal transition that encompasses endothelial cell division, separation, and migration from the endothelial layer and expression of smooth muscle cell markers. Muscularization of pulmonary arteries normally extends to the level of the terminal bronchiole and is minimal to absent in blood vessels surrounding respiratory bronchioles. Abnormal extension of smooth muscle along arterioles supplying acinar structures occurs in infants dying from persistent pulmonary hypertension of the newborn and in severe bronchopulmonary dysplasia. CHAPTER 42 Lung Development: Embryology, Growth, Maturation, and Developmental Biology 577 TABLE 42-3 Composition of Human Fetal Lung Fluid Compared With Other Body Fluids Component Lung Fluid Interstitial Fluid Plasma Amniotic Fluid Sodium (mEq/L) 150 147 150 113 Potassium (mEq/L) 6.3 4.8 4.8 7.6 Chloride (mEq/L) 157 107 107 87 Bicarbonate (mEq/L) 3 25 24 19 pH 6.27 7.31 7.34 7.02 Protein (g/dL) 0.03 3.27 4.09 0.10 MECHANISMS OF LUNG DEVELOPMENT Fetal and postnatal lung development depend on several key developmental processes: branching morphogenesis to promote branching of the lung bud into the surrounding mesenchyme, static and cyclic stretching of the lung that assist in promoting lung branching, alveolarization to enhance the expansion of the gas exchange surface area, and vasculogenesis and angiogenesis to ensure that the developing epithelial surface area is invested with a similarly extensive vascular supply. BRANCHING MORPHOGENESIS Branching morphogenesis is the fundamental mechanism of lung development. Branching is mediated by the accelerated growth of epithelial cells lateral to branch points with concomitant growth arrest at the branch point (Affolter et al, 2009). This process requires extensive communication between the cells lining the tubular lung bud and cells contained within the invested mesenchyme. Classic tissue recombination experiments in which mesenchyme from proximal airways was transplanted to distal airways (and vice versa) indicate that the mesenchyme has an important inductive role in dictating the branching pattern and cell fate of the expanding epithelium. More recently, studies of mouse lung development indicate that three modes of branching—domain branching, planar bifurcation, and orthogonal bifurcation—are the basic mechanisms that characterize the complex three-dimensional development of the respiratory tree through the pseudoglandular phase (Metzger and Krasnow, 1999). These routines occur repetitively during lung development and appear to be set by a genetic clock dictating when side branches form, a bifurcation program determining when a branch bifurcates, and a rotator function controlling the planar orientation of the bifurcation. STRETCH AND MECHANOTRANSDUCTION The role of physical factors in modulating lung size is well established; normal lung growth requires adequate space in the chest cavity and appropriate tonic and cyclic distending forces. Genetic defects that compromise the thoracic skeleton and space-occupying lung masses like congenital cystic adenomatoid malformations are associated with pulmonary hypoplasia because of the restriction of intrathoracic space. Denervation of the diaphragm to eliminate fetal breathing movements is also associated with pulmonary hypoplasia, as is the manipulation of fetal lung fluid volume. Static Stretch: Fetal Lung Fluid Production Fetal lung fluid is a product of the epithelial lining of the developing lung (Wilson et al, 2007), averaging 4 to 6 mL/kg/h. Because of the resistance imparted by laryngeal abduction, fluid accumulates to a total volume of 20 to 30 mL/kg during gestation, providing end-­expiratory pressure of approximately 2.5 cm H2O pressure (Kitterman, 1996). The composition of fetal lung fluid is distinct from both amniotic fluid and plasma, as shown in Table 42-3. The increased chloride content of fetal lung fluid compared with serum is the result of active chloride secretion by the tracheal and distal pulmonary epithelium, largely because of the chloride channel CLC-2/CLCN2. Fetal lung fluid secretion can be enhanced by prolactin, KGF, PGE2 and PGF2, whereas it is inhibited by a variety of mediators, including β-adrenergic agonists, vasopressin, serotonin, and glucagon. Fetal lung fluid is an essential component of lung development, but it presents a significant obstacle to the transition to air breathing upon delivery. Three important events must occur to decrease the amount of fetal lung fluid and its potential effect on alveolar surface tension: absorption, bulk removal, and maturation of pulmonary surfactant. The transition to air breathing must be coupled with a conversion from a secretory pulmonary epithelium to one that is absorptive. Enhanced sodium transport across the alveolar epithelium is in part responsible for this change. Much evidence suggests that induction of components of the epithelial sodium channels (ENaC) around the time of birth is a major factor in promoting sodium transport with water passively following the movement of sodium. Absence of the α-subunit of ENaC in transgenic mice is perinatal lethal because of the failure of fetal lung fluid clearance. Induction of ENaC components occurs at a transcriptional level in response to changes in extracellular matrix components, glucocorticoids, aldosterone, and oxygen. By comparison, agents that increase intracellular cAMP levels (i.e., β-agonists, phosphodiesterase inhibitors, and cAMP analogues), while not increasing the number of sodium channels, increase the probability of a channel being open to sodium transport. In addition, glucocorticoid and thyroid hormones have important roles in priming the lung epithelium to be responsive to the actions of β-adrenergic agonists on sodium transport across lung epithelia near term. Water channels consisting of aquaporins 578 PART X Respiratory System are also induced during the late fetal period to facilitate fluid movement, but their contribution is unclear because of the perinatal survival of mice in which aquaporin 5 or both aquaporins 5 and 1 were absent. Conversion to an absorptive surface is not sufficient to minimize the fetal lung fluid at the time of term delivery. The absence of uterine contractions is associated with an increased incidence of retained fetal lung fluid in infants delivered by cesarean section without the benefit of labor. On delivery of the head and neck, continued uterine contractions on the fetal thorax promote expulsion of bulk fluid from the fetal lung. However, animal studies have shown that the magnitude of the benefit of thoracic compression during labor is modest (Bland, 2001). The primary mechanism by which labor facilitates clearance of lung fluid is through hormonal effects on fluid clearance, especially through catecholamine-induced changes in the ENaC opening. The onset of air breathing, associated with increased intrathoracic negative pressure, assists in the clearance of residual fetal lung fluid into the loose interstitial tissues surrounding alveoli. Fluid is then reabsorbed via lymphatics and pulmonary blood vessels. It is generally accepted that the amount of residual liquid in the lung after complete transition is approximately 0.37 mL/kg body weight. Cyclic Stretch: Fetal Breathing Movements Fetal breathing is an essential stimulus for lung growth (Kitterman, 1996). Fetal breathing is readily detectable as early as 10 weeks’ gestation. Fetal breathing occurs for 10% to 20% of the time at 24 to 28 weeks’ gestation, increasing to 30% to 40% after 30 weeks’ gestation. Originating from the diaphragm, fetal breathing is erratic in frequency and amplitude, and it changes throughout gestation. The volume of fluid moved is small and insufficient to be cleared from the trachea. Respiratory rates range from 30 to 70 breaths/min, and periods of apnea of up to 2 hours have been recorded. Sustained periods of fetal breathing increase in duration with advancing gestation. The frequency of fetal breathing varies with sleep state (inhibited during quiet sleep) and exhibits diurnal variation, with the lowest rates recorded early in the morning. Fetal breathing is hormonally responsive and the inhibition of fetal breathing with the onset of labor is attributed to the action of increased circulating prostaglandins. Maternal medications can influence the frequency of fetal breathing movements. Central nervous system stimulants are associated with increased fetal breathing (i.e., caffeine, amphetamines), whereas depressants are associated with decreased fetal breathing (i.e., anesthetics, narcotics, ethanol). Maternal smoking is associated with reduced fetal breathing, largely because of increased fetal hypoxemia. Animal studies have shown that permanent cessation of fetal breathing, regardless of the insult, is associated with impaired fetal lung growth. However, the effects of short-term alterations in fetal breathing frequency and amplitude on fetal lung development are unknown. Together, constant distention from the production and retention of fetal lung fluid, and episodic cyclic fetal breathing, are important mechanisms for lung growth during fetal life. ALVEOLARIZATION Branching morphogenesis is the primary developmental program that establishes the conducting airways of the lung, but it is important to remember that alveolarization is the developmental program that will establish the large surface area involved in gas exchange (Galambos and Demello, 2008). This process will result in a 20-fold increase in surface area between birth (with between 0 and 50 million alveoli) and adulthood (>300 million alveoli). Primitive saccules develop low ridges (primary septa) that subdivide the saccule into an alveolar duct containing primary alveoli and outpouchings between the ridges (secondary septa/crests) that establish secondary alveoli (Figure 42-3). Regions destined for secondary septation exhibit increased elastin deposition (Mariani et al, 1997), and elastin localizes to the tips of the secondary crests as they form. Septae contain a connective tissue core separating two capillary membranes, suggesting that the septum is formed by the folding of a capillary on itself, as mentioned previously. Septation also leads to the development of the pores of Kohn, allowing gaseous continuity between acini. The process of alveolarization is poorly understood, but is receiving much attention because of observations that infants who die after severe bronchopulmonary dysplasia exhibit alveolar simplification with little evidence of secondary septation. It remains unclear to what extent the process of alveolarization is disrupted by preterm birth and whether this developmental program can be resurrected after preterm birth in the absence or presence of lung injury. INTERDEPENDENCE OF ALVEOLAR AND VASCULAR DEVELOPMENT Recent evidence suggests that the pulmonary capillary bed actively promotes normal alveolar development and contributes to the maintenance of alveolar structures throughout life (Thebaud and Abman, 2007). The observation that combined abnormalities in the airways and vasculature occur in bronchopulmonary dysplasia supports this hypothesis. Intraacinar arteries and veins continue to develop after birth by angiogenesis as long as alveoli continue to increase in number and size. This process may be reciprocal because vascular growth around the distal airspaces suggests an inductive influence from the alveolar epithelial cells as well. MOLECULAR BASIS FOR LUNG DEVELOPMENT The developmental processes that contribute to lung organogenesis are under the regulation of interdependent signaling pathways mediated by secreted growth factors that are themselves under the control of large networks of transcription factors controlling gene expression. Gene regulatory networks common to other organs that depend on branching morphogenesis during development, most notably in the kidney and mammary gland, are also found in the lung. Selected regulatory networks are highlighted in the following sections to illustrate these systems. For more detailed reading, additional references are provided (Kaplan, 2000; Kumar et al, 2005; van Tuyl and Post, 2000; Warburton et al, 2000;). CHAPTER 42 Lung Development: Embryology, Growth, Maturation, and Developmental Biology Conducting airway Primary septa 1. Wrinkling of primary septa 3. Transformation of conducting airways into respiratory airways 2. Outgrowth of secondary septa FIGURE 42-3 Development of primary versus secondary alveoli. The development of secondary alveoli from primary alveoli allows for expansion of the total surface area available for gas exchange. Continued growth of primary septa results in a wrinkled appearance that is followed by elastin deposition at points where secondary septae, also known as secondary crests, will form by an in-folding of the epithelium that pulls underlying capillaries into the septum. Finally the more proximal epithelium becomes differentiated into alveolar epithelial cells capable of participating in gas exchange. Shaded cells represent bronchiolar epithelial cells. GROWTH FACTORS IN LUNG DEVELOPMENT The initiation of branching morphogenesis is the result of the interplay of signals between the developing lung epithelial tube and its surrounding mesenchyme. Central to this process is the family of fibroblast growth factors (FGFs) that are produced and secreted by mesenchymal cells, and bind to receptors on the plasma membrane of epithelial cells, establishing a system of mesenchymalepithelial cross talk. In particular, the growth factor FGF10 secreted by mesenchymal cells binds to the receptor FGFR2b on nearby epithelial cells. This signal will be strongest in the closest epithelial cells because of a gradient that develops with diffusion of the secreted FGF10. Binding of ligand to receptor results in signal transduction within the epithelial cell that induces expression of a group of genes via the mitogen-activated protein kinase pathway. One of the genes induced by FGF10/FGFR2b signaling, Sprouty 2 (Spry2), inhibits the mitogen-activated protein kinase pathway, resulting in inhibition further FGF10/ FGFR2b signaling. Therefore a signal propagated by the mesenchyme has an effect that is subsequently dampened within the epithelial cell. Examination of genes expressed early in the formation of a branch point in response to FGF10 revealed a large number of genes that regulate cell adhesion, cytoskeleton, and cell polarity—all essential elements of cell migration (Lu et al, 2005). Surprisingly, bud initiation was not associated with the expression of genes that promote proliferation, which appears to be more important for branch elongation. Still, animals expressing reduced FGF10 develop pulmonary hypoplasia with reduced numbers of large airways (Abler et al, 2009; Ramasamy et al, 2007). Furthermore, increased FGF10 signaling during fetal lung development in mice by intrapulmonary injections of recombinant FGF10 produced cystic structures with epithelial characteristics dependent 579 on the location of the injection: proximally with Clara cells, distally with alveolar type 2 cells (Gonzaga et al, 2008). These data provide strong evidence that FGF10 signaling has diverse responses—from initiation of branching to differentiation of epithelial cells—depending on the temporal or spatial context of signaling. Like branching morphogenesis, lung vascular development is a complex and highly organized process that requires multiple vascular signaling molecules to interact in a specific temporospatial sequence. Vascular endothelial growth factor (VEGF) is a critical growth factor in angiogenesis and vasculogenesis. The expression of the VEGF ligand by epithelial cells and VEGF receptors by endothelial cells of the developing human fetal lung reinforces the interdependence of the airspace and vascular development. The expression of VEGF mRNA and protein is localized to the epithelial cells at the distal tips of developing lung branches, and expression levels increase with time (Galambos and Demello, 2007; Hislop, 2005). VEGF gene expression is induced in epithelial cells by the hypoxic environment of the growing fetal lung through the actions of the oxygen-sensing hypoxia-inducible factor (HIF) family of transcription factors. From the single VEGF gene, five different VEGF protein isoforms can be produced, although VEGFA (VEGF165) is the most studied. Each isoform has different affinities for each of the three VEGF receptors (Flt-1/VEGFR1, Flk-1/ KDR/VEGFR2, and Flt-4/VEGFR3). Vascular endothelial cells express primarily VEGFR-1 and VEGFR-2, whereas VEGFR-3 is on the plasma membrane of the lymphatic endothelium. VEGF receptors are expressed on the plasma membrane endothelial cells surrounding the developing airways from very early in gestation, and expression of VEGFR2/Flk1 is considered the earliest marker of an endothelial progenitor cell. In vitro and in vivo experiments have shown that VEGFA induces endothelial cell proliferation and migration, both key elements of vascular sprouting, as well as tube formation through interactions with VEGFR2. VEGFR1 appears to have more importance in transforming primitive endothelial tubes into more stable vascular networks, in part by reducing endothelial proliferation through downregulation of VEGF production. The embryonic lethality of animals with reduced VEGF expression attests to the critical importance of VEGF and VEGFR signaling to vascular development in the fetus, though not limited to the developing pulmonary vasculature. TRANSCRIPTION FACTORS IN LUNG DEVELOPMENT The ligand-receptor interactions important to branching morphogenesis and pulmonary vascular development are in part determined by the actions of transcription factors on facilitating or reducing gene expression. Transcription factors are also critical in the differentiation of the lung epithelium from the most rudimentary lung bud out to the type 1 and 2 alveolar epithelial cells. The more important transcription factors in epithelial cell determination in the lung include the Gli, hepatocyte nuclear factor (HNF), GATA, and Hox families of transcription factors (Kumar et al, 2005; Warburton et al, 2000). 580 PART X Respiratory System The most important transcription factor in the lung is thyroid transcription factor-1 (TTF-1), a product of the Nkx2.1 gene. TTF-1 is considered a master regulator of lung development, as transgenic mice null for Nkx2.1 exhibit complete absence of lung branching. However, Nkx2.1 also has a prominent role in establishing cell fate proximally to distally along the branching lung epithelium. TTF-1 expression is detected as early as 11 weeks’ gestation in humans, and continues to be expressed by epithelial cells at the distal tips of the branching lung epithelium, ultimately becoming more restricted to Clara cells and alveolar type 2 cells. TTF-1 is critical for the expression of genes that are unique to differentiated epithelium, such as CC10 expression in Clara cells and surfactant proteins in alveolar type 2 cells. DNA binding sites for TTF-1 are found in the promoter regions of all four surfactant proteins, CC10, and Nkx2.1 itself, creating a positive feedback loop for sustained TTF-1 expression. TTF-1 function is highly dependent on phosphorylation of critical amino acids, although it remains unclear which kinase is involved in this process (DeFelice et al, 2003). Interestingly, VEGFA expression is reduced in animals unable to phosphorylate TTF-1, providing another important link between epithelial and vascular development. TTF-1 is itself regulated by other transcription factors that bind to the promoter region of the Nkx2.1 gene, specifically HNF-3β and GATA-6. Therefore the ability of networks of transcription factors to bind to gene regulatory elements of DNA in a coordinated fashion during fetal lung development enables the temporospatial expression of growth factor networks that foster branching morphogenesis, the process of differentiation that ultimately gives rise to the approximately 40 cell types that constitute the human lung, as well as critical coordination of both epithelial and vascular development. OTHER TOPICS IN LUNG DEVELOPMENT DEVELOPMENT OF PULMONARY HOST DEFENSE The adult human lung takes in approximately 7 L/min of air contaminated with a variety potential pathogens that can cause epithelial injury. The continuous exposure of the epithelial surface of the conducting airway to inhaled pathogens requires the presence of an efficient innate immune response system to prevent infections. The proximal and distal airway epithelia have a major role in clearing pathogens by secreting antimicrobial as well as antiinflammatory molecules. Components of mucociliary clearance appear as early as 11 weeks’ gestation with the differentiation of ciliated epithelial cells and the expression of mucus in mucus-secreting goblet cells within the epithelium and submucosal glands, as discussed previously. The number of goblet cells peaks in mid-gestation (30% to 35% of total airway epithelial cells), decreasing toward term to levels lower than in adults, and then increasing after birth. The postnatal increase in goblet cells occurs after preterm birth as well, giving premature infants a greater number of goblet cells than in term infants. Therefore premature infants are prone to more mucus production in smaller airways than are term infants, and premature infants have relatively fewer ciliated cells to assist in mobilization of secretions. A number of microbial defense molecules are produced and secreted by epithelial cells into the airways (Bartlett et al, 2008). They include lysozyme, C-reactive protein, lactoferrin, collectins, β-defensins, and the only human member of the cathelicidin family, hCAP-18/LL-37 (Hiemstra, 2007). Two lung collectins were originally identified as surfactant-related proteins—the hydrophilic surfactant proteins SP-A and SP-D. Although originally identified as products secreted by epithelial cells lining airways (Clara cells) and alveoli (type 2 cells), SP-A and SP-D have been found in other sites associated with epithelial surfaces exposed to the external environment (Haagsman et al, 2008). They interact with microorganisms, inflammatory cells, and leukocytes to facilitate clearance of microorganisms from the airspace, and they modulate allergic responses. The basis for the interactions of the lung collectins with microbes and antigens centers on the binding of sugars by the carbohydrate recognition domains of these proteins (Wright, 2005). Both collectins bind a variety of fungi and Pneumocystis carinii, and they have an important role in inhibiting a variety of respiratory viruses, including influenza A and respiratory syncytial virus. Differences in the structure of the carbohydrate recognition domain provide SP-A and SP-D with altered affinities for different sugar molecules, allowing complementary functions and improving the diversity of microbial interactions. Interactions with gram-negative organisms frequently depends on the ability of SP-A and SP-D to bind lipopolysaccharide, whereas the mechanism of SP-A interactions with grampositive organisms, including group B beta-hemolytic streptococci, are not as clear. The lung collectins also modulate the functions of a variety of immune cells, including macrophages, neutrophils, eosinophils, and lymphocytes (Wright, 2005). In addition to opsonizing microorganisms, functions of the lung collectins include stimulating chemotaxis of macrophages and neutrophils, enhancing cytokine production by macrophages and eosinophils, attenuating lymphocyte responses by inhibiting T cell proliferation, and modulating the production of reactive oxygen and nitrogen species used in killing microorganisms. Like the hydrophobic surfactant proteins, SP-A and SP-D exhibit both developmental and hormonal regulation of expression (Mendelson, 2000). In human fetuses, SP-A mRNA is undetectable before 20 weeks’ gestation, and SP-A protein is first detectable in amniotic fluid by 30 weeks’ gestation, increasing toward term (Pryhuber et al, 1991). In humans, SP-A gene expression is induced by cAMP and glucocorticoids, although the response to glucocorticoids in biphasic, showing attenuation at higher doses. Retinoids, insulin, and growth factors such as TGFβ and TNF-α inhibit SP-A gene expression. Like SP-A, SP-D levels in human lung are low during the second trimester (Dulkerian et al, 1996), are detectable in amniotic fluid, and increase toward term (Whitsett, 2005). Levels of both SP-A and SP-D increase markedly in the first days after preterm birth. CHAPTER 42 Lung Development: Embryology, Growth, Maturation, and Developmental Biology DEVELOPMENT OF DETOXIFICATION SYSTEMS Although essential to cellular processes, oxygen concentrations beyond the physiologic limits may be hazardous to cells. The lung is particularly susceptible to reactive forms of oxygen and free radicals, because it is the organ with the highest exposure to atmospheric oxygen. The fetal lung is exposed to oxygen tensions of 20 to 25 mm Hg in utero, and the transition to air breathing is associated with a fourfold to sevenfold increase in oxygen tension, presenting a significant oxidant stress. Oxygen free radicals arise from endogenous production through metabolic reactions, or by exogenous exposure, from air pollutants and cigarette smoke, and they can result in lung injury from oxidation of proteins, DNA, and lipids. Therefore it is imperative for the lung to develop an antioxidant detoxification system, and for this system to be functional upon the fetal transition to air-breathing. Oxygen free radicals are highly toxic substances. Superoxide, produced by the reduction of molecular oxygen by the addition of an electron, is formed by all cells and occurs in particularly high concentrations in phagocytic cells to facilitate the killing of microorganisms. Hydrogen peroxide is generated from the transfer of a single electron to superoxide, and hydroxyl radicals are generated from the interaction of hydrogen peroxide with superoxide. The free electrons of free radicals interact with membrane lipids, resulting in lipid peroxidation, with sulfhydryl and other groups on exposed amino acids in proteins and with DNA causing direct damage. These lipid, protein, and nucleic acid modifications damage airway and alveolar epithelial cells as well as capillary endothelial cells, leading to altered epithelial integrity, interstitial and airspace edema, and infiltration of inflammatory cells. Increased reactive oxygen species levels have also been implicated in initiating inflammatory responses through the activation of transcription factors such as NFκB and AP-1, signal transduction pathways, and chromatin remodeling that foster the expression of proinflammatory mediators (Rahman et al, 2006). Oxygen can have additional effects beyond the direct toxicity of free radicals. Oxygen levels regulate the activity of plasminogen activator inhibitor-1 and other protease-antiprotease systems within the airspaces, thereby modulating the destructive effects of proteases elaborated from inflammatory cells. Oxygen can also regulate cellular growth responses by altering the secretion of growth factors and DNA synthesis, generally through oxygen-sensing transcription factors, such as the HIF family. Antioxidants attenuate the effects of oxygen free radicals in the lungs, and in the lung there are both nonenzymatic and enzymatic antioxidants. The major nonenzymatic antioxidants are gluthatione (GSH), vitamins C (ascorbate) and E (primarily α-tocopherol), β-carotene, and uric acid. Enzymatic antioxidants include superoxide dismutases 1, 2, and 3, catalase, and a variety of peroxidases. Animal studies indicate that many of the antioxidant enzymes are induced before term delivery, and limited data suggest that the same is true of human fetuses (Weinberger et al, 2002). Premature animals fail to induce antioxidant enzymes in response to oxidative lung injury (Thibeault, 2000). Therefore preterm infants are significantly more 581 compromised in antioxidant defenses and, because of the need for oxygen in the treatment of respiratory distress syndrome, they are more susceptible to oxygen toxicity. NOVEL CONCEPTS IN LUNG DEVELOPMENT STEM AND PROGENITOR CELLS IN THE LUNG The ability of lung epithelium to replace cells damaged from normal aging or injury has become the focus of recent attention (Rawlins et al, 2008; Warburton et al, 2008). Stem cells are undifferentiated and have an unlimited capacity for self-renewal. Asymmetric divisions allow for self-renewal through one daughter cell while enabling the other daughter cell to become more terminally differentiated. Progenitor cells are more committed, and although capable of self-renewal, they have more restricted cell fates. Thus far, three cell types in the lung have been identified as progenitor cells, having the capacity for self-renewal and for replacement of a variety of specialized lung epithelial cells. Basal cells in large airways and submucosal glands, identified by the expression of Trp63/p63 and cytokeratin 5, are able to self-renew and give rise to ciliated and secretory cells. Clara cells in smaller airways, identified by the expression of CC10/Scgb1a1, seem to be more committed progenitor cells because they are able to self-renew or differentiate into ciliated cells. A subset of Clara cells at the bronchoalveolar duct junction are both CC10/Scgb1a1 and SP-C positive, and thus have the potential to produce either Clara cells or alveolar type 2 cells. The most limited lung epithelial progenitor cell is the alveolar type 2 cell, which divides rapidly to reestablish epithelial continuity in damaged alveoli and then transdifferentiates into alveolar type 1 cells. The mesenchymal progenitor cells that provide for vascular and muscular components of the developing lung have been studied less. Candidate cells have been identified that give rise to endothelial cells in the process of vasculogenesis and airway smooth muscle cells along the branching lung epithelial tubes. Evidence for the existence of stem-progenitor cells in the lung is strong, but limited largely to mouse models of lung development and lung repair; therefore extrapolation to humans should be done with caution. The capacity for self-renewal is tantalizing, but it means that such cells have a risk for autonomous growth such as cancer. Controversy exists around the potential to harness these populations of cells as a means for correcting errors in lung development (pulmonary hypoplasia), genetic diseases of the lung (cystic fibrosis), and abnormal repair of injured lungs (bronchopulmonary dysplasia). EPITHELIAL TO MESENCHYMAL TRANSITION Fibrosis often occurs as the result of severe injury to the lung and has historically been a component of bronchopulmonary dysplasia (BPD). Emerging evidence suggests that expansion of fibroblast populations in fibrotic lesions is not simply the result of increased proliferation of local fibroblast populations, rather the transformation of epithelial cells into mesenchymal components, a process known as epithelial-mesenchymal transition (EMT) (Guarino 582 PART X Respiratory System TABLE 42-4 Potential Effects of Premature Birth on Lung Development and Maturation Event Effect of Preterm Birth Potential Consequences Development of ­conducting airways Branching: no effect; completed to the level of the respiratory bronchi by 24 wk gestation Tone: increased secondary to lung disease in part reflecting ­developmental deficiency of nitric oxide None Increased airways resistance Alveolarization Variable depending on timing of delivery and severity of lung disease; may also be compromised by excess glucocorticoids Reduced lung growth and lung surface area with increased alveolar size; impaired ­pulmonary function Development of ­alveolocapillary ­membrane Minimal, reaches adult diameter by 24 wk gestation; ­glucocorticoids induce precocious thinning Gas exchange largely dependent upon surface area, not alveolocapillary diameter Type II cell ­differentiation Variable immaturity and deficient surfactant production ­depending on timing of delivery; improved with antenatal ­glucocorticoids Developmental deficiency of surfactant ­content and composition results in RDS Type I cell differentiation Variable depending on timing of delivery; cells develop from type II cells Gas exchange largely dependant upon surface area Hydrophobic surfactant proteins (SP-B, SP-C) Variable depending on timing of delivery and other factors, such as infection, that can impair gene transcription High alveolar surface tension; RDS Hydrophilic surfactant proteins (SP-A, SP-D) Variable depending on timing of delivery; both proteins appear relatively late in third trimester Comprised host defense: ability to clear microorganisms from airways, alveolar space, or both; impaired ability to modulate ­inflammatory responses Clara cell differentiation Variable depending on timing of delivery; these cells appear in middle 2nd trimester, but antioxidant products appear late in 3rd trimester Impaired antioxidant and antimicrobial defenses; may contribute to chronic lung disease and pneumonia Induction of antioxidant systems Variable depending on timing of delivery; expression of ­antioxidants occurs late in third trimester Lung injury from oxidant stress exacerbated by need for increased oxygen; may contribute to chronic lung disease Mucociliary clearance Variable depending on timing of delivery; goblet cells decrease in number toward term Increased mucous production may obstruct small airways Development of the ­pulmonary capillary bed Variable depending on timing of delivery in parallel to alveolar development Variable degrees of impaired gas exchange commensurate with impaired alveologenesis and any superimposed lung injury; pulmonary hypertension Pulmonary arteries Variable depending on presence and severity of associated lung disease Pulmonary hypertension associated with chronic lung disease Fetal lung liquid Fluid loss: variable effects depending on magnitude and duration of fluid loss (i.e., prolonged premature rupture of membranes) as well as timing of delivery Fluid retention: variable effects depending on timing of ­delivery, because hormone surges near term and in labor promote ­reabsorption before delivery Pulmonary hypoplasia Transient tachypnea of the newborn Fetal breathing ­movements Variable depending on timing of delivery, but also ­depending on maternal exposure to substances that reduce fetal ­breathing movements Unlikely to have effects in preterm infants unless coexisting conditions severely limit fetal breathing Respiratory drive Variable depending on timing of delivery Apnea of prematurity RDS, Respiratory distress syndrome. et al, 2009; Thiery and Sleeman, 2006). EMT is an essential part of gastrulation and the development of cardiac valves. Neural crest cells are the best model system for EMT, because these epithelial cells must lose their local attachments and travel long distances before locating their final niche and differentiating accordingly. The evidence is clear that the changes in cell phenotype characteristic of EMT—from sedentary, interconnected epithelial cell expressing epithelial marker proteins to mobile, proliferative mesenchymal cell expression markers of fibroblasts and secreting collagen and other components of extracellular matrix—are under the control of local growth factors. In the lung, the most prominent growth factor in EMT associated with pulmonary fibrosis is TGF-β, a growth factor essential for normal branching morphogenesis during early lung development. It remains uncertain whether EMT has a similarly important role in lung development as it does in cardiac development, and whether EMT alone is important to the abnormal repair response to lung injury. EMT has the potential to reduce epithelial populations, resulting in lung destruction, while expanding mesenchymal fibroblast pools and enhancing local matrix deposition that reduces lung elasticity. Many of the models of EMT are derived from lung cancer cell lines or mouse models, limiting their applicability to humans. However, improved understanding of the processes controlling EMT could lead to novel therapies for limiting or reversing pulmonary fibrosis after CHAPTER 42 Lung Development: Embryology, Growth, Maturation, and Developmental Biology lung injury. Furthermore, given the potential for EMT to contribute to lung destruction, a more common feature of the “new BPD,” early events in EMT may be particularly important targets of preventive therapy in premature infants. ROLE OF miRNA IN LUNG DEVELOPMENT AND MATURATION Regulation of normal development and consequences of abnormal development are at the heart of understanding the implications of preterm birth and developing potential protective lung therapies. The central tenet of DNA to RNA to protein is being challenged by new epigenetic mechanisms—histone modifications, modification of DNA and RNA, silencing RNA, and micro RNA (miRNA)—for regulating gene expression. The evidence that miRNA has an important role in normal fetal development is strong (Nana-Sinkam et al, 2009). miRNAs are small, noncoding RNAs (generally 19 to 25 nucleotides long) found within cells that target genes for RNA degradation or inhibition of protein synthesis. There are more than 500 recognized miRNA, some of which are particularly enriched in lung cell populations. MiRNAs are generated from a process that begins in the nucleus. Long primary miRNA are transcribed, processed, and exported from the nucleus, followed by further cytoplasmic maturation before the final miRNA is able to interact with regions of messenger RNA, usually in the 3ˈ untranslated region of RNA that extends beyond the RNA sequence used for protein translation. Emerging evidence indicates that miRNA is essential for normal lung development, because targeted deletion of Dicer, a key enzyme in miRNA processing, results in abnormal airway development and excessive apoptosis in the lungs. The miR-17-92 cluster of miRNA is highly expressed in embryonic mouse lung, decreasing into adulthood as lung development progresses. Altered expression of the miRNA cluster miR-17-92 suggests a primary role for these miRNA in maintaining a population of undifferentiated lung epithelium during lung development. Because of their role in regulating developmental and pathologic processes, miRNAs are increasingly seen as targets for therapeutic interventions. Obstacles to miRNAs as therapeutic agents are similar to obstacles encountered in other gene therapies, including mode of delivery, cell and tissue specificity, and the potential for off-target effects. SUMMARY Lung branching morphogenesis is coordinated with pulmonary vascular development to provide a large surface area and thin alveolocapillary membrane for adequate gas 583 exchange in the transition to air breathing and to meet the needs of a growing infant and child. Although maturation occurs late in fetal lung development, it is similarly critical in the transition to air breathing. Although the fetal lung developmental program requires an array of transcription factors, hormones, and growth factors promoting branching morphogenesis, lung growth is equally dependent on intact neural input to modulate fetal breathing, stability of an appropriately sized thorax, and the presence of adequate lung and amniotic fluid. Furthermore the maturation of the host defense and detoxification systems minimize the effects of increased oxygen tension and exposure to potential pathogens accompanying the transition to air breathing. Premature birth affects all these functions as illustrated in Table 42-4. Bronchopulmonary dysplasia is the net result of multiple injuries to the underdeveloped lungs of premature newborns that compromise postnatal growth and development, and thus impairing function. Integrated approaches to therapy that reflect the interdependency of these lung functions have the most promise for minimizing the effects of premature birth on childhood and, ultimately, adult lung function. SUGGESTED READINGS Burri PH: Structural aspects of postnatal lung development: alveolar formation and growth, Biol Neonate 89:313-322, 2006. Galambos C, Demello DE: Regulation of alveologenesis: clinical implications of impaired growth, Pathology 40:124-140, 2008. Hislop A: Developmental biology of the pulmonary circulation, Paediatr Respir Rev 6:35-43, 2005. Jeffrey PK: The development of large and small airways, Am J Respir Crit Care Med 157:S174-S180, 1998. Kaplan F: Molecular determinants of fetal lung organogenesis, Mol Genet Metab 71:321-341, 2000. Kitterman JA: The effects of mechanical forces on fetal lung growth,, Clin Perinatol 23:727-740, 1996. Kumar VH, Lakshminrusimha S, El Abiad MT, et al: Growth factors in lung development, Adv Clin Chem 40:261-316, 2005. Mallampalli RK, Acarregui MJ, Snyder JM: Differentiation of the alveolar epithelium in the fetal lung. In McDonald JA, editor: Lung growth and development, vol 100, New York, 1997, Marcel Dekker, pp 119-162. Mendelson CR: Role of transcription factors in fetal lung development and surfactant protein gene expression, Annu Rev Physiol 62:875-915, 2000. Thibeault DW: The precarious antioxidant defenses of the preterm infant, Am J Perinatol 17:167-181, 2000. Warburton D, Schwarz M, Tefft D, et al: The molecular basis of lung morphogenesis, Mech Dev 92:55-81, 2000. Zuo YY, Veldhuizen RA, Neumann AW, et al: Current perspectives in pulmonary surfactant–inhibition, enhancement and evaluation, Biochim Biophys Acta 1778:1947-1977, 2008. Complete references used in this text can be found online at www.expertconsult.com C H A P T E R 43 Control of Breathing Estelle B. Gauda and Richard J. Martin The developmental aspects of respiratory control are of considerable interest to physiologists and clinicians for a multitude of reasons. The transition from fetal to neonatal life requires a rapid conversion from intermittent fetal respiratory activity not associated with gas exchange to continuous breathing upon which gas exchange is dependent. In addition, the circuitry that regulates respiratory control serves as a unique link between the maturing lung and brain. The term infant has the repertoire of respiratory and cardiovascular responses that meets his or her metabolic demands during wakefulness, sleeping, ­crying, and feeding even though the breathing pattern may be punctuated with apneic pauses and sighs (augmented breaths), and the neurocircuitry that controls breathing is not completely developed. In some term infants who appear well, developmental abnormalities in the neurocircuitry and neurochemistry that regulate respiratory rhythmogenesis can lead to profound disorders of breathing, placing them at increased risk for sudden death. Infants who have died of sudden infant death syndrome (SIDS), and those infants with central congenital hypoventilation and Rett syndrome, have provided a window of opportunity for careful epidemiologic, genetic, neurochemical, and anatomic study that has allowed researchers to better understand which genes regulate the development of the respiratory system and how environmental factors in fetal and early neonatal life may adversely affect normal development. The infant who is born prematurely has also provided a unique opportunity to observe the natural developmental trajectory of respiratory control. Breathing in the most premature infants is akin to fetal breathing, which is episodic, punctuated by periods of disturbingly long apneic pauses interspersed with frequent periods of hyperventilation and sighs (augmented breaths). The purpose of fetal breathing movements is not gas exchange, rather it is lung development by inducing sufficient stretch of the chest wall. For the infant who is born prematurely, the frequent apneic events and hypoventilation associated with oxygen desaturations and bradycardia are of significant concern. The purpose of this chapter is to present what is currently known about the neuroanatomy, neurocircuitry, and neurochemistry that controls breathing during development to better understand how infants breathe, why premature infants have apnea of prematurity, why term infants have apnea of infancy, and how genetic errors and environmental influences modify mechanisms that control breathing during fetal and early neonatal life. ANIMAL MODELS OF CONTROL OF BREATHING Much of our understanding of basic mechanisms that lead to a stable respiratory pattern is derived from animal studies, both adult and newborn. Earlier studies used the 584 newborn pig, dog, and cat, as well as fetal and newborn sheep models, to better understand developmental physiology, and much has been gained from these models. However, a more detailed understanding of the neuroanatomy, neurocircuitry, and neurochemistry has been obtained using in vitro models from newborn rats and mice. Of particular relevance, the stage of respiratory development of the rat born at term is similar to that of the human infant born at 25 to 29 weeks’ gestation; therefore the newborn rodent model is applicable to breathing in the premature infant. Studies using either of two in vitro models (brainstem slices that include the area that contains the “pacemaker cells mediating rhythmogenesis” and the isolated brainstem spinal cord preparation from fetal and newborn rodents) have led to an explosion of information about all aspects of maturation of respiratory control within the last decade (Richter and Spyer, 2001). Because the stability and viability of the in vitro preparations are best when tissues are used from late embryonic or early postnatal rodents (within first week of postnatal life), data from these in vitro models are relevant to respiratory control during early development. OVERVIEW OF RESPIRATORY CONTROL The leading hypothesis regarding autonomic control of breathing is that respiratory rhythm is generated from a core group of synaptically coupled excitatory neurons in the brainstem located in the pre-Bötzinger complex (PBC). These neurons depolarize during the three phases of respiration: inspiration, postinspiration, and expiration. The timing of the respiratory cycle is controlled by inhibitory interneurons that discharge during specific phases of the respiratory cycle. Collectively this complex network of neurons is called the central pattern generator (Duffin, 2004; Smith et al, 2009). Inspiratory, postinspiratory, and expiratory neurons depolarize with each phase of the respiratory cycle coincident with activity motoneurons innervating the muscles of respiration (Figure 43-1). Whereas inspiration is always associated with phrenic nerve activity and contraction of the diaphragm, the first phase of expiration (E1), known as postinspiratory activity, can be associated with phrenic nerve activity, and the second phase of expiration (E2) is not (see Figure 43-1). With development, the intrinsic properties and neurotransmitter profiles of respiratory-related neurons and synaptic inputs from (1) higher brain centers (frontal and insular cortex, hypothalamus, reticular activating system and amygdala, (2) mechanoreceptors in the lungs and upper airways, (3) peripheral chemoreceptors in the carotid body, and (4) central chemoreceptors on the ventral medullary surface modify the excitability of respiratory-related CHAPTER 43 Control of Breathing neurons. In addition, the intrinsic “sensing” properties of mechanoreceptors and chemoreceptors are also changing. Last, the integrated respiratory output is also dependent on the strength of the synapse between the premotor respiratory neurons and the respiratory motoneurons innervating the diaphragm, intercostals and muscles of the upper airway, and the corresponding neuromuscular junctions (Figure 43-2). The effect of development on respiratory motoneurons and the neuromuscular junction will not be discussed in depth in this chapter; for further discussion of this topic the reader is referred to two excellent reviews (Greer and Funk, 2005; Mantilla and Sieck, 2008). 55 Vl mV 70 1 Insp.N. 3 Exp.N. Post-Insp.N. 2 Ve INSP Output phrenic Cycle phases PI MUSCLES OF RESPIRATION EXP aug dec E1 2s 585 E2 FIGURE 43-1 Simultaneous triple recordings of electrical activities of the main populations of respiratory neurons: 1, an inspiratory neuron (large spikes, middle tracing); 2, a postinspiratory neuron (small spikes, middle tracing); 3, an expiratory neuron (top tracing) and the phrenic nerve (bottom tracing). Note the augmenting (aug) activity of the phrenic nerve during inspiration and the decrementing (dec) activity of the phrenic nerve during the postinspiratory (PI) phase of the respiratory cycle. The postinspiratory phase is the first phase of expiration (E1) and is associated with phrenic nerve activity. In the second phase of expiration (E2), the phrenic nerve is silent. (Reproduced from Richter DW, Spyer KM: Studying rhythmogenesis of breathing: comparison of in vivo and in vitro models, Trends Neurosci 24:464-472, 2001.) The muscles of respiration include the pump muscles (diaphragm, intercostal muscles, and abdominal muscles) and muscles of the upper airway (alae nares, pharyngeal muscles, and laryngeal muscles). Upper airway muscles modulate the rate of inspiratory and expiratory airflow. During the inspiratory phase of the respiratory cycle the diaphragm, external intercostals (in infants), and posterior cricoarytenoid (laryngeal dilator) contract. During the postinspiratory phase, the diaphragm and the thyroarytenoid (laryngeal constrictor) contract. Diaphragmatic and thyroarytenoid postinspiratory activity is common in newborns, both human (Eichenwald et al, 1993) and animals (England et al, 1985). Because the chest wall of newborn infants, particularly premature infants, is highly compliant, diaphragmatic and laryngeal postinspiratory activity retards expiration to maintain functional residual ParabrachialKölliker-Fuse complex Bötzinger complex Nucleus of the solitary tract PreBötzinger complex Rostral VRG Caudal VRG Nucleus ambiguus Ventral respiratory column Medullary raphe Arcuate nucleus FIGURE 43-2 Schematic illustrating the anatomic relationship between the regions in the human brainstem that compose the respiratory network. These regions include specialized neurons in the dorsolateral pons (parabrachial and Kölliker–Fuse nuclei) nucleus of the solitary tract (nTS), and ventral respiratory column. The ventral respiratory column is organized rostrocaudally extending from the level just below the facial nucleus to the C1 level of cervical cord. The ventral respiratory column consists of the Bötzinger complex, pre-Bötzinger complex, and the rostral and caudal ventral respiratory groups (VRGs). Vagal motor neurons of the nucleus ambiguus innervate the laryngeal muscles. The medullary raphe, arcuate nucleus, located just underneath the ventral medullary surface, contains neurons that depolarize in response to hypercapnia and hypoxia. The retrotrapazoid nucleus (not shown) are CO2/H+ chemosensors, and is located rostrally below the facial nucleus on the ventral medullary surface. (Modified with permission from Benarroch EE: Brainstem respiratory control: substrates of respiratory failure of multiple system atrophy, Mov Disord 22:155-161, 2007.) 586 PART X Respiratory System capacity. This postinspiratory activity is often audible, known as grunting, and is heard in infants with low lungvolume disease states, such as surfactant deficiency and atelectasis. As outlined previously, the second phase of expiration (E2) is often quiescent regarding muscle activity; the diaphragm relaxes and the lungs recoil. During forced expiration often seen during exercise and significant hypercapnia, the internal intercostals and abdominal muscles may contract during E2. Brainstem nuclei that are involved in the control of upper airway muscles include the upper airway motoneurons of the nucleus ambiguous, the dorsal motor nucleus of the vagus, and the hypoglossal nucleus (­Jordan, 2001; Nunez-Abades et al, 1992). Projections from ­respiratory-related neurons synapse onto these upper airway motoneurons, thereby regulating the activity of these nerves and muscles during the respiratory cycle. RESPIRATORY CENTER NEUROANATOMY As shown in the anatomic illustration (see Figure 43-2) and the schematic (Figure 43-3), the respiratory-related neurons are located in three main areas in the brainstem: (1) the dorsal respiratory group within the nucleus tractus solitarii, (2) the ventral respiratory column (VRC) that extends from the facial nucleus to the ventrolateral medulla at the spinal-medullary junction, and (3) the pontine respiratory group within the dorsolateral pons (Alheid and McCrimmon, 2008). The VRC can be subdivided into a rostral part, involved in rhythmogenesis, and a caudal part involved in pattern formation. The rostral VRC contains both the rostral ventral respiratory group (VRG), which contains a large proportion of bulbospinal inspiratory neurons that project directly to the phrenic and external intercostal motoneurons, and the caudal VRG, which contains bulbospinal expiratory neurons that project to abdominal and internal intercostal motoneurons. Bulbospinal neurons are neurons that originate in the medulla and synapse onto motoneurons in the spinal column. Of importance, within the rostral VRC are two areas that are essential to formation of respiratory rhythm: the Bötzinger complex, and PBC. The Bötzinger complex contains propriobulbar expiratory neurons that provide strong inhibitory inputs onto inspiratory and expiratory bulbospinal neurons in the VRC. Propriobulbar neurons are neurons originating in the brainstem that send projections to other neurons in the brainstem. The PBC contains a core group of synaptically coupled excitatory neurons that have pacemaker properties, similar to the pacemaker cells in the AV node of the heart. These pacemaker cells are rostral to the nucleus SARs RARs C-fiber Inspiratory neurons Expiratory neurons Peripheral arterial chemoreceptors Laryngeal chemoreflex CO2/H chemosensors RONTINE PONTINE RESPIRATORY GROUP (phase switching) Caudal nucleus tractus solitarii (nTS) Dorsal Respiratory Group Inspiratory neurons VENTRAL RESPIRATORY COLUMN (VR Column) Facial nucleus Rostral (rhythm generating) BötzC expiratory neurons PBC pacemaker neurons inspiratory & expiratory Caudal (pattern forming) Ventral Respiratory Group VRG Rostral VRG (rVRG) Caudal VRG (cVRG) RTN/pFRG CO2/H Bulbopspinal neurons Bulbopspinal neurons synapse on to respiratory motoneurons (phrenic, intercostals, upper airway, abdominals) FIGURE 43-3 Simplified schematic illustrating the major pontine-medullary brainstem network controlling respiration and afferent and efferent projections. The pontine respiratory group contains two nuclei that send neuronal connections to respiratory-related neurons in the ventral respiratory column (VRC). The pontine respiratory group mediates phase switching between inspiration and expiration. Within the nucleus tractus solitarii (nTS) are the inspiratory neurons of the dorsal respiratory group with projections to motoneurons in the spinal column. The nTS also receive monosynaptic inputs from vagally mediated reflexes in the lung and upper airways (including slowly adapting receptors [SARs], rapidly adapting receptors [RARs], and C-fibers receptors), laryngeal chemoreceptors, and peripheral arterial chemoreceptors. Projections from second-order neurons in the nTS then synapse on to neurons in the rostral and caudal VRC. The VRC extends from the level of the facial nucleus to C1 in the cervical spinal cord. The rostral VRC is involved in respiratory rhythmogenesis and contains the expiratory neurons of the Bötzinger complex (BötzC) and the pacemakers cells of the pre-Bötzinger complex (PBC). The BötzC and PBC contain propriobulbar neurons that project to inspiratory neurons in the rostral ventral respiratory group (VRG) and expiratory neurons in the caudal VRG. The BötzC also contains bulbospinal neurons that synapse on to phrenic motoneurons in the spinal cord, whereas the PBC only contains propriobulbar neurons. Neurons in the VRG are responsible for shaping the respiratory pattern and receive inputs from second-order neurons in the nTS and from rhythm-generating neurons in the BötzC and PBC. Bulbospinal neurons from the DRG, BötzC, rVRG, and cVRG synapse onto motoneurons that control the activity of the muscles of respiration. CHAPTER 43 Control of Breathing ambiguous, have both intrinsic inspiratory and expiratory bursting properties, and are essential to maintaining respiratory rhythm (Smith et al, 1991). Progressive destruction of the PBC disrupts rhythmogenesis, leading to death in an animal model (Ramirez et al, 1998). Another important group of neurons are those within the retrotrapezoid nucleus (RTN) located along the ventral medullary surface beneath the facial nucleus (see ­Figure 43-3). These neurons have chemosensitive properties and depolarize in response to increasing CO2 and decreasing pH, and they are presumed to synapse on rhythm and pattern generating neurons in the VRC (Guyenet et al, 2008). All these neuronal groups and networks for rhythmogenesis are present in newborn animals born at term (e.g., sheep, cats, pigs) or born prematurely (e.g., rodents) in which rhythmogenesis is well established before birth. Episodic spontaneous fetal breathing movements occur in human fetuses as early as 10 weeks’ gestation (de Vries et al, 1985). In rodents, respiratory rhythmogenesis is first detected at embryonic day 15 in rats and day 17 in mice (Thoby-Brisson and Greer, 2008). The emergence of this respiratory related activity in rats is coincident with the characteristic expression of neurokinin 1 receptors of the PBC (Thoby-Brisson and Greer, 2008). In summary, respiratory rhythm and inspiratory-­ expiratory patterns emerge from dynamic interactions between: (1) excitatory neuron populations in the PBC and rostral VRG, which are active during inspiration and form the inspiratory motor output; (2) inhibitory neuron populations in the PBC that provide inspiratory inhibition within the network; and (3) inhibitory neuron populations in the Bötzinger complex, which are active during expiration and provide expiratory inhibition within the network and to phrenic motor neurons (see Figure 43-3) (Smith et al, 2009). NEUROCHEMISTRY MEDIATING RESPIRATORY CONTROL Glutamate is the major neurotransmitter mediating excitatory synaptic input to brainstem respiratory neurons and respiratory premotor and motor neurons through binding to the α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid kainite (Bonham, 1995) and metabotropic glutamate receptors (Pierrefiche et al, 1994). GABA (gamma-­ aminobutyric acid) and glycine are the two major inhibitory neurotransmitters mediating inhibitory synaptic input in the respiratory network; they have a key role in pattern generation and termination of inspiratory activity (Haji et al, 2000). GABA (via GABAA receptors) and glycine (via glycine receptors mediate fast synaptic inhibition via activation of chloride channels (Bianchi et al, 1995). GABAB receptors, which are metabotropic G-protein coupled receptors, have a greater role in inhibiting respiratory rhythm in adult mammals (Kerr and Ong, 1995). Throughout development, glutamate always functions as an excitatory neurotransmitter; however, it is not the case that GABA and glycine are always inhibitory neurotransmitters. During early development, GABA and glycine mediate excitatory neurotransmission in many neuronal networks, including the respiratory network (Putnam et al, 2005). The expression of K± ­chloride cotransporters (KCC2 and NKCC2) reduces the intracellular chloride concentrations that 587 change the effect of GABA binding to GABAA receptors from depolarizing during late embryonic and early postnatal to hyperpolarizing with maturation (Rivera et al, 1999). NEUROMODULATION OF RESPIRATORY PATTERN AND RHYTHM The baseline excitatory and inhibitory influences mediated by glutamate and GABA–glycine, respectively, on major neuronal networks are further altered by many endogenously released neuromodulators that shape and fine tune respiratory pattern and rhythm throughout development, as outlined in Table 43-1. For example, acetylcholine, substance P, cholecystokinin, cholecystokinin (CCK) and thyrotropin-­releasing hormone all exert an excitatory drive, whereas opioids and somatostatin exert an inhibitory drive on respiratory-related neurons. Dopamine, adenosine, serotonin, and norepinephrine can have excitatory and inhibitory influences depending on the specific receptors that the neuromodulator binds. Neurons within the PBC that are the kernel of rhythmogenesis are also distinctly identified by being immunopositive for the glutamate transporter, NK1, μ-opioid, and GABAB-receptors. Although these cells are primarily excitatory and release glutamate, they can be modulated by synaptic inputs that release substance P, opioids, and GABA (Doi and Ramirez, 2008). Recently these rhythmogenic neurons have been found to produce, and be excited by, brain-derived neurotrophic factor (­Bouvier et al, 2008). GENETIC MUTATIONS AFFECTING BIOGENIC AMINES Some neuromodulators may be more critical for supporting respiratory rhythmogenesis than others. By identifying the genetic mutations that are associated with marked abnormalities in respiratory control, a better understanding of the key role of several neuromodulator systems has been elucidated. For example, serotonergic neurons in the caudal medullary raphe nuclei have extensive projections to phrenic and hypoglossal motoneurons, the nucleus tractus solitarii (nTS), the RTN, and the PBC (Pilowsky et al, 1990; Voss et al, 1990). Thus, as reviewed by Kinney et al (2009), the serotonergic system has a significant influence on the modulation and integration of diverse homeostatic functions. Medullary serotonergic neurons are also CO2 sensitive (Richerson et al, 2001). In genetically modified mice that do not develop medullary serotonergic neurons, CO2 sensitivity is reduced by 50% (Hodges et al, 2008). Individuals with Prader–Willi syndrome, who may exhibit breathing abnormalities at birth) have mutations in the Necdin gene associated with abnormalities in the brainstem serotonergic system (Zanella et al, 2008a). Mice lacking the Necdin gene also have abnormal brainstem serotonergic neurochemistry (Zanella et al, 2008b). In some infants who have died of SIDS, disruptions of the brainstem serotonergic system have been identified (Kinney et al, 2009; Paterson et al, 2006b). Abnormalities in norepinephrine production also disrupt normal breathing. Rett syndrome is an X-linked disorder with mutations in the methyl CpG binding protein 2 (MECP2) gene. Affected individuals have severe respiratory disturbances that can be fatal. Genetically 588 PART X Respiratory System TABLE 43-1 Neurotransmitters and Neuromodulators That Mediate Respiratory Rhythm Neuromodulator Glutamate Receptor ­Subtype Source of the Endogenous Ligand NMDA, AMPA GluR Excitatory or ­Inhibitory on Respiratory Rhythm Comment Excitatory Major excitatory neurotransmitter Ach M3 PAG, LC, X Excitatory NE α1-adrenergic LC Excitatory Serotonin 5-HT2A2B, 5-HT3, 5-HT4 Raphe Excitatory Dopamine Likely D1 PVN, hypothalamus Excitatory ATP P2X2 Ventral medulla; CO2/H+ cells in the RTN Excitatory Adenosine P2Y1 Ventral medulla Excitatory Substance P NK1 nTS, NA Excitatory CCK CCK1 nTS, raphe Excitatory TRH TRH-R, R2 raphe Excitatory GABA GABAA GABAB Glycine GlyR NE α2-adrenergic Pons Dopamine D4 PVN, hypothalamus Adenosine A1, A2 Ubiquitous from metabolism of ATP that increases during hypoxia Inhibitory Contributes to respiratory depression at baseline (A1), and mediates HVD Opioid μ, δ, κ nTS, PBN, PVN, raphe Inhibitory Prominent inhibitory effect during early development PDGF PDGF-β nTS, Inhibitory Contributes to HVD Inhibitory Major inhibitory neurotransmitter (can be excitatory during fetal life) Inhibitory Can be excitatory during fetal life Inhibitory Data combined from Doi A, Ramirez JM: Neuromodulation and the orchestration of the respiratory rhythm, Respir Physiol Neurobiol 164:96, 2008; and Simakajornboon N, Kuptanon T: Maturational changes in neuromodulation of central pathways underlying hypoxic ventilatory response, Respir Physiol Neurobiol 149:273, 2005. Ach, Acetylcholine; ATP, Adenosine triphosphate; CCK, cholecystokinin; GABA, gamma-aminobutyric acid; HVD , hypoxic ventilatory depression; LC locus ceruleus; NA, nucleus ambiguous; NE, norepinephrine; nTS, nucleus tractus solitarii; PAG, periaqueductal gray; PBN, parabrachial nucleus; PDGF, platelet-derived growth factor; PVN, paraventricular nucleus; RTN, retrotrapezoid nucleus; TRH, thyrotropin-releasing hormone; X, vagal nucleus. modified mice that lack the Mecp2 gene have reduced levels of norepinephrine and serotonin in the medulla and have breathing patterns similar to humans with Rett syndrome (Roux et al, 2008). Pharmacologic treatment to increase brain norepinephrine and serotonin levels stabilizes breathing and prolongs the life of these mice (Roux et al, 2007). PERIPHERAL MECHANORECEPTOR INPUTS THAT MODULATE RESPIRATORY PATTERN AND RHYTHM The nTS in the brainstem (see Figure 43-3) is where sensory afferents that are transmitting pressure signals from the upper airway, volume signals from the lung, and chemical signals from the blood and cerebrospinal fluid, synapse onto second-order neurons that send projections either monosynaptically or polysynaptically to phrenic motor neurons via respiratory-related bulbospinal neurons. VAGALLY MEDIATED REFLEXES Essentially all bronchopulmonary reflexes that modify depth and duration of inspiration and expiration are mediated through the vagal nerve. The vagal nerve has both myelinated and unmyelinated fibers. Myelinated vagal afferent fibers are activated via (1) slowly adapting stretch receptors (SARs), which are activated by volume and stretch of the lung (mediating the Breuer-Hering Reflex), or (2) rapidly adapting receptors (RARs), which are activated in response to inhaled irritants (e.g., ammonia, cigarette smoke) and large inflations or deflations of the lung (Kubin et al, 2006). Changes in respiratory timing of inspiration and expiration are the respiratory response to activation of SARs, whereas sighing (i.e., augmented breaths) and coughing are the characteristic ventilatory responses to activation of RARs. Unmyelinated vagal afferents, specifically C fibers, are activated by a multitude of chemical stimuli, including CO2 and capsaicin, in addition to lung edema and elevated temperature. Rapid shallow breathing and apnea are the characteristic ventilatory responses to the activation of C-fibers in the airways. BREUER–HERING REFLEX DURING DEVELOPMENT The duration of inspiration and expiration is greatly influenced by lung inflation, and the most well-characterized bronchopulmonary reflex is the pulmonary stretch reflex mediated through SARs, discovered by Josef Breuer in 1868. In adult cats, Breuer showed that expansion of the lungs reflexively inhibits inspiration and promotes CHAPTER 43 Control of Breathing expiration, and that deflation of the lungs promotes inspiration and inhibits expiration (Widdicombe, 2006). In the nTS, these afferents monosynaptically synapse on to second-order neurons called pump cells and inspiratory-β neurons. These second-order neurons then send projections to respiratory-related bulbospinal neurons in the VRC. Bulbospinal neurons synapse on the phrenic motoneurons in the cervical spinal cord. SARs also influence the activity of a subset of propriobulbar VRC neurons that are involved in rhythmogenesis (for review, see Kubin et al, 2006). The Breuer–Hering (B-H) reflex does not appear to be important in regulating fetal breathing movements, because vagotomy performed in fetal sheep has little effect on the incidence, frequency, or amplitude of these breathing movements (Hasan and Rigaux, 1992). However, the B-H reflex is important in establishing continuous breathing and adequate gas exchange at birth (Wong et al, 1998). The reflex maintains functional residual capacity in newborns and infants, because vagotomy within 48 hours of birth results in respiratory failure associated with marked atelectasis in newborn sheep (­Lalani et al, 2001; Wong et al, 1998). Vagal innervation in utero was initially thought to be necessary for the development of surfactant (Alcorn et al, 1980), but this belief has recently been challenged by new findings from Gahlot et al (2009). These studies showed that vagal denervation performed at 110 to 113 days’ gestation (term gestation, 147 days) in fetal sheep had no effect on alveolar architecture, number of type II cells, morphology of lamellar bodies, or the level of surfactant proteins A and B and total phospholipids in lung tissue (Gahlot et al, 2009). In humans, the contribution of the B-H to tidal breathing is determined by occluding the airway at either end expiration, where the next occluded inspiratory effort is prolonged and expiratory effort is shortened, or end inspiration, where the next occluded expiratory effort is prolonged and inspiratory effort is shortened. The inspiratory and expiratory time of the occluded effort is compared to the inspiratory and expiratory time of the preceding nonoccluded breath to determine the percentage increase or decrease of the inspiratory or expiratory times. With this technique, the B-H reflex has been shown to contribute significantly to tidal breathing in infants, which is strongest at birth and then decreases during the 1st year of life (Rabbette et al, 1994). It is reasoned that the strength of the B-H reflex is inversely related to gestational and postnatal age because of the excessively compliant chest wall in newborns, which collapses at lung volumes less than functional residual capacity. With decreasing lung volumes during expiration, the B-H deflation reflex will become activated and thereby shorten expiratory time and prolong inspiratory time. Several factors increase the strength of the B-H reflex, including premature birth (De Winter et al, 1995; Kirkpatrick et al, 1976), prone sleeping position (Landolfo et al, 2008) and active sleep (Hand et al, 2004) and respiratory distress syndrome (RDS) (Rabbette et al, 1994). RAPIDLY ADAPTING RECEPTORS DURING DEVELOPMENT Lung deflation, mechanical stimulation, and chemical irritants also stimulate vagal afferents of RARs causing augmented breaths and increased mucous production 589 (Widdicombe, 2006). RAR afferents synapse on RAR cells and inspiratory Iγ neurons throughout the nTS (Kubin et al, 2006). RAR cells receive monosynaptic excitatory inputs from vagal afferents, which are excited by irritants, particularly ammonia. RAR cells then send projections to the pons respiratory group and bulbospinal inspiratory and expiratory neurons. Stimulation of RARs as a result of lung deflation monosynaptically activates inspiratory Iγ neurons in the nTS, which then send axonal projections to bulbospinal neurons (Kubin et al, 2006). RARs are particularly important in restoring lung inflation in premature and term infants, because the excessive compliance of the chest wall in newborn infants causes low lung volumes during tidal breathing. RARs become active at low lung volumes, leading to threshold activation of vagal afferents that results in augmented breaths, restoring lung volume. The frequency of augmented breaths is inversely related to gestational age, with premature infants having the greatest number (Alvarez et al, 1993), and the characteristic pattern of the augmented breath differs between newborns and adults. Augmented breaths in infants have a biphasic pattern with two large inspiratory efforts in succession, whereas in adults only one large inspiratory effort is seen. Augmented breaths in preterm and term infants are also relatively larger than those in adults; immediately after the augmented breath, preterm and term infants often hypoventilate or have apnea. In contrast, ventilation often increases after the augmented breath in adults (Qureshi et al, 2009). The increased frequency of augmented breaths and the hypoventilation and apnea after augmented breaths in premature infants suggest that peripheral arterial chemoreceptor inputs may have a greater influence on respiration in infants than in adults. Peripheral arterial chemoreceptors reflexively alter ventilation in response to acute changes in arterial CO2 and O2 tensions (discussed later). In response to an augmented breath, Pao2 rapidly increases and PaCO2 rapidly decreases, which reduces the excitatory input from peripheral arterial chemoreceptors, leading to hypoventilation or apnea. Peripheral arterial chemoreceptors are also key in inducing augmented breaths, because carotid sinus nerve denervation in animals is associated with decreased frequency of augmented breaths (Matsumoto et al, 1997). As a result, increased activity of RARs during lung deflation and increased sensitivity of peripheral arterial chemoreceptors in early development both likely contribute to the increased frequency and ventilatory consequences of augmented breaths in premature infants. C-FIBER RECEPTORS AND RESPONSES DURING DEVELOPMENT Pulmonary and bronchial C-fiber receptors are unmyelinated vagal fibers located throughout the respiratory tract, extending from the nose to the lung parenchyma. Pulmonary C-fibers are accessible from the pulmonary circulation, whereas bronchial C-fibers are accessible from the bronchial circulation and have similar sensitivity to various stimuli (Coleridge and Coleridge, 1984). C-fibers are activated by a variety of substances: inflammatory mediators, capsaicin, lobeline, and phenylbiguanidine. Capsaicin and phenylbiguanidine are often used experimentally to 590 PART X Respiratory System identify vagal afferents as C-fibers and characterize stimulus-response profiles. C-fiber simulation induces central and local effects: cough, apnea, and laryngospasm, followed by rapid shallow breathing, bradycardia, and hypotension mediated by the central reflex pathways. Bronchoconstriction, increased mucous secretion, and bronchial and nasal vasodilation are mediated by local or axon reflexes (Carr and Undem, 2003). The central effects involve transmission of impulses to interneurons in the central nervous system, which influences the activity of autonomic or somatic efferent nerves. The local, direct effects are mediated by the release of neuropeptides, particularly substance P, from C-fiber endings. By far the most common respiratory response from C-fiber stimulation is reflex apnea characterized by prolongation of expiratory time from excitation of postinspiratory neurons and continuous firing of central expiratory neurons (Coleridge and Coleridge, 1984). Central integration of pulmonary C-fiber afferent information also occurs in the subnucleus of the nTS. Unlike the relay pathways that have been carefully delineated for the SARs, and to a lesser extent for the RARs, the specific secondorder neurons in the nTS that are activated in response to bronchopulmonary C-fiber stimulation have not been completely identified. It is likely, however, that the central integration of pulmonary C-fiber stimulation occurs within the nTS, because the interruption of synaptic transmission within the subnucleus of the nTS ablates reflex response of pulmonary C-fibers from intraatrial injections of phenylbiguanidine (Bonham and Joad, 1991). In newborns, the stimulation of pulmonary C-fibers by chemical stimulants causes bronchoconstriction and apnea (Frappell and MacFarlane, 2005). Capsaicin-induced apneic response and the sensitivity of the reflex was greatest in newborn rat pups younger than 10 postnatal days (Wang and Xu, 2006). Bronchopulmonary C-fibers are also stimulated by acidosis, adenosine, reactive oxygen species, hyperosmotic solutions, and lung edema. Furthermore, inflammatory mediators in the local environment sensitize C-fibers to other stimuli (Lee and Pisarri, 2001). Pulmonary C-fiber–mediated respiratory inhibition may be causing persistent apnea beyond term postconceptional age in infants born at the limit of viability who have significant lung disease (Eichenwald et al, 1997). As proposed by Lee and Pisarri (2001), C-fiber activation may also account for the increased frequency of apnea observed in infants with viral infections, especially caused by respiratory syncytial virus (Pickens et al, 1989). LARYNGEAL REFLEXES Receptors that respond to changes in upper airway pressure and chemical compounds are abundantly distributed throughout the laryngeal mucosa. These receptors can be slowly adapting, rapidly adapting irritant receptors, or C-fibers. Water receptors that are simulated by hyposmolarity and low chloride content may also be involved. Stimulation of upper airway mechanoreceptors and chemoreceptors modifies activity of upper airway muscles as well as the pattern and timing of diaphragmatic activity. The upper airway reflex that mediates significant cardiorespiratory effects that occur in newborns is the laryngeal chemoreflex (LCR). The LCR is one of the most potent defensive reflexes protecting the respiratory tract from inadvertent aspiration (Harding et al, 1978). These receptors are stimulated by liquid in the airway, which induces coughing, swallowing, and arousal in mature models. However, the response in immature models is apnea followed by hypoventilation, laryngeal constriction, and swallowing. In addition to respiratory inhibition, bradycardia, peripheral vasoconstriction, and redistribution of blood flow also occurs. Of importance, the associated apnea and bradycardia can be life threatening in newborns (Boggs and Bartlett, 1982; Sasaki et al, 1977; Thach, 2001; Wetmore, 1993), and in newborn infants baseline hypoxemia enhances the severity of the apnea and bradycardia induced by the LCR (Wennergren et al, 1989). Afferent fibers for this reflex travel in the superior laryngeal nerve, a branch of the vagus. These afferents synapse onto neurons in the nTS which then send (1) excitatory projections to motoneurons of the recurrent laryngeal nerve in the nucleus ambiguous, causing constriction of the thyroarytenoid muscle (laryngeal constrictor) resulting in laryngospasm; (2) inhibitory projections to phrenic motoneurons in the cervical spinal cord, inhibiting diaphragmatic contraction resulting in apnea; and (3) an excitatory pathway to cardiac vagal neurons in the nucleus ambiguous causing bradycardia. However, the circuitry from second-order neurons in the nTS to cardiac vagal afferent neurons has not been demonstrated clearly. The LCR occurs in the fetus and likely functions to prevent aspiration of amniotic fluid, which contains approximately half the chloride content of pulmonary fluid (Bland, 1990; Reix et al, 2007). With premature birth, the reflex may be involved in the apnea and bradycardic responses associated with feeds and gastroesophageal reflux that reaches the larynx or nasopharynx. Whether the immature response is still present in term infants or how the maturation of the reflex is affected by premature birth has not been determined. Because of its profound inhibitory cardiorespiratory effects, stimulation of the LCR may be an important reflex that is operative in some SIDS cases and infants with acute life-threatening events (Duke et al, 2001; Gauda et al, 2007; Richardson and Adams, 2005; Thach, 2001). CHEMICAL CONTROL OF BREATHING (CO2/H+ and O2) In air-breathing animals, respiratory rhythmogenesis is primarily driven by the level of Pco2 in the blood and cerebrospinal fluid and, to a lesser extent, by oxygen tension. In fact, for every increase of 1mm Hg in Pco2, ventilation will increase by 20% to 30%. Specialized chemosensitive cells in the brainstem depolarize in response to changes in CO2/H+; they drive breathing through synaptic inputs to respiratory-related neurons (Spyer and ­Gourine, 2009). Although peripheral arterial chemoreceptors in the carotid body also depolarize in response to increasing CO2/H+, these receptors are primarily responsible for modifying breathing in response to changes in oxygen tension (reviewed later). As a result of careful anatomic, physiologic, neurochemical, and genetic studies, the location and the development of central chemoreceptors and some of the genetic factors that drive the development of these receptors in health and disease have been determined. Several groups of neurons in the brainstem, specifically in CHAPTER 43 Control of Breathing the medullary raphe, RTN, nTS, locus ceruleus, and the fastigial nucleus are responsive to CO2/H+ in vitro and in vivo; therefore they are characterized as chemosensitive. The demonstration that (1) focal acidification either with acetazolamide (inhibits carbonic anhydrase) or local elevations of CO2 in these regions in either nonanesthetized or anesthetized adult animals increases ventilation, and (2) focal ablation or disruption of the region inhibits the ventilatory response, is essential for a region to be identified as chemosensitive (Feldman et al, 2003). Although several regions are CO2/H+ sensitive, there appears to be some specificity in the contribution of each of the regions in chemical control of breathing wakefulness and sleep (Feldman et al, 2003). The greatest density of CO2/H+– sensitive neurons in the brainstem is in serotonergic cells of the raphe of the caudal medulla and glutamatergic neurons of the RTN, located just below the ventral medullary surface. The serotonergic neurons in the caudal raphe project to phrenic motoneurons, where they modulate neuronal plasticity in response to hypoxia (Feldman et al, 2003). The RTN receives afferent polysynaptic excitatory inputs from peripheral arterial chemoreceptors cartoid body (Takakura et al, 2006) and sends projections to neurons in the VRC, including the PBC (Guyenet et al, 2008). In animals older than 7 postnatal days, the activity of the RTN depends on intrinsic pH sensitivity and synaptic drive. The parafacial respiratory group may be the precursor to the RTN, and these neurons have intrinsic bursting properties starting at embryonic day 19 (Guyenet et al, 2008). In humans the arcuate nucleus in the medulla (not to be confused with the arcuate nucleus in the hypothalamus) is believed to be the homologous chemosensitive region, as described in animals based on the following findings: (1) the arcuate nucleus is located along the ventral medullary surface, (2) it contains a large population of glutamatergic neurons and a smaller population of serotonergic neurons (Paterson et al, 2006a), (3) it depolarizes in response to hypercapnia (Gozal et al, 1994), and (4) the absence of the arcuate nucleus in a human infant was associated with lack of CO2 sensitivity during life (Folgering et al, 1979). DEVELOPMENT OF CENTRAL CO2/H+ SENSITIVITY In fetal sheep, hypercapnia causes an increase in the depth of fetal breathing movements, with no change in inspiratory or expiratory time. In humans, maternal exposure to CO2 also increases fetal breathing (Ritchie and Lakhani, 1980). Ventilatory responses to CO2 are present immediately after birth in most mammalian species, and CO2 sensitivity increases with maturation. However, in the newborn rat, CO2 sensitivity is robust during the first several days of postnatal life; it declines markedly in the next 2 weeks of life and then gradually increases to reach adult levels by the end of the 3rd week (Stunden et al, 2001). Premature and term infants tested at 2 days of postnatal age have modest ventilatory responses to 2% and 4% CO2, although the strength of the ventilatory response is less in the more immature infants (Frantz et al, 1976). In premature and late preterm infants, CO2 sensitivity increases with postnatal age, reaching a mature response within 4 weeks of postnatal development (Figure 43-4) (Frantz et al, 1976; 591 FIGURE 43-4 The relationship between ventilatory sensitivity to carbon dioxide and gestational age (top panel), postnatal age (middle panel), and the concentration of inspired oxygen (bottom panel). (From Rigatto H, Brady J, Verduzco RT: Chemoreceptor reflexes in preterm infants: the effect of gestational and postnatal age on the ventilatory response to inhaled carbon dioxide, Pediatrics 55:614-620, 1975; and Rigatto H, Verduzco RT, Cates DB: Effects of O2 on the ventilatory response to CO2 in preterm infants, J Appl Physiol 39:896-899, 1975.) Rigatto et al, 1975a, 1975c). Similar to the response in the fetus, the increase in ventilation is predominately due to an increase in tidal volume and not respiratory rate. Using the increase in inspiratory effort against an occluded airway as an indicator of central respiratory drive, when exposed to 2% and 4% CO2, the increase in CO2 sensitivity with postnatal development is due to an increase in central respiratory drive in human infants (Frantz et al, 1976). Premature infants with apnea of prematurity have reduced ventilatory responses to CO2 compared with control infants at the same postconceptional age (Durand et al, 1985; Gerhardt and Bancalari, 1984). This finding suggests that infants with apnea of prematurity have reduced central respiratory drive to breathe when compared with infants who do not have apnea of prematurity at the same postconceptional age. It is unknown whether maturation of synaptic inputs from chemosensitive neurons to respiratory related neurons in the brainstem, or maturation of intrinsic properties of chemosensitive neurons, accounts for the increase in CO2 sensitivity with early postnatal development. Although such 592 PART X Respiratory System studies of human infants are impossible, data from studies performed in neonatal rats show that intrinsic responses of chemosensitive neurons in the nTS and locus ceruleus are already mature at birth. It is less clear whether there is a developmental increase in the sensitivity of chemosensitive neurons in the medullary raphe, an increase in the number of chemosensitive neurons in the RTN, or both (­Putnam et al, 2005). However, within the first several weeks of postnatal development in the rat, the size of brainstem neurons changes, and both dendritic arborization in the nTS and astrocyte proliferation increase. Astrocytes contribute substantially to the pH of the extracellular milieu surrounding chemosensitive neurons (Putnam et al, 2005). It is likely that all these morphologic and neurochemical changes within and between neurons and astrocytes in the brainstem contribute to maturation of CO2 sensitivity within the early weeks of postnatal development. MATURATION OF PERIPHERAL CO2/H+ CHEMOSENSITIVITY IN THE CAROTID BODY The arterial chemoreceptors in the carotid body, located at the bifurcation of the carotid artery, are primarily responsible for reflex control of ventilation in response to changes in arterial oxygen tension. Specialized cells within the carotid body also depolarize in response to changes in blood CO2/H+, reflexively increasing ventilation in response to acidosis and hypercapnia and reflexively decreasing ventilation in response to hypocapnia. Histologically, the carotid body chemoreceptors consist of (1) type I or glomus cells, similar to presynaptic neurons, that are chemosensitive and contain neurotransmitters and autoreceptors; (2) postsynaptic afferent nerve fibers from the carotid sinus nerve, which oppose glomus cells (Gonzalez et al, 1994), contain neurotransmitters and postsynaptic receptors, and have cell bodies in the petrosal ganglion; (3) type II cells similar to glial cells, which are not chemosensitive; (4) microganglion cells that express cholinergic traits (Gauda et al, 2004); and (5) blood vessels and sympathetic fibers innervating these vessels. The commissural nucleus of the nTS is the primary target for afferents from peripheral arterial chemoreceptors. Although these afferent processes contain glutamate, dopamine, and substance P, it is glutamate, binding to both NMDA and non-NMDA receptors on second-order neurons in the nTS, that is responsible for chemical transmission of excitatory inputs from the peripheral arterial chemoreceptors (Vardhan et al, 1993). These second-order neurons then send tonic excitatory projections to CO2 sensitivity neurons in the RTN (see Development of Central CO2/H+ Sensitivity, earlier), bulbospinal neurons in the dorsal respiratory group, and VRG that synapse onto respiratory motoneurons. In order to separate the contribution of peripheral arterial chemoreceptors from that of central chemoreceptors on ventilatory control, studies performed in animals have either perfused the carotid body separately from the systemic circulation or directly measured the neuronal output from the carotid body (in vivo or in vitro) in response to changes in arterial gas tension. With these approaches, it has been determined that peripheral chemoreceptors sense changes in arterial CO2/H+ more rapidly than do central chemoreceptors (Smith et al, 2006). Sensitivity of the peripheral arterial chemoreceptors to CO2 increases with postnatal development in newborn animals (Carroll et al, 1993); however, in human infants, the contribution of peripheral arterial chemoreceptors to ventilatory response to CO2 is difficult to delineate. Inferences can be made from the ventilatory response that is seen within a few seconds of exposure to a particular concentration of CO2, O2, or both, because the response time of the peripheral arterial chemoreceptors is faster than that of the central chemoreceptors. What is inferred is that, after the first 2 days after birth, peripheral arterial chemoreceptors in newborns are highly responsive to changes in Pco2. Hypoventilation and apnea are frequently seen in newborn infants after a sigh or augmented breath, in which the Pco2 is rapidly reduced. This response is secondary to a high-gain system that is regulated at the level of the peripheral arterial chemoreceptors. In premature infants, the Pco2 apneic threshold is close to the Pco2 that mediates regular breathing (Khan et al, 2005). The level of Pco2, however, increases the response of the peripheral arterial chemoreceptors to any level of arterial O2 tension (Pao2), and this O2-CO2 interaction at the carotid body increases with development (Carroll et al, 1993). MATURATION OF HYPOXIC CHEMOSENSITIVITY The contribution of afferent inputs from peripheral arterial chemoreceptors during maturation is more easily accessed by determining the acute change in ventilation in response to changes in oxygen tension. This determination is often made according to hypoxic gas exposure in newborn animals; an acute increase in ventilation within 30 seconds of hypoxic exposure is a measure of the strength of the peripheral arterial chemoreceptors. Whereas hypoxic gas exposure is occasionally used in human infants to test peripheral arterial chemoreceptor function, the reduction in ventilation in response to short acute exposure to hyperoxia (Dejours test) is more frequently performed to assess the strength of the reflex during postnatal development. Although the peripheral arterial chemoreceptors can respond to lower than baseline Pao2 in the fetus, peripheral arterial chemoreceptors do not contribute significantly to fetal breathing, and functioning peripheral arterial chemoreceptors are not necessary for continuous breathing to occur at birth (for a historical overview, see Walker, 1984). However, studies using newborn animals (e.g., cats, dogs, rats, sheep, pigs) show that carotid body denervation shortly after birth destabilizes breathing, often leading to persistent apneic periods and death of the animal within days to weeks (Gauda and Lawson, 2000). Therefore it is believed that trophic factors from peripheral arterial chemoreceptors acting on central mechanisms that control breathing during early postnatal development are the key to stable rhythmogenesis throughout life. The critical period for these trophic influences appears to be within the first 2 weeks of postnatal development (Gauda et al, 2007). Exposures during this critical period of development that can lead to lifelong alterations in chemoreceptor function include environmental exposure to the extremes of oxygen tension (chronic hypoxia and hyperoxia), intermittent hypoxia (associated with apnea of prematurity) (Carroll, CHAPTER 43 Control of Breathing 2003; Gauda et al, 2007), nicotine exposure (Gauda et al, 2001; Hafstrom et al, 2005), and maternal separation in male rats (Genest et al, 2004). Acute insensitivity to changes in oxygen tension associated with birth and then a gradual increase in hypoxic chemosensitivity during the first 2 to 3 weeks of postnatal development are two phases that affect the contribution of peripheral arterial chemoreceptors on breathing. Although oxygen tension less than 25 mm Hg stimulates the carotid body in exteriorized fetal sheep, peripheral arterial chemoreceptors are not functional at any level of hypoxic exposure for the first several days after birth in most mammalian species, including human infants (term and preterm). It is speculated that the resetting of peripheral arterial chemoreceptors occurs at birth because of the increase in arterial oxygen tension during the transition from fetal to neonatal life. After birth, there is a gradual increase in hypoxic chemosensitivity that occurs during the first 2 weeks of postnatal life (Rigatto et al, 1975b). Mechanisms accounting for the increase in hypoxic chemosensitivity of peripheral arterial chemoreceptors with maturation in most mammalian species have been reviewed recently by Gauda et al (2009). Although premature infants may have reduced peripheral chemoreceptor responses soon after birth, by 2 weeks of postnatal age they often have enhanced peripheral arterial chemoreceptor influences on eupneic breathing. Furthermore, premature infants who have a greater frequency of apnea and periodic breathing have a greater reduction in ventilation when exposed to short bouts of hyperoxia (Dejours test) (Rigatto and Brady, 1972), suggesting a greater influence of peripheral ­arterial chemoreceptors on eupneic breathing in this infant ­population (Al Matary et al, 2004; Nock et al, 2004). GENETIC REGULATION OF CENTRAL AND PERIPHERAL CHEMORECEPTOR DEVELOPMENT Development of central and peripheral chemoreceptors is genetically regulated by the expression of the transcription factor PHOX2B. PHOX2B is a homeobox gene located on chromosome 4 that is specifically expressed in limited types of neurons involved in autonomic processes (Dauger et al, 2003). Its expression is required for the development of the carotid body, nTS, and catecholaminergic neurons, and it is expressed in chemosensitive glutamatergic neurons in the RTN that receive polysynaptic inputs from peripheral arterial chemoreceptors (Guyenet et al, 2008). Mutations in the PHOX2B gene affect the development of key structures that regulate chemical control of breathing. Congenital central hypoventilation syndrome (CCHS) is characterized by impaired ventilatory responses to CO2 and hypoxia and other abnormalities of autonomic control (Berry-Kravis et al, 2006). More than 90% of individuals with CCHS have mutations in the PHOX2B gene (for a review, see Amiel et al, 2009; Dubreuil et al, 2009. NEONATAL HYPOXIC VENTILATORY DEPRESSION Although the peripheral arterial chemoreceptors function to increase ventilation in response to hypoxia, minute ventilation significantly declines after 2 to 3minutes of hypoxic 593 exposure. This decline is commonly referred to as hypoxic roll-off, hypoxic ventilatory decline, or hypoxic ventilatory depression. Hypoxic ventilatory depression occurs in individuals at all ages, but it is most pronounced in the fetus and newborn (Bissonnette, 2000). Whereas the hypoxic ventilatory decline is usually still above baseline ventilation in mature models, the hypoventilatory response in newborns is usually below baseline ventilation and is often associated with apnea. Mechanisms accounting for hypoxic respiratory depression are most well characterized in the fetal animals in which the central brainstem nuclei mediating this response are located in the pons. Transverse section of the upper pons results in a sustained hyperventilatory response to hypoxia in fetal and newborn sheep (Gluckman and Johnston, 1987). Hypoxia activates expiratory neurons in the ventrolateral pons, and chemical blockade of this area blocks the hypoxic respiratory depression in newborn rats (Dick and Coles, 2000). Several neuromodulators have been implicated in mediating hypoxic ventilatory decline, including norepinephrine, adenosine, GABA, serotonin, opioids, and platelet-derived growth factor (see Table 43-1) (Simakajornboon and ­Kuptanon, 2005). All these neuromodulators have been shown to contribute to the ventilatory depression in newborns, but particular attention has been given to adenosine. Degradation of intracellular and extracellular ATP is the main source of extracellular adenosine, which then mediates its cellular effects by binding to A1, A2a, A2b, and A3 adenosine receptors. In response to hypoxia, brain adenosine levels can increase 2.3-fold in fetal sheep (Koos et al, 1994), and 100-fold in rats in response to ischemia (Winn et al, 1981). Nonspecific adenosine receptor blockers, particularly caffeine and methylxanthine, are commonly administered to premature infants to increase central respiratory drive, and aminophylline inhibits hypoxic ventilatory depression in newborn infants (­Darnall, 1985). A1-adenosine inhibitory receptors are found on respiratory related neurons (Bissonnette and Reddington, 1991). Specific A1-adenosine receptor agonists depress phrenic output in a reduced brainstem spinal cord preparation, whereas A1-adenosine receptor blockers reverse this inhibitory effect (Dong and Feldman, 1995). In fetal sheep, the hypoxic respiratory depression appears to be mediated by excitatory A2a receptors, because blockade of A2a receptors eliminates hypoxic ventilatory roll-off in conscious newborn sheep. Xanthine (e.g., caffeine, aminophylline) blocks both A1 and A2a adenosine receptors; therefore the effectiveness of methylxanthine in stabilizing ventilation and decreasing the frequency of apnea in premature infants may be directly altering the ventilatory response to hypoxia as well as nonspecifically increasing respiratory drive (see Apnea of Prematurity, later). SLEEP STATE AND BREATHING Sleep state has a profound influence on breathing in the fetus and newborn, and most disorders of breathing that affect the young and old are worse during sleep. Active sleep (AS) and quiet sleep (QS) in infants is equivalent to rapid eye movement (REM) and non-REM sleep, respectively in older children and adults. Breathing during AS is mostly driven by inputs from the reticular activating 594 PART X Respiratory System system with less influence from Pco2, whereas breathing during QS is driven by chemical control. Similar to REM sleep in adults, AS is associated with paralysis of striated muscles. Although this paralysis may be necessary to prevent acting out dreams, paralysis of striated muscles that are involved in breathing can be problematic for the newborn. AS and REM sleep are characterized by rapid eye movements, increase in cerebral blood flow and metabolic rate, and irregular breathing because of inhibition of the intercostals and upper airway dilating muscles. The discoordination between chest wall muscles and the diaphragm during active sleep causes paradoxic breathing: the chest wall moves in during inspiration with the abdomen moving outward. The more compliant the chest wall, the greater propensity for paradoxic breathing, which is common in the most immature infants. In addition, during inspiration, intrathoracic pressure becomes more negative, and this “suction pressure” causes narrowing or collapse of the compliant upper airway, particularly pharyngeal structures, leading to upper airway obstruction. Paradoxic breathing movements seen on physical examination or detected on inductive plethysmography are often interpreted as a sign of upper airway obstruction. During QS, breathing is characterized by smooth, regular breaths of consistent frequency and depth associated with tonic and phasic activity of the muscles of respiration that are in phase with each other. The chest wall and the abdomen move outward during inspiration, whereas they move inward during expiration. AS and QS can be reliably assessed at 30 to 32 weeks’ gestation (Curzi-Dascalova et al, 1993). At this gestation, premature infants spend 80% of their sleep time in SA. This time decreases to 50% by term, and in adulthood REM sleep accounts for only 20% of sleep time. Sleep state in normal infants also modifies the time to arousal and the ventilatory responses. The time to arousal upon exposure to a hypoxic, somatosensory or auditory stimuli is greater in AS compared with QS in infants during the first 6 months of life (Horne et al, 2005). Of interest, the arousal latency in response to hypoxic stimulus in AS is longer in preterm infants at 2 to 5 weeks postnatal age than that of term infants at the same postnatal age (Verbeek et al, 2008). The level of oxygen desaturation at the time hypoxic arousal occurs is similar in the two sleep states (Richardson et al, 2007). Respiratory pauses and periodic breathing are more common during AS in both term and premature infants, but the ventilatory response to CO2 is greater during QS (Cohen et al, 1991). Because of the complexity of the ventilatory response to hypoxia and the frequent occurrence of arousals induced by hypoxic exposure, assessing the affect of sleep state on the ventilatory response to hypoxia in newborns is more difficult. Other than the clear difference in arousal in response to hypoxic stimulus between the two sleep states, differences between sleep states on other respiratory parameters are more variable (Richardson et al, 2007). Although most disorders of breathing, such as obstructive sleep apnea, become more severe during sleep, the breathing disorder that is most significantly influenced by sleep state in the newborn is CCHS. As noted previously, CCHS is characterized by abnormalities in central chemoreception and chemical control of breathing; therefore, with maturation, as the frequency of QS increases, so does the severity of the disorder. During QS, CO2 sensitivity is markedly impaired in affected individuals, and exposure to hypercapnia does not significantly increase minute ventilation (Paton et al, 1989). In infants who died of SIDS, their sleep state during their terminal event cannot be known. However, adequate arousal mechanisms are key in preventing respiratory failure and death, and impaired arousal mechanisms are hypothesized to be causative in SIDS. Prone sleeping position increases the percent of QS (Horne et al, 2002), and QS is associated with increased time to hypoxic arousal in human infants (Richardson et al, 2007). A hypoxic microenvironment from rebreathing with defected arousal and autoresuscitative mechanisms is hypothesized to have occurred in infants who have died of SIDS in the face-down (i.e., prone) sleeping position (Patel et al, 2001). Therefore sleep state can have a significant influence on control of breathing during health and disease, especially in the newborn infant. This chapter has outlined the neurocircuitry and neurochemistry of the respiratory network along with its synaptic inputs that undergo significant maturation during the newborn period. Because these pathways are less developed in premature infants, premature infants have apnea of prematurity, which often requires active therapeutic intervention and can delay hospital discharge. At preterm gestational ages, the components of the respiratory network, similar to other developing organ systems, are plastic and uniquely vulnerable to pathologic processes. Therefore the episodes of intermittent hypoxemia and bradycardia that accompany apnea of prematurity may be a cause of acute and chronic morbidities in this high-risk population. What follows is a discussion of the clinical presentation of physiology and pathophysiology of breathing seen in premature infants and therapeutic interventions. APNEA OF PREMATURITY Respiratory pauses are universal features of preterm birth and are most prominent in infants of the lowest gestational age. There is no consensus as to when a respiratory pause can be defined as an apneic episode. It has been proposed that apnea be defined by its duration (e.g., longer than 15 seconds) or accompanying bradycardia and desaturations. However, even the 5- to 10-second pauses that occur in periodic breathing can be associated with bradycardia or desaturation. It should be emphasized that periodic breathing—ventilatory cycles of 10 to 15 seconds with pauses of 5 to 10 second—is a normal breathing pattern that should not require therapeutic intervention. It is thought to be the result of dominant peripheral chemoreceptor activity responding to fluctuations in arterial oxygen tension. Episodic bradycardia and desaturation in preterm infants is almost invariably secondary to apnea or hypoventilation (Figure 43-5) (Martin and Abu-Shaweesh, 2005). The rapidity of the fall in oxygen saturation after a respiratory pause is directly proportional to baseline oxygenation, which is related to lung volume and severity of lung disease. Apnea is classified traditionally into three categories based on the absence or presence of upper airway obstruction: central, obstructive, and mixed. Central apnea is characterized by total cessation of inspiratory efforts with no CHAPTER 43 Control of Breathing DECREASED RESPIRATORY DRIVE Apnea, hypoventilation Increased vagal tone Decreased O2 delivery Bradycardia Desaturation Carotid body FIGURE 43-5 Schematic representation of the sequence of the events whereby apnea results in various combinations of desaturation and bradycardia. (From Martin RJ, Abu-Shaweesh JM: Control of breathing and neonatal apnea, Biol Neonate 87:288-295, 2005.) evidence of obstruction. In obstructive apnea, the infant tries to breathe against an obstructed upper airway, resulting in chest wall motion without airflow through the entire apneic episode. Mixed apnea consists of obstructed respiratory efforts, usually following central pauses. The site of obstruction in the upper airways is primarily in the pharynx, although it also may occur at the larynx and possibly at both sites. It is assumed that there is an initial loss of central respiratory drive, and its recovery is accompanied by a delay in activation of upper airway muscles superimposed on a closed upper airway (Gauda et al, 1987). Mixed apnea typically accounts for more than 50% of long apneic episodes, followed in decreasing frequency by central and obstructive apnea. Purely obstructive spontaneous apnea in the absence of a positional problem is probably uncommon. Because standard impedance monitoring of respiratory efforts via chest wall motion cannot recognize obstructed respiratory efforts, mixed versus obstructive apnea is frequently identified by the accompanying bradycardia or denaturation. THERAPEUTIC APPROACHES FOR APNEA OF PREMATURITY Presentation of apnea can reflect a nonspecific alteration in either the environment (e.g., thermal) or general well being of preterm infants. For example, neonatal sepsis can manifest as an increase in frequency or severity of apnea, and the underlying cause must be treated. Studies using a rat pup model suggest that the systemically released cytokine interleukin-1β binds to its receptor on vascular endothelial cells at the blood-brain barrier. This binding induces synthesis of prostaglandin E2, which induces respiratory depression in the brainstem (Hofstetter et al, 2007). These studies provide new insight into mechanisms whereby sepsis often manifests as apnea of prematurity. Anemia, presumably via decreased oxygen delivery, is also frequently implicated as a cause of apnea, although transfusion of packed red cells is of variable benefit for apnea of prematurity. 595 CONTINUOUS POSITIVE AIRWAY PRESSURE Continuous positive airway pressure (CPAP) at 4 to 6 cm H2O is a relatively safe and effective therapy. Because longer episodes of apnea frequently involve an obstructive component, CPAP appears to be effective by splinting the upper airway with positive pressure and decreasing the risk of pharyngeal or laryngeal obstruction (Miller et al, 1985). CPAP also benefits apnea by increasing functional residual capacity, thereby improving oxygenation status. At a higher functional residual capacity, time from cessation of breathing to desaturation and resultant bradycardia is prolonged. High-flow nasal cannula therapy has been suggested as an equivalent treatment modality to allow CPAP delivery while enhancing infant mobility. Although this approach is used widely, its efficacy for apnea of prematurity has not been studied in depth. Noninvasive ventilatory strategies, using a nasal mask to deliver intermittent positive pressure, avoid the need for full ventilatory support in some infants. For severe or refractory episodes, endotracheal intubation and mechanical ventilation may be needed. Minimal ventilator settings should be used to allow for spontaneous ventilatory efforts and to minimize the risk of barotrauma. XANTHINE Methylxanthine has been the mainstay of pharmacologic treatment of apnea of prematurity for several decades. Both theophylline and caffeine are used and have multiple physiologic and pharmacologic mechanisms of action. Xanthine therapy appears to increase minute ventilation, improve CO2 sensitivity, decrease hypoxic depression of breathing, enhance diaphragmatic activity, and decrease periodic breathing. The likely major mechanism of action is through competitive antagonism of adenosine receptors. Adenosine acts as an inhibitory neuroregulator in the central nervous system via activation of adenosine A1 receptors (Herlenius et al, 1997). In addition, activation of adenosine A2A receptors appears to excite GABAergic interneurons, and released GABA may contribute to the respiratory inhibition induced by adenosine (Mayer et al, 2006). Methylxanthine has some well-documented acute adverse effects. Toxic levels can produce tachycardia, cardiac dysrhythmias, feeding intolerance, and seizures (infrequently), although these effects are seen less commonly with caffeine at the usual therapeutic doses. Mild diuresis is caused by all methylxanthines. The observation that xanthine therapy causes an increase in metabolic rate and oxygen consumption of approximately 20% suggests that caloric demands can be increased with this therapy at a time when nutritional intake already is compromised. A recent large, international, multicenter clinical trial was designed to test short- and long-term safety of caffeine therapy for apnea of prematurity. In the neonatal period, caffeine treatment was associated with a significant reduction in the postmenstrual ages at which both supplemental oxygen and endotracheal intubation were needed (Schmidt et al, 2007). Of even greater interest was the significant decrease in cerebral palsy and cognitive delay in the caffeine-treated group (Schmidt et al, 2007). This finding raises interesting questions regarding possible mechanisms underlying this beneficial effect of caffeine on 596 PART X Respiratory System Xanthines Less apnea Less BPD Altered CNS neurotransmission (adenosine, GABA) Less hypoxemic events ? ? Downregulation of cytokine release from immune cells ? Protection against hypoxia-induced PVL and ventriculomegaly ? Improved neurodevelopmental outcome FIGURE 43-6 Multiple proposed mechanisms are demonstrated whereby xanthine therapy for apnea of prematurity improves neurodevelopmental outcome. These outcomes include functional changes in neurotransmitters in the brain, a decrease in hypoxemic episodes that accompany apnea, especially in the presence of BPD, a proposed protective effect of adenosine receptor inhibition on hypoxia induced white matter injury, and the benefit of adenosine receptor blockade on cytokine mediated lung or brain injury. (Modified from: Abu-Shaweesh JM, Martin RJ: Neonatal apnea: what’s new? Pediatr Pulmonol 43:937-944, 2008.) neurodevelopmental outcome (Figure 43-6). These beneficial effects include the observation in animal models that loss of the adenosine A1 receptor gene is protective against hypoxia-induced loss of brain matter (Back et al, 2006) and a potential benefit of caffeine on immune mechanisms that mediate lung and brain injury (Chavez-Valdez et al, 2009). RELATIONSHIP TO GASTROESOPHAGEAL REFLUX [GER] GER is often incriminated as a cause for neonatal apnea. Despite the frequent coexistence of apnea and GER in preterm infants, investigations of the timing of reflux in relation to apneic events indicate that they are rarely related temporally. When these events coincide, there is no evidence that GER prolongs the concurrent apnea (Di Fiore et al, 2005). Although physiologic experiments in animal models reveal that reflux of gastric contents to the larynx induces reflex apnea, there is no clear evidence that treatment of reflux affects the frequency of apnea in most preterm infants. Therefore pharmacologic management of reflux with agents that decrease gastric acidity or enhance gastrointestinal motility generally should be reserved for preterm infants who exhibit signs of emesis or regurgitation of feedings, regardless of whether apnea is present. Therapy for such infants should begin with nonpharmacologic approaches, such as thickened feeds, because acid suppression therapy has been shown to increase the risk of lower respiratory infection in infants (Orenstein et al, 2009). Recent data indicate considerable differences of opinion among neonatologists, pediatric gastroenterologists, and pediatric pulmonologists regarding diagnosis and management of this problem (Golski, 2010). RESOLUTION AND CONSEQUENCES OF NEONATAL APNEA Apnea of prematurity generally resolves by 36 to 40 weeks’ postconceptional age; however, apnea frequently persists beyond this time in more immature infants. Available data indicate that cardiorespiratory events in such infants return to the baseline normal level at 43 to 44 weeks’ postconceptional age (Ramanathan et al, 2001). In other words, beyond 43 to 44 weeks’ postconceptional age, the incidence of cardiorespiratory events in preterm infants does not significantly exceed that in term infants. The persistence of cardiorespiratory events may delay hospital discharge for a subset of infants. In these infants, apnea longer than 20 seconds is rare; rather they exhibit frequent bradycardia to less than 70 to 80 beats/min with short respiratory pauses (Di Fiore et al, 2001). The reason that some infants exhibit marked bradycardia with short pauses is unclear, but available data suggest a vagal phenomenon and benign outcome. For a few of these infants home cardiorespiratory monitoring, until 43 to 44 weeks’ postconceptional age, is offered in the United States as an alternative to a prolonged hospital stay. The apparent lack of a relationship between persistent apnea of prematurity and SIDS has significantly decreased the practice of home monitoring, with no increase in the SIDS rate. Infants born prematurely experience multiple problems during their time in the neonatal intensive care unit, and many of these conditions can contribute to poor neurodevelopmental outcomes. For example, a history of prior hyperbilirubinemia has been associated with persistent apnea of prematurity in preterm infants and animal models (Amin et al, 2005; Mesner et al, 2008). The problem of correlating apnea with outcome is compounded by the fact that nursing reports of apnea severity may be unreliable, and impedance monitoring techniques will fail to identify mixed and obstructive events. Despite these reservations, available data suggest a link between the number of days that apnea and assisted ventilation were recorded during hospitalization and impaired neurodevelopmental outcome (Janvier et al, 2004). A relationship has also been shown between delay in resolution of apnea and bradycardia beyond 36 weeks’ corrected age and a higher incidence of unfavorable neurodevelopmental outcome (Pillekamp et al, 2007). Finally, a high number of cardiorespiratory events recorded after discharge via home cardiorespiratory CHAPTER 43 Control of Breathing 597 INTERMITTENT EPISODIC HYPOXIA ? ? ? ? Acute morbidity Cardiovascular Respiratory instability Neurodevelopmental [e.g., retinopathy instability [e.g., sleep disordered disability of prematurity] [e.g., hypertension] breathing] FIGURE 43-7 Proposed acute and longer-term morbidities that might be a consequence of intermittent hypoxic episodes in early postnatal life. (Modified from Martin RJ, Wilson CG: What to do about apnea of prematurity, J Appl Physiol 107:1015-1016, 2009.) monitoring appear to correlate with less favorable neurodevelopmental outcome (Hunt et al, 2004). Future studies might focus more on the incidence and severity of desaturation events, because techniques for long-term collection of pulse oximeter data are now more advanced. Furthermore, it is likely that recurrent hypoxia is the detrimental feature of the breathing abnormalities exhibited by preterm infants. Figure 43-7 summarizes proposed morbidities that might be attributable to intermittent hypoxic episodes in early life. Recurrent episodes of desaturation during early life and resultant effects on neuronal plasticity related to peripheral and central respiratory control mechanisms may serve as an important future direction for study. Ongoing investigation into the genetic background of infants with CCHS may also serve to enhance our understanding of the problem of apnea in preterm infants (Berry-Kravis et al, 2006). SUGGESTED READINGS Abu-Shaweesh JM, Martin RJ: Neonatal apnea: what’s new? Pediatr Pulmonol 43:937-944, 2008. Alheid GF, McCrimmon DR: The chemical neuroanatomy of breathing, Respir Physiol Neurobiol 164:3-11, 2008. Bianchi AL, Denavit-Saubie M, Champagnat J: Central control of breathing in mammals: neuronal circuitry, membrane properties, and neurotransmitters, Physiol Rev 75:1-45, 1995. Gauda EB, Carroll JL, Donnelly DF: Developmental maturation of chemosensitivity to hypoxia of peripheral arterial chemoreceptors: invited article, Adv Exp Med Biol 648:243-255, 2009. Kinney HC, Richerson GB, Dymecki SM, et al: The brainstem and serotonin in the sudden infant death syndrome, Annu Rev Pathol 4:517-550, 2009. Martin RJ, Wilson CG: What to do about apnea of prematurity, J Appl Physiol 107:1015-1016, 2009. Nock ML, DiFiore JM, Arko MK, et al: Relationship of the ventilatory response to hypoxia with neonatal apnea in preterm infants, J Pediatr 144:291-295, 2004. Richter DW, Spyer KM: Studying rhythmogenesis of breathing: comparison of in vivo and in vitro models, In TRENDS in Neurosciences 24:464-472, 2001. Schmidt B, Roberts RS, Davis P, et al: Long-term effects of caffeine therapy for apnea of prematurity, N Engl J Med 357:1893-1902, 2007. Thach BT: Maturation and transformation of reflexes that protect the laryngeal airway from liquid aspiration from fetal to adult life, Am J Med 111:69S-77S, 2001. Complete references and supplemental color images used in this text can be found online at www.expertconsult.com C H A P T E R 44 Pulmonary Physiology of the Newborn Robert M. DiBlasi, C. Peter Richardson, and Thomas Hansen LUNG MECHANICS AND LUNG VOLUMES The lungs possess physical, or mechanical, properties that resist inflation, such as elastic recoil, resistance, and inertance. The dynamic interaction between these properties determines the effort that must be exerted during spontaneous breathing and the resting and extreme values for the volume of gas in the lung. ELASTIC RECOIL The lung contains elastic tissues that must be stretched for lung inflation to occur. Hooke’s law requires that the pressure needed to inflate the lung must be proportional to the volume of inflation (Figure 44-1). Conventionally, volume of inflation is plotted on the y-axis, and the distending pressure is plotted on the x-axis. In this way, the constant of proportionality is volume divided by pressure, or lung compliance. Throughout the range of tidal ventilation, the relationship between pressure and volume is linear. At higher lung volumes, as the lung reaches its elastic limit (i.e., total lung capacity), this relationship plateaus, making the pressure-volume relationship nonlinear. The lungs and the chest wall function as a unit (the respiratory system) coupled by the interface between the parietal and visceral pleura. Therefore, respiratory system compliance (CRS) can be partitioned into lung compliance (CL) and chest wall compliance (CCW), where CRS = 1/CL + 1/CCW. The tendency for the lung to collapse inward at end-exhalation is balanced by the outward recoil of the chest wall resulting in a negative (subatmospheric) intrapleural pressure. The functional residual capacity (FRC) is the volume of gas in the lungs when the elastic forces of these two structures reach equilibrium (Greenspan et al, 2005). Inflation of the respiratory system above FRC requires a positive distending pressure that must overcome the elastic recoil of both the lung (transpulmonary pressure; alveolar-intrapleural pressure) and the chest wall (transthoracic pressure; intrapleural pressure-atmospheric pressure) (Gappa et al, 2006). Deflation below FRC requires an active expiratory maneuver. Residual volume (RV) is defined as the volume of air that cannot be expired with a forced deflation. As depicted in Figure 44-1, the relative compliance of the lung of the newborn is similar to that of the adult (Krieger, 1963). However, the infant’s chest wall is composed primarily of cartilage, whereas in adults the chest wall is completely ossified and therefore, CCW is greater in infants than adult (see Figure 44-1) and pleural pressure is only slightly subatmospheric. Measurements of lung and chest wall compliance suggest that the newborn should have a lower percent RV and a lower percent FRC 598 than the adult. In fact, the percent FRC in the newborn is equal to the adult’s, and the infant’s percent RV is slightly greater. This seeming paradox exists because FRC and RV are measured while the infant is breathing, and predictions from the pressure volume curves assume that there is no air movement and passive relaxation of all respiratory muscles (Bryan and England, 1984). Establishment of the FRC is vital to the role of maintaining adequate lung mechanics and gas exchange. The highly compliant newborn chest wall exerts very little outward distending pressure, and thus the lung is more prone to collapse at end-exhalation (Hülskamp et al, 2005. Data suggest three mechanisms by which the newborn can limit expiratory flow and increase the intrapulmonary pressure to maintain a normal FRC during spontaneous breathing: (1) by increasing expiratory resistance through laryngeal adduction (glottic narrowing), (2) by maintaining inspiratory muscle activity throughout expiration, and (3) by initiating high breathing frequencies to limit the expiratory time (Magnenant et al, 2004; te Pas et al, 2008). RESISTANCE Resistance to gas flow arises because of friction between gas molecules and the walls of airways (i.e., airway resistance) and because of friction between the tissues of the lung and the chest wall (i.e., viscous tissue resistance). Airway resistance represents approximately 80% of the total resistance of the respiratory system, and tissue resistance and inertial forces account for the remaining 20% (Polgar and String, 1966). In the newborn, nasal resistance represents nearly half the total airway resistance; in the adult, it accounts for about 65% of the airway resistance (Polgar and Kong, 1965). Gas flows only in response to a pressure gradient (Figure 44-2). During laminar flow, the pressure difference needed to move gas through the airway is directly related to the flow rate times a constant—airway resistance. During turbulent flow, however, this pressure is directly proportional to a constant times the flow rate squared. Gas flow becomes turbulent at branch points in airways, at sites of obstruction, and at high flow rates. Turbulence occurs whenever flow increases to a point that Reynolds’ number exceeds 2000. This dimensionless number is directly proportional to the volumetric flow rate and gas density, and it is inversely proportional to the radius of the tube and gas viscosity. Obviously, turbulent flow is most likely to occur in the central airways where volumetric flow is high, rather than in lung periphery where flow is distributed across a large number of airways. Both types of flow exist in the lung, so the net pressure drop is calculated as follows (Pedley et al, 1977): 2 ΔP = ( K 1 × V̇ ) + (K2 × ˙V ) (1) CHAPTER 44 Pulmonary Physiology of the Newborn A 599 B FIGURE 44-1 This is an idealized plot of volume as a function of distending pressure for the lung, chest wall, and respiratory system (lung plus chest wall) of an adult (A) and an infant (B). These curves are derived by instilling or removing a measured volume of gas from the lung and allowing the respiratory system to come to rest against a shuttered airway. At this point only elastic forces are acting on the respiratory system, and airway pressure is equal to alveolar pressure. Intrapleural pressure can be measured using an esophageal balloon. Because airway pressure is equal to alveolar pressure, the distending pressure for the lung can be measured as airway pressure – intrapleural pressure. The distending pressure for the chest wall is intrapleural pressure – atmospheric pressure, and the distending pressure for the respiratory system is airway pressure – atmospheric pressure. Compliance is the change in volume divided by the change in distending pressure. The shaded area is the resting intrapleural pressure at functional residual capacity. Lung volumes depicted include residual volume (  RV), functional residual capacity (  FRC), and total lung capacity (  TLC). LAMINAR FLOW Pin Pout ∇ P = Pin - Pout • ∇ P = V x K1 TURBULENT FLOW Pin Pout ∇ P = Pin - Pout • P = V2 x K2 FIGURE 44-2 Gas flow ( V̇ ) through tubular structures occurs only in the presence of a pressure gradient (Pin > Pout). For laminar flow, P is directly proportional to V̇ . ∇ Δ P = V̇ × 8×L×μ π × r4 In this case, the constant of proportionality (K1) is directly related to the length of the airway (L) and the viscosity of the gas (μ) and indirectly proportional to the fourth power of the radius of the airway (r). For turbulent flow, ΔP is proportional to V̇ . The constant of proportionality (K2) is directly proportional to the length of the airway and the density of the gas and inversely proportional to the fifth power of the radius of the airway. It is possible to take advantage of the differences between laminar and turbulent flow to determine the site of airway obstruction in the lung. If obstruction to gas flow is in the central airways, turbulent flow is affected the most. Because turbulent gas flow is density dependent, allowing the patient to breathe a less dense gas (such as helium mixed with oxygen) reduces the resistance to gas flow. If the site of obstruction is peripheral, the mixture of helium and oxygen does not appreciably affect resistance. Inflation of the lung increases the length of airways and might therefore be expected to increase airway resistance; however, lung inflation also increases airway diameter. Because airway resistance varies with the fourth to fifth power of the radius of the airway, the effects of changes in airway diameter dominate, and resistance is inversely proportional to lung volume (Rodarte and Rehder, 1986). Similarly, airway resistance is lower during inspiration than during expiration because of the effects of changes in intrapleural pressure on airway diameter. During inspiration, pleural pressure becomes negative, and a distending pressure is applied across the lung. This distending pressure increases airway diameter as well as alveolar diameter and decreases the resistance to gas flow. During expiration, pleural pressure increases and airways are compressed. Collapse of airways is opposed by their cartilaginous support and by the pressure exerted by gas in their lumina. During passive expiration, these defenses are sufficient to prevent airway closure. When intrapleural pressure is high, during active expiration, airways may collapse, and gas may be trapped in the lung. This problem may be accentuated in the small preterm infant with poorly supported central airways. INERTANCE Gas and tissues in the respiratory system also resist accelerations in flow. Inertance is a property that is negligible during quiet breathing and physiologically significant only at rapid respiratory rates. DYNAMIC INTERACTION Compliance, resistance, and inertance all interact during spontaneous breathing (Figure 44-3). This interaction is described by the equation of motion for the respiratory system: ( ) 1 (2) P (t) = V [t] × + (V̇ [t] × R) + (V̇ [t] × I )  C 600 PART X Patm = Respiratory System >> = << = Palv –100 –80 Slope = –1/RC • Mid-volume Mid-volume Volume V(t) (mL/sec) Flow Pleural pressure –60 –40 C= Volume Pressure –20 Ptot Pel Pfr 0 Ptot Pfr Pel FIGURE 44-3 Gas flows from the atmosphere into the lung only if atmospheric pressure (Patm) is greater than alveolar pressure (Palv). At end exhalation, when Patm equals Palv, there is no gas movement in or out of the lung. During a spontaneous inspiration, the diaphragm contracts, the chest wall expands, and the volume in the intrathoracic space increases. As a result, pleural pressure (Ppl) decreases relative to Patm, and a gradient is created between Ppl and Palv, distending the lung, increasing alveolar volume, and decreasing Palv. A gradient is also created between Patm and Palv, and gas flows from the atmosphere into the alveolar space. The rate of gas flow increases rapidly, reaches a maximum (peak flow), then decreases as the alveolus fills with gas and Palv approaches Patm. At peak inspiration, Palv equals Patm, and lung volume is at its maximum, as is Ppl. The curved solid line connecting end expiration to end inspiration is the total driving pressure for inspiration (Ptot). The dotted line represents the pressure needed to overcome elastic forces alone (Pel). The difference between the two lines is the pressure dissipated overcoming flow resistive forces (Pfr). During exhalation, this cycle is reversed. 25 P (t) = (V [t] × 1/C) and C = V (t) /P (t) (4) This series of equations and Figure 44-3 demonstrate that at points of no gas flow (end expiration and end inspiration), only elastic forces are operating on the lung. During inflation or deflation of the lung, however, both elastic and resistive forces are important. Although the solution to the equation of motion for the respiratory system is beyond the scope of this discussion, the behavior of the respiratory system during passive exhalation is a special situation for which a solution can be obtained relatively easily using the occlusion technique (Lesouef et al, 1984; McIlroy et al, 1963). Before a passive exhalation maneuver, the infant is given a positive pressure breath, and the airway is occluded—invoking the Hering-Breuer reflex and a brief apnea. Airway pressure is measured, and the occlusion is released. Expired gas flow is measured using a pneumotachometer and integrated to volume; flow is then plotted as a function of volume (Figure 44-4, A). During a passive exhalation, there are no external forces acting on 15 A 10 5 0 V(t) (mL) 30 Vo 24 V(t) (mL) 18 12 Pressure (cm H2O) 5.0 5.0 Compliance (mL/cm H2O) Resistance (cm H2O/mL/sec) 0.05 Time Constant (sec) 0.25 63% 6 V(t) = Vo x e 86% 95% –t/RC 98% 99% 0 0 where P(t) is the driving pressure at time, t, V[t] is the lung volume above FRC, C is the respiratory system compliance, V̇[t] is the rate of gas flow, R is the resistance of the respiratory system, V̇[t] is the rate of acceleration of gas in the airways, and I is the inertance of the respiratory system. If I is neglected, the equation simplifies to: ( ) 1 (3) P (t) = V [t] × +(V̇ [t] ×R) C  At times of zero gas flow (end expiration and end inspiration), the equation further simplifies to: 20 B 0.50 0.75 1.0 1.25 Time (sec) FIGURE 44-4 A, Plot of flow of gas out of the lung versus volume of gas remaining in the lung V(t) for a passive exhalation. Flow of gas out of the lung is negative by convention. After an initial sharp increase, flow decreases linearly as the lung empties. Static compliance of the respiratory system is obtained by dividing the exhaled volume by the airway pressure at the beginning of the passive exhalation. Resistance is calculated from the slope of the flow-volume plot −(1/RC) and the compliance. This technique has the advantage of not requiring measurements of pleural pressure and being relatively unaffected by chest wall distortion. B, V(t) is plotted as a function of time for a passive exhalation. The graph is an exponential with the equation V (t) = Vo × e−t / RC V0 is the starting volume, and e is mathematical constant (roughly 2.72). For this example, the time constant of the respiratory system (Trs) is roughly 0.25 second. Calculations show that when exhalation persists for a time equal to one time constant (t = 0.25 sec = 1 × Trs), 63% of the gas in the lung is exhaled. For t = 2 Trs, 86% of the gas is exhaled; for 3 Trs, 95%; for 4 Trs, 98%; and for 5 Trs, 99%. If expiration is interrupted before a time t = 3Trs, gas is trapped in the lung. the respiratory system (P[t] = 0), so the equation of motion simplifies to a first-order differential equation: ( ) 1 V [t] × + (V̇ [t] × R) = 0 C Rearranging yields the linear equation ( ) 1 (5) V̇ (t) =− × V (t) [RC]  where the slope is −1/RC, which can be determined using a linear regression of V̇ (t) versus V(t). CHAPTER 44 Pulmonary Physiology of the Newborn This equation states that during passive exhalation, flow plotted against volume is a straight line with slope −1/(RC). The quantity RC has the units of time and is termed the respiratory system time constant (Trs). Trs defines the rate at which the lung deflates during a passive exhalation (see Figure 44-4). Time constants affect the rate of lung inflation in the same manner in which they affect lung deflation (see Chapter 45). WORK OF BREATHING The work of breathing is a reflection of the amount of energy required to overcome the elastic and resistive elements of the respiratory system and move gas into and out of the lung during spontaneous breathing. Work of breathing is defined as the cumulative product of distending pressure and the given volume displaced during inhalation or exhalation (Figure 44-5): WOB = ∫ PdV C Tidal volume (change in mL) Expiration 20 D Compliance line B 10 Inspiration FRC A where P is pleural pressure in time relative to pleural pressure at end exhalation and V is the volume in time relative to the volume at end exhalation. The work of breathing required to ventilate the lungs of normal newborns is approximately 10% of that required in adults (McIlroy and Tomlinson, 1955). However, infants have been shown to have a higher oxygen cost and lower mechanical efficiency associated with the work of breathing than adults (Thibeault et al, 1966). In healthy infants, the majority of the work of breathing is being done by the diaphragm during inhalation. Approximately one third of the total inspiratory work of breathing is related to overcoming the resistance to gas flow in the airways (Mortola et al, 1982). Exhalation is usually passive due to potential energy stored in the lung and the chest wall at endinhalation but may become active as expiratory resistance increases or lung volumes decrease below FRC. MEASUREMENTS OF RESPIRATORY SYSTEM MECHANICS TLC E 601 5 Tidal pressure (change in cm H2O) RV FIGURE 44-5 Pressure-volume loop showing the compliance line (AC, joining points of no flow); work done in overcoming elastic resistance (ACEA), which incorporates the frictional resistance encountered during expiration (ACDA); work done in overcoming frictional resistance during inspiration (ABCA); and total work done during the respiratory cycle (ABCEA), or the entire shaded area. (Figure and legend taken from Wood BR: Physiologic principles, In Goldsmith JP, Karotkin EH, editors: Assisted ventilation of the neonate, ed 4, Philadelphia, 2003, WB Saunders, pp 15-30.) Respiratory system mechanics are objective measurements used to determine the severity of lung disease, changes in pathophysiology, response to therapeutic interventions, and progression of lung growth and development. Table 44-1 shows mechanics values for well and sick newborn infants. Lung mechanics in infants are typically measured during spontaneous breathing (dynamic) and by applying maneuvers during passive breathing conditions (static). As mentioned previously, the mechanical behavior of the respiratory system (chest wall and lung) can be decoupled effectively by measuring pleural pressure. Because measuring pleural pressure is not feasible, esophageal pressure measured in the distal third of the esophagus, using an air-filled catheter attached to a pressure transducer, can be used to estimate pleural pressure. Thus, transpulmonary pressure is estimated by the difference between airway pressure and esophageal pressure and is useful in measuring lung and chest wall mechanics. Dynamic compliance takes advantage of gas flow transiently being equal to zero at end inspiration and end expiration (see Figure 44-3). Dynamic compliance is calculated by dividing the change in volume between these two points in time by the concomitant change in distending pressure. Static compliance is the compliance measured when the infant is completely passive and can be estimated using an inspiratory hold at end inhalation during assisted ventilation (McCann et al, 1987). TABLE 44-1 Lung Volumes and Mechanics of Well and Sick Neonates Measurements Units Normal RDS BPD 4-6 4-7 Tidal volume mL/kg 5-7 FRC mL/kg 25-30 20-33 20-30 Compliance mL/cm H2O 1-2 0.3-0.6 0.2-0.8 Resistance cm H2O/L/sec 25-50 60-160 30-170 FRC, Functional residual capacity. Data from Choukroun et al, 2003; Cook et al, 1957; Gerhardt and Bancalari, 1980; McCann et al, 1987; Polgar and Promadhat, 1971; Polgar and String, 1966; Reynolds and Etsten, 1966. 602 PART X Respiratory System Another estimate of static respiratory system compliance requires instilling a known volume of gas into the lung, then measuring airway pressure at equilibrium in the absence of airflow and respiratory muscle activity (see Figure 44-1). This single occlusion technique is used to measure compliance during the passive exhalation maneuver described previously (see Figure 44-4). In the normal infant, it is generally assumed that dynamic compliance is equal to static compliance. However, in infants that are tachypneic or who have elevated airway resistance, the dynamic compliance may underestimate the static compliance of the lung (Katier et al, 2006). As was mentioned earlier, measurements of compliance are affected by lung size. For example, if a 5 cm H2O distending pressure results in a 25-mL increase in lung volume in a newborn, calculated lung compliance is 5 mL/cm H2O. In an adult, the same 5 cm H2O distending pressure increases the lung volume by roughly 500 mL, and calculated compliance is 100 mL/cm H2O. Although the calculated lung compliances are different, the forces needed to carry out tidal ventilation are similar (i.e., lung function is normal in both circumstances). This example points out that if lung compliances are to be compared, they must be corrected for size. This is usually performed by dividing compliance by resting lung volume to get specific compliance. For the newborn, resting lung volume is roughly 100 mL, so specific compliance is 0.05 mL/cm H2O/mL lung volume. For the adult, resting lung volume is nearly 2000 mL, so specific compliance is 0.05 mL/cm H2O/mL lung volume—identical to that of the newborn. Thus, one might expect an infant born small for gestational age to have low lung compliance and normal specific compliance. Lung compliance changes with volume history, meaning that it decreases with fixed tidal volumes and increases after deep breaths that recruit air spaces that may have been poorly ventilated or collapsed. The periodic sigh in spontaneous breathing is typically associated with an increase in lung compliance and in oxygenation (Frappell and MacFarlane, 2005). However, sighs in premature infants are often followed by apnea and hypoventilation that could lead to destabilization in infants affected with lung disease (Qureshi et al, 2009). Many respiratory disorders result in nonhomogeneous increases in small airway resistance in the lung (see Table 44-1). Therefore, if lung compliance remains relatively uniform, the product of resistance and compliance (Trs) varies throughout the lung. During lung inflation, units with normal resistance have the lowest Trs and fill rapidly. Units with high resistance have a longer Trs and fill more slowly. At rapid respiratory rates when the duration of inspiration is short, only those lung units with a short Trs are ventilated. In effect, the ventilated lung becomes smaller. As discussed earlier, as the lung becomes smaller, its measured compliance decreases. Therefore, in infants with ventilation inhomogeneities, dynamic lung compliance decreases as respiratory rate increases. This decrease in lung compliance with increasing respiratory rate is termed frequency dependence of compliance, and it is suggestive of inhomogeneous small airway obstruction. Resistance of the total respiratory system can be measured using the passive exhalation technique described previously (see Figure 44-4), or it can be calculated from measurements of distending pressure, volume, and flow (see Figure 44-3). Points of equal volume are chosen during inspiration and expiration. The gas flow and the distending pressure are measured at each point. The pressure needed to overcome elastic forces should be the same for inspiration and expiration and therefore cancel out. Total resistance, consequently, is equal to distending pressure at the inspiratory point minus distending pressure at the expiratory point, divided by the sum of the respective inspiratory and expiratory point gas flows. Investigators have calculated compliance and resistance by measuring distending pressure, gas flow, and volume (see Figure 44-3), then fitting these measurements to the equation of motion (see Equation 3), using multiple linear regression techniques, and solving for the coefficients 1/C and R (Bhutani et al, 1988). The forced oscillation technique is used to estimate RRS in spontaneously breathing subjects (Goldman et al, 1970). A loudspeaker is used to generate a sinusoidal pressure wave (PRS) to the infant’s nose via a face mask, with the mouth occluded, and the resulting gas flow is measured using a pneumotachograph. The PRS / V̇ relationship, called impedance (Z), is expressed as an amplitude ratio, called modulus (|Z|), and a phase shift (Φ) of both signals (Desager et al, 1996). Applying the simple RLC lung model described earlier in Equation 2 yields: ZRS = RRS + j[ωIRS − 1/ωCRS ] where j = √−1 and Φ = 2π (frequency). The imaginary term on the right in the equation is termed reactance. At low frequencies, 1/ωCRS >> ωIRS, and Φ is negative. At high frequencies, the inertance term predominates, and Φ is positive. At an intermediate frequency, called resonant frequency, the effects of compliance and inertance are equal and cancel each other. The forced oscillations can be superimposed on spontaneous breaths (Jackson et al, 1996), making this technique suitable for infants. Forced oscillations methods can also be used with more complex lung models separating airway impedance from tissue impedance (Peslin et al, 1972). FRC is measured by inert gas dilution techniques (helium dilution) or inert gas displacement (nitrogen washout) (Figure 44-6). Both of these techniques measure gas that communicates with the airways. The total volume of gas in the thorax at end expiration (thoracic gas volume [TGV]) can be measured using a body plethysmograph and applying Boyle’s law. This technique measures all gas in the thorax—even trapped gas that is not in contact with the airways. Obviously, FRC measured by inert gas dilution is less than TGV if significant volumes of trapped gas are present. ALVEOLAR VENTILATION The tissues of the body continuously consume O2 and produce CO2 (Figure 44-7). The primary function of the circulation is to pick up O2 from the lungs and deliver it to the tissues, then to pick up CO2 from the tissues and deliver it to the lungs. The exchange of O2 and CO2 with the blood occurs within the alveolar volume of the lungs. The alveolar volume acts as a “large sink” from which O2 is continuously extracted by the blood and to which CO2 is continuously added. This mechanism for acquiring O2 from CHAPTER 44 Pulmonary Physiology of the Newborn A 603 B FIGURE 44-6 A, Measurement of functional residual capacity (FRC) by helium dilution. At end exhalation, the infant breathes from a bag containing a known volume (Vbag) and concentration of helium (Hei) in oxygen. The gas in the infant’s lungs dilutes the helium oxygen mixture to a new concentration (Hef): FRC = Bag volume × (HeI − Hef)/ Hef. B, Measurement of thoracic gas volume (TGV) using a plethysmograph. The infant breathes spontaneously in a sealed body plethysmograph. At end exhalation, the airway is closed with a shutter. As the infant attempts to inspire against the shutter, the volume of the thorax increases and airway pressure decreases. The increase in volume of the thorax can be measured from the change in the pressure inside of the plethysmograph (Pbox). By Boyle’s law: P × TGV = (P − ΔP) × (TGV + ΔV), where P is atmospheric pressure, (P − ΔP) is airway pressure during occlusion, and (TGV + ΔV) is thoracic volume during occlusion. Therefore, TGV = (P − ΔP) × ΔV/ΔP. Because ΔP is small compared with P, this can be simplified to TGV = P × ΔV/ΔP. the atmosphere and excreting CO2 into the atmosphere is the alveolar ventilation (Slonim and Hamilton, 1987). The alveolar volume of the lung includes all lung units capable of exchanging gas with mixed venous blood: respiratory bronchioles, alveolar ducts, and alveoli. Because the conducting airways do not participate in gas exchange, they constitute the anatomic dead space (Vd). At end exhalation, the FRC is the sum of the volume of gas in the alveolar volume and in the anatomic dead space. During normal breathing, the amount of gas entering and leaving the lung with each breath is the tidal volume (Vt): VT × respiratory rate (RR) = minute ventilation (V̇) Part of each Vt is wasted ventilation because it moves gas in and out of the Vd. Therefore, alveolar ventilation ( V̇A ) can be expressed as: (6) V̇A = (VT − VD ) × RR Alveolar ventilation is an intermittent process, whereas gas exchange between the alveolar space and the blood occurs continuously. Because arterial O2 and CO2 tensions (Pao2 and Paco2) are roughly equal to the O2 and CO2 tensions within the alveolar space, these fluctuations in breathing could result in intermittent hypoxemia and hypercarbia. Fortunately the lung has a large buffer—the FRC. The FRC is four to five times as large as the Vt; therefore, only a fraction of the total gas in the lung is exchanged during normal breathing. This large buffer continues to supply O2 to the blood during expiration and acts as a sump to accept CO2 from the blood, so alveolar O2 and CO2 tensions (PaO2 and PaCO2) change little throughout the ventilatory cycle. FIGURE 44-7 Schematic showing coupling of alveolar ventilation to tissue oxygen consumption. Alveolar ventilation is linked tightly to metabolism. When alveolar ventilation is uncoupled from the body’s metabolic rate, hypoventilation or hyperventilation results. During hypoventilation, less O2 is added to the alveolar space than is removed by the blood, and less CO2 is removed from the alveolar space than is added by the blood. As a result, PaO2 decreases and PaCO2 increases. The net result of hypoventilation is hypoxemia and hypercapnia. Administering supplemental O2 increases the quantity of O2 in each breath delivered to the alveolar space, and it may prevent arterial hypoxemia. For example, suppose a 1-kg male infant has a Vt of 6 mL, an anatomic Vd of 2 mL, and a respiratory rate of 40 breaths/min. His alveolar ventilation is 160 mL per minute ([6 mL − 2 mL] × 40/ min). If he breathes room air (21% O2), he delivers 33.6 mL of O2 to the alveolar space every minute (160 mL/min × 0.21). If he maintains the same Vt but breathes only 20 times per minute, his alveolar ventilation decreases to 80 mL per minute, only 16.8 mL of O2 (80 mL/min × 0.21) is delivered to the alveolar space each minute, and his PaO2 and Pao2 decrease. If he is allowed to breathe 50% O2, O2 delivery to the alveolar space increases to 40 mL per minute (80 mL/min × 0.50), and both his PaO2 and Pao2 increase. Because O2 administration has no effect on the accumulation of CO2, it does not prevent hypercapnia. Hyperventilation delivers more O2 to the alveolar space than can be removed by the blood and removes more CO2 than can be added by the blood. As a result, PaO2 increases and PaCO2 decreases. 604 PART X Respiratory System Measurements of alveolar ventilation and anatomic Vd in the infant rely on the relationship between CO2 ­production (VCO2 ), V̇A , and PaCO2. The mathematical expression of this relationship states (Cook et al, 1955): V̇CO2 FA CO2 = (7) V̇A  FACO2 is the fraction of CO2 in total alveolar gas, or FA CO2 = PACO2/(PB − 47) (8) Pb is the barometric pressure, and 47 mm Hg is the vapor pressure of water at body temperature. Therefore, V̇A = [ V̇CO2 × (PB − 47) ] PA CO2  (9) If minute ventilation ( V̇) is measured, dead space ventilation ( V̇D ) is calculated as: (10) V̇D = V̇ − V̇A  and Vd is calculated by dividing by the respiratory rate. This method measures the anatomic Vd in the lung. As seen in the next section, portions of some gas exchanging units in the lung can also function as Vd; therefore, the total Vd, or the physiologic Vd, may be greater than the anatomic Vd. Physiologic Vd is calculated by substituting PaCO2 into Equation 9 for PaCO2. When PaCO2 = PaCO2, all the Vd is anatomic Vd, and the gas-exchanging units are all functioning normally. As physiologic Vd increases, however, PaCO2 increases relative to PaCO2. Therefore, the difference between PaCO2 and PaCO2 (the aA.DCO2) is a measure of efficiency of gas exchange in the lung. For clinical purposes, V̇CO2 in Equation 9 is assumed to be a constant so that V̇A is proportional to 1/PaCO2. Thus, increased PaCO2 means that alveolar ventilation has decreased; decreased PaCO2 means that alveolar ventilation has increased. VENTILATION-PERFUSION RELATIONSHIPS Under ideal circumstances, ventilation and perfusion of the lung are evenly matched ( V̇/Q̇ = 1) , both in the lung as a whole and in each individual air space. The air spaces receive O2 from the inspired gas and CO2 from the blood. O2 is transported into the blood, while CO2 is transported to the atmosphere. Even though V̇/Q̇ is 1, CO2 and O2 are exchanged in the lung at the same ratio at which they are exchanged in the tissues: A little less CO2 is transported out than O2 is transported in, so the respiratory exchange ratio R equals 0.8. If there is no diffusion defect, the gas composition of the air spaces and the blood comes into equilibrium. N2 makes up the balance of dry gas. The sum of partial pressures of all gases in the air spaces must equal atmospheric pressure. The ideal alveolar gas composition is PO2 = 100, PCO2 = 40, PN2 = 573, and PH2O = 47 (all in mm Hg) at an atmospheric pressure of 760 mm Hg. The ideal arterial blood composition is the same. Therefore, differences between alveolar and arterial gas composition under ideal circumstances are all zero. Knowing the values for PaCO2 and inspired gas, ideal alveolar gas composition can be calculated from the alveolar gas equations (Farhi, 1966): PAO2 = PI O2 − PACO2 × [FI O2 + (1 − FI O2 ) / R] (11) where PI O2 = FI O2 × (PB − PH2 O) (12) PA N2 = FI N2 × [PACO2 × (1 − R) / R + (PB − PH2 O) ] (13) Under normal circumstances, and certainly in the presence of lung disease, this ideal situation is not the case; some air spaces receive more ventilation than perfusion, and others receive more perfusion than ventilation. A reduction of ventilation may occur because of atelectasis, alveolar fluid, or airway narrowing. Reduced ventilation in one part of the lung may cause increased ventilation elsewhere. A reduction of perfusion may occur if air spaces are collapsed or overdistended or because of gravitational effects, and increased perfusion may occur in congenital heart disease. As with ventilation, reduced perfusion in one part of the lung may cause increased perfusion in other regions. If an air space is relatively overventilated (high V̇/Q̇ ), its gas composition trends toward that of inspired gas, which in the case of room air is PO2 = 150 mm Hg and PCO2 = 0 mm Hg. If an air space is relatively underventilated (low V̇/Q̇ ), its gas composition tends toward that of mixed venous blood, which is PO2 = 40 mm Hg and PCO2 = 46 mm Hg. What counts is the V̇/Q̇ ratio, not absolute values of V̇ or Q̇ (West, 1986). To understand V̇/Q̇ imbalance, it is common to view the lung as a three-compartment model (Figure 44-8): V̇/Q̇ = 0 (Figure 44-8, A); V̇/Q̇ = 1 (Figure 44-8, B); and V̇/Q̇ = infinity (Figure 44-8, C). The O2 saturation of blood in each compartment depends on the PO2 and the O2 dissociation curve. For illustrative purposes, in a badly diseased lung, 50% of ventilation goes to V̇/Q̇ = 1 and 50% to V̇/Q̇ = infinity , whereas 50% of perfusion goes to V̇/Q̇ = 1 and 50% to V̇/Q̇ = 0. Perfusion of V̇/Q̇ = 0 causes venous admixture, whereas ventilation of V̇/Q̇ = infinity causes alveolar Vd. The mixed alveolar gas composition is easily calculated as the mean. For mixed arterial blood, the PO2 must be read from the O2 dissociation curve, but because the CO2 dissociation curve is fairly linear, the values for CO2 are easily calculated as the mean. The abnormalities in distribution of V̇ and Q̇ have created an Aa.DO2 = 70, aA.DCO2 = 23, and aA.DN2 = 32 mm Hg (see Figure 44-8). The Aa.DO2 is greater than the sum of the other two because the O2 dissociation curve is not linear. Of course, the situation in most lungs is not as extreme as the one illustrated. From this illustration, however, it can be seen that: 1. Open low V̇/Q̇ units produce increased Aa.DO2, significant hypoxemia, and increased aA.DN2, but because they are poorly ventilated and have a PCO2 close to the ideal value, they do not change the aA.DCO2 significantly. 2. High V̇/Q̇ units produce increased Aa.DO2 without hypoxemia and increased aA.DCO2, but because they are poorly perfused and have a PN2 close to the ideal value, they do not change the aA.DN2 significantly. For the calculation of Aa.DO2 and aA.DN2, it is customary to calculate the ideal alveolar gas composition for O2 and N2 from the alveolar gas equations and use these values with those measured for arterial PO2 and PN2. This CHAPTER 44 Pulmonary Physiology of the Newborn where Q̇va/Q̇t = venous admixture, CO2 = oxygen content, ċ = pulmonary capillary, a = arterial, and v = mixed venous blood. For practical application, Cv O2 is calculated from a constant a v .O2 difference, which does introduce an error. If an infant breathes 100% O2 for 15 minutes, most N2 is washed out of the lung, and the PO2 in open low V̇/Q̇ Q units becomes so high that associated blood is 100% saturated with O2. The remaining venous admixture is attributed to true right-to-left shunt ( Q̇s/ Q̇t). If an infant has the total venous admixture Q̇va/ Q̇t measured while breathing room air, then true shunt ( Q̇s/ Q̇t) measured while breathing 100% O2, the venous admixture caused by open low V̇/Q̇ units ( Q̇o/ Q̇t) can be calculated as the difference. The venous admixture caused by open low V̇/Q̇ units can also be calculated from the aA.DN2 (Markello et al, 1972): PAO2 = 125 PACO2 = 20 PAN2 = 568 A B C Q̇o Q̇t PAO2 = 55 PACO2 = 43 PAN2 = 600 FIGURE 44-8 Three-compartment model of the lung with V̇ V̇ V̇ = O (A) , = 1 (B) , = infinity (C) Q̇ Q̇ Q̇ The inspired gas is room air, and B is the ideal compartment. The sum of alveolar gas partial pressures is always 713 mm Hg. SO2 is oxygen saturation in capillary blood. PaO2 is read from the oxygen dissociation curve for a saturation of 86%. By calculated differences, Aa.DO2 is 70 mm Hg, aA.DCO2 is 23 mm Hg, and aA.DN2 is 32 mm Hg. emphasizes that part of the Aa.DO2 and aA.DN2 responsible for hypoxemia. For aA.DCO2, both an arterial and mixed alveolar sample are required. In the newborn, a fourth compartment in the model is important. A significant part of the venous return may be shunted from right to left at the foramen ovale, ductus arteriosus, pulmonary arteriovenous vessels, or lung mesenchyme without airway development, thus adding mixed venous to mixed arterial blood. This substantially increases the Aa.DO2 but has little effect on aA.DCO2 and no effect on aA.DN2. The last mentioned is because there is no significant exchange of N2 in the body, so venous and arterial PN2 are the same. The effect on aA.DCO2 is small because venous PCO2 is only slightly higher than arterial. From this analysis, it can be seen that hypoxemia is produced by a true right-to-left shunt and open low V̇/Q̇ units. Diffusional problems are not thought to be important in the newborn. Hypoxemia may be modeled as a venous admixture, the part of mixed venous blood, expressed as a fraction of cardiac output, that when added to blood equilibrated with an ideal lung would produce the measured arterial oxygen saturation. It is calculated as follows: Q̇va Q̇t = Cc O2 − Ca O2 Cc O2 − Ca O2  605 (14) = PAN2 − PAN2 PON2 − PAN2  (15) where Pon2 is the Pn2 in the units (see Figure 44-8), PaN2 is measured, and PaN2 is the ideal value calculated from the alveolar gas equation. In newborns with a significant value for true shunt, this value really represents venous admixture as a fraction of effective pulmonary blood flow ( Q̇o/ Q̇ c .). A better estimate for Q̇o/ Q̇t can be obtained from simple arithmetic (Corbet et al, 1974): ( ) Q̇Va ( ) 1− Q̇t Q̇o Q̇o ) = × ( Q̇o Q̇t Q̇ċ (16) 1− Q̇ċ  The true right-to-left shunt can then be estimated without 100% O2 breathing using the equation: Q̇s Q̇va Q̇o = − (17) Q̇t Q̇t Q̇t  The normal values for the various indices of ventilationperfusion imbalance in normal newborn infants are shown in Table 44-2. HEART-LUNG INTERACTION EFFECTS OF THE LUNG ON THE HEART There exists considerable potential for the lung to affect the heart. Because they share the thoracic cavity, changes in intrathoracic pressure accompanying lung inflation are transmitted directly to the heart. In addition, all of the blood leaving the right ventricle must traverse the pulmonary vascular bed, so changes in pulmonary vascular resistance may greatly affect right ventricular function. EFFECTS OF CHANGES IN INTRATHORACIC PRESSURE ON THE HEART Negative Intrathoracic Pressure During spontaneous inspiratory efforts, the chest wall and diaphragm move outward, intrathoracic volume increases, and intrathoracic pressure decreases (Figure 44-9, A). The 606 PART X Respiratory System TABLE 44-2 Indices of Ventilation‑Perfusion Imbalance in the Normal Newborn Breathing Room Air Aa.DO2 mm Hg Q̇va Q̇t aA.DN2 mm Hg Q̇o Q̇s Q̇t Q̇t Newborn 25 0.25 Adult 10 0.07 10 0.10 0.15 1 7 0.05 0.02 1 aA.DCO2 mm Hg Adapted from Nelson NM: Respiration and circulation after birth. In Smith CA, ­Nelson NM, editors: The physiology of the newborn infant. Springfield, Ill, 1976, Charles C Thomas. heart also resides within the thoracic cavity and is subject to the same negative intrathoracic pressure during inspiration. With a decrease in intrathoracic pressure, the heart increases in volume, and the pressure within its chambers decreases relative to atmospheric pressure. Analogous to the lung, when the pressure within the heart decreases, blood is literally sucked back into the heart from systemic veins and arteries. On the right side of the heart, the phenomenon serves to increase the flow of blood from systemic veins into the right atrium, increasing right ventricular preload and ventricular output. On the left side of the heart, ventricular ejection is impaired. During systole, the left ventricle must overcome not only the load imposed by the systemic vascular resistance, but also the additional load imposed by the negative intrathoracic pressure (McGregor, 1979). In infants with normal lungs, spontaneous respiratory efforts result in relatively small swings in pleural pressure (2 to 3 mm Hg) that have little effect on the pressure within the heart. With airway obstruction or parenchymal lung disease, however, swings in pleural pressure can be much greater (5 to 20 mm Hg), and systemic arterial pressure may fluctuate as much as 5 to 20 mm Hg depending on where in the respiratory cycle ventricular systole occurs. In older children with asthma or some other form of airway obstruction, these fluctuations in blood pressure constitute pulsus paradoxus and are indicative of severe airway obstruction. Positive Intrathoracic Pressure During positive-pressure ventilation, the lung inflates and pushes the chest wall and diaphragm outward (Figure 44-9, B). This outward push generates a pressure in the thoracic space that is greater than atmospheric pressure. The magnitude of the increase (relative transmission of airway pressure to the pleural space) is determined by the volume of lung inflation (which, in turn, is determined by the airway pressure and lung compliance) and by the compliance of the chest wall and diaphragm. If the lung is compliant and the chest wall rigid, little airway pressure is lost inflating the lung, but considerable pressure is generated in the thoracic cavity as the lung attempts to push the rigid chest wall outward. In this instance, intrathoracic pressure (intrapleural pressure) is much greater than atmospheric and in fact nearly equal to airway pressure. If the lung is poorly compliant and the chest wall highly compliant, most of the airway pressure is dissipated trying to inflate the lungs, and little is transmitted to the thoracic cavity. The effects of positive intrathoracic pressure on the heart are opposite to those of negative intrathoracic pressure. A B FIGURE 44-9 A, Negative intrathoracic pressure increases the volume of the heart and decreases the pressure within the chambers. This facilitates return of blood from the superior vena cava (SVC) and inferior vena cava ( IVC) to the right atrium ( RA) and impedes ejection of blood from the left ventricle ( LV ) into the extrathoracic aorta. B, Positive intrathoracic pressure decreases the volume of the heart and increases pressure within its chambers. This impedes blood return to the right atrium and augments ejection of blood from the left ventricle. The heart is compressed by the lungs and chest wall, and blood is squeezed out of the heart and the thoracic cavity. Return of blood from systemic veins is impaired, and right ventricular preload and output decrease. If the increase in intrathoracic pressure coincides with ventricular systole, the effect is to augment left ventricular ejection and reduce the load on the left ventricle. In the infant undergoing positive-pressure ventilation, the degree to which lung inflation compromises venous return is related to the relative compliances of the lung and chest wall. If the infant’s lung is poorly compliant and the chest wall is compliant, as in hyaline membrane disease, there is little effect of lung inflation on venous return. If the infant’s lung is normally compliant but tight abdominal distention prevents descent of the diaphragm, intrathoracic pressure increases dramatically during positive pressure ventilation, and venous return and cardiac output can be impaired. This mechanism may help explain the circulatory instability of infants after repair of gastroschisis or omphalocele. A similar situation may arise in the preterm infant with pulmonary interstitial emphysema and massive lung overinflation. In these infants, the heart is tightly compressed between the hyperinflated lungs, the other structures of the mediastinum, and the diaphragm. Venous return may be severely limited and venous pressures so increased that massive peripheral edema often accompanies the reduction in cardiac output. Although the effects of increased pleural pressure on the right atrium are detrimental, the effects on the left ventricle may be extremely beneficial (Niemann et al, 1980). During cardiopulmonary resuscitation, the chest wall is compressed against the lung, and intrathoracic pressure increases. Because the left ventricle is in the thorax, left ventricular pressure increases as well. A gradient is created, favoring flow of blood out of the ventricle and thorax and into the extrathoracic systemic circulation. Between chest compressions, elastic recoil causes the chest wall to pull CHAPTER 44 Pulmonary Physiology of the Newborn A C B FIGURE 44-10 A, Effects of lung inflation on extra-alveolar ­vessels. B, Effects of lung inflation on alveolar vessels. C, Effect of lung volume on pulmonary vascular resistance (  PVR; solid line). Inflation is from residual volume ( RV) to functional residual capacity (  FRC) to total lung capacity (  TLC). Dashed line represents alveolar vessels; dotted line ­represents extraalveolar vessels. away from the lung and heart, decreasing pleural pressure, favoring return of venous blood, and priming the heart for the next chest compression. A similar phenomenon may result in augmentation of systemic pressure when ventilator breaths coincide with ventricular systole. EFFECT OF LUNG INFLATION ON PULMONARY VASCULAR RESISTANCE The pulmonary interstitium comprises three different interconnected connective tissue compartments, each containing a different element of the pulmonary circulation (Fishman, 1986). The first—the perivascular cuffs—­consists of a sheath of fibers that contain the preacinar pulmonary arteries, lymphatics, and bronchi. The second consists of the intersegmental and interlobular septa and contains pulmonary veins and additional lymphatics. The third connects these two within the alveolar septa and contains the majority of the pulmonary capillaries. The first and second compartments represent the extra-alveolar interstitium, whereas the third represents the alveolar interstitium. The perivascular cuffs are surrounded by alveoli and expand during lung inflation (Figure 44-10, A). As a result, pressure within each cuff decreases, distending extra-alveolar blood vessels and decreasing their resistance to blood flow. The alveolar interstitium lies between adjacent alveoli and contains the majority of gas-exchanging vessels in the lung. These vessels are exposed to alveolar pressure on both sides and during lung inflation (Figure 44-10, B) are compressed so that their resistance to blood flow increases. Therefore, during lung inflation (Figure 44-10, C ), the resistance in extra-alveolar vessels decreases, whereas resistance in alveolar vessels increases. As a result, the overall pulmonary vascular resistance decreases initially, with lung inflation reaching its nadir at FRC, and then increases with further inflation. If transition from intrauterine life to extrauterine life is to be successful, after birth all of the right ventricular output must traverse the pulmonary vascular bed. To some extent, this adaptation is facilitated by a reduction in pulmonary vascular resistance that occurs with inflation 607 of the lungs (see Figure 44-10) to a stable FRC. Inflation of the lung beyond FRC increases pulmonary vascular resistance. If care is not taken during positive pressure ventilation, it is possible to inflate the lung to the point that alveolar vessels close and blood flow through the lung is impaired. When this occurs, either cardiac output decreases or the blood bypasses the lung via the foramen ovale or ductus arteriosus. Clinically, this is manifest as circulatory insufficiency from impaired right ventricular output or hypoxemia from right-to-left shunting of blood, or both. EFFECTS OF THE HEART ON THE LUNG Pulmonary Edema Pulmonary edema is the abnormal accumulation of water and solute in the interstitial and alveolar spaces of the lung (Bland and Hansen, 1985; Staub, 1974). In the lung, fluid is filtered from capillaries in the alveolar septa into the alveolar interstitium (Figure 44-11, A) and then siphoned into the lower pressure extra-alveolar interstitium. The extraalveolar interstitium contains the pulmonary lymphatics, and under normal conditions, they remove fluid from the lung so that there is no net accumulation in the interstitium. Pulmonary edema results only when the rate of fluid filtration exceeds the rate of lymphatic removal. There are only three mechanisms by which this can occur (Figure 44-11, B): (1) the driving pressure for fluid filtration (filtration pressure) increases, (2) the permeability of the vascular bed (hence, the filtration coefficient Kf) increases, or (3) lymphatic drainage decreases. Increased Driving Pressure Filtration pressure can be increased by increased intravascular hydrostatic pressure, decreased interstitial hydrostatic pressure, decreased intravascular oncotic pressure, or increased interstitial oncotic pressure (Figure 44-11, C and D). By far the most common cause of increased filtration pressure is increased intravascular hydrostatic pressure (Box 44-1). In the newborn, intravascular hydrostatic pressure increases with increased left atrial pressure from volume overload or a number of congenital and acquired heart defects. In the preterm and term newborn, evidence suggests that alterations in pulmonary blood flow that are independent of any change in left atrial pressure may also influence fluid filtration in the lung. Preterm infants with patent ductus arteriosus and left-to-right shunts exhibit signs of respiratory insufficiency before they develop any evidence of heart failure, and experiments performed in newborn lambs show that fluid filtration in the lung can be increased by increasing pulmonary blood flow without increasing left atrial pressure (Feltes and Hansen, 1986). In the newborn with a reduced pulmonary vascular bed, either from lung injury or from hypoplasia, cardiac output appropriate for body size may represent a relative overperfusion to the lung and can result in increased fluid filtration. This phenomenon has been invoked to explain the lung edema that often complicates the course of the infant with bronchopulmonary dysplasia. The exact cause of pulmonary edema that accompanies severe hypoxia or asphyxia in the newborn is still a 608 PART X Respiratory System B A C D FIGURE 44-11 A, In the lung, fluid is continuously filtered out of vessels in the microcirculation into the interstitium and then returned to the intravascular compartment by the lymphatics. Only when the rate of filtration exceeds the rate of lymphatic removal can fluid accumulate in the interstitium. Spillover of fluid into the alveolar space occurs only when the interstitial space fills or when the alveolar membrane is damaged. B, Fluid flows out of vessels at a flow rate (Q̇f ) that is equal to the driving pressure for fluid flow (ΔP) times the filtration coefficient (Kf): (Q̇f ) = Kf × ΔP. Kf can be thought of as the relative permeability of the vascular bed to fluid flux. Kf in the normal lung is a small number so that despite a driving pressure of roughly 5 mm Hg, the net rate of fluid filtration is approximately 1 to 2 mL/kg per hour. C, The driving pressure for fluid flow out of the microvascular bed represents a balance of two sets of pressures. Within the blood vessel, hydrostatic pressure tends to push fluid out of the vessel into the interstitium. This pressure is partially opposed by a smaller hydrostatic pressure within the interstitium pushing fluid back into the blood vessel. Within the blood vessel, there also exists a discrete oncotic pressure that results predominantly from intravascular albumin that tends to draw fluid from the interstitium back into the blood vessel. This pressure is partially opposed by an interstitial oncotic pressure tending to draw fluid from the blood vessel into the interstitium. D, The intravascular hydrostatic pressure must be less than pulmonary artery pressure (Ppa) for blood to flow into the microvascular bed and greater than left atrial pressure (Pla) for blood to flow out. Intravascular pressure within the microvascular bed is roughly equal to 0.4 (Ppa − Pla) + Pla. The interstitial hydrostatic pressure is roughly equal to alveolar pressure. The intravascular oncotic pressure can be calculated from the plasma albumin concentration. The interstitial oncotic pressure is roughly two thirds of the intravascular oncotic pressure. The balance of these pressures favors filtration out of the vessel (in the normal lamb, this pressure is roughly 5 mm Hg). BOX 44-1 Increased Intravascular ­Hydrostatic Pressure INCREASED LEFT ATRIAL PRESSURE Intravascular volume overload Overzealous fluid administration Overtransfusion Renal insufficiency Heart failure Left-sided obstructive lesions Left-to-right shunts Myocardiopathies INCREASED PULMONARY BLOOD FLOW Normal pulmonary vascular bed Patent ductus arteriosus Increased cardiac output Reduced pulmonary vascular bed Bronchopulmonary dysplasia Pulmonary hypoplasia controversial issue. Data suggest that it is the result of increased filtration pressure and not the result of any alteration in permeability. Heart failure accounts for some of the increased filtration pressure following severe asphyxia. In addition, there may be some element of pulmonary venous constriction. Finally, there is evidence that hypoxia and acidosis may redistribute pulmonary blood flow to a smaller portion of the lung and result in relative overperfusion and edema, similar to that seen with anatomic loss of vascular bed (Hansen et al, 1984). Several investigators have suggested that upper airway obstruction may cause pulmonary edema by decreasing interstitial hydrostatic pressure relative to intravascular hydrostatic pressure. Other data suggest, however, that with airway obstruction, vascular pressures decrease with intrapleural pressure in such a way that filtration pressure remains unchanged (Hansen et al, 1985). Hypoproteinemia in infants results in a decrease in intravascular oncotic pressure. Its effects on filtration pressure, however, are blunted by the simultaneous decrease in protein concentration in the interstitial space of the lung. As a result, edema is unlikely to occur unless hydrostatic pressure also increases (Hazinski et al, 1986). Increased Permeability Another possible mechanism for increased fluid filtration in the lung is a change in the permeability of the microvascular membrane to protein—high-permeability CHAPTER 44 Pulmonary Physiology of the Newborn pulmonary edema. In this form of edema, the sieving properties of the microvascular endothelium are altered so that Kf increases and patients may develop pulmonary edema despite relatively normal vascular pressures (Albertine, 1985). Furthermore, even small changes in vascular pressures can result in a dramatic worsening of pulmonary status. High-permeability pulmonary edema usually implies either direct or indirect injury to the capillary endothelium of the lung. Direct injuries result from local effects of an inhaled toxin such as oxygen. Indirect injuries imply that the initial insult occurs elsewhere in the body and that the lung injury occurs secondarily. An example of indirect lung injury is sepsis: Neutrophils activated by bacterial toxins attack endothelial cells in the lung and increase permeability to water and protein (Brigham et al, 1974). Indirect injuries usually involve bloodborne mediators, such as leukocytes, leukotrienes, histamine, or bradykinin. Alveolar overdistention can also cause high-permeability pulmonary edema, presumably by direct injury of the pulmonary vascular bed. This type of vascular injury probably accounts for some of the edema that accompanies diseases such as hyaline membrane disease and bronchopulmonary dysplasia, in which maldistribution of ventilation results in areas of alveolar overdistention (Carlton et al, 1990). Decreased Lymphatic Drainage In the normal lung, the rate of lung lymph flow is equal to the net rate of fluid filtration, and as long as lymphatic function can keep up with the rate of fluid filtration, water does not accumulate in the lung. Although lymphatics can actively pump fluid against a pressure gradient, studies show that this ability is limited and that lung lymph flow varies inversely with the outflow pressure (pressure in the superior vena cava). Several groups of investigators have demonstrated that, in the presence of an increased rate of transvascular fluid filtration, the rate of fluid accumulation in the lung is substantially greater if systemic venous pressure is increased (Drake et al, 1985). Recent data suggest that the ability of the lymphatics to pump against an outflow pressure is impaired in the fetus and newborn. In fact, in fetal lambs, lymph flow ceases at an outflow pressure of roughly 15 mm Hg (Johnson et al, 1996). This explains why pulmonary edema often complicates the course of infants with bronchopulmonary dysplasia and cor pulmonale and explains the particular problem of edema with pleural effusions complicating the postoperative course of patients following cavopulmonary shunts. More recently, investigators have shown that the ability of the lymphatics to pump can be affected by other mediators with pumping increased by α-adrenergics and certain leukotrienes and impaired by nitric oxide, β-adrenergics (Von Der Weid, 2001), and products of hemolysis (Elias et al, 1990). CONGENITAL PULMONARY LYMPHANGIECTASIS Congenital pulmonary lymphangiectasis is a rare form of pulmonary lymphatic dysfunction that can be characterized into two groups: (1) cases associated with congenital heart disease and (2) cases not associated with congenital 609 heart disease. The cardiac anomalies may include hypoplastic left heart syndrome, total anomalous pulmonary venous drainage, and pulmonary stenosis, including Noonan syndrome (France and Brown, 1971). The group that does not include associated cardiac anomalies may be of early or late onset and has a wide spectrum of severity. In some individuals, the lesion is asymptomatic, whereas in others it can lead to severe respiratory failure, usually in the first hours after birth but sometimes during the first weeks or months of life. Most infants with this condition die early in the neonatal period. Pulmonary lymphangiectasis has been reported twice as often in males and has been seen in families (Scott Emaukpor et al, 1981). Usually the radiologist is the first to suggest the diagnosis after observing dilated lymphatic vessels and sometimes small accumulations of pleural fluid on the chest radiograph. The older infant may have no symptoms or mild to moderate tachypnea with various degrees of hypoxemia. If pleural fluid has accumulated, examination of the fluid is important. If the infant has received milk feedings, pleural fluid is chylous, and if not, there is an elevation in mononuclear cells and moderate protein of up to about 4%. These findings in the absence of fever or other signs of systemic illness are diagnostic of impaired lymphatic drainage. Lung biopsy is probably not indicated and can be hazardous because once the distended lymphatic channels are severed, they can leak fluid for weeks. Only if the diagnosis is in doubt is open lung biopsy appropriate. In noncardiac associated diffuse lymphangiectasis, only supportive treatment is available. The long-term prognosis depends on the severity of the lesion, but this form of pulmonary lymphangiectasis is compatible with asymptomatic life as an adult (Wohl, 1989). SYMPTOMS OF PULMONARY EDEMA As discussed previously, fluid filtered into the alveolar interstitium ordinarily moves rapidly along pressure gradients into the extra-alveolar interstitium where it is removed by the lymphatics. A delay in this process at birth can result in clinical transient respiratory distress (see Chapter 47). The extra-alveolar interstitium has a large storage capacity. Fluid does not begin to spill over into the alveoli and airways until total lung water is increased more than 50%, unless the alveolar membrane is damaged. Therefore, the first signs and symptoms of pulmonary edema are related to the presence of extra fluid in the interstitial cuffs of tissue that surround airways. As fluid builds up in these cuffs, airways are compressed, and the infants develop signs of obstructive lung disease. The chest may appear hyperinflated, and auscultation reveals rales, rhonchi, and a prolonged expiration. Early in the course, chest radiographs reveal lung overinflation and an accumulation of fluid in the extra-alveolar interstitium—linear densities of fluid that extend from the hilum to the periphery of the lung (the so-called sunburst appearance) and fluid in the fissures. With more severe edema, fluffy densities appear throughout the lung as alveoli fill with fluid (Figure 44-12). Heart size may be increased in infants with edema from increased intravascular pressure. Initially, infants present with increased Paco2 secondary to impaired ventilation. Later PaO2 decreases secondary to ventilation-perfusion 610 PART X Respiratory System mismatching and alveolar flooding. In adults, a ratio of protein concentration in tracheal aspirate to that in plasma greater than 0.5 may help to differentiate permeability pulmonary edema from high-pressure pulmonary edema (Fein et al, 1979). (Rowe and Avery, 1966). Fedrick and Butler (1971) judged massive pulmonary hemorrhage to be the principal cause of death in about 9% of neonatal autopsies. TREATMENT OF PULMONARY EDEMA Pulmonary hemorrhage usually occurs between the 2nd and 4th days of life in infants who are being treated with mechanical ventilation. It has been associated with a wide variety of predisposing factors, including prematurity, asphyxia, overwhelming sepsis, intrauterine growth retardation, massive aspiration, severe hypothermia, severe Rh hemolytic disease, congenital heart disease, and coagulopathies. It is often associated with central nervous system injury, such as asphyxia or intracranial hemorrhage. Cole et al (1973) studied a group of infants with pulmonary hemorrhage to determine the clinical circumstances under which the illness occurred as well as the hematocrit and protein composition of fluid obtained from lung effluent and arterial or venous blood. Their results indicated that the lung effluent was, in most cases, hemorrhagic edema fluid and not whole blood (i.e., as indicated by hematocrit values significantly lower than those of whole blood). In addition, they did not find that coagulation disorders initiated the condition but probably served to exacerbate it in some cases. They postulated that the important precipitating factor was acute left ventricular failure caused by asphyxia or other events that might increase the filtration pressure and so injure the capillary endothelium of the lung. Thus, pulmonary hemorrhage may be considered as the extreme form of high-permeability pulmonary edema. Pulmonary edema following central nervous system injury probably results from increased hydrostatic pressure and some increase in vascular permeability (Malik, 1985). With the massive sympathetic discharge that accompanies central nervous system injury, left atrial pressure increases, and pulmonary arteries and veins constrict. As a result, microvascular pressure increases dramatically and causes dramatic damage to the microvascular endothelium, increasing its permeability to proteins and red blood cells. In infants with overwhelming sepsis and endotoxin production, increased microvascular permeability is apparent in the pulmonary circulation as well, undoubtedly contributing to the massive pulmonary hemorrhage sometimes seen in this group of infants. Pulmonary hemorrhage also has Treatment of pulmonary edema is directed at relieving hypoxemia and lowering vascular pressures. Hypoxemia should be treated with the administration of oxygen and, if necessary, positive pressure ventilation. Positive end expiratory pressure frequently improves oxygenation in individuals with pulmonary edema by improving ventilation-perfusion matching within the lung. Available evidence suggests that positive pressure ventilation does not reduce the rate of transvascular fluid filtration in the lung (Woolverton et al, 1978). Optimal treatment of pulmonary edema requires correction of the underlying cause. In infants with patent ductus arteriosus (see Figure 44-12) or other heart disease amenable to surgery, this is often easily accomplished. In cases of permeability edema or edema from nonsurgical heart defects, correction of the underlying cause may not be possible. In these instances, the only remaining option is to lower vascular pressures (even in permeability edema, lowering vascular pressures lowers the rate of fluid filtration and may also improve lymphatic function). This can be accomplished by lowering circulating blood volume by use of diuretics and fluid restriction, by improving myocardial function with the use of digitalis or other inotropic agents, or in severe cases by using a systemic vasodilator to reduce afterload and lower vascular pressures directly. More recent data suggest that clearance of fluid from the alveolar space may be accelerated by β-adrenergic agents (Frank et al, 2000) and by dopamine (Saldias et al, 1999). Whether these agents will have any clinical efficacy remains to be determined. PULMONARY HEMORRHAGE Landing (1957) described pulmonary hemorrhage in 68% of lungs of 125 consecutive infants who died in the 1st week of life; massive pulmonary hemorrhage was found in 17.8% of neonatal autopsies at the Johns Hopkins Hospital FIGURE 44-12 A, Preterm infant with a large patent ductus arteriosus and pulmonary edema. B, The same infant 24 hours after the ductus arteriosus was closed by surgical ligation. A Etiology and Pathogenesis B CHAPTER 44 Pulmonary Physiology of the Newborn been described occasionally in the presence of a large patent ductus arteriosus, with a left-to-right shunt that results in high flow and high pressure injurious to the vascular bed. Pulmonary hemorrhage is also associated with surfactant replacement therapy. Presumably the hemorrhage results from the rapid increase in pulmonary blood flow that accompanies improved lung function after surfactant therapy. The contribution of the patent ductus arteriosus to this increased blood flow remains to be determined. A meta-analysis suggests that surfactant replacement may be associated with an increased risk of pulmonary hemorrhage. This risk, however, is still extremely small compared to the known benefits of surfactant replacement (Pappin et al, 1994; Raju and Langenberg, 1993). Diagnosis Infants with any of the conditions mentioned previously should be observed carefully for possible pulmonary hemorrhage. Particular note should be made of any occurrence of blood-stained fluid from endotracheal tube aspirates, especially if repeated suctioning shows an increase in the amount of hemorrhagic fluid. The infant’s chest radiograph may show the fluffy appearance of pulmonary edema in addition to the underlying pathology, and the infant may have increased respiratory distress. Frank pulmonary hemorrhage, when it occurs, is an acute emergency, and the fluid has the appearance of fresh blood being pumped directly from the vascular system, although hematocrit values of the fluid are at least 15 to 20 points lower than the hematocrit of circulating blood, in keeping with hemorrhagic pulmonary edema. Treatment Effective treatment of pulmonary hemorrhage requires (1) clearing the airway of blood to allow ventilation; (2) use of adequate mean airway pressure, particularly end expiratory pressure; (3) resisting the temptation to administer 611 large volumes of blood because in most cases the infant has not had a large loss of volume, and thus, administration of excessive volume exacerbates the increase in left atrial pressure and hemorrhagic pulmonary edema; rather, red cell replacement should be done as a slow administration of packed cells after the infant’s pulmonary status has been stabilized; and (4) evaluation of the possibility of coagulopathy and administration of vitamin K and platelets, if appropriate. SUGGESTED READINGS Choukroun ML, Tayara N, Fayon M, Demarquez JL: Early respiratory system mechanics and the prediction of chronic lung disease in ventilated preterm neonates requiring surfactant treatment, Biol Neonate 83:30-35, 2003. Cook CD, Cherry RB, O’Brien D, et al: Studies of respiratory physiology in the newborn infant: I. Observations on normal premature and full term infants, J Clin Invest 34:975-982, 1955. Corbet AJ, Ross JA, Beaudry PH, et al: Ventilation-perfusion relationships as assessed by aADN2 in hyaline membrane disease, J Appl Physiol 36:74-81, 1974. Frappell PB, MacFarlane PM: Development of mechanics and pulmonary reflexes, Respir Physiol Neurobiol 149:143-154, 2005. Gappa M, Pillow JJ, Allen J, et al: Lung function tests in neonates and infants with chronic lung disease: lung and chest-wall mechanics, Pediatr Pulmonol 41:291317, 2006. Greenspan JS, Miller TL, Shaffer TH: The neonatal respiratory pump: a developmental challenge with physiologic limitations, Neonatal Netw 24:15-22, 2005. Hülskamp G, Pillow JJ, Stocks J: Lung function testing in acute neonatal respiratory disorders and chronic lung disease of infancy: a review series, Pediatr Pulmonol 40:467-470, 2005. Katier N, Uiterwaal CS, de Jong BM, et al: Passive respiratory mechanics measured during natural sleep in healthy term neonates and infants up to 8 weeks of life, Pediatr Pulmonol 41:1058-1064, 2006. Lesouef PN, England SJ, Bryan AC: Passive respiratory mechanics in newborns and children, Am Rev Respir Dis 129:552-556, 1984. McCann EM, Goldman SL, Brady JP: Pulmonary function in the sick newborn infant, Pediatr Res 21:313-325, 1987. McIlroy MB, Tomlinson ES: The mechanics of breathing in newly born babies, Thorax 10:58-61, 1955. Qureshi M, Khalil M, Kwiatkowski K, Alvaro RE: Morphology of sighs and their role in the control of breathing in preterm infants, term infants and adults, Neonatology 96:43-49, 2009. Complete references and supplemental color images used in this text can be found online at www.expertconsult.com C H A P T E R 45 Principles of Respiratory Monitoring and Therapy Eduardo Bancalari and Nelson Claure RESPIRATORY MONITORING One of the most important aspects of newborn intensive care is monitoring of cardiorespiratory function, ventilation, and oxygenation. This can be accomplished by bedside monitoring devices and laboratory analysis. The information provided by these monitoring techniques is an essential tool in the diagnosis and treatment of respiratory problems in the newborn. BREATHING FREQUENCY, APNEA, AND HEART RATE MONITORING Transthoracic impedance is the standard method used to monitor neonatal respiration. This technique is based on changes in the thorax’s electrical impedance caused by changes in gas volume during respiration measured by surface electrodes. Transthoracic impedance is mainly used to monitor breathing frequency and to detect apnea (Olsson and Victorin, 1970). Breathing is detected when the change in impedance exceeds a set threshold. A low threshold decreases sensitivity for apnea, and small disturbances such as cardiogenic oscillations can be falsely considered as breathing. Conversely, a high threshold may lead to false apnea alarms during shallow breathing. Monitors can automatically adjust this threshold or the impedance amplitude, or use sophisticated methods for detecting breathing or apnea. Transthoracic impedance is not reliable in assessing absolute tidal volume, but it can be used to assess relative changes. This technique is more effective in detecting apneas of central origin than obstructive apnea because the latter can produce internal displacement of gas volumes that still change impedance. More specific techniques are used for diagnosis and classification of apnea, including thermistors and CO2 monitors to detect gas exhalation and respiratory inductance plethysmography that measures expansion of the chest and abdomen. Transthoracic impedance monitors can also measure heart rate and are used to detect brady- and tachycardia. More sophisticated devices or built-in algorithms are utilized to detect cardiac rhythm anomalies. VENTILATION MONITORING The basic clinical determination of the adequacy of ventilation in the mechanically ventilated neonate consists of assessment of chest expansion, breathing frequency and auscultation. This can be complemented by monitoring tidal volume (VT) and minute ventilation utilizing flow sensors available in most neonatal ventilators. This allows for monitoring of spontaneous breathing, assessment of 612 respiratory system mechanics and detection of excessive or insufficient VT, hypoventilation, and gas trapping (Becker and Donn, 2007). These flow sensors are either mainstream (connected between the endotracheal tube and the ventilator circuit) or built into the ventilator. In small infants, mainstream flow sensors have better accuracy than those built into the ventilator because VT and flow are only a fraction of the gas volumes compressed in the circuit and the circulating bias flow (Cannon et al, 2000; Chow et al, 2002). Although mainstream flow sensors are usually small and typically have a dead space volume <1 mL, they can induce rebreathing of exhaled gases and affect CO2 elimination in small preterm infants (Claure et al, 2003; Figueras et al, 1997). BLOOD GAS MONITORING Determination of arterial blood gas status is key to the management of respiratory failure and lung disease in the neonate. Measurements of oxygen (Pao2) and carbon dioxide (Paco2) tension in arterial blood are considered the reference standard by which the efficacy and adequacy of ventilation and oxygenation are assessed. In order to obtain repeated arterial blood samples, placement of invasive catheters is required. In neonates, umbilical artery catheters (UACs) are commonly used during the acute phases of respiratory failure. Samples obtained from a UAC are the most accurate when proper sample handling and laboratory procedures are followed. However, there are important issues that must be considered before their placement and during their use to ensure that the risk-to-benefit ratio remains favorable. UAC have risks during placement and use, including perforation of the vessels, formation of thrombi and emboli, vasospasm and infection. UAC lines with the tip above the celiac plexus are associated with fewer complications than those below the renal or mesenteric artery (Barrington, 2000). Alternatively, percutaneous lines in the radial, ulnar, or dorsalis pedis artery are used when placement of a UAC is not possible. When an invasive line is not available, arterial punctures to the radial, ulnar, temporal, posterior tibialis, or dorsalis pedis artery are done to obtain blood samples. These latter results should be interpreted carefully because the procedure frequently disturbs the infant and alters blood gases. An alternative is to obtain blood samples from the peripheral capillary bed of the medial or lateral plantar surface. These are obtained after warming of the area, which produces hyperemia or “arterialization.” These samples provide a gross estimate of arterial blood gases, and they should also be interpreted cautiously (Courtney et al, 1990). Erroneous estimation usually occurs due to CHAPTER 45 Principles of Respiratory Monitoring and Therapy contamination by venous blood or air. This procedure can also disturb the infant and alter the basal status. Errors in blood gas measurement are often technique related. Contamination by an air bubble can reduce Pco2 and alter the Po2 in the direction of the air’s oxygen tension. Contamination by fluids can reduce Po2 and Pco2 while also affecting pH. Because blood cell metabolism continues, reducing the time from sampling to analysis and cold temperature storage and transport attenuate its effects. During conditions such as transport or when the turnaround time is important, bedside portable blood analysis devices have proven to be most effective (Murthy et al, 1997). Another alternative measurement method is that of indwelling electrodes inserted through a UAC that provide a continuous stream of blood gas information (RaisBahrami et al, 2002). TRANSCUTANEOUS BLOOD GAS MONITORING Transcutaneous measurement of O2 (TcPo2) and CO2 (TcPco2) tensions provide a noninvasive estimate of Pao2 and Paco2, respectively. Local hyperperfusion of the skin induced by heating creates a skin-electrode unit under the transcutaneous sensor, and electrochemical measurements in an electrolyte solution within this unit determine the partial pressure of O2 and CO2. During measurement of TcPo2, oxygen diffusing from the capillaries to the skin is reduced by the electrode, and the resulting electrical current is proportional to the Po2 of the capillary bed. Because skin metabolism continues, local hyperemia is needed to maintain skin perfusion sufficiently high that the measured TcPo2 is not affected by the skin’s O2 consumption. For this reason, accuracy of TcPo2 depends on electrode temperature with improved accuracy at or above 43° C (Huch et al, 1976). Under adequate conditions, TcPo2 correlates with Pao2 (Peabody et al, 1978). However, conditions such as arterial hypotension and acidosis often result in underestimation due to insufficient skin perfusion (Versmold et al, 1979). During TcPco2 measurement, CO2 molecules diffusing from the capillary bed change the pH of the electrolyte solution in the skin-electrode unit. This changes the electric potential between a reference and the electrode that is then translated into TcPco2. Metabolism in the skin produces CO2, and if the local perfusion is not adequate, the measured values can exceed those of Paco2. TcPco2 measurements can also be affected by conditions affecting peripheral perfusion (Peabody and Emery, 1985) and tend to overestimate Paco2 in hypercapnia as local perfusion decreases (Martin et al, 1988). Recent reports have demonstrated improved TcPco2 measurement accuracy in preterm infants of <29 weeks’ gestation, but inconsistent precision due to large betweenpatient variability (Aliwalas et al 2005; Bernet-Buettiker et al, 2005; Tingay et al, 2005). TcPco2 measurements have also been shown to be tightly correlated to capillary Pco2. Although capillary blood gases may not be the optimal reference, these are often the only available method for long-term monitoring because indwelling lines are available only during the acute phase of respiratory failure. In preterm infants, TcPco2 may reduce the need for 613 blood sampling and the number of painful punctures, but the main benefit is the ability to monitor continuously. For this reason, TcPco2 is commonly used as an adjunct to standard blood gas sampling to provide information on trends and respiratory stability. This is particularly useful in the management of invasive ventilatory support where close tracking of the effects of ventilator changes is required. The thin epidermal skin layer of the neonate has a relatively low metabolism. Nonetheless, some transcutaneous monitors include a metabolic correction factor. Measurements of TcPo2 and TcPco2 require a period of stabilization after sensor application until skin perfusion increases. A tight seal around the skin-electrode unit is required for accuracy. Similar to blood gas sampling, an air bubble transiently lowers TcPco2 and shifts TcPo2 toward the partial pressure of O2 in room air until the O2 is reduced. The transcutaneous electrode is usually applied on the thorax or the thigh. The need for high electrode temperature combined with the premature infant’s skin sensitivity requires frequent change of the application site to avoid thermal injury. Transcutaneous measurements have an intrinsic delay with respect to changes occurring in the arterial blood. END-TIDAL CARBON DIOXIDE MONITORING Monitoring of the partial pressure of CO2 in end-tidal gases (PetCo2) is based on the assumption that gases measured at the airway opening at the end of exhalation represent alveolar gases and that these match the arterial levels. PetCo2 is obtained by infrared sensors placed mainstream or by side-stream gas sampling. PetCo2 measurements are dependent on tidal volume size because the exhaled gas has to carry alveolar gas. The accuracy of PetCo2 is affected by lung disease with an increased arterial to alveolar CO2 gradient that results in underestimation of Paco2 (Sivan et al, 1992). For this reason, PetCo2 is more often used in term or near-term neonates and pediatric patients who require CO2 monitoring for reasons other than lung disease. ARTERIAL OXYGEN SATURATION BY PULSE OXIMETRY Estimation of the oxygen saturation in arterial blood (Sao2) by pulse oximetry (Spo2) is based on the differences in the rates of light absorption between oxygenated and deoxygenated or reduced hemoglobin (Hb) in the red and infrared regions of light. Deoxygenated Hb absorbs more red light and less infrared light than oxygenated Hb. As Sao2 increases, the ratio of the absorption of red light to that of infrared light decreases. It is assumed that in the circulation, changes in this ratio can only be produced by pulsating arterial blood. In neonates, Spo2 has been shown to correlate well with measured saturation in arterial samples (Hay et al, 1989). The absorption by pulsatile blood is only a small fraction of the light absorbed by tissue and venous blood. Thus, changes in pulse amplitude or patient movement that disrupt the optical pathway from transmitter to receiver side of the probe, or produce venous blood fluctuations, can affect Spo2 accuracy, although newer techniques have 614 PART X Respiratory System reduced the effect of the latter (Hay et al, 2002). The accuracy of Spo2 is also affected by conditions such as low perfusion or by inappropriate placement such as excessive tightening of the probe (Bucher et al, 1994). Data indicate reliable detection of hypoxemia spells by pulse oximetry (Bohnhorst et al, 2000; Hay et al, 2002). Nonetheless, some hypoxemia episodes detected by pulse oximetry are considered artifactual because of their temporal association with infant movement. However, increased infant activity leading to heart rate, lung volume, and ventilation changes has been shown to trigger hypoxemia (Bolivar et al, 1995; Dimaguila et al, 1997) with increased frequency during periods when the infants are awake compared to periods of active or quiet sleep (Lehtonen et al, 2002). OXYGEN THERAPY PRINCIPLES Neonates presenting with acute respiratory failure or with some degree of respiratory distress suffer abnormalities of gas exchange that almost invariably result in hypoxemia. Depending on the severity and duration of hypoxemia and the metabolic demands for oxygen, this can lead to reduced O2 availability and tissue hypoxia. Hypoxemia in the neonate can result from reduced alveolar oxygen content, low ventilation-perfusion ratio, reduced diffusion capacity, and extrapulmonary right-to-left shunts. The most common form of respiratory therapy for the neonate with hypoxemia consists of oxygen supplementation. The increased fraction of inspired oxygen (Fio2) increases the alveolar O2 tension (PAo2) in both welland partially ventilated areas of the lung. The resulting increase in the alveolar-arterial O2 gradient (A-aDo2) in part compensates for the conditions producing hypoxemia mentioned earlier. The proportion of neonates requiring supplemental O2 increases with lower gestational ages because premature birth is associated with many of the factors that contribute to hypoxemia. The primary goal of oxygen therapy is to maintain adequate O2 availability to the tissues, especially to the central nervous system and the heart, and to improve an incomplete hemodynamic adaptation to extrauterine life evident by persistently elevated pulmonary vascular resistance and patency of the ductus arteriosus. These goals, however, need to be attained without inducing the side effects of O2 toxicity on the eye, brain, and other organs that are common in the premature infant. Normal Sao2 for room-air-breathing term or healthy preterm infants is reportedly greater than 93%, with Pao2 levels above 70 mm Hg (Fenner et al, 1975; O’Brien et al, 2000; Richard et al, 1993). Maintenance of such oxygenation levels in premature infants with lung disease and immaturity would almost invariably require high inspired O2 levels. Earlier strategies to “normalize” oxygenation with use of high Fio2 in this population resulted in high rates of retinopathy of prematurity (ROP) and blindness (Campbell, 1951; Cross, 1973; Crosse and Evans, 1952). Conversely, strict curtailment of supplemental O2 regardless of the oxygenation level was associated with increased rates of neurologic damage and death (Avery, 1960; Bolton and Cross, 1974; Patz et al, 1952). The preterm infant is at risk for O2-induced injury because of an immature antioxidant system that is unable to balance the oxidative effects of O2 radicals. In the past, severe neonatal lung injury was only partly attributed to exposure to high Fio2. However, animal experiments have showed that lung damage was caused by high alveolar O2 independent of Pao2 (Miller et al, 1970; Northway et al, 1967; Taghizadeh and Reynolds, 1976). In preterm infants, hyperoxia has been linked to neurologic damage and impairment (Ahdab-Barmada et al, 1980; Collins et al, 2001; Haynes et al, 2003). For this reason, when supplemental O2 is administered to hypoxemic neonates, oxygenation is continuously monitored to avoid hyperoxemia. METHODS OF ADMINISTRATION In neonates, supplemental O2 is usually administered by means of a head box, mask, nasal cannula, nasal continuous positive airway pressure (CPAP), or a mechanical ventilator. In mechanically ventilated infants or in infants receiving nasal CPAP, supplemental O2 is administered by the mixture of air and O2 in the ventilator or CPAP device. In all four methods, verifying that correct mixing in the airO2 blender produces the desired Fio2 (by means of an O2 analyzer) is recommended. A head box, the least invasive of these methods, is generally used for infants who only need supplemental O2. Depending on the size of the infant, a minimum flow is required to flush exhaled gases and to minimize entrainment of ambient air. Gas warming and humidification is recommended to avoid drying of the airways and secretions as well as to avoid convective heat losses. Nasal cannulas deliver a constant flow of the air-O2 mixture to the nostrils. The actual Fio2 is determined by the delivered flow and the infant’s inspiratory flow. With increasing flows the actual inspired O2 concentration approaches the mixture delivered by the cannula, whereas higher inspiratory flows, in larger infants or during periods of increased demands, reduce the actual inspired O2 by entraining more room air (Walsh et al, 2005). The actual Fio2 during oral breathing has not been determined. Cannula flows >1 lpm can produce positive pressure at the nose at levels similar or above typical CPAP levels (Locke et al, 2003), which can become very high if the prongs fit tightly in the nostrils. Gas conditioning is recommended to avert drying of the nose and the risk of mucosal damage (Kopelman and Holbert, 2003). Nasal cannulas have gained popularity because they are flexible and facilitate access to and mobility of the infant. TREATMENT STRATEGIES The effects of the introduction of continuous monitoring of oxygenation in the care of preterm infants in relation to ROP have been inconsistent (Bancalari et al, 1987; Grylack, 1987; Yamanouchi et al, 1987). Data showing that ROP severity was associated with duration of hyperoxemia emphasized the importance of monitoring and curtailing hyperoxemia (Flynn et al, 1987). However, infants with severe ROP were also found to spend considerable periods of time in hypoxemia. Continuous monitoring of arterial oxygen saturation by pulse oximetry (Spo2) has become an important and CHAPTER 45 Principles of Respiratory Monitoring and Therapy standard component of neonatal intensive care. The use of Spo2 has recently been extended to the delivery room for titration of Fio2 during resuscitation (Dawson et al, 2009; Escrig et al, 2008). Although the optimal range of Spo2 in premature infants has not been fully defined, existing data suggest deleterious effects of hyperoxemia, with higher rates of ROP and worse respiratory course in neonatal centers that have tolerant policies toward high Spo2 levels (Anderson et al, 2004; Tin et al, 2001). Observational data indicate that more restrictive policies toward high Spo2 can reduce rates of ROP (Chow et al, 2003; Wright et al, 2006). Tolerance of high Spo2 levels in convalescent infants beyond the neonatal period has been proposed to improve growth and neurologic outcome, as well as to stop the progression of threshold ROP. However, clinical trials showed minimal benefits that were outweighed by deterioration in lung function caused by the additional O2 required to maintain higher Spo2 (Askie et al, 2003; STOP-ROP Study Group, 2000). Implementation of policies to curb hyperoxemia should also consider the potential deleterious effects of insufficient oxygenation. Hypoxemia can increase patency of the ductus arteriosus (Noori et al, 2009; Skinner et al, 1999) as well as increase the resistance of the pulmonary vasculature and airways, particularly in infants with established lung disease (Abman et al, 1985; Cassin et al, 1964; Halliday et al, 1980; Tay-Uyboco et al, 1989; Teague et al, 1988). Observational data showed that policies targeting lower Spo2 ranges were associated with better outcomes (Chow et al, 2003; Wright et al, 2006). However, it is unknown how closely such ranges were actually maintained and is possible that the side effects of lower Spo2 levels were not detected because infants were maintained above those ranges most of the time. Those findings should be interpreted with caution in face of recent evidence from a large blinded randomizedcontrolled trial showing increased mortality when targeting a lower Spo2 range (Carlo et al, 2010). The use of pulse oximetry to avoid hyperoxemia and hypoxemia must be done in the context of the sigmoidshaped oxygen dissociation curve between Pao2 and Spo2. Although Spo2 levels in the upper range can be associated with a wide range of Pao2 levels (Brockway and Hay, 1998), data indicate that a Spo2 threshold of 93% or 94% can avoid most Pao2 values >80 mm Hg (Bohnhorst et al, 2002; Castillo et al, 2008; Hay et al, 1989; Poets et al, 1993). On the other end, most Spo2 values <80% or 85% are associated with Pao2 <40 mm Hg. Thus, a moderate and stepwise increase in Fio2 when Spo2 decreases below this threshold can adequately correct hypoxemia and is unlikely to produce hyperoxemia unless the additional oxygen is excessive or prolonged unnecessarily. When Fio2 is increased in response to a hypoxemia spell, Spo2 should be continuously observed, and Fio2 should be weaned as the spell ends. Policies of oxygenation monitoring should clearly identify both intended range and the alarm limits of Spo2. Although in many centers they coincide, these ranges serve different purposes. The intended range usually defines a prescribed basal level of oxygenation to be maintained by the staff, whereas the alarm limits usually define specific conditions that require immediate intervention. In practice, the use of the low and high Spo2 alarms differs. The low Spo2 alarm, typically set between 85% and 88%, is generally used to detect acute hypoxemia. Setting 615 Spo2 alarms below these levels is often aimed at ensuring that the caregiver will respond only to the most severe events, thereby indirectly avoiding overuse of supplemental O2. The high Spo2 alarm is primarily used to avoid hyperoxemia and is typically set between 93% and 95%. The high Spo2 alarm level is quite important because it has been shown to be closely linked to the actual mean Spo2 observed in preterm infants receiving supplemental O2 (Hagadorn et al, 2006). Observational data have shown that preterm infants spend only half the time within the intended range of Spo2, with the remaining 30% of the time above and 20% of the time below this range (Hagadorn et al, 2006; Laptook et al, 2006). Setting Spo2 alarm limits within 2% of the intended range produces some improvement, but not strikingly so. Staff compliance to Spo2 alarms plays an important role, with the prescribed high Spo2 alarm level being frequently altered by the staff (Clucas et al, 2007). Insufficient staff education and communication can also influence the maintenance of Spo2 within the intended range. However, a recent survey showed that only a third of the caregivers were aware of and could identify their center’s oxygenation policies (Nghiem et al, 2008). Policies for oxygen supplementation for preterm infants who reach term-corrected postmenstrual ages (e.g., 36 weeks) vary significantly between centers (Ellsbury et al, 2002). Centers with policies that target higher Spo2 have greater proportions of infants on supplemental O2, and many of those infants could be off O2 if lower Spo2 levels are tolerated or if their actual needs are frequently evaluated (Walsh et al, 2004). This is also the case during the discharge period as well as during home oxygen therapy. Many preterm infants present with hypoxemia spells, and these are more frequent in infants with evolving chronic lung disease (Bolivar et al, 1995; Dimaguila et al, 1997; Garg et al, 1988). These spells often require a transient increase in Fio2, but a delayed response can prolong the hypoxemia spell, whereas a delayed weaning of Fio2 after hypoxemia ends can induce hyperoxemia. It is also evident that caregivers often tolerate or maintain high Spo2 levels with the purpose of reducing the frequency of the spells or attenuating their severity (Claure et al, 2009, 2011). However, this is not truly effective and may actually increase the risk of ROP (McColm et al, 2004). Neonatal center policies should clearly define the response of the caregiver to these events to minimize both excessive and inadequate oxygenation. Currently, automated systems to adjust the inspired oxygen for maintenance of Spo2 within a set range are being developed and tested (Claure et al, 2001, 2009, 2011; Urschitz et al, 2004). Reports indicate improvements in terms of maintenance of an oxygenation range as well as reductions in hyperoxemia, exposure to supplemental O2, and staff workload with these automated systems. Large trials are necessary to determine the effects of this form of Fio2 control on short- and long-term neonatal outcomes. At the present time, most clinical effort is focused on avoiding the extreme high and low ranges of Spo2. Forthcoming are the findings of multicenter trials being conducted to compare the effects of maintaining Spo2 within different ranges in terms of ophthalmic, neurologic, and respiratory outcome. The information obtained from these trials will further refine the oxygen management strategies. 616 PART X Respiratory System NONINVASIVE RESPIRATORY SUPPORT Because of the severe complications associated with invasive mechanical ventilation, there has been a persistent search for less invasive alternatives to support infants with respiratory failure. This is especially relevant in the small preterm infant, who is more susceptible to acute complications and chronic pulmonary sequelae. Alternatives have ranged from transdermal oxygenation to lung volume maintenance by sternal traction. The strategies that have shown to be most clinically effective are nasal continuous positive airway pressure and, more recently, nasal intermittent positive pressure ventilation. These strategies can be used to support infants with respiratory failure in lieu of invasive mechanical ventilation. NASAL CONTINUOUS POSITIVE AIRWAY PRESSURE CPAP was introduced in 1971 by George Gregory to stabilize lung volume in preterm infants with respiratory distress syndrome (RDS) (Gregory et al, 1971). Initially, the constant pressure was delivered through a mask or head box and later was applied through nasal prongs. With the introduction of surfactant therapy, more infants were intubated and treated with mechanical ventilation, and the use of nasal CPAP declined for a period of several years; in the past decade, however, it has been reintroduced as a safer alternative to mechanical ventilation. Physiologic Effects Nasal CPAP (N-CPAP) prevents alveolar collapse and stabilizes functional residual capacity, thereby reducing pulmonary shunts and resulting in improved oxygenation. N-CPAP also stabilizes large and small airways and can decrease the work of breathing and obstructive apnea episodes. By preventing alveolar closure and reopening with each breath, N-CPAP can also preserve surfactant. Indications Nasal CPAP is used primarily in infants with RDS who because of surfactant deficiency have decreased functional residual capacity (FRC). N-CPAP is also effective in the transition period immediately after birth while reabsorption of fetal lung fluid and establishment of FRC is occurring. N-CPAP is also effective in reducing apneic episodes and in stabilizing respiratory function after extubation (Locke et al, 1991; Martin et al, 1977; Miller et al, 1985). Nasal CPAP is also used in infants with increased pulmonary blood flow and pulmonary edema secondary to heart lesions with left-to-right shunting such as patent ductus arteriosus (PDA). Nasal CPAP can also be used effectively in infants with airway obstruction. Devices for Nasal CPAP Nasal CPAP has been applied using a variety of systems and devices. Most of the evidence suggests that best results are obtained using double short nasal prongs (De Paoli et al, 2002). The devices used to generate the pressure can be a water column such as the one used for “bubble CPAP” or an adjustable valve at the end of a continuous-flow system or a variable-flow system. Although there are data suggesting that some of these systems are superior to others, in clinical practice the stability and permeability of the nasal interface is probably the most important factor determining the success or failure of this form of respiratory support. Recently, nasal cannulas have been used as a method for generating nasal CPAP. The limitation of this method is that it is difficult to control and measure the actual pressures and inspired oxygen concentrations that are delivered to the infant because both depend on the gas flow and the leaks around the cannulas and through the mouth. Clinical Application As mentioned earlier, the major indication for N-CPAP use in preterm infants is RDS. There is considerable debate as to the ideal time to start N-CPAP and about the population of infants in which it should be used. In recent years most centers have used N-CPAP soon after birth in infants who have sufficient respiratory drive. The exceptions are infants who are depressed at birth or those who have severe respiratory failure and are intubated to administer surfactant. The success rate with this strategy depends on the maturity of the infant and the severity of the respiratory failure, but more than 50% of preterm infants who are not depressed at birth can be managed successfully with N-CPAP, avoiding the use of invasive ventilation (Ammari et al, 2005; Dani et al, 2004; Kamper and Ringsted, 1990; Reininger et al, 2005). The sooner N-CPAP is started, the better the results are in infants with RDS (Gittermann et al, 1997; Hegyi and Hiatt, 1981; Jonsson et al, 1997; Verder et al, 1999). It has been suggested that the early use of N-CPAP instead of invasive ventilation could lead to a reduction in the incidence of bronchopulmonary dysplasia (BPD). However, two large prospective randomized trials addressing this question did not show that early use of N-CPAP reduces BPD (Finer et al, 2010; Morley et al, 2008). The other frequent indication for N-CPAP is to stabilize respiratory functions after weaning from mechanical ventilation and extubation (Andreasson et al, 1988; Engelke et al, 1982; Higgins et al, 1991). Although it is not clear whether the use of N-CPAP in this situation reduces the need for reintubation, it clearly prevents a deterioration of respiratory function (Davis and Henderson-Smart, 1999; Peake et al, 2005). The other indication for N-CPAP is to reduce the incidence of apneic episodes in preterm infants, in which it has been shown to be very effective (Martin et al, 1977; Miller et al, 1985). The levels of N-CPAP that are used in clinical practice range from 3 to 8 cm H2O, but there are few data to justify a specific level. In practice, the level of N-CPAP used is determined by the severity of the lung disease, reflected by the inspired oxygen concentration and the degree of lung expansion in the chest radiograph. The more severe the disease is, the higher the level of N-CPAP that is used. As the respiratory condition improves and the oxygen requirement decreases the level of N-CPAP is reduced to 3 to 4 cm H2O before removing the prongs. CHAPTER 45 Principles of Respiratory Monitoring and Therapy Complications 617 The complications of N-CPAP can be related to the pressure that is applied or to the interface with the nose. The application of excessive pressure can cause overdistention and alveolar rupture with pulmonary interstitial emphysema and pneumothorax. It can also reduce venous return, increase pulmonary vascular resistance, and reduce cardiac output. Overdistention of the lung can also reduce compliance and induce hypoventilation. When the nasal prongs are too large or are applied with too much pressure over the nasal septum, they can produce erosions or pressure necrosis that sometimes makes it impossible to continue the N-CPAP. Avoiding these complications and keeping the nasal prongs in place is a task that requires considerable time and skill. During N-CPAP there is a risk of gas being pushed into the stomach. A nasogastric catheter is often used to avoid accumulation of gas and gastric distention. A/C or PSV. Internal flow sensors are available in many new ventilators. However, data on their use in neonatal noninvasive ventilation are lacking, and it is unlikely that small infants are able to produce flows large enough to be detected by the internal sensors. Devices that alternate between two levels of positive airway pressure (Bi-PAP) are also available. In these devices the increase in pressure is achieved by an increase in circulating flow instead of a valve. For this reason, in Bi-PAP the increase in pressure is small compared to that in conventional ventilators, and it is achieved over a relatively long time. Synchronization of the increase in airway pressure with the infant’s inspiration using the Graseby capsule is available in some countries. Another approach has been that of applying high-frequency ventilation nasally, but it is unclear what proportion of the oscillating pressure is transmitted to the infant’s airways. NONINVASIVE VENTILATION Clinical Application Noninvasive forms of respiratory support have been introduced with the aim of avoiding the use of invasive ventilation in infants in respiratory failure. N-CPAP is effective in improving gas exchange by stabilizing lung volume and the airways. However, it is not fully effective in maintaining ventilation in infants with weak inspiratory effort and inconsistent respiratory drive. Intermittent positive pressure ventilation via nasal devices was among the original forms of support used in preterm infants in respiratory failure (Llewellyn et al, 1970). Recently, it has been reintroduced in neonatal care for indications including respiratory distress and apnea and to facilitate weaning from invasive mechanical ventilation. In infants with apnea, the cycling positive pressure at the upper airway may produce an intermittent stimulus that prevents or attenuates the duration of breathing pauses of central origin. In infants with central apnea, resumption of breathing following ventilator cycles has been described, and the reduction of apnea in comparison to N-CPAP is more striking among infants with more frequent apnea spells (Bisceglia et al, 2007; Lin et al, 1998; Ryan et al, 1989). Noninvasive ventilation can improve gas exchange and ventilation compared to N-CPAP in infants with respiratory insufficiency during the first hours after birth as well as during weaning from mechanical ventilation (Bisceglia et al, 2007; Moretti et al, 1999). In contrast, infants with mild respiratory distress who are stable on N-CPAP and are able to maintain adequate gas exchange only reduced their spontaneous breathing effort when receiving noninvasive ventilation (Aghai et al, 2006; Ali et al, 2007; Chang et al, 2011). In the preterm infant, the chest wall is excessively compliant, and the breathing effort is in part dissipated by an inward motion of the chest. Hence, the reduced breathing effort with noninvasive ventilation can also be explained by a reduction in chest distortion (Ali et al, 2007; Kiciman et al, 1998). In spite of applying cycles of relatively small positive pressure, Bi-PAP can improve CO2 elimination and oxygenation (Migliori et al, 2005) compared to N-CPAP. Noninvasive application of high-frequency ventilation was also shown to be useful as a rescue for avoidance of intubation in a group of infants failing N-CPAP (van der Hoeven et al, 1998). In addition to the effects on apnea, ventilation, and breathing effort, it is possible that the increase in mean airway pressure during noninvasive ventilation can better maintain lung volume than N-CPAP alone. It is also possible that CO2 removal is improved during noninvasive ventilation by clearance of exhaled gases from the upper airway. The use of noninvasive ventilation to achieve a reduction in the need for intubation or the duration of invasive mechanical ventilation is aimed at reducing the associated Devices for Noninvasive Ventilation The devices used for neonatal noninvasive positive pressure ventilation consist of conventional time-cycled pressure-limited ventilators that utilize the same circuits and gas conditioning devices used for invasive ventilation, with the main difference being that in lieu of the endotracheal tube, positive pressure is applied through the same interfaces used for N-CPAP. Noninvasive positive pressure can be delivered in the intermittent mandatory ventilation mode (N-IMV) where the positive pressure cycles are delivered at fixed intervals. In some ventilators, ventilator cycles can be synchronized to the neonate’s inspiration to provide nasal synchronized IMV (N-SIMV). Synchronized noninvasive ventilation can also be provided in the assist/control (N-A/C) and pressure support (N-PSV) modes to assist every spontaneous inspiration. The most commonly reported method for synchronization in noninvasive ventilation is the Graseby pressure capsule placed on the abdomen (Barrington et al, 2001; Bhandari et al, 2007; Friedlich et al, 1999; Khalaf et al, 2001). Mainstream flow sensors have been used for synchronization in noninvasive ventilation, but they are likely to require frequent sensitivity adjustments to avoid autocycling due to variable gas leaks around the interface and the mouth. This is particularly important during noninvasive 618 PART X Respiratory System risks of lung injury. This is obviously more relevant in the smaller and more immature infants in whom N-CPAP more frequently fails. In infants with RDS, nasally delivered IMV has not consistently reduced the rate of intubation compared to N-CPAP (Kugelman et al, 2007; Meneses et al, 2011), but it has been shown to facilitate early extubation after surfactant administration (Bhandari et al, 2007). In some of these trials, noninvasive ventilation reduced BPD appreciably, but this was not a consistent finding. The possible beneficial effects of nasal ventilation during the initial respiratory failure still need to be confirmed. During weaning from mechanical ventilation, adequate maintenance of lung volume is often achieved by N-CPAP. However, many infants fail because of insufficient ventilation resulting from central apnea, a weak respiratory pump, or poor lung mechanics due to the underlying lung disease. Noninvasive ventilation has consistently been shown to be an effective way to reduce extubation failure (Barrington et al, 2001; Friedlich et al, 1999; Khalaf et al, 2001; Moretti et al, 2008), mainly by reducing apnea and improving gas exchange. Smaller infants and those with poor lung function at extubation were more likely to benefit from nasal ventilation than larger infants or infants with better lung function (Khalaf et al, 2001). Ventilator cycles delivered at fixed intervals can fall toward the end of spontaneous inspiration or during exhalation and disturb the infant’s breathing pattern, whereas delivery of the ventilator cycle when the upper airway is patent may improve transmission of the pressure and reduce the risk of gas being pushed into the esophagus. Both nonsynchronized and synchronized modes of nasal ventilation have been shown to be more effective than N-CPAP, but data are lacking on the superiority of synchronization in terms of efficacy or safety. Data on the most effective mode, frequency, and duration of the cycle and, most importantly, peak pressures during noninvasive ventilation are also lacking. Until more data are available, a conservative approach with relatively low ventilator settings and in ranges near those used for intubated infants recovering from RDS or near extubation is recommended. This is particularly important in setting peak pressures, because VT monitoring is not available during noninvasive ventilation. Potential Drawbacks The risks for gastrointestinal complications observed with N-CPAP may be greater during noninvasive ventilation because of the additional positive pressure. An earlier report indicated an increased rate of gastrointestinal complications (Garland et al, 1985), but more recent clinical trials have not confirmed this. Because of the additional pressures, the risks for pneumothorax and pulmonary interstitial emphysema may also be increased. Although there are no data on these side effects, caution should be exercised and high peak pressures or ventilator rates should be avoided. The risks for nasal damage and obstruction often observed during N-CPAP are also present in noninvasive ventilation. Proper application and maintenance of the nasal interface and avoidance of excessive force on the nasal septum are important to avoid these complications. INVASIVE MECHANICAL VENTILATION Because of the high incidence of respiratory failure in the neonate, mechanical ventilation has been one of the main therapies responsible for the progress in neonatal critical care. This is especially relevant in the small preterm infant who, besides lung immaturity, has a soft chest wall and poor central respiratory drive, making the need for mechanical ventilation very common. INDICATIONS The decision to initiate invasive mechanical ventilation in the newborn is very important because of the serious complications associated with this mode of therapy and because in smaller infants, it is often difficult to wean them from respiratory support. There is considerable variation between different centers in the criteria used to initiate mechanical ventilation. Most often, this decision is based on the gestational age of the infant, the severity of the respiratory failure, and the disease process that is underlying the respiratory failure. This is also done considering the alternatives available to support the infant’s respiratory function. The experience of the team and the outcomes of infants exposed to mechanical ventilation in each institution should also be an important consideration. In units with vast experience and good outcomes, ventilation may be used more liberally, whereas in units with limited experience and high rates of complications, other alternatives should be considered before embarking on invasive ventilation. The initiation of mechanical ventilation is usually based on the clinical condition of the infant and the evaluation of arterial blood gases. In the preterm infant, mechanical ventilation is frequently started because of recurrent episodes of apnea and hypoxemia that require some intervention to recover. In more immature infants, ventilation is often begun in the delivery room because of severe respiratory depression and bradycardia not responsive to stimulation. The other common indication for mechanical ventilation is when levels of Paco2 rise rapidly, indicating alveolar hypoventilation. Although there are no specific levels of Paco2, most centers initiate mechanical ventilation when Paco2 rises acutely above 55 to 65 mm Hg and the pH decreases below 7.25 to 7.20. The introduction of positive pressure ventilation is associated with complications and seldom results in improved lung function. In fact, intermittent positive pressure ventilation (IPPV) frequently results in further deterioration in pulmonary function due to the negative effects of high inspired oxygen concentrations, ventilation-associated infections, and overdistention of the lung. The effectiveness of mechanical ventilation is primarily due to the support of the infant’s failing respiratory pump and the reduction in the work of breathing. One exception to this is the infant with RDS in whom the positive airway pressure produces recruitment of distal air spaces with improvement in ventilation-perfusion matching and gas exchange. In other instances, mechanical ventilation is started because of hypoxemia not responsive to continuous positive airway pressure. Although there are no set levels of Pao2 or Fio2 to start ventilation in preterm infants with CHAPTER 45 Principles of Respiratory Monitoring and Therapy RDS, the initiation of mechanical ventilation is frequently associated with the decision to administer exogenous surfactant. The indications for surfactant vary between institutions, but there is good evidence that early administration of surfactant results in better outcomes, and therefore, in an infant with RDS, surfactant is usually given when the inspired oxygen concentration required to maintain acceptable O2 saturation levels increases above 30% to 40%. If the infant has hypercapnia or clinical signs of significant distress and impending failure, ventilation may be started earlier. In many centers surfactant is given as prophylaxis to all infants below certain gestational age, usually 26 to 28 weeks. After the infant becomes more stable, he or she is extubated to nasal CPAP. In full-term infants, the indication for mechanical ventilation is usually more conservative because these infants are better able to cope with increased work of breathing. The indication also varies depending on the underlying cause of the respiratory failure. For example, in an infant with respiratory failure due to a congenital diaphragmatic hernia, ventilation is usually started immediately after birth, whereas in an infant with a congenital pneumonia or meconium aspiration, a more conservative approach can be taken, and ventilation may not be started until there is evidence of rising Paco2 and hypoxemia requiring inspired oxygen concentrations up to 40% to 60%. The clinical evaluation of the infant is critical in deciding whether invasive ventilation should be started. INSPIRED GAS CONDITIONING During normal breathing, inspired gas is heated and humidified in the nasal passages and airways. By the time it reaches the distal airways, it is fully saturated with water vapor at core body temperature, that is, 100% relative humidity at 37° C for an absolute humidity of 44 mg H2O per liter of gas. The isothermal saturation point region in the respiratory tract, where inspired gas equilibrates with core body temperature and is fully saturated, is near the main bronchi. As inspiratory flow increases, this point moves distally into the airways. The nose and airways function as heat and moisture exchangers. There, a portion of heat and water added to the inspired gas is recovered during exhalation, with the net loss depending on the temperature and relative humidity of the gas. This requires continuous replenishment of water to the aqueous mucosal layer by the airway epithelium and loss of heat. In contrast to ambient air at 22° C and 50% saturated (absolute humidity of 9. 7 mg H2O per liter of gas) medical gases are typically colder (15° C or less) and dry (<2% relative humidity) and therefore require more heat energy and humidity. Hence, all forms of respiratory support where medical gases are used require conditioning of the inspired gas. This is a key component of mechanical ventilation because the endotracheal tube (ETT) bypasses the nose and upper airway. Inadequate conditioning of the inspired gas can increase water and heat loss (Fonkalsrud et al, 1980). Exposure to dry and cold inspired gas can also produce inflammation of the airway epithelium and increase the risk of airway damage (Marfatia et al, 1975; Todd et al, 1991). Insufficient humidification can also affect the mucociliary transport 619 system, reducing clearance of secretions, pathogens, and foreign particles as well as increase the risk of airway blockage by mucus plugs (Fonkalsrud et al, 1975). These effects are likely to be more striking in small infants, infants with impaired thermoregulation, and those who are fluid and energy limited. In small preterm infants, a few minutes of mechanical ventilation with inadequately conditioned gases can increase airway resistance and reduce lung compliance (Greenspan et al, 1991). Inadequate gas conditioning has also been associated with increased risk for air leaks and augmented need for O2 (Tarnow-Mordi et al, 1989). The standard method for conditioning of the inspired gases consists of a heater/humidifier (H/H) device and heated breathing circuits. Dry and cold medical gases are heated in the H/H chamber to 37° C to increase the water carrying capacity to 44 mg per liter of gas at 100% relative humidity. The gas travels through the ventilator circuit, where it is heated to 39° C to prevent condensation. The gas temperature decreases as it travels through the ETT, and gas is delivered at approximately 37° C and near 100% saturated at the distal end. Many conditions can affect gas conditioning. Air temperature in an incubator or radiant warmer above the 39° C set point can inadvertently reduce heating at the ventilator circuit and result in condensation. To avoid this, the temperature probe can be insulated or place outside the incubator or radiant warmer. If gas heating at the ventilator circuit is not adequate, low ambient temperatures and low circulating gas flows can also produce condensation. In general, presence of water condensate in ventilator circuit and minimal consumption of the humidifier water indicate inadequate conditioning of the gas. Water condensation can occur in the ETT and connector when incubator temperature is low and result in water droplets being pushed into the airway. Humidity of the inspired gas can also be increased by water nebulization. In contrast to the size of the water vapor molecules (0.0001 μm), the size of aerosolized water particles ranges between 0.5 and 5 μm, which can potentially transport virus or bacteria. Thus, water nebulization is not an efficient or safe method, particularly for prolonged use. Gas conditioning is also recommended during nasal CPAP or nasal ventilation and oxygen head box or nasal cannula use because the nasal passages and airways may not achieve adequate conditioning of the cold and dry medical gases. These gases are typically heated to 32° C, but heating at or above room temperature may be sufficient. On the other hand, passing the gas through a water bath may not produce a sufficient gain in humidity because of the low water carrying capacity of cold gas. Thick and dry nasal mucus and airway secretions are observed when gas conditioning is insufficient. Lack of conditioning may also affect body temperature and water losses. CONVENTIONAL POSITIVE PRESSURE VENTILATION Principles Neonatal positive pressure ventilation can be described in its basic form as the cycled application of two distinct levels of positive pressure at the infant’s airway via an ETT. 620 PART X Respiratory System 5 lpm 5 lpm Flow Flow 0 0 20 mL 20 mL VT VT 0 Low bias flow High bias flow 0 25 cm H2O Ti Te PAW 25 cm H2O PIP PEEP PAW 0 FIGURE 45-1 Time-cycled pressure-limited ventilation. In this example, recordings of flow, tidal volume ( VT), and airway pressure ( PAW) show that ventilator cycles, occurring at intervals determined by Te, increase PAW from the PEEP level to PIP during the set Ti. The increase in pressure with respect to the alveolar pressure drives gas into the lung to achieve VT. The inspiratory flow, which determines the rate of lung inflation, peaks initially and subsequently declines to zero as the lung is inflated. The positive end-expiratory pressure (PEEP) provides a continuous distending pressure to the lung and is aimed at maintaining lung volume. This is important because the ETT bypasses the upper airway mechanism that normally prevents lung volume loss by active closure of the glottis. The positive pressure applied at the airway opening is intermittently increased to a predetermined peak inspiratory pressure (PIP) during a set inspiratory time (Ti). The rise in airway pressure produces a gradient with respect to the alveolar pressure that drives a tidal volume (VT) of gas into the lung. This form of ventilation is known as timecycled pressure-limited ventilation. Neonatal ventilators utilize a constant flow of conditioned gas, also known as bias flow, through the breathing circuit to produce positive pressure. A controlled obstruction by a valve at the expiratory port of the ventilator produces PEEP, and the intermittent increase to PIP at intervals determined by the set expiratory time (Te) as shown in Figure 45-1. The circulating gas flow determines the profile of the airway pressure by modulating the rise to the set PIP. The airway pressure rise is faster and PIP is reached earlier at higher circulating flow rates. A fraction of this bias flow is driven into the infant’s airways during tidal inflation. The infant’s inspiratory flow, which indicates how fast is the lung being inflated, is in part determined by the profile of the airway pressure. The rapid rise in airway pressure produces a rapid increase in the infant’s inspiratory flow 0 FIGURE 45-2 Effect of circulating bias flow. Recordings of flow, tidal volume ( VT) and airway pressure ( PAW) show how increasing bias flow rates in the ventilator circuit change the profile of PAW with a more rapid rise toward the peak pressure. This produces a higher peak flow that indicates faster lung inflation. An insufficient bias flow rate does not permit generating the desired peak pressure with each ventilator cycle and results in a smaller VT. to a high peak inspiratory flow that decays as the lung is inflated. In this pattern of rapid lung inflation, most of the VT is delivered early in the inspiratory phase, as shown in Figure 45-2. In contrast, a slow rising airway pressure produces a slower inflation and a smaller peak inspiratory flow. A low bias flow may not be sufficient to reach the set PIP within a fixed Ti and in consequence produce a smaller VT, also shown in Figure 45-2. In most neonatal ventilators, the bias flow is constant during both the inspiratory and expiratory phases, whereas other ventilators self-adjust the flow necessary to produce a desired profile. In other ventilators, the caregiver can set a maximal flow to be delivered by the ventilator to the circuit during the cycle, which also modifies the airway pressure profile. The bias flow should be sufficient to sustain the PEEP level when the infant’s spontaneous respiratory flow demands increase. Otherwise, it may create an inspiratory load that the infant has to overcome with each breath. This is noted as a decline in PEEP during spontaneous inspiration toward zero pressure, or it can even become negative. The bias flow is also responsible for removal of exhaled gases, and an insufficient bias flow may cause rebreathing. On the other hand, a very high bias flow produces an almost square inspiratory airway pressure waveform that increases the velocity of lung inflation. This may have undesirable consequences because the lung will be expanded much faster than during a normal physiologic inflation. In general, circulating bias flow rates for ventilated preterm infants are set between 5 and 8 liters per minute, with higher circulating flow rate settings to accommodate gas leaks around the airway or the demands of larger neonates. CHAPTER 45 Principles of Respiratory Monitoring and Therapy 5 lpm Respiratory System Mechanics during Positive Pressure Ventilation The neonate’s respiratory system mechanical properties (i.e., compliance and airway resistance) determine its time constant as their product. The respiratory time constant, which is a measure of the time to achieve equilibrium between the applied pressure and the alveolar pressure, varies with different lung diseases and their severity. The compliance of the respiratory system (CRS) is a measure of the recoil pressure that opposes expansion volume and is determined by the compliance of lung (CL) and the chest wall. In the neonate, and most particularly in the preterm infant, CRS is mainly determined by the compliance of lung because the chest wall is highly distensible. Lung diseases characterized by lung restriction, such as respiratory distress syndrome, where the increased lung recoil is indicated by a low CRS and a short time constant, are quite prevalent in preterm infants. In addition, they have a respiratory pump that is often too weak to produce the pressure required to achieve an adequate VT, and thus they require mechanical ventilation. During positive pressure ventilation, a decreased CRS is characterized by a smaller VT for a given peak pressure and a relatively brief duration of inflation. As shown in Figure 45-3, the shorter time constant is illustrated by a shorter duration of inflation with an earlier return of the inspiratory flow to zero. This marks the point when the airway and alveolar pressures are at equilibrium. At this point, the PIP equals the lung’s recoil pressure, with CRS as the ratio between VT and the applied pressure. When CRS decreases, prolonging Ti is ineffective, and a higher PIP is required to maintain VT constant. The resistance of the airways opposes the flow of gas, which dissipates part of the driving pressure during lung inflation. In diseases characterized by increased airway resistance, this attenuates the infant’s inspiratory flow and results in a slower inflation rate. The longer time constant indicates the increased time required for alveolar pressure to rise and equilibrate with the applied pressure. When airway resistance increases, the set Ti may not be long enough to achieve VT, as shown in Figure 45-4. Although setting a longer Ti may be reasonable, this should be done cautiously because the increased airway resistance also attenuates the expiratory flow and prolongs the time required to achieve complete exhalation. A long Ti combined with a Te that does not allow complete exhalation will result in gas trapping. Modes of Conventional Ventilation Modes of conventional neonatal ventilation are generally classified according to the parameter controlled by the ventilator in each cycle as well as by the timing and duration of the cycle. Ventilator cycles can be pressure or volume controlled depending on whether the ventilator targets a set peak pressure or volume, respectively. Ventilator cycles can be delivered at fixed intervals regardless of the timing with respect to the infant’s spontaneous breathing or they can be delivered in synchrony with the spontaneous inspiration. The duration of the inspiratory phase of each ventilator cycle can be constant or it can adapt to the time 621 End of inflation Flow 0 20 mL Reduced CRS VT 0 CRS 0.66 mL/cm H2O CRS 0.33 mL/cm H2O 25 cm H2O PAW 0 FIGURE 45-3 Effect of lung compliance. In this example, recordings of flow, tidal volume ( VT ), and airway pressure ( PAW) show the effects of a decrease in lung compliance (CL ). A reduction in CL results in a proportional decrease in VT for the same driving PAW. The effect is also noted by a shorter time constant indicated by an earlier return of the inspiratory flow to zero marking the end of lung inflation. required to complete lung inflation, deliver the set volume, or to reach the end of spontaneous inspiration. Intermittent Mandatory Ventilation The time-cycled pressure-limited (TCPL) ventilation mode described earlier and illustrated in Figure 45-1 has been most commonly known as intermittent mandatory ventilation (IMV). This mode can be classified as pressure controlled because ventilator cycles deliver the PIP set by the caregiver. These cycles are of constant inspiratory duration, that is, Ti, and are delivered at fixed intervals, that is, Te is determined by the set ventilator frequency. For many years IMV has been one of the most commonly used modes of neonatal ventilation during the acute as well as the more chronic phases of neonatal respiratory disease. In IMV, the increase in airway pressure due to PIP in each ventilator cycle produces a given lung inflation during the set Ti. If Ti is too short, complete lung inflation may not be achieved, whereas a long Ti that maintains a pressure plateau does not achieve a larger VT and instead 622 PART X Respiratory System Increased resistance 5 lpm End of inflation 5 lpm End of inflation Flow Flow 0 0 End of exhalation 10 mL 20 mL Volume plateau VT VT 0 0 25 cm H2O Ti PAW 0 FIGURE 45-4 Effect of airway resistance. In this example, recordings of flow, tidal volume ( VT), and airway pressure ( PAW) show the effects of an increase in airway resistance. The increase in resistance results in a decrease in inspiratory flow that does not permit delivery of VT in the set Ti. The prolonged time constant is noted by the increased time required to achieve full inflation and to complete exhalation. produces an inspiratory hold while the lung is kept inflated as illustrated in Figure 45-5. In infants with respiratory failure Ti is usually set in the range of 0.25 and 0.4 second. Caution should be exercised when setting ventilator cycles of longer Ti or high ventilator frequencies that could result in gas trapping due to an insufficient Te to allow full exhalation as shown in Figure 45-6. This is particularly important when the infant’s time constant is long because of a high airway resistance or when the ventilator frequency results in Te <0. 5 seconds. During IMV, total minute ventilation results from the ventilation produced by the ventilator and the contribution of the infant’s spontaneous breathing effort. During clinical use, PIP is usually adjusted to maintain VT between 3 and 5 mL/kg of body weight which is considered adequate for infants with lung disease. Ventilator frequency is adjusted depending on the infant’s ability to contribute to minute ventilation and maintain sufficient gas exchange. PaCO2 levels between 40 and 50 mm Hg are considered adequate, but higher levels are often accepted in infants with chronic ventilator dependency. Although the management of IMV is relatively simple, there are some important limitations. With IMV, cycles are often delivered out of synchrony with the infant’s inspiration (Greenough et al, 1983a, 1983b). Asynchronous ventilator cycles delivered toward the end of the infant’s 25 cm H2O Long Ti Short Ti PAW 0 Pressure plateau FIGURE 45-5 Inspiratory duration. Recordings of flow, tidal volume ( VT), and airway pressure ( PAW) illustrate the effects of inspiratory time ( Ti). A Ti that prolongs the inspiratory phase beyond the time needed to achieve inflation does not increase VT and only keeps the lung expanded. A Ti of insufficient duration does not permit achieving full inflation, as indicated by an inspiratory flow that does not return to zero and leads to a decrease in VT. inspiration can prolong or increase lung inflation. IMV cycles delivered during the infant’s exhalation can prolong its duration and in some cases elicit active exhalation against the positive pressure. Reports suggest associations between asynchrony and the occurrence of air leaks and intraventricular bleeding (Greenough and Morley, 1984; Greenough et al, 1984; Perlman et al, 1983, 1985; Stark et al, 1979). In older IMV ventilators, VT cannot be measured, and PIP is adjusted based on the expansion of the chest, which does not permit determination of hypoventilation and gas trapping and assessment of spontaneous ventilation. Synchronized Intermittent Mandatory Ventilation Synchronized intermittent mandatory ventilation (SIMV) is for the most part similar to IMV except that ventilator cycles are delivered in synchrony with the onset of spontaneous inspiration. In SIMV, ventilator management is similar to that of IMV during the different phases of respiratory failure. SIMV was rapidly accepted and has become CHAPTER 45 Principles of Respiratory Monitoring and Therapy 5 lpm Flow 0 Exhalation ends Interrupted exhalation 20 mL VT 0 Gas trapping 25 cm H2O PAW 0 Ti  Te Ti  Te FIGURE 45-6 Gas trapping. Recordings of flow, tidal volume ( VT), and airway pressure ( PAW) illustrate the potential for gas trapping at high ventilator rates. As the ventilator rate increases, Te becomes insufficient to complete exhalation where every exhalation is interrupted by a new ventilator cycle. This produces a volume of gas trapped in the airways and alveoli and a reduction in VT. one of the most common modes of neonatal ventilation. During SIMV, spontaneous breathing is continuously monitored by the ventilator, and the first spontaneous inspiration during a time window of constant duration triggers a ventilator cycle. These time windows are opened consecutively, and their duration is similar to the interval between cycles in IMV. If no spontaneous inspiration is detected by the time the window elapses, a backup ventilator cycle is delivered. In this manner the number of ventilator cycles the infant receives every minute is the same as in IMV, but the interval between them is not constant. Different techniques have been incorporated into neonatal ventilators to achieve synchronization, including the Graseby abdominal pressure capsule, esophageal balloons, transthoracic impedance, airway pressure changes, and 623 flow sensors. The last are used by most neonatal ventilators for synchronization and VT monitoring. SIMV produces a greater and more consistent VT in comparison to IMV because of the increased transpulmonary pressure in ventilator cycles that are in synchrony with spontaneous inspiration (Bernstein et al, 1994; Hummler et al, 1996). Thus, lower PIP and slightly shorter Ti are often sufficient because the infant’s inspiratory effort contributes to generating the VT. As lung disease subsides, the infant’s lung mechanics improve, and spontaneous breathing effort becomes more consistent, PIP is gradually adjusted to maintain a VT between 3 and 5 mL/Kg. During acute respiratory failure, a ventilator SIMV frequency ranging between 40 and 60 b/m is usually required to maintain Paco2 between 40 and 50 mm Hg. At this frequency, SIMV is likely to assist almost all spontaneous breaths. During weaning, SIMV frequency is gradually reduced as spontaneous unassisted breathing can better maintain acceptable Paco2 levels. In infants with chronic lung disease, Paco2 levels between 50 and 65 mm Hg are often tolerated for weaning of the ventilator settings as long as respiratory acidosis is not observed. The relative advantages of SIMV in comparison to IMV include faster weaning and a shorter ventilator dependency with improved respiratory outcome in the smaller infants (Chen et al, 1997; Bernstein et al, 1996). As the SIMV frequency is weaned, the spontaneous breathing effort must be able to sustain most of the ventilation, thus requiring a greater breathing effort. It has been suggested that the work of breathing in unassisted spontaneous breaths between SIMV cycles is counterproductive and can potentially lead to diaphragmatic fatigue. Conversely, it has also been suggested that exposure to fewer ventilator cycles may reduce the risk of lung injury and improve diaphragmatic fitness toward an eventual extubation. Failure to synchronize leads the ventilator to cycle at the backup IMV rate. Delayed cycling prolongs the duration of inspiration and can disrupt the breathing pattern (Beck et al, 2004), whereas autocycling produce asynchrony between ventilator and infant. Assist-Control Ventilation In A/C, every spontaneous inspiration is assisted by a synchronous ventilator cycle. If the ventilator does not detect spontaneous breaths because of either apnea or very shallow breathing, it provides an IMV rate with the same PIP as the assisted breaths. A/C can also be used during the acute and weaning phases of respiratory failure. A/C is primarily managed by adjustment of PIP to maintain VT within an adequate range. As the respiratory failure resolves with improving lung mechanics and stronger spontaneous breathing effort, PIP is gradually weaned as VT remains within an acceptable range. In the more immature infants, the respiratory drive is not consistent, and such infants have frequent apnea. Thus, the “controlled” or backup IMV rate becomes more relevant in the maintenance of ventilation during apnea. In infants with respiratory failure, the Ti required to achieve lung inflation usually ranges between 0.25 and 0.4 seconds. Longer Ti is not recommended because 624 PART X Respiratory System of the potential for gas trapping with ventilator cycles delivered at the infant’s often high breathing frequency. A long Ti can also disrupt spontaneous breathing (Beck et al, 2004; Dimitriou et al, 1998; Upton et al, 1990). Some ventilators provide flow cycling where Ti terminates at the end of the spontaneous inspiration or at the end of lung inflation. The advantages of A/C include a reduction in spontaneous breathing effort relative to IMV or SIMV because in contrast to these modes, A/C assists every inspiration (Bernstein et al, 1994; Kapasi et al, 2001). It has been suggested that assisting every spontaneous inspiration in A/C prevents diaphragmatic fatigue and can reduce the duration of weaning and ventilator dependency compared to IMV (Baumer, 2000; Beresford et al, 2000; Chan and Greenough, 1993; Donn et al, 1994). However, A/C has not been shown to be more effective than SIMV (Chan and Greenough, 1994; Dimitriou et al, 1995). In A/C, PIP should be adjusted to avoid a large VT that can produce hyperventilation, whereas the backup rate should be just sufficient to prevent hypoventilation during apnea. Autocycling is a problem in A/C because neonatal ventilators do not have a limit on cycling frequency and therefore can produce hyperventilation or gas trapping. Pressure Support Ventilation In PSV, the ventilator also provides a positive pressure breath with every spontaneous inspiratory effort. The start of the support pressure is in synchrony with the onset of the infant’s inspiration and termination of the support pressure occurs toward the end of spontaneous inspiration or when lung inflation is completed, that is, flow cycling. The use of PSV is aimed at unloading the respiratory pump with support pressure that helps overcome the elastic and resistive loads imposed by the preterm infant’s underlying lung disease. PSV can be used as a stand-alone mode with an IMV backup rate for apnea or in some ventilators as an adjunct to SIMV to assist the spontaneous breaths, whereas the SIMV cycles with a higher peak pressure are used to maintain lung volume. Although PSV can be used during both the acute and weaning phases of respiratory failure, it is most commonly used during weaning. A consistent respiratory drive is required for the use of PSV as a stand-alone mode. For this reason, a backup IMV rate is recommended particularly in preterm infants to prevent hypoventilation. During PSV, gas leaks around the ETT can produce autocycling and extend the duration of the cycle. To avoid the risk of hyperventilation or gas trapping, trigger and termination sensitivity levels must be adjusted and limits on cycle duration must be set. When PSV is used as an adjunct to SIMV, PSV can reduce the spontaneous breathing effort compared to SIMV alone even with support pressures set at a fraction of the peak pressures of the SIMV cycles (Gupta et al, 2009; Osorio et al, 2005; Patel et al, 2009). More importantly, by reducing the need for high SIMV rates, PSV can facilitate weaning in comparison to SIMV alone in preterm infants of birthweight (BW) less than 1000 g (Reyes et al, 2006). In that study PSV set at 30% to 50% of the peak pressure of the SIMV cycles were sufficient to maintain acceptable Paco2 levels with significantly lower SIMV rates. Volume-Targeted Ventilation In volume-targeted ventilation, automatic adjustments to the peak positive pressure or the duration of ventilator cycle are done to maintain a target VT. Volume targeted ventilation modes have been proposed as means to reduce ventilator-associated lung injury caused by ventilation with excessive or insufficient tidal volumes during conventional pressure-controlled ventilation. A number of volume-targeted modes are available in neonatal ventilators. These differ in the timing of the adjustment or in the duration of the mechanical cycle, that is, whether it occurs as the cycle is delivered to the infant or in the subsequent cycle. These modes also differ in the volume parameter that is controlled—whether it is the tidal volume received by the infant or the volume delivered by the ventilator to the circuit, and whether this is measured during the inspiratory phase or during exhalation. Volume-Controlled Ventilation In volume-controlled ventilation (VC), the ventilator delivers a set volume of gas into the ventilator circuit in each cycle. VC cycles are delivered in the IMV, SIMV, or A/C modes described earlier. The time required to deliver the set volume depends on the ventilator flow rate, which can be constant during the cycle or variable with an initial peak followed by a gradual decline. The flow continues until the set volume is delivered. The cycle ends before the set volume is delivered if its duration exceeds the set Ti or the airway pressure exceeds the set PIP. In small infants, most of the volume delivered by the ventilator is compressed in the circuit, and the actual VT delivered to the infant is only a fraction. Some ventilators use algorithms to correct the measured volume by the compressed volume to estimate true VT. VC has been proposed as a strategy to facilitate weaning and reduce the complications of positive pressure ventilation. Clinical trials showed that compared to pressure limited ventilation in the A/C mode, VC reduced the weaning time and the duration of mechanical ventilation in infants weighing at least 1200 g at birth (Sinha et al, 1997). Weaning was also faster with VC in infants of birthweight <1000 g, but the total duration of mechanical ventilation did not change and respiratory outcome did not differ significantly (Singh et al, 2006). Pressure-Regulated Volume-Controlled Ventilation In pressure-regulated volume-controlled ventilation (PRVC), the peak pressure of pressure-controlled ventilator cycles is adjusted from one cycle to the next to maintain a target volume. The targeted volume can be that delivered by the ventilator or the actual estimated or measured VT. During PRVC, gas leaks around the ETT can produce overestimation of volume during the inspiratory phase and consequently lead to inappropriate reductions in peak inspiratory pressure. Compared to conventional IMV, PRVC in A/C mode was effective in reducing the duration of mechanical ventilation CHAPTER 45 Principles of Respiratory Monitoring and Therapy and the incidence of intraventricular hemorrhage (IVH) in infants with RDS of BW <1000 g (Piotrowski et al, 1997). These advantages can be attributed, in addition to volume targeting, to synchronized delivery of PRVC cycles. These advantages were not evident when PRVC was compared to SIMV (D’Angio et al, 2005). Volume Guarantee Ventilation In volume guarantee ventilation (VG), the peak pressure of each ventilator cycle is adjusted to maintain a target VT based on exhaled VT measured in previous cycles. Measurement of exhaled VT is aimed at circumventing the effects of gas leaks around the ETT during the inspiratory phase. VG can be used in combination with A/C, PSV, SIMV, or IMV modes. VG was proposed as a potential alternative to avoid both extremes of VT and to achieve a consistent weaning of peak pressure. The stability of VT and gas exchange in infants with RDS can be improved by VG when combined with A/C or PSV as noted by fewer breaths with too small or large VT and less hypocapnia (Abubakar and Keszler, 2001; Cheema et al, 2007; Herrera et al, 2002; Keszler and Abubakar, 2004). In infants with RDS or who had received surfactant, VG achieved the proper reduction in PIP, but this was dependent on setting a lower target VT than the VT attained with the pressure-controlled modes (Abd El-Moneim et al, 2005; Abubakar and Keszler, 2001; Cheema and Ahluwalia, 2001; Herrera et al, 2002; Nafday et al, 2005; Olsen et al, 2002). However, it must be noted that the smaller target VT was in some cases not sufficient, resulting in increased Pco2 and spontaneous breathing effort (Herrera et al, 2002). VG was proposed as a means to attenuate hypoxemia spells triggered by hypoventilation during periods of agitation and decreased compliance in preterm infants. VG reduced the duration of the hypoxemia spells compared to SIMV, but only when the target VT was increased which resulted in a considerable increase in airway pressure (Polimeni et al, 2006). New and Experimental Modes Proportional Assist Ventilation In proportional assist ventilation (PAV), the ventilator pressure is increased in proportion to the measured volume, flow, or both generated by the infant’s inspiratory effort. This achieves a perceived reduction of the elastic and resistive loads imposed by lung disease that commonly prevent the infant from producing an adequate VT. The proportionality factors by which the positive pressure increases in relation to volume or flow are the elastic (volume proportional) and resistive (flow proportional) gains. The elastic and resistive gain factors must be individualized to each infant’s lung compliance and airway resistance. Elastic gain factors that exceed the lung elastance (inverse of compliance) can lead to a runaway increase in pressure, whereas excessive resistive gain factors can produce oscillations in airway pressure. For this reason, PAV devices provide peak pressure and volume limits. Because PAV amplifies only the spontaneous breathing effort, a backup IMV rate is needed to prevent hypoventilation during apnea. 625 By compensating for the loads induced by lung disease, PAV reduced the breathing effort in infants recovering from RDS (Musante et al, 2001). Compared to conventional modes that deliver a constant pressure during inspiration, PAV produced similar ventilation with lower ventilator and transpulmonary pressures (Schulze et al, 1999, 2007). Neurally Adjusted Ventilatory Assist Neurally adjusted ventilatory assist (NAVA) is a mode where the ventilator pressure is adjusted in proportion to the electrical activity of the crural diaphragm measured by esophageal electrodes. NAVA was developed to improve the coupling of the infant’s inspiration and the ventilator by overcoming conditions that delay or limit pneumatic triggering of the ventilator cycle. A study in preterm infants approaching extubation showed NAVA can produce comparable gas exchange with lower pressures than PSV, but no physiologic effects of the better synchrony, including breathing effort, were noted. This study showed the feasibility of noninvasive NAVA after extubation (Beck et al, 2009). Data are lacking on the effects of different NAVA proportionality factors between diaphragmatic electrical activity and ventilator pressure within the same infant or between infants, because electrical activity of the diaphragm cannot be normalized. A backup ventilator rate is needed to prevent hypoventilation during apnea. Targeted Minute Ventilation Targeted minute ventilation (TMV) is an experimental mode where the ventilator rate is adjusted to maintain minute ventilation at a target level. If minute ventilation exceeds or decreases below the target level, the ventilator rate is reduced or increased, respectively. If spontaneous breathing can maintain a normal minute ventilation, the ventilator rate is reduced. In preterm infants recovering from RDS, TMV reduced the ventilator rate to half without impairing gas exchange compared to SIMV. Although these infants were able to sustain their ventilation for long periods, at times they required increased ventilator rates (Claure et al, 1997). Mandatory minute ventilation (MMV) is a form of targeted minute ventilation where the ventilator rate is turned off when spontaneous breathing maintains ventilation. If minute ventilation falls below the target level, a constant ventilator rate of VC cycles is provided. In nearterm infants ventilated for reasons other than lung disease, MMV weaned the rate and reduced the mean airway pressure compared to SIMV (Guthrie et al, 2005). Adaptive backup ventilation is a form of backup support. In this mode, a ventilator rate is provided during apnea as well as during the occurrence of hypoxemia detected by Spo2. In preterm infants recovering from RDS, this hybrid backup mode reduced the incidence and duration of hypoxemia spells compared to a backup ventilator rate for apnea alone (Herber-Jonat et al, 2006). HIGH-FREQUENCY VENTILATION Because of the association between pulmonary overdistention, lung injury, and the development of BPD, the possibility of achieving gas exchange with very small tidal 626 PART X Respiratory System volumes has been of great interest to neonatologists for many years. This led to the development of instruments that can generate changes in airway pressure at rates in excess of 10 Hertz (600/min) with the aim of producing ventilation with very small tidal volumes, usually much smaller than the dead space. Gas Transport during High-Frequency Ventilation During conventional ventilation, gas exchange occurs by introducing of fresh gas into the distal airspaces with each inspiration. In contrast, during high-frequency ventilation (HFV), the volume of fresh gas delivered by each cycle is very small and does not reach the most distal portions of the lung. Therefore, different mechanisms must explain alveolar ventilation and gas exchange. These include bulk flow into the more proximal portions of the lung, enhanced mixing of gas within the conducting airways, and out-of-phase movement between different regions of the lung that have different time constants. There is also enhanced diffusion of gas in large and medium-sized airways due to asymmetrical velocity profiles during inspiration and expiration. Finally, there is molecular diffusion in the more distal air spaces that moves the different gas molecules from areas of higher to lower concentrations. These mechanisms have been mostly explored in adult lung models, and therefore it is not clear how well they apply to the immature or sick neonatal lung. Devices for HFV Several types of high-frequency (HF) ventilators have been used in the neonate, including jet ventilators, oscillators, and flow interrupters. Jet ventilators generate a high-velocity gas flow that is injected through a small-diameter tube that opens into the airway connector. Expiration is passive, and the cycling rate is determined by an electrically operated valve that opens and closes the jet at a predetermined rate and timing. The high velocity of the gas injected into the airway produces a Venturi effect that pulls additional gas from the ventilator circuit. This is known as “gas entrainment.” Tidal volume is determined primarily by the driving pressure of the gas, the inspiratory time, and the resistance of the injection port. Because expiration is passive during high-frequency jet ventilation (HFJV), there is a risk of gas trapping and lung overdistention (Bancalari et al, 1987). This risk is higher when the time constant of the respiratory system is increased by airway obstruction and when large tidal volumes are delivered. Because of this, jet ventilators are used at lower rates (4 to 10 Hz). Jet ventilators are used in combination with conventional ventilators that provide PEEP and may also provide conventional positive pressure cycles. Oscillatory ventilators use a piston or a membrane driven by an electromagnetic force and connected to the ventilator circuit. The mean airway pressure is determined by the gas flow through the circuit and a variable resistance, whereas the tidal volume is generated by the size of the excursion of the piston or membrane. With these devices, the expiratory phase is active because during expiration, airway pressure falls below baseline. This reduces the risk of gas trapping. Flow interrupters are a hybrid between jets and oscillators. They produce airway pressure changes by interrupting the gas source at very high rates using a standard ventilator circuit rather than an injection cannula. They are relatively simple and are usually offered as an additional mode in conventional ventilators. Most do not have enough power to generate sufficient tidal volumes to effectively ventilate large infants or infants with very stiff lungs. Clinical Experience The indications for HFV vary widely among different centers. Although some use HFV routinely as a primary mode of support, most centers use HFV as rescue when conventional ventilation has failed. This includes preterm infants who require increasing ventilator settings to maintain Co2 elimination and oxygenation within acceptable limits and those with evidence of pulmonary interstitial emphysema. In larger infants, HFV is also indicated in situations where conventional ventilation is not sufficient to maintain acceptable Co2 elimination or oxygenation, and especially in infants with persistent pulmonary hypertension secondary to hypoplastic lungs due to congenital diaphragmatic hernia or oligohydramnios or to severe lung disease due to meconium aspiration or pneumonia. HFV in RDS The clinical results with the use of HFV in infants with RDS have been inconsistent. Whereas some studies have shown better outcomes such as increased survival with no BPD (Courtney et al, 2002), others have not shown differences (Johnson et al, 2002). Some studies have suggested a higher risk of air leaks (Thome et al, 1999), and others higher incidence of intracranial hemorrhage with HFV (HIFI Study Group, 1989; Moriette et al, 2001). Many of the earlier studies were performed before exogenous surfactant was available, and therefore the results are not applicable to the present situation where most preterm infants are exposed to antenatal steroids and, when indicated, receive exogenous surfactant. A meta-analysis of the most recent trials suggests a possible advantage for HFV in the outcomes “BPD at 36 weeks” or “death or BPD” at 36 weeks corrected age. However, when the results were analyzed using adjustments for heterogeneity between trials, the beneficial effect disappeared (Thome et al, 2005). The results of this meta-analysis revealed a significant increase in the risk of air leaks and a trend for higher risk of IVH grades III and IV with HFV. When the results of HFV were compared with conventional ventilation, including only studies where HFV was used with high-volume strategies and conventional ventilation with an optimized low positive pressure and tidal volume strategy, there were no significant differences in any of the outcomes (Bollen et al, 2007; Thome et al, 2005). Because of the inconsistency of the results and the lack of solid evidence of benefits of HFV over conventional ventilation in infants with RDS, the selection of one modality over the other is mostly based on individual preference. CHAPTER 45 Principles of Respiratory Monitoring and Therapy HFV in PPHN The clinical evidence with the use of HFV in infants with PPHN is not solid and comes from a few, relatively small trials. Some of these studies suggested a decreased need for extracorporeal membrane oxygenation (ECMO) in infants with PPHN treated with HFV compared to those managed with IPPV (Clark et al, 1994). However, other studies have not shown clear advantage of HFV over conventional management (Kinsella et al, 1997; Rojas et al, 2005). HFV may offer some advantage over conventional ventilation in infants with hypoplastic lungs secondary to congenital diaphragmatic hernia (Desfrere et al, 2000). Other Indications HFV has also been used in infants with bronchopleural fistulas in an attempt to reduce the amount of gas leak into the pleural space (Gonzalez et al, 1987; Walsh and Carlo, 1989). It can also be used during bronchoscopy or during airway surgery because in contrast to conventional ventilation, it can produce adequate gas exchange with a partially open airway and large gas leaks (Nutman et al, 1989). Side Effects of HFV Because during HFV the tidal volumes are extremely small, when HFV is used in unstable lungs there is a possibility of progressive loss of lung volume and atelectasis. For this reason it is necessary to use mean airway pressures that are usually higher than those used during conventional IPPV. In fact there is evidence that during HFV the use of an open lung strategy utilizing recruiting maneuvers and high mean airway pressures produces better outcomes than utilizing lower pressures (Thome et al, 2005). As a result of these higher pressures, it is likely that HFV may negatively influence cardiovascular function (Osborn and Evans, 2003; Trindade et al, 1985; Truog and Standaert, 1985; Weiner et al, 1987) and may also explain the higher incidence of pneumothorax reported in some studies with HFV (Thome et al, 2005). Because of the effectiveness of HFV in enhancing alveo­ lar ventilation it is easy to drive Paco2 to very low values within a very short period of time. For this reason it is important to monitor Paco2 values closely, especially when HFV is initiated or when settings are changed. Because of the very short inspiratory and expiratory times, it is extremely important to maintain the airway as patent as possible, ensuring a correct position of the endotracheal tube. Any obstruction will produce a decrease in pressure transmission and in tidal volume and can lead to gas trapping. This is even more important with jet ventilation. During jet ventilation, it is also critical to ensure proper humidification of the inspired gas to prevent airway damage that can be produced by the high velocity of gas injected into the airway. Ventilator Settings during HFV The ventilator settings during HFV are simpler to adjust than with conventional ventilation. The mean airway pressure determines the lung volume and is adjusted to achieve 627 better ventilation-perfusion ratio and oxygenation. This is done considering the severity and type of lung disease. Most infants with RDS are managed with mean airway pressures between 8 and 15 cm H2O. Higher levels may be necessary in some cases with severe lung disease and poorly compliant lungs. The tidal volume is determined by the delta pressure, and this, in combination with the frequency, determines the Co2 elimination. In contrast to conventional ventilation, during HFV the tidal volume is reduced as the frequency increases because the shorter times for inspiration and expiration prevent the equilibration of pressures between the ventilator circuit and the distal portions of the lung. Therefore, during HFV a reduction in rate may result in larger tidal volumes, more Co2 elimination, and lower Paco2 levels. During HF oscillation or with flow interrupters the frequencies used range from 8 to 15 Hz, whereas during jet ventilation lower frequencies between 4 and 10 Hz are used. The pressure transmission to the distal airways is greatly influenced by the resistance of the tube and the airway, so only a small fraction of the delta pressure generated by the ventilator is transmitted to the terminal airspaces. For this reason, the delta pressure during HFV represents more of a relative value used to adjust the ventilator than the real pressure change in the distal airways. WEANING FROM MECHANICAL VENTILATION Weaning from IPPV is extremely difficult in very immature infants who have poor inconsistent respiratory drive, a weak respiratory pump, and immature and many times damaged lungs. This combination of factors explains why they frequently become ventilator dependent for long periods of time. For many years, infants who were ventilated received most of their minute ventilation from the ventilator and were not allowed to have effective spontaneous ventilation. In recent years this has changed, and ventilators are used as an assist support, preserving the patient’s respiratory drive and effort. This has been possible by the introduction of synchronized patient-triggered ventilation (PTV) and has resulted in better outcomes and shorter times on mechanical ventilation. Because of the high rate of complications associated with prolonged mechanical ventilation, weaning should start as soon as respiratory function is stabilized. The order in which the different ventilator parameters are decreased is determined by the relative risk of complications associated with each of them. With the possibility of measuring VT, it has become much simpler to define the appropriate PIP that is required to generate an adequate VT of 3 to 5 mL/kg body weight. As lung compliance improves PIP can be reduced to keep VT within this range. The level of PEEP is usually kept between 4 and 8 cm H2O depending on the type of underlying lung disease and the level of oxygenation and Fio2 requirement. The PEEP level is decreased gradually as the oxygenation improves until a level of 4 to 5 cm H2O is reached, and this level is maintained until extubation. The inspired oxygen concentration is adjusted according to the level of arterial oxygen 628 PART X Respiratory System tension or O2 saturation measured by pulse oximetry. The ideal ranges of oxygenation have not been defined, but most clinicians accept saturations between 88% and 95% in preterm infants and up to 98% in term infants. In infants with evidence of pulmonary hypertension, higher levels are targeted to prevent pulmonary vasoconstriction. If VT measurement is not available, the reduction in PIP is based on the observation of chest movement, the degree of aeration on chest radiograph, and Paco2 levels. The ventilator rate is adjusted depending on the type of ventilation strategy being used. When the infant is ventilated with synchronized modes such as A/C or pressure support (PS), the infant determines the rate of the mechanical breaths so that the set rate is only the backup that the ventilator will provide when the infant’s own rate falls below that level. Therefore, this rate is only relevant when the infant becomes apneic or hypoventilates. When the infant is controlled, the rate in the ventilator is not determined by the infant, and the adjustment in mechanical rate is based on the arterial Pco2 level. During weaning, it is advisable to do gradual changes and adjust one parameter at a time to evaluate the response of the infant to each change. With the availability of continuous oxygen and Co2 monitoring, it is not always necessary to wait for results of arterial gas measurement to change ventilator settings, and the weaning can proceed faster. Synchronized Patient-Triggered Ventilation The use of patient-triggered synchronized ventilation has become common practice in neonatal units. Most randomized trials comparing PTV with nonsynchronized ventilation have shown reduction in the duration of mechanical ventilation in infants treated with synchronized modes (Greenough, 2001). This may be because, during PTV, the infant retains more control of ventilation, and the effectiveness of the mechanical and spontaneous breaths is enhanced by the summation of the ventilator positive pressure and the negative pressure generated by the respiratory muscles. A/C and SIMV are the most common modes of synchronized ventilation in the neonate. It has been suggested that assisting each spontaneous inspiration in A/C may avoid respiratory muscle fatigue and facilitate weaning. However, the duration of weaning has not been consistently shorter with A/C than with SIMV (Beresford et al, 2000; Chan and Greenough, 1994). On the other hand, a randomized trial comparing the use of pressure support ventilation as an adjunct to SIMV to assist every spontaneous inspiration revealed a faster weaning and shorter duration of ventilation compared with SIMV alone in preterm infants (Reyes et al, 2006). Volume Monitoring and Targeting The continuous monitoring of VT allows rapid weaning of PIP as the mechanical conditions of the lung improve. This can be achieved by a manual decrease of PIP as VT increases, or automatically by using volume-targeted ventilation where weaning is achieved automatically independent of the clinician, who only sets the VT targeted by the ventilator. Evidence from randomized trials using volume-targeting strategies suggest that faster weaning from mechanical ventilation can be achieved, although the results have not been entirely consistent (Singh et al, 2006; Sinha et al, 1997). Nasal CPAP and Nasal Ventilation After extubation, the infant is exposed to a number of mechanical impediments that explain the frequent need for reintubation in the smaller preterm infants. These include upper airway damage and retained secretions leading to obstruction and atelectasis, loss in lung volume due to poor respiratory effort, and a highly compliant chest wall. For these reasons, the use of continuous positive airway pressure applied through the nose can significantly reduce the deterioration that occurs frequently after extubation. Surprisingly, despite this improvement in respiratory function, the need for reintubation has not been consistently shown to be reduced by the use of N-CPAP after extubation (Davis and Henderson-Smart, 2003). In contrast with N-CPAP, the use of nasal ventilation after extubation has been shown to significantly reduce extubation failure (Davis et al, 2001; De Paoli et al, 2003). Although these studies have included small numbers of infants, the effects have been consistent. This is a promising therapeutic alternative that needs further evaluation and the development of suitable equipment to provide synchronized noninvasive support. Respiratory Stimulants Respiratory stimulants such as aminophylline and caffeine have been shown to be effective to increase respiratory center activity in preterm infants and to decrease the incidence of severe apneic episodes. These drugs have also been shown to facilitate successful weaning from mechanical ventilation and decrease the need for reintubation. For this reason most preterm infants receive a loading dose of caffeine or aminophylline before extubation, and they are maintained on these stimulants at least during the first days after extubation while they are also maintained on N-CPAP or nasal ventilation (Henderson-Smart and Davis, 2003). Permissive Hypercapnia Tolerance of higher carbon dioxide levels may reduce the need for support and reduce the duration of ventilation. However, the results of clinical trials have been inconsistent. Whereas initial trials suggested faster weaning from mechanical ventilation in the group with higher Co2 levels (Carlo et al, 2002; Mariani et al, 1999), a more recent study showed no benefit in duration of ventilation and a possible increase in mortality and central nervous system (CNS) impairment in infants with higher Paco2 (Thome et al, 2006). Although these results shed doubt on the benefits of high Co2 levels in premature infants during the acute stages of their clinical course, in infants with chronic lung disease it is necessary to tolerate high Co2 levels to wean them from mechanical ventilation. CHAPTER 45 Principles of Respiratory Monitoring and Therapy Dead Space Reduction The large anatomical dead space in preterm infants in addition to the instrumental dead space can also delay weaning from mechanical ventilation (Figueras et al, 1997). Continuous tracheal gas insufflations (CTGI) pumped through small capillaries to the distal end of the endotracheal tube to produce a continuous washout can reduce arterial CO2 and shorten the weaning process (Dassieu et al, 2000). This method is limited by the need for special ventilators and endotracheal tube. Continuous washout of the flow sensor by a controlled gas leak can also improve CO2 elimination and facilitate weaning (Claure et al, 2003; Estay et al, 2010). Extubation from IMV versus Endotracheal CPAP For many years infants on mechanical ventilation were extubated only after they had tolerated some time on CPAP via endotracheal tube. This practice was changed after a small, simple trial demonstrated that infants extubated from a low IMV rate had better success rates than those kept on CPAP for 6 hours before extubation (Kim and Boutwell, 1987). Weaning from High-Frequency Ventilation Infants ventilated with HFV are frequently switched to conventional ventilation before extubation. However, infants can be extubated directly from HFV following similar steps to those used to wean from conventional ventilation. The reduction in mean airway pressure (MAP) is done after oxygenation and lung expansion is estimated by chest radiographs while the Fio2 is adjusted to maintain the desired oxygenation levels. The oscillatory amplitude is gradually reduced in response to the levels of Paco2. If the MAP is weaned to low levels before the infant has spontaneous respiratory effort, this can lead to hypoventilation, loss in lung volume, and atelectasis. For this reason, it is advisable to lower the delta pressure and allow the CO2 to rise and stimulate spontaneous respiration before the MAP is lowered below 10 or 8 cm H2O. Prediction of Successful Extubation It is difficult to decide the best time for extubation in ventilated infants. It is obvious that many infants remain intubated for longer than necessary. This is evidenced by the fact that many infants who are accidentally extubated tolerate it well without needing further respiratory support. Various tools have been evaluated to predict successful extubation. These include measurement of lung mechanics, minute ventilation, inspiratory effort strength and ability to cope with mechanical loads, and stability of respiratory pattern during periods when the ventilator cycling is stopped (Kamlin et al, 2006). Some of the tools have been shown to predict with some accuracy successful extubation, but these tests have not been widely accepted in neonatal clinical practice. The decision to extubate an infant is usually based on the level of inspired oxygen and ventilator support that the infant is requiring to maintain acceptable arterial blood 629 gas levels. In general terms, if an infant needs less than 30% to 40% oxygen, a ventilator rate less than 15 per minute, and peak airway pressures below 15 cm H2O and keeps acceptable blood gases, most clinicians attempt extubation. The lower the gestational age, the more likely it is that the infant will not tolerate extubation and will require reintubation. In most cases this failure is because of poor respiratory effort or severe apneic episodes. Automated and Computer-Assisted Weaning In the targeted minute ventilation mode described earlier, the ventilator rate is automatically reduced during periods of consistent spontaneous breathing where minute ventilation is maintained at or above the target level. In a study of preterm infants recovering from RDS, this experimental mode reduced the ventilator SIMV rate set by the clinical team by half while arterial blood gases remained unchanged (Claure et al, 1997). A similar reduction in rate was observed in nearterm infants without lung disease when supported by mandatory minute ventilation, a mode where the ventilator rate is turned off if minute ventilation exceeds a set level or delivers a set rate of volume-controlled breaths when minute ventilation decreases below this level (Guthrie et al, 2005). Preterm infants often need supplemental O2, which increases their risk for eye and lung injury, particularly when exposure to oxygen is prolonged. In these infants, hyperoxemia is induced by an excessive Fio2, and therefore it is modifiable by appropriate weaning. Automated weaning of supplemental oxygen was achieved by systems developed to adjust Fio2 in a continuous manner. Systems of automated Fio2 control have been shown to be as or more effective than routine or dedicated manual control in maintaining oxygenation within a desired range, with most of the improvement due to significant reductions in hyperoxemia (Claure et al, 2009, 2011). Ventilator management involves adjustments of several parameters that affect the infant’s ventilation and gas exchange. Computerized algorithms for ventilator management have been proposed to achieve efficient and consistent weaning. In infants with RDS, ventilator management assisted by one of these algorithms led to improvements in gas exchange and avoided unnecessary increases in ventilator settings (Carlo et al, 1986). Data indicated that during routine care, hypoxemia and hypercapnia were more diligently corrected than hyperoxia and hypocapnia, whereas computer-assisted management was similarly effective in correcting both extremes. The potential benefits of computer-assisted weaning on long-term outcome have not been explored. ACUTE COMPLICATIONS OF RESPIRATORY SUPPORT PULMONARY GAS LEAKS Some of the most serious complications of mechanical ventilation are pulmonary gas leaks. These include pulmonary interstitial emphysema (PIE), pneumothorax, pneumomediastinum, pneumopericardium, pneumoperitoneum, and intravascular gas. 630 PART X Respiratory System involved areas in the lung. If PIE is unilateral, this can be accomplished by positioning the infant with the involved side down (Swingle et al, 1984). Mechanical ventilation using short inspiratory times (0.1 to 0.2 seconds), low inflation pressures, and small tidal volumes can reduce the gas leak (Meadow and Cheromcha, 1985). Unfortunately, it is difficult to maintain oxygenation and ventilation while using low volumes. A multicenter controlled trial found that HFJV allowed the use of lower peak and mean airway pressures in infants with PIE than did rapid-rate conventional ventilation and led to more rapid improvement in PIE (Keszler et al, 1991). Pneumothorax FIGURE 45-7 Pulmonary interstitial emphysema. A grossly hyperinflated lung with coarse radiolucencies extending from the pleura to the hilum. These radiolucencies represent gas bubbles in the perivascular and peribronchial interstitial cuffs. Pulmonary Interstitial Emphysema PIE is the result of rupture of air spaces from overdistention. Although it may be seen in infants breathing spontaneously it is much more common in preterm infants undergoing mechanical ventilation and in infants with evidence of pulmonary infection. Once alveolar rupture occurs, gas is forced from the air spaces into the connective tissue sheaths surrounding airways and vessels and into the interlobular septa containing pulmonary veins. The air follows a track along these sheaths to the hilum of the lung, producing the characteristic radiographic appearance of PIE (Figure 45-7). The PIE occupies space within the lung parenchyma, decreasing lung compliance. Gas trapped within the interstitial cuffs compresses airways and increases airway resistance. In addition, gas in the interstitial space impairs lymphatic drainage, increasing interstitial fluid. This leads to significant deterioration in gas exchange with increase in Paco2 and decrease in Pao2. The main cause of PIE is airspace overdistention and rupture. Therefore, minimizing pulmonary overdistention should reduce the risk of PIE. Shorter inspiratory times also decrease the incidence of PIE by reducing overdistention and avoiding an inspiratory hold (Heicher et al, 1981; OCTAVE Study Group, 1991). PIE is a serious complication of mechanical ventilation. Infants with PIE have a significantly increased risk of developing chronic lung disease as well as higher mortality rates (Gaylord et al, 1985; Powers and Clemens, 1993). Management Because the gas leak can behave like a check valve, gas trapping occurs, resulting in further alveolar overdistention and rupture. Therefore, the first step in treatment must be to interrupt this cycle by putting to rest the more severely Pneumothorax can occur spontaneously in healthy infants because of the very high transpulmonary pressures produced at birth. However, the incidence is much higher in the presence of underlying lung disease and in infants exposed to mechanical ventilation. Pneumothorax develops in 5% to 10% of spontaneously breathing infants with hyaline membrane disease. Although positive pressure ventilation increases the risk dramatically, treatment with surfactant has markedly lowered this risk. Pneumomediastinum occurs when gas tracks through the perivascular and peribronchial cuffs to the hilum and then ruptures into the mediastinum. From there, gas can rupture into the pleural space, producing a pneumothorax. If the leak is large, it produces a tension pneumothorax that collapses the lung and results in severe hypoxia and hypercapnia. In addition, by compressing mediastinal structures, venous return to the heart may be impeded, resulting in circulatory collapse. Diagnosis Because of the severe consequences of a tension pneumothorax, it is important to diagnose it as early as possible. In the spontaneous breathing infant, pneumothorax usually produces tachypnea, grunting, pallor, and cyanosis. The cardiac sounds may be shifted away from the side of the pneumothorax, and the affected hemithorax and abdomen may appear to be bulging. In the infant on positive pressure ventilation, signs are frequently more dramatic, with sudden onset of hypoxemia and cardiovascular collapse (Ogata et al, 1976). Transillumination of the chest is positive over the affected side, and the pneumothorax is confirmed by chest radiograph that shows gas in the pleural space and partial collapse of the lung with shift of the mediastinum to the opposite side (Figure 45-8). Management A small pneumothorax in a spontaneously breathing infant may be observed closely until spontaneous resolution occurs. Infants with larger symptomatic pneumothoraces and all infants receiving positive pressure ventilation require thoracostomy and pleural tube placement. The tube may be inserted at the midaxillary line and directed anteriorly, or placed in the second intercostal space in the midclavicular line and directed toward the diaphragm so that the tip lies between the lung and the anterior chest CHAPTER 45 Principles of Respiratory Monitoring and Therapy FIGURE 45-8 Tension pneumothorax. The lung on the involved side is collapsed, and the mediastinum is shifted to the opposite side, with bulging of the pleura into the intercostals spaces. wall. When placing the tube, the operator must avoid puncturing the lung, especially when a trocar, rather than a curved hemostat, is used to direct the tube. It also is possible to drain pneumothoraces using pigtail catheters that are placed percutaneously into the pleural space. This technique produces less trauma but may not be effective when large gas leaks are present. The thoracostomy tube is connected to a water seal with 10 to 20 cm H2O negative pressure and is left in place until it ceases to drain. The negative pressure should be discontinued, and the tube should be left under the water seal for 12 to 24 hours before removal. Pneumopericardium Pneumopericardium results from direct tracking of interstitial gas along the great vessels into the pericardial sac. Gas under tension in the pericardium impairs atrial and ventricular filling, decreases stroke volume, and ultimately decreases cardiac output and systemic blood pressure. Infants present with increasing cyanosis, decreased heart sounds, and decreased systemic blood pressure and pulse pressure. The chest radiograph is diagnostic, showing gas surrounding the cardiac silhouette (Figure 45-9). Needle aspiration alleviates the acute symptoms, but because recurrence rate is high, continuous tube drainage is necessary in most cases. Pneumoperitoneum Pneumoperitoneum results from dissection of air from the mediastinum along the sheaths of the aorta and vena cava into the peritoneal cavity. Infants with this condition present sudden abdominal distention and a typical abdominal radiograph. Occasionally, the pneumoperitoneum may be large enough to cause respiratory embarrassment by compromising descent of the diaphragm and may require 631 FIGURE 45-9 Pneumopericardium. A thin rim of pericardium is visible and clearly separated from the heart by gas within the pericardial sac. drainage. This cause of peritoneal free gas must be distinguished from a primary gastrointestinal perforation. Intravascular Gas It has been suggested that the intravascular gas is pumped under high pressure through the pulmonary lymphatics into the systemic venous circulation (Booth et al, 1995). The intravascular gas results in immediate cardiovascular collapse and is often diagnosed when gas is seen in vessels or in the heart chambers when a chest radiograph is taken to determine the cause of cardiovascular collapse. Intravascular gas is usually a fatal complication. AIRWAY COMPLICATIONS Prolonged endotracheal intubation can produce airway damage and subglottic stenosis. The risk is increased by a too snugly fitting endotracheal tube, prolonged duration of intubation, and traumatic intubation. Some infants may require tracheostomy and even surgery to repair the stenosis. Inadequate humidification of the inspired gas can produce necrotizing tracheobronchitis, a necrotic inflammatory process involving the trachea and main bronchi. Infants present with acute respiratory deterioration due to airway obstruction, hyperexpansion on chest radiograph, and poor chest movement. Emergency bronchoscopy may be necessary to relieve airway obstruction. Atelectasis frequently occurs during prolonged ventilation and after extubation from mechanical ventilation. This can also be secondary to airway injury due to trauma produced by suction catheters and by inadequate conditioning of the inspired gas. VENTILATOR-ASSOCIATED PNEUMONIA Ventilator-associated pneumonia is a very common complication of mechanical ventilation at any age, and it is likely to occur even more frequently in the 632 PART X Respiratory System immune-compromised premature infant. Unfortunately, it is not commonly diagnosed in this population because of the lack of specific criteria and the difficulty of differentiating pneumonia from other acute and chronic pulmonary pathologies observed previously in infants requiring prolonged mechanical ventilation. Despite these limitations, it should be suspected any time there is deterioration in lung function with radiographic changes suggestive of pneumonia. Changes in the amount and quality of the secretions obtained from the airway and colonization with pathogens is another indication of possible pulmonary infection that may require antibiotic therapy. Some evidence indicates that a large proportion of ventilated premature infants have aspiration of gastric contents into their airways. This is more common in fed infants and those receiving methylxanthines (Farhath et al, 2006). The consequences of this are not clear, but this could further contribute to the risk of pulmonary infection and chronic lung damage. SUGGESTED READINGS Avery ME: Recent increase in mortality from hyaline membrane disease, J Pediatr 57:553-559, 1960. Bancalari A, Gerhardt T, Bancalari E, et al: Gas trapping with high-frequency ventilation: jet versus oscillatory ventilation, J Pediatr 110:617-22, 1987. Bancalari E, Flynn J, Goldberg RN, et al: Influence of transcutaneous oxygen monitoring on the incidence of retinopathy of prematurity, Pediatrics 79:663669, 1987. Beck J, Reilly M, Grasselli G, et al: Patient-ventilator interaction during neurally adjusted ventilatory assist in low birth weight infants, Pediatr Res 65:663-668, 2009. Becker MA, Donn SM: Real-time pulmonary graphic monitoring, Clin Perinatol 34:1-17, 2007. Bollen CW, Uiterwaal CS, van Vught AJ: Meta-regression analysis of highfrequency ventilation vs conventional ventilation in infant respiratory distress syndrome, Intensive Care Med 33:680-688, 2007. Carlo WA, Finer NN, Walsh MC, et al: Target ranges of oxygen saturation in extremely preterm infants. SUPPORT Study Group of the Eunice Kennedy Shriver NICHD Neonatal Research Network, N Engl J Med 362:1959-1969, 2010a. Chang HY, Claure N, D’ugard C, et al: Effects of synchronization during nasal ventilation in clinically stable preterm infants, Pediatr Res 69:84-89, 2011. Chow LC, Vanderhal A, Raber J, et al: Are tidal volume measurements in neonatal pressure-controlled ventilation accurate, Pediatr Pulmonol 34:196-202, 2002. De Paoli AG, Davis PG, Lemyre B: Nasal continuous positive airway pressure versus nasal intermittent positive pressure ventilation for preterm neonates: a systematic review and meta-analysis, Acta Paediatr 92:70-75, 2003. Finer NN, Carlo WA, Walsh MC, et al: Early CPAP versus surfactant in extremely preterm infants, N Engl J Med 362:1970-1979, 2010. Gregory GA, Kitterman JA, Phibbs RH, et al: Treatment of the idiopathic respiratory-distress syndrome with continuous positive airway pressure, N Engl J Med 284:1333-1340, 1971. Hay WW Jr, Rodden DJ, Collins SM, et al: Reliability of conventional and new pulse oximetry in neonatal patients, J Perinatol 22:360-366, 2002. Keszler M, Abubakar K: Volume guarantee stability of tidal volume and incidence of hypocarbia, Pediatr Pulmonol 38:240-245, 2004. Moretti C, Giannini L, Fassi C, et al: Nasal flow-synchronized intermittent positive pressure ventilation to facilitate weaning in very low-birthweight infants: unmasked randomized controlled trial, Pediatr Int 50:85-91, 2008. Morley CJ, Davis PG, Doyle LW, et al: Nasal CPAP or intubation at birth for very preterm infants, N Engl J Med 358:700-708, 2008. Reyes ZC, Claure N, Tauscher MK, et al: Randomized, controlled trial comparing synchronized intermittent mandatory ventilation and synchronized intermittent mandatory ventilation plus pressure support in preterm infants, Pediatrics 118:1409-1417, 2006. Walsh M, Engle W, Laptook A, et al: Oxygen delivery through nasal cannulae to preterm infants: Can practice be improved? National Institute of Child Health and Human Development Neonatal Research Network, Pediatrics 116:857861, 2005. Complete references used in this text can be found online at www.expertconsult.com C H A P T E R 46 Respiratory Distress in the Preterm Infant J. Craig Jackson INTRODUCTION This chapter on the causes of respiratory distress in the preterm infant will focus primarily on respiratory distress syndrome (RDS) also known as hyaline membrane disease and on surfactant replacement therapy. However, the chapter also includes discussion of two less frequent causes - pneumonia/sepsis and pulmonary hypoplasia. Other causes of respiratory distress that are more commonly seen in term infants, such as transient tachypnea of the newborn (TTNB), are discussed in Chapter 47. Table 46-1 lists causes of respiratory distress in newborns of all gestational ages and demonstrates the relative frequency of each diagnosis. This table provides a useful differential diagnosis for respiratory distress in preterm neonates, particularly if polycythemia and hypoglycemia are included. More than half of extremely low birthweight newborns will have some type of respiratory distress (Figure 46-1). In that population respiratory distress syndrome is by far the most common diagnosis (50.8%), followed by transient tachypnea of the newborn (4.3%) and pneumonia/sepsis (1.9%). In higher birthweight preterm newborns, the incidence of any type of respiratory distress is much lower, but in symptomatic babies, the incidence of RDS and pneumonia/sepsis is lower, whereas TTNB is more common. Compared to the three diagnoses featured in Figure 46-1, the incidence of other causes of respiratory distress (see Table 46-1) is very low in preterm infants. However, these less common diagnoses should be considered in preterm infants with respiratory distress who follow an atypical clinical course. RESPIRATORY DISTRESS SYNDROME During breathing (either spontaneous or assisted), shear stresses in the alveoli and terminal bronchioles occur due to the repetitive reopening of collapsed alveoli and the overdistention of open alveoli (Nilsson et al, 1978). These forces can quickly damage the fragile lung architecture, leading to leakage of proteinaceous debris into the airways (i.e., hyaline membranes). This debris (Figure 46-2) may impair the function of what little surfactant is present, leading to a downward spiral that may end in respiratory failure and death if not interrupted. If supportive therapy is successful, the repair phase begins during the 2nd day after birth with the appearance of macrophages and polymorphonuclear cells. Debris is phagocytosed and the damaged epithelium is regenerated. Edema fluid in the interstitium is mobilized into lymphatics, leading to the “diuretic” phase of RDS characterized by high urine output. With uncomplicated RDS, the patient improves by the end of the first week after birth. However, infants born at <1250 g and larger newborns needing high concentrations of oxygen and positive-pressure ventilation for severe RDS may develop inflammation and inappropriate repair of the growing lung, leading to emphysema and fibrosis (see Chapter 48). PURPOSE OF SURFACTANT Surface tension is generated from molecular attractive forces within a liquid that oppose spreading; it is the reason that water ‘beads up” on a clean surface. If you try to inflate a bubble under water with a straw (Figure 46-3, A), the RISK FACTORS The main risk factor for RDS, by far, is prematurity (see Figure 46-1). Other factors that increase the risk of RDS include perinatal asphyxia, maternal diabetes, lack of labor, absence of antenatal steroid administration to the mother, male gender, and White race. The central feature of RDS is surfactant deficiency due to lung immaturity, commonly a result of premature birth or delayed lung maturation associated with maternal diabetes or male gender. Surfactant dysfunction can also be caused by perinatal asphyxia, pulmonary infection, or excessive fetal lung liquid due to delivery without labor. PATHOPHYSIOLOGY OF RDS Because alveoli with insufficient (or dysfunctional) surfactant are unstable and tend to collapse, patients with RDS develop generalized atelectasis, ventilation-perfusion mismatching, and subsequent hypoxemia and respiratory acidosis. TABLE 46-1 Causes of Respiratory Distress Respiratory distress syndrome Transient tachypnea of newborn Pneumonia/sepsis Meconium aspiration syndrome Congenital cardiac malformation Chromosomal disorder/multiple congenital anomalies Spontaneous pneumothorax Perinatal asphyxia Pulmonary hemorrhage Persistent pulmonary hypertension Diaphragmatic hernia Apnea of prematurity Pulmonary hypoplasia Pulmonary dysplasia Hydrothorax Postsurgical diaphragmatic palsy 46% 37% 5% 2% 2% 1.4% 1.2% 1.1% 1.0% 0.8% 0.8% 0.6% 0.3% 0.2% 0.2% 0.2% 100% Data from Rubaltelli FF, Dani C, Reali MF, et al: Acute neonatal respiratory distress in Italy: a one-year prospective study, Acta Paediatr 87:1261-1268, 1998. 633 634 PART X Respiratory System spreading of the water’s surface and enlargement of its surface area will be opposed by its surface tension (T). According to Laplace’s law, the pressure (P) required to inflate the bubble is proportional to T divided by the radius of the bubble (r). If you simultaneously try to inflate two interconnected bubbles, the smaller one will deflate into the larger one, because of its smaller radius (Figure 46-3, B). Surfactants are surface active materials that lower surface tension; a detergent is a type of surfactant which causes water to spread out on a surface rather than beading up. The surface tension of clean water resists the creation 60.0% 50.0% of bubbles, but the addition of detergent allows bubbles to form. Because the surfactant properties of detergents do not persist, all of the small bubbles eventually collapse into the large ones, and finally even the large ones collapse. Lung surfactant has the miraculous property of reducing surface tension as the size of the bubble decreases. In clean water with high surface tension, the pressure required to keep open a bubble as the radius decreased would be greater than the pressure keeping open a larger bubble, so the smaller bubbles would eventually collapse into larger ones (or small alveoli would collapse into larger 1.9% 4.3% Pneumonia/sepsis TTNB RDS 0.84% Incidence 40.0% 9.2% 30.0% 50.8% 20.0% 32.7% 0.28% 10.0% 0.0% 6.1% 6.6% 1000g 1000-1500g 1500-2500g Birthweight FIGURE 46-1 Incidence of respiratory problems in preterm newborns. (Data from Rubaltelli FF, Dani C, Reali MF, et al: Acute neonatal respiratory distress in Italy: a one-year prospective study, Acta Paediatr 87:1261-1268, 1998.) A A 100 m 200 m A B FIGURE 46-2 Photomicrographs of lung tissue from experimental animal with RDS, flash frozen during inflation. Note liquid-air interface (arrows). The alveolar debris forms hyaline membranes. A, Low power. B, Higher power. (From Jackson JC, MacKenzie AP, Chi EY, et al: Mechanisms for reduced total lung capacity at birth during hyaline membrane disease in premature newborn monkeys, Am Rev Respir Dis 142:413-419, 1990. American Thoracic Society.) CHAPTER 46 Respiratory Distress in the Preterm Infant ones), as shown in Figure 46-3, B. However, if the bubbles are lined with good-quality surfactant, the surface tension falls quickly as the radius gets smaller because the surfactant molecules become crowded during deflation (Figure 46-4). When the radius is very small, the surface tension P P r A r2 r1 B T T T FIGURE 46-3 Effect of surface forces generated by inflating bubbles under water. A, Single bubble of radius (r) resists inflation and thus requires pressure (P) to overcome the surface tension (T). B, If the surface tension is the same in two bubbles of unequal size, the smaller one will collapse into the larger one, because of the Laplace relationship, P = T/r (small r requires larger pressure to stay inflated). P P r T T r Expiration FIGURE 46-4 Surfactant molecules on surface crowding during deflation. (From Jackson JC: Respiratory distress syndrome (figure 212–1). In: Osborn LM, DeWitt TG, First LR, Zenel JA, editors: Pediatrics, Philadelphia, 2005, Elsevier Mosby.) 635 falls almost to zero, and the pressure required to keep the smaller bubble open is negligible. Thus, it does not collapse. During inflation, as the radius of the bubble increases, surface tension increases even faster. This means that the amount of pressure in the larger bubbles will be higher than in smaller ones; in the lung where bubbles are interconnected, this pressure difference will cause flow to the smaller alveoli, thus keeping all the alveoli about the same size. The cumulative result of the alveoli remaining open during deflation is a nonlinear pressure-volume relationship. The lung with sufficient surfactant retains gas during expiration, compared to rapid and almost complete loss of gas in the surfactant-deficient lung (Figure 46-5, A). During inflation, more pressure is required to achieve similar tidal volume (Figure 46-5, B), because of poor compliance (∆V/∆P) from having to reopen collapsed alveoli. For instance, to achieve a tidal volume of 5 mL/kg, an infant with RDS may require a pressure increase of 25 cm H2O; dividing the volume change, ∆V, by pressure change, ∆P, we calculate that the compliance is only 0.25 mL/kg/cm H2O, which is about one third of normal. In the absence of adequate amounts of functional surfactant in the newborn lung, there is widespread alveolar collapse with overdistention of open alveoli (Figure 46-6). Because reopening collapsed alveoli requires high pressure, the spontaneously breathing newborn with surfactant deficiency must generate highly negative intrathoracic pressure. Clinically, this is manifested by retractions of the respiratory muscles during inspiration. Newborns may also attempt to prevent alveolar collapse by grunting. This partial closure of the glottis during expiration helps maintain an end-expiratory pressure that may keep unstable alveoli open. A consequence of widespread alveolar collapse is intrapulmonary shunting of blood past atelectatic lung, without the opportunity for blood in pulmonary capillaries to pick up oxygen from, or deliver carbon dioxide to, the alveoli. In addition, lungs that are poorly inflated have widespread collapse of pulmonary vessels, leading to pulmonary hypertension. The elevated pulmonary artery pressures lead to right-to-left shunting of unoxygenated Healthy With surfactant Without surfactant Volume Volume V P RDS V P A Pressure B Pressure FIGURE 46-5 Effect of surface forces on pressure-volume relationships. During deflation (A), the lungs with surfactant retain gas even at very low pressures, because of falling surface tension as the alveoli get smaller. The alveoli without surfactant collapse as the alveoli get smaller. During inflation of RDS lungs (B), the starting lung volume (functional residual capacity) is lower, and much more pressure is required during inflation, compared to healthy lungs. 636 PART X Respiratory System O  N O O P O O O O O FIGURE 46-7 The main ingredient of lung surfactant, dipalmitoylphosphatidylcholine (DPPC). A for low viscosity for optimal spreading and redistribution along the smallest airways with the need for a stable and low surface tension. Surfactant phospholipids are assembled in the type II pneumocytes of the lung epithelium into lamellar bodies in the form of bilayered membranes (Figure 46-9). Surfactantassociated proteins SP-B and SP-C are essential for the transition to a monolayer at the air-liquid interface. The molecular structure of the hydrophobic SP-B is complex and it interacts with the phospholipid monolayer as shown in Figure 46-10. Its absence is associated with fatal neonatal respiratory failure. Surfactant-associated proteins SP-A and SP-D are hydrophilic and have roles in immune defense. SP-A is involved in reuptake and reuse of secreted surfactant (see Chapter 42). CLINICAL SIGNS OF RDS B FIGURE 46-6 Histology of RDS. Scanning electron micrograph of lung frozen during inflation with air, in premature monkey healthy premature monkey (A) compared to one with RDS (B). Lungs affected by RDS have collapsed alveoli full of liquid and proteinaceous debris, with overdistended terminal airways. (From Jackson JC, Truog WE, Standaert TA, et al: Effect of high-frequency ventilation on the development of alveolar edema in premature monkeys at risk for hyaline membrane disease, Am Rev Respir Dis 143:865-871, 1991. American Thoracic Society.) blood across the patent ductus arteriosus to the descending aorta (see Chapter 52). ORIGIN AND COMPOSITION OF SURFACTANT Pulmonary surfactant is composed of approximately 90% lipids and 10% proteins. The main phospholipid in surfactant is dipalmitoylphosphatidylcholine (DPPC), also known as lecithin. It is surface active because of its hydrophilic head and hydrophobic tails (Figure 46-7). However, DPPC by itself does not adsorb efficiently at the air-liquid interface and is in the form of a gel at body temperature. The presence of some unsaturated phospholipids and cholesterol helps to make it more fluid (Mingarro et al, 2008). The remainder of surfactant’s ingredients is shown in Figure 46-8. One of them, phosphatidylglycerol (PG), is sometimes used as a marker of lung maturation; it interacts with the hydrophobic surfactant proteins to improve biophysical activity. Even minor components of pulmonary surfactant play important roles; for instance, free fatty acids improve the stability of the interfacial film, especially after repeated compression. The composition of surfactant is complex because it has evolved to balance the need The cardinal clinical signs of RDS are tachypnea, grunting, and increased work of breathing (Box 46-1). The newborn respiratory rate is elevated in an attempt to increase the exchange of oxygen and carbon dioxide, but with exhaustion, the rate may decline or even stop. Grunting is used to create positive pressure in the lungs to reduce the collapse of air sacs. Signs of increased work of breathing include nasal flaring and retraction of respiratory muscles, especially the intercostal and subcostal muscles. Because the ribcage in premature infants is so flexible, the sternum may deeply retract during inspiration. Cyanosis results from inadequate oxygenation, and pallor from acidosis due to poor elimination of carbon dioxide. The combination of increased work of breathing, cyanosis, and acidosis causes lethargy and disinterest in feeding, and eventually apnea. Rather than progressing through these signs during the first hours of life, newborns with intrapartum asphyxia or extreme prematurity may present with apnea immediately following birth. On auscultation, breath sounds may be distant or shallow from the fast inspiratory rate and low tidal volume, and fine inspiratory rales may be heard due to reopening of moist, collapsed air sacs. The onset of symptoms is always within hours after birth and, in severe cases, may occur with the first few breaths after delivery. In general, the respiratory distress from RDS tends to get worse over the first 1 to 3 days after birth, and then usually improves gradually over a few days (although the natural course may be interrupted by exogenous surfactant therapy). Differential Diagnosis Clinical improvement during first 12 hours after birth suggests TTNB, and onset after the first 24 hours suggests pneumonia and sepsis. Tachypnea without increased work of breathing suggests cyanotic heart disease; the diagnosis of obstructed pulmonary veins from total 637 CHAPTER 46 Respiratory Distress in the Preterm Infant Phosphatidylglycerol Other phospholipids SP-A 5% SP-B 2% SP-C 2% SP-D 1% 8% Neutral Lipids 8% 8% 6% Unsaturated Phosphatidylcholine 20% Saturated Phosphatidylcholine 50% FIGURE 46-8 Composition of surfactant. (From Jobe AH, Ikegami M: Biology of surfactant, Clin Perinatol 28:655-669, 2001, WB Saunders.) Alveolar type II cell Alveolar type I cell SP-B SP-A Lamellar body Tubular myelin Surfactant layer SP-C Air space Alveolar fluid Alveolar macrophage FIGURE 46-9 Assembly of surfactant. (Republished and adapted from Hawgood S, Clements JA: Pulmonary surfactant and its apoproteins, J Clin Invest 86:1-6, 1990. The American Society for Clinical Investigation.) anomalous venous return is occasionally confused with RDS. Hypoventilation without increased work of breathing suggests a central nervous system problem such as intracranial hemorrhage or asphyxia. Asymmetric breath sounds may be due to pneumothorax (which is a complication of RDS), congenital diaphragmatic hernia, or unilateral pleural effusion. Meconium staining of amniotic fluid suggests the possibility of meconium aspiration syndrome, but this is rare in premature infants—green-stained amniotic fluid in this population is more likely to be due to infection or to bile refluxed into the esophagus because of intestinal obstruction, rather than meconium. LABORATORY FEATURES OF RDS Initially, the arterial blood gases or oxygen saturation monitor may show only hypoxemia or desaturation. The Paco2 may be normal because of tachypnea, but is almost FIGURE 46-10 Structure and interactions of surfactant and proteins. (From Pérez-Gil J: Molecular interactions in pulmonary surfactant films, Biol Neonate 81:6-15, 2002. Karger AG Basel.) BOX 46-1 Clinical Signs of RDS Tachypnea Grunting Increased work of breathing Nasal flaring Retraction of respiratory muscles (intercostal, subcostal, sternal) Cyanosis Pallor Lethargy Disinterest in feeding Apnea 638 PART X Respiratory System always elevated. Later, as the infant tires, the Paco2 will rise further and cause respiratory acidosis. With imminent respiratory failure, there may be metabolic acidosis due to inadequate oxygen delivery to tissues, and from poor peripheral perfusion due to respiratory acidosis. Differential diagnosis: Extremely elevated Paco2 within minutes of birth suggests pulmonary hypoplasia, tension pneumothorax, congenital diaphragmatic hernia, or obstruction of the airways due to debris or an anatomic cause. The tachypneic, cyanotic newborn with low Paco2 may have transient tachypnea of the newborn or cyanotic congenital heart disease. A positive blood culture suggests pneumonia and sepsis. Low blood glucose (<40) suggests symptomatic hypoglycemia, and high hematocrit (>65) suggests symptomatic polycythemia. RADIOGRAPHIC FEATURES OF RDS The classic radiographic findings of RDS include a reticulogranular (i.e., ground-glass) pattern and air bronchograms (Figure 46-11). The lungs are diffusely and homogeneously dense because of widespread collapse of alveoli. The appearance is reticular (i.e., netlike) because the small airways are open (black) and surrounded by interstitial and alveolar fluid (white); in severe cases, the lungs may appear complete white on the film. Air bronchograms are commonly seen because the large airways beyond the second or third generation are more visible than usual as a result of radiodensity from engorged peribronchial lymphatics and fluid-filled or collapsed alveoli. Another cardinal feature is low lung volume (e.g., the diaphragms are at the eighth rib level or higher) due to widespread alveolar collapse and low functional residual capacity. Differential diagnosis: Normal or high lung volumes, especially with prominent interstitial fluid pattern, suggest TTNB. In this case, there are coarse white lines (engorged lymphatics and interstitial water) radiating from the hilum rather than the crisp black lines (air bronchograms) of RDS. Other causes of a coarse (rather than diffuse) fluid pattern include pneumonia with sepsis, and obstructed pulmonary venous drainage due to total anomalous pulmonary venous return. An abnormal cardiac silhouette or size should suggest congenital heart disease, and asymmetry of the lungs suggests pneumothorax, congenital diaphragmatic hernia, or lung anomaly. Very low lung volumes, especially with pneumothorax, may indicate pulmonary hypoplasia. INITIAL DIAGNOSTIC EVALUATION FOR POSSIBLE RDS The initial diagnostic evaluation (Box 46-2) is shaped by the differential diagnoses listed in Table 46-1, and by the urgency to intervene quickly for serious but infrequent conditions such as bacterial pneumonia and sepsis, hypoglycemia, and polycythemia. It is often helpful to obtain both anteroposterior and lateral radiographs for the initial evaluation. If congenital heart disease is suspected on clinical grounds, an echocardiogram is indicated. TREATMENT OF PATIENTS WITH RDS The treatment of patients with RDS is outlined in Box 46-3. All patients with RDS need the basics of warmth, hydration, and nutrition appropriate for the degree of prematurity, as described in other chapters. Because pulmonary edema contributes to surfactant dysfunction in RDS, it is important to avoid excessive intravenous (IV) fluid administration. Newborns with significant tachypnea (e.g., more than 60 breaths per minute) or increased work of breathing BOX 46-2 Initial Diagnostic Evaluation for Possible RDS Arterial or capillary blood gas Blood glucose Complete blood count Blood culture AP and lateral chest radiograph Echocardiogram (if clinically indicated) BOX 46-3 Treatment of Established RDS FIGURE 46-11 Radiograph of RDS. Note low lung volumes and reticulogranular pattern. (From Welty S, Hansen TN, Corbet A: Respiratory distress in the preterm infant. In Taeusch HW, Ballard RA, Gleason CA, editors: Avery’s diseases of the newborn, ed 8, Philadelphia, 2005, Elsevier.) Warmth by radiant warmer or incubator Hydration at approximately 60–80 mL/kg/d Nutrition Initially D5W or D10W (with protein, if possible) NPO if respiratory rate over 60 or moderate/severe work of breathing Gavage feeds if stable Consider parenteral nutrition if enteral feeds are delayed Antibiotics if at risk for pneumonia and sepsis Supplemental oxygen Oxygen saturation monitoring, with appropriate target for infants at risk for retinopathy of prematurity Exogenous surfactant CPAP or mechanical ventilation, as needed CPAP, Continuous positive airway pressure. CHAPTER 46 Respiratory Distress in the Preterm Infant (moderate or severe) often do not have the energy required for oral feeding, and there is some risk of aspiration if nipple feeding is attempted. Initially, IV fluids and nothing by mouth (NPO) status may be appropriate, with consideration of small gavage tube gastric feedings if the baby is otherwise stable. Parenteral nutrition may be indicated because of the increased caloric expenditures associated with work of breathing. Antibiotics should be considered unless the risk of pneumonia and sepsis is negligible. The techniques of respiratory support are described in other chapters. See Chapter 45 regarding risks and benefits of different modes of respiratory therapy. RESPIRATORY COMPLICATIONS OF RDS Air leak complications occur in patients with RDS because of the asymmetry of alveolar inflation and the sheer stresses in terminal bronchioles, leading to dissection of air into the interstitium (causing pulmonary interstitial emphysema) and through the visceral pleura (causing pneumothorax). The former can be seen in up to 50% of patients with RDS, and the latter in 5% to 10%, even in those treated with exogenous surfactant. Hemorrhagic pulmonary edema (i.e., pulmonary hemo­ rrhage) occurs more frequently in the most premature infants, probably due to left ventricular failure and excessive left-to-right flow through a patent ductus arteriosus, with resultant disruption of the pulmonary capillaries (Cole et al, 1973). It may also be related to insufficient attention to reducing mechanical ventilator settings after lung compliance improves following exogenous surfactant treatment. Onset is typically at 1 to 3 days of age, with sudden respiratory deterioration associated with pink or red frothy fluid in the ET tube, and widespread white-out on chest radiograph. Treatment considerations include increased positive end-expiratory pressure (PEEP); there is also evidence that exogenous surfactant may be beneficial (Pandit et al, 1995). Bronchopulmonary dysplasia (BPD), also known as chronic lung disease, is probably due to abnormal lung repair following lung injury from RDS (see Chapter 48). PREVENTION OF RDS Box 46-4 outlines the key measures for prevention and treatment of RDS. Antenatal Steroids Antenatal steroids were shown in 1972 to be effective in reducing the risk of RDS (Liggins and Howie, 1972). A consensus panel convened by the National Institutes of BOX 46-4 Prevention of RDS Antenatal steroids Prevention of asphyxia Continuous positive airway pressure Exogenous surfactant Surfactant, then CPAP CPAP, Continuous positive airway pressure. 639 Health concluded that “antenatal corticosteroid therapy is indicated for women at risk of premature delivery with few exceptions and will result in a substantial decrease in neonatal morbidity and mortality, as well as substantial savings in health costs” (NIH Consensus Panel, 1995). They recommended that the treatment be used from 24 to 34 weeks’ gestation without limitation by gender or race, and whether or not surfactant therapy was available. Although the beneficial effects were found to be greatest if treatment was begun more than 24 hours before delivery, there was also a benefit when given for less than 24 hours. A metaanalysis in 2006 (Roberts and Dalziel, 2006) reviewed 21 published trials and concluded that antenatal corticosteroids did not increase a mother’s risk of death, chorioamnionitis, or puerperal sepsis. Treatment was associated with a 31% reduction in neonatal death, 34% reduction in RDS, 46% reduction in cerebroventricular hemorrhage, and 54% reduction in necrotizing enterocolitis (all significant at p <0.05 or less). Although there is now widespread consensus on use of antenatal steroids, many issues remain controversial, including the type of corticosteroid to use; the dose, frequency, and timing of use; and the route of administration. A metaanalysis in 2008 (Brownfoot et al, 2008) reviewed 10 trials and found that antenatal dexamethasone appears to decrease the incidence of intraventricular hemorrhage but possibly increases the rate of NICU admissions, compared with betamethasone. Oral antenatal dexamethasone was found in one study to increase the incidence of neonatal sepsis compared to intramuscular drug (Egerman et al, 1998). More research is needed to optimize antenatal steroid administration. Because the effectiveness appears to wane if antenatal steroids are given more than 1 week before premature delivery, several trials have been conducted to determine whether one or more repeat doses at weekly intervals was beneficial. Although a metaanalysis (Crowther and Harding, 2007) found modest improvements in occurrence and severity of RDS, there were concerning findings of smaller birthweight and head circumference, indicating that there is insufficient evidence to recommend repeated doses (Bonanno and Wapner, 2009). Prevention of Asphyxia Prevention of asphyxia may decrease the incidence and severity of RDS, because asphyxia leads to hypoxemia and acidosis, which reduce surfactant synthesis. Leakage of fluids from capillaries into alveoli may also impair surfactant function. Thus, delivery by cesarean section should be considered for signs of fetal distress if the fetus is deemed to be viable, or if the fetus is in the breech presentation during labor. Compliance with consensus neonatal resuscitation techniques outlined by the American Academy of Pediatrics and American Heart Association (Kattwinkel, 2006) is critical, especially because premature infants are at much higher risk for needing intervention at birth. Maternal transfer to a center experienced in management of premature infants, if it can be accomplished safely, is associated with improved neonatal outcome. 640 PART X Respiratory System Continuous Positive Airway Pressure Gregory et al (1971) first introduced continuous positive airway pressure (CPAP) for newborns with RDS, primarily by endotracheal tube. Since then, a variety of devices have been developed, including short nasal prongs that do not add much to the work of breathing (De Paoli et al, 2002). Avery et al (1987) reported that centers that used early nasal CPAP for RDS had a lower incidence of chronic lung disease. The goals of CPAP are to prevent end-expiratory alveolar collapse, reduce the work of breathing, and better match ventilation to perfusion. If started in the delivery room, it may help the newborn establish functional residual capacity, in addition to stabilizing the chest wall and reducing airway resistance. Furthermore, adequate expansion of the lungs at birth improves pulmonary blood flow. Some have speculated that prophylactic use of CPAP in the delivery room might make intubation for exogenous surfactant unnecessary (De Klerk and De Klerk, 2001), and that avoidance of intubation and mechanical ventilation may lower the risk of BPD (see Chapter 48). The COIN Trial randomized 25- to 28-weeks’ gestation newborns who were breathing spontaneously at 5 minutes of age either to CPAP or to prophylactic intubation with early extubation to CPAP if possible. The investigators found that 46% of the CPAP group eventually required intubation anyway during the first 5 days, and the rate of pneumothorax was 9% compared to only 3% in those intubated routinely (p <0.001) (Morley et al, 2008). This suggests that routine intubation may be safer than immediate CPAP in the extremely premature infant, but it also demonstrated that some patients in this extremely highrisk population do well with early CPAP. Success with CPAP, especially in this population, requires considerable experience and skill (Bohlin et al, 2008). Recently, some centers have begun using humidified high-flow nasal cannulas in infants with RDS (Lampland, 2009) to achieve the benefits of positive airway pressure with fewer of the perceived disadvantages of CPAP (challenges in keeping the nasal prongs in the nares, easier handling of the patient, and less risk of pressure necrosis of the nasal septum). However, the pressure generated from this therapy, although proven to produce clinical effect, is variable, unpredictable, and unregulated, and the commercially available systems are not approved by the U.S. Food and Drug Administration (FDA) for this indication. It should only be used by practitioners aware of the balance of risks and benefits, and those prepared to recognize and treat pneumothorax and respiratory failure. The use of CPAP following tracheal administration of exogenous surfactant is discussed later in the section on timing of surfactant administration. Exogenous Surfactant Historical Summary The development of exogenous surfactant for treatment of RDS is one of the most important advances in the history of newborn medicine. This history is well told elsewhere (e.g., Halliday, 2008). The key milestones include von Neergaard’s discovery in 1929 that surface tension contributes to lung recoil, Gruenwald’s demonstration in 1947 that lungs of stillborn infants have high surface tension, Prattle’s speculation in 1955 that absence of surfactant active material contributes to RDS, Clement’s description in 1957 of surfactant dysfunction in experimental animals, and Avery and Mead’s demonstration in 1959 that RDS in human infants is due to surfactant deficiency. There was an increase in research interest and funding for RDS treatment after U.S. President Kennedy’s son, born at 34 weeks’ gestation, died of RDS in 1963. In 1972, Enhorning and Robertson used natural surfactant to delay the progression of RDS in preterm rabbits, and in 1980 Fujiwara et al demonstrated the first successful use of exogenous surfactant in human infants. By 1990, exogenous surfactant was widely used throughout the developed world, and many large clinical trials have been conducted since then to refine and improve surfactant treatment and prevention of RDS. Compared to standard therapy without surfactant, a metaanalysis of 13 randomized controlled trials suggests that animal-derived exogenous surfactant reduces the risk of pneumothorax by 58%, pulmonary interstitial emphysema by 55%, mortality by 32%, and the combined outcome of BPD or death by 17% (Seger and Soll, 2009). Types of Surfactant Available A list of commonly used surfactant preparations can be found in Table 46-2. The first generation of commercially available artificial surfactants (e.g. Exosurf [colfosceril]) was composed mainly of DPPC and did not have SP-B or SP-C. However, a metaanalysis of 11 randomized controlled trials showed that natural surfactants were faster acting than artificial surfactants with lower incidence of pneumothorax and mortality (Soll and Blanco, 2001). TABLE 46-2 Surfactant Preparations and Their Sources Brand Name Generic Name Constituents Nonprotein Synthetic Surfactants Adsurf Pumactant (ALEC) DPPC, PG Exosurf Colfosceril palmitate DPPC Protein-Containing Animal Surfactants Curosurf Poractant α Porcine lung tissue Alveofact SF-RI 1 Bovine lung lavage BLES Bovine lipid extract surfactant Bovine lung lavage Infasurf Calactant CLSE Bovine (calf) lung lavage Surfacten Surfactant-TA Bovine lung homogenate Survanta Beractant Bovine lung tissue Peptide-Containing Synthetic Surfactants Venticute rSP-C surfactant DPPC, POPG, PA, rSP-C Surfaxin Lucinactant DPPC, POPG, PA, KL4 Adapted from Sinha S, Moya F, Donn SM: Surfactant for respiratory distress syndrome: are there important clinical differences among preparations? Curr Opin Pediatr 19:150-154, 2007. ALEC, Artificial lung-expanding compound; DPPC, dipalmitoylphosphatidylcholine; PG, phosphatidylglycerol; rSP-C, recombinant human SP-C; POPG, palmitoyl-oleoyl phosphatidylglycerol; PA, palmitic acid. CHAPTER 46 Respiratory Distress in the Preterm Infant Clinical trials comparing the available natural surfactants have been inconclusive. Speer et al (1995) showed no differences in a small study, whereas Baroutis et al (2003) showed differences in days of intubation, oxygen use, and duration of hospitalization, but not survival. A comparison of Infasurf and Survanta by Bloom and Clark (2005) was stopped because of lack of statistical significance in the primary outcome variable. Ramanathan et al (2004) compared Curosurf and Survanta, finding that the former was less costly because of fewer doses, but there were no important clinical differences in outcome. A new generation of synthetic surfactants is under development to lower cost and because of theoretical concerns about immunologic or infectious complications from animal-derived surfactant. These new products have synthetic peptides or proteins, such as KL4 in lucinactant, which mimics the actions of natural surfactant-associated proteins SP-B and SP-C. Lucinactant has been shown to be superior to older synthetic surfactants (Sinha et al, 2005), but its superiority to animal surfactants has not been conclusively proven (Moya et al, 2005). In addition, there are uncertainties regarding the metabolic fate of lucinactant and its component chemicals, and there may be risks from the requirement to convert the lucinactant gel into liquid by using a special warming cradle immediately before instillation (Engle et al, 2008); the authors added that “efforts to develop more effective and safer surfactant formulations continue to be warranted because of concerns with animal-derived surfactants for transmission of microbes, exposure to animal proteins and inflammatory mediators, susceptibility to inactivation, and inconsistent content.” Surfactant Selection The important issue for clinical providers and managers is which surfactant to stock in their pharmacy, because of the expense of the preparations and the need to standardize dosing guidelines. Compared to first-generation synthetic surfactants, current studies strongly support use of natural surfactants because of reduced risk of air leak and more rapid response to treatment. Among natural surfactants that are commercially available, no differences in longterm outcomes have been conclusively proven. However, poractant alfa (Curosurf) when given as prophylaxis with an initial dose of 200 mg/kg appears to have some advantages when compared to beractant (Survanta) and calfactant (Infasurf), such as fewer repeat doses, quicker oxygen weaning, and potentially less cost. The benefit may be related to the larger amount of phospholipid in each dose. For hospitals currently using beractant and calfactant, the advantages of poractant alfa must be considered in light of the logistical issues and potential drug errors when a hospital changes from one preparation to another, especially when there may be better products on the horizon. routinely intubated. The treatment appeared feasible and safe, but the sample size (23) was too small to prove that the approach was beneficial compared to surfactant instillation via endotracheal tube during the first minutes after delivery. Prophylactic surfactant generally refers to the administration of doses within the first 15 minutes after birth. To avoid giving an unnecessary, expensive medication and subjecting a newborn to the risks of tracheal intubation, prophylactic endotracheal surfactant should ideally be given only to patients who would have eventually developed RDS and met treatment criteria for surfactant anyway. However, in practice, the clinician can only select those patients at highest risk and for whom clinical trials indicate that the benefits of prophylactic surfactant outweigh the risks. One such population is the baby born at <30 weeks’ gestation whose mother was not given antenatal steroids (Table 46-3). In the highest risk populations, a metaanalysis of eight clinical trials suggests that prophylactic surfactant reduces mortality (by 39%), frequency and severity of RDS, pneumothorax (by 38%), and the combined outcome of BPD and death, compared to infants who received placebo or rescue surfactant (Soll and Morley, 2001). However, the criteria for selecting the high risk infants remain controversial. Table 46-3 provides one practical approach for the clinician who must make a decision in the delivery room as to whether to begin surfactant immediately or to wait. However, some centers manage even 25-week gestational newborns with initial CPAP, with only one third eventually needing surfactant (Aly et al, 2004). A randomized study of 27 to 31 weeks’ gestation newborns with signs of early RDS demonstrated that infants intubated for surfactant during the first hour of life followed immediately by extubation and CPAP had a lower rate of subsequent ventilation, incidence of pneumothorax, and chronic lung disease compared to newborns treated with early CPAP alone (Rojas, 2009). This observation may be especially important for hospitals not equipped to provide mechanical ventilation for this population. Early rescue surfactant (usually defined as treatment at 1 to 2 hours after birth) may be indicated for infants of <30 weeks’ gestation at the first signs of RDS. The clinical trials showing superiority of prophylactic surfactant over rescue (Soll and Morley, 2001) compared prophylaxis to late rescue (needing mechanical ventilation and >40% Fio2 and typically treated 4 to 6 hours after birth). TABLE 46-3 Sample Algorithm for When to Begin Exogenous Surfactant When to Give First Dose of Surfactant Populations to Consider Prophylactically, within 15 minutes after birth <26 weeks’ gestation and 26 to 30 weeks if (a) no antenatal steroids or (b) needs intubation anyway Early rescue, preferably in first 60 minutes after birth <30 weeks’ gestation at first signs of RDS Treatment of established RDS, preferably started within 12 hours after birth All newborns with established RDS, regardless of gestational age, if they need ventilator and at least 30–40% O2 Timing of Surfactant Administration In theory, surfactant would ideally be given with the first breath. This concept was evaluated by Kattwinkel et al (2004) by delivering the head of the infant, suctioning the nasopharynx, and then instilling calfactant into the airway before delivery of the shoulders. CPAP was initiated immediately after delivery, but the trachea was not 641 642 PART X Respiratory System One clinical trial showed clinical advantages of prophylactic compared to early-rescue surfactant (Kattwinkel et al, 1993) but a historically controlled study showed that aggressive use of CPAP led to a large reduction in the need for any surfactant at all and was associated with lower rates of BPD and other morbidity (De Klerk and De Klerk, 2001). Furthermore, immediate intubation in the delivery room may be associated with significant procedure morbidity, including apnea and bradycardia. Thus, there is no compelling evidence favoring prophylactic over early rescue surfactant for preterm infants at the time of this writing, except in those patients at highest risk (see Table 46-3) such as extremely preterm infants whose mothers did not receive antenatal steroids (Sweet and Halliday, 2009). Late rescue surfactant: Clinical trials have firmly established the benefits of surfactant compared to no surfactant in preterm infants with established RDS (Engle et al, 2008). However, as noted previously, prophylactic or early rescue surfactant is more beneficial than late rescue in the highest-risk populations. In lower-risk populations (e.g., near-term infants) it is likely that the late rescue approach offers a better balance of benefit, risk of intubation, and cost efficiency than treatment before or at the earliest onset of RDS. Entry criteria in clinical trials of rescue surfactant compared to no surfactant usually required that RDS be severe enough to require mechanical ventilation and at least 30 to 40% supplemental oxygen. Many of the trials were conducted before nasal CPAP was widely used, so the indications for intubating and administering surfactant in infants tolerating CPAP are unknown. However, those with a requirement for very high concentration of supplemental oxygen (e.g., >80%) or with sustained severe hypercarbia (e.g., Paco2 >60 torr) will likely benefit. Combining surfactant therapy with CPAP: Because of advances in use of CPAP since the first clinical trials of surfactant, some have advocated immediate extubation to CPAP after surfactant dosing (Verder et al, 1999) to avoid complications from the endotracheal tube (including excessive tidal volumes and airway inflammation that may lead to BPD). Stevens et al (2007) reviewed six randomized controlled clinical trials of this treatment approach in infants with clinical signs of RDS. The “early rescue/ CPAP” group was intubated for early surfactant therapy followed by immediate extubation to nasal CPAP. In comparison to the late rescue group, early rescue/CPAP was associated with a 33% reduction in need for mechanical ventilation, 48% reduction in air leak syndrome, and 49% reduction in incidence of BPD. They also noted that initiation of surfactant treatment at an Fio2 of less than 0.45 was associated with lower incidence of air leak, compared to waiting until the Fio2 was higher. Methods of Surfactant Dosing Animal studies suggest that surfactant distribution is better distributed when administered as a bolus rather than by infusing it slowly over several minutes (Fernandez et al, 1998). The package inserts of many surfactant preparations recommend moving the infant into multiple positions for better distribution of surfactant, but there are no clinical trials supporting one particular method. Therefore, surfactant should be given as quickly as tolerated, and with the least disruptive infant positioning. Exogenous surfactant is generally given via an endotracheal tube. However, tracheal intubation is potentially risky, and sometimes difficult to accomplish, so administration through a laryngeal mask airway (LMA) may be considered. In a pilot study of 8 preterm infants (body weight range 880 to 2520 g) receiving nasal CPAP for RDS, the LMA administration approach appeared to be safe and potentially effective (Trevisanuto et al, 2005). Number of Surfactant Doses and Dosing Intervals A metaanalysis of two clinical trials that compared one versus multiple doses of animal-derived surfactant suggests a 49% reduction in incidence of pneumothorax and a trend toward a 37% reduction in mortality when multiple doses are used (Soll and Ozek, 2009). There are no data to suggest that dosing should continue once the patient’s ventilator and oxygen requirement are at minimal levels, or beyond four doses. The interval between doses is usually at least 6 hours, and most protocols recommend discontinuation of dosing after 48 hours. In an attempt to reduce the incidence of BPD, clinical trials of surfactant administration at several days of age are underway. Clinical Care after Dosing Because natural surfactants may work quickly, the clinician must be prepared after dosing to immediately lower the Fio2 while carefully monitoring the pulse oximeter. The tidal volume, as measured by the ventilator and/or by careful observation of chest wall movement, may gradually increase, resulting in a need to lower inspiratory pressures to avoid air leak syndrome, lung injury, and possibly pulmonary hemorrhage. Blood gases should be monitored by transcutaneous monitors and/or intermittent sampling. PEEP should be maintained but may be reduced if the starting levels at dosing were high, given that functional residual capacity increases shortly after surfactant administration. A poor response to exogenous surfactant may occur because the patient does not have surfactant deficiency but rather, lung hypoplasia, pneumonia, or congenital heart disease. Other causes for a lack of response may be poor distribution of the surfactant, such as administration down the right stem bronchus due to malposition of the endotracheal tube, plugging of the tube, or malposition of the tube in the esophagus. A less likely reason is an inadequate dose of surfactant. The rapid improvement in lung compliance after exogenous surfactant therapy may lead to excessive pulmonary blood flow from left-to-right shunting from a patient ductus arteriosus (Raju and Langerberg, 1993). It is uncertain whether early and aggressive intervention to close the ductus with medication or surgery will reduce the risk of pulmonary hemorrhage. PULMONARY HYPOPLASIA DEFINITION AND INCIDENCE One or both lungs of newborns with pulmonary hypoplasia are smaller than normal, including reduced numbers of lung cells, airways, blood vessels, and alveoli. CHAPTER 46 Respiratory Distress in the Preterm Infant Pulmonary hypoplasia is the cause of respiratory distress at birth in only 0.3% of newborns with respiratory symptoms (Rubaltelli et al, 1998), but it is commonly fatal, especially in preterm infants. The incidence is about 1 per 1000 live births, and 90% are associated with congenital anomalies or pregnancy complications (Churg et al, 2005). There is a continuum of severity of pulmonary hypoplasia—from negligible to severe—and therefore all but the most severe cases are difficult to diagnose. This is especially true when attempting to diagnose prenatally by imaging studies, but is also true during clinical assessment of the newborn, and even after postmortem examination of the lungs. When considering the causes of pulmonary hypoplasia, it is useful to categorize them into associated conditions, each with representative diagnoses, as in Table 46-4. This chapter on respiratory distress in preterm infants focuses on the problem of pulmonary hypoplasia associated with oligohydramnios from preterm premature rupture of membranes (PPROM) because this cause of pulmonary hypoplasia occurs almost exclusively in preterm infants. When PPROM occurs at 15 weeks’ gestation, the incidence TABLE 46-4 Categories of Conditions Associated With Pulmonary Hypoplasia Category Representative Diagnoses Restriction of thoracic space Diaphragmatic hernia or eventration Intrathoracic mass Congenital cystic adenomatoid malformation Bronchogenic cyst Extralobar sequestration Thoracic neuroblastoma Pleural effusions Chylothorax Hydrothorax Oligohydramnios Renal Bilateral renal agenesis or dysplasia Bladder outlet obstruction (­posterior urethral valves) Nonrenal Prolonged preterm rupture of membranes Skeletal anomalies Chondroectodermal dysplasia Osteogenesis imperfecta Thanatophoric dwarfism Hydrops fetalis Rh isoimmunization Neuromuscular and central nervous system anomalies Fetal akinesia Anencephaly Arnold-Chiari malformation Cardiac anomalies Hypoplastic right or left heart Pulmonary stenosis Ebstein’s anomaly Abdominal wall defects Omphalocele Gastroschisis Syndromes Trisomy 13, 18, 21 Larsen’s Cerebrocostomandibular Jarcho-Levin Roberts Lethal multiple pterygium Adapted from Churg AM, Myers JL, Tazelaar H, Wright J, editors: Thurlbeck’s pathology of the lung, ed 3, New York, 2005, Thieme. 643 of pulmonary hypoplasia is 80% and at 19 weeks it is 50%, whereas after 26 weeks it is near zero (Rothschild et al, 1990). Other conditions associated with pulmonary hypoplasia are covered in other chapters. PATHOLOGY The easiest method of defining pulmonary hypoplasia during postmortem exam is to calculate the ratio of lung weight to body weight. At 28 weeks’ gestation, a ratio of 0.015 is at the 5th percentile, whereas at 35 weeks’, the 5th percentile is at a ratio of 0.012. The lung-to-body weight ratio may be artificially elevated if the lungs are wetter than usual from edema, hemorrhage, inflammation, or lymphangiectasia—this may lead to the false conclusion that the patient does not have pulmonary hypoplasia. Conversely, the ratio may be artificially low if the body is heavier than usual because of renal cystic disease, hydrops, ascites, tumors, hydrocephaly, and so forth. Therefore, a better method for postmortem diagnosis of pulmonary hypoplasia is to measure the lung volume by inflating the lung with fixative at physiologic pressure and then measuring the displacement of fluid when the lung is immersed. Pulmonary hypoplasia assessed by this method is defined as lung volume less than the 5th percentile of standards for each gestational age (Churg et al, 2005). This method facilitates comparison of postmortem to in utero estimates of lung volume made during antenatal imaging. However, low lung volume does not necessarily correlate with deficiency of lung structure and function. A postmortem technique more physiologically relevant than lung volume is the radial alveolar count, which is proportional to alveolar surface complexity (Churg et al, 2005) and thus gas exchange surface area; however, this procedure is complex and time-consuming. Impairment of lung development before 16 weeks causes reduced airway branching, reduced cartilage development, reduced acinar complexity and maturation, delayed vascularization, and delayed thinning of the air-blood barrier. Impairment after 16 weeks typically causes reduced acinar complexity and maturation. These outcomes are predictable, given the time in gestation when these structures are developing (Figure 46-12, A). Because the growth of lung blood vessels parallels the development of the airways, pulmonary hypoplasia causes decreased total size of the pulmonary vascular bed, decreased number of vessels per unit of lung tissue, and increased pulmonary arterial smooth muscle. This last phenomenon accounts for persistent pulmonary hypertension after birth. CLINICAL SIGNS The preterm newborn with pulmonary hypoplasia often presents with immediate signs of respiratory distress and cyanosis indistinguishable from those in the newborn with severe respiratory distress syndrome (see earlier section on clinical signs of RDS). However, respiratory failure from severe pulmonary hypoplasia often becomes apparent within minutes of birth, whereas respiratory failure from RDS usually progresses over the first few hours after birth. The thorax may appear small or bell-shaped 644 PART X Respiratory System and, if oligohydramnios was severe, there may be flattening of the face and deformation (such as contractures of the extremities). Hypercarbia may be severe on the earliest blood gas measurement, despite aggressive mechanical ventilation. The hypoxemia from surfactant deficiency and lung immaturity may be compounded by right-to-left shunting of deoxygenated blood due to pulmonary hypertension, leading to severe desaturation. If there is also a tension pneumothorax, there may be asymmetry of breath sounds or malposition of heart sounds, as well as impaired cardiac output from impaired venous return to the thorax. bronchograms, as with RDS, but the lung volumes may be smaller and the diaphragms higher than in RDS. With severe pulmonary hypoplasia, the thorax may appear bellshaped, and early pneumothorax is common. TREATMENT After birth, treatment for pulmonary hypoplasia is similar to that provided to patients with RDS, including assisted ventilation and exogenous surfactant. Even with cautious ventilation, however, tension pneumothoraces are common and are often the proximate cause of death, so the neonatal team should be prepared for urgent needle decompression of the chest and insertion of a chest tube. Permissive hypercapnia is appropriate, and high-frequency ventilation may be necessary for adequate ventilation and removal of very high arterial Pco2 levels. If pulmonary RADIOGRAPHIC SIGNS Because lung immaturity and surfactant deficiency accompany pulmonary hypoplasia, particularly in preterm infants, the lungs may be radiographically dense with air cell lineage lung cell fate Embryonic Pseudoglandular Canalicular Saccular Alveolar lung bud airway branching capillarization sac complexity alveoli Type I & Type II cells surfactant production tracheo-esophageal septation neuroendocrine, basal ciliated & secretory cells cartilage & smooth muscle 0.0 4.0 8.0 Human Weeks Gestation 16.0 26.0 0.0 9.0 12.0 Mouse Days Gestation 16.5 17.5 A 36.0 Birth Birth 2.0 Postnatal-Years 5.0 Postnatal-Days 30.0 Total Lung Volume—Gestational Age Ratio 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 B Nonsurvivor Survivor C1 C2 FIGURE 46-12 A, Stages of lung development. B, Measurement of fetal lung volume by magnetic resonance imaging and outcomes after delivery. C, Lung growth after occlusion of the fetal trachea for 6 days: (C1) before tracheal occlusion and (C2) after tracheal occlusion. (A from Kimura J, Deutsch GH: Key mechanisms of early lung development, Pediatr Dev Pathol 10:335-347, 2007. Allen Press, Inc. C from Kohl T, Geipel A, Tchatcheva K, et al: Life-saving effects of fetal tracheal occlusion on pulmonary hypoplasia from preterm premature rupture of membranes, Obstet Gynecol 113:480-483, 2009, Lippincott Williams and Wilkins.) CHAPTER 46 Respiratory Distress in the Preterm Infant hypertension is present, as it commonly is, inhaled nitric oxide for pulmonary vasodilation is logical, but clinical trials with sufficient sample size to prove better survival are lacking (Chock et al, 2009). Extracorporeal membrane oxygenation may be appropriate for larger preterm infants if the pulmonary hypoplasia, surfactant deficiency, and pulmonary hypertension are expected to improve within a few days. However, it is very difficult to select appropriate candidates with pulmonary hypoplasia for extracorporeal membrane oxygenation, and it is often unsuccessful with severe disease or in combination with other major lifethreatening malformations. PRENATAL DIAGNOSIS An accurate prenatal test for pulmonary hypoplasia is important because it may affect the obstetric management. It would be particularly helpful to be able to discriminate lethal from nonlethal pulmonary hypoplasia, especially before 24 weeks’ gestation when termination of the pregnancy may be an option. This is an easier task than quantifying the degree of fetal pulmonary hypoplasia, which—as noted previously—is challenging even for the pathologist. A metaanalysis of clinical parameters for prediction of fatal pulmonary hypoplasia following prelabor rupture of membranes before 28 weeks’ gestation concluded that gestational age at rupture was a better predictor of pulmonary hypoplasia than latency or oligohydramnios (van Teeffelen et al, 2010). In 1992, ultrasound measurement of the fetal chest circumference was shown to be helpful in predicting fatal pulmonary hypoplasia (Ohlsson et al, 1992), whereas in 2002 the best prediction was achieved by combining clinical, biometric, and Doppler parameters (Laudy et al, 2002). These parameters included thoracic, cardiac, and abdominal circumference, the largest vertical amniotic fluid pocket, and pulsed Doppler measurements of the arterial pulmonary branches. Nonsurvivors with pulmonary hypoplasia due to fetal urinary anomalies were noted to have lower in utero lung volumes as determined by fetal MRI, adjusted for gestational age (Zaretsky et al, 2005), but there was no clear separation between the survivors and nonsurvivors (Figure 46-12, B). Furthermore, there was considerable overlap of the confidence intervals before 26 weeks’ gestation, which limits the usefulness of MRI assessment of fetal volume for prenatal counseling. Threedimensional ultrasound measurements of lung volume had better diagnostic accuracy for predicting pulmonary hypoplasia, compared to two-dimensional measurements of thoracic/heart area ratio (Gerards et al, 2008). Although there will continue to be advances in fetal imaging to assess lung volume, it will likely remain difficult to differentiate lethal from nonlethal pulmonary hypoplasia. The management of the preterm pregnancy with premature rupture of membranes near the limit of fetal viability is beyond the scope of this chapter. However, a recent review article by Waters and Mercer (2009) offers a management algorithm that illustrates the areas of controversy. For the patient who desires conservative management of premature rupture before the limit of viability, the algorithm recommends serial assessments for signs of infection or labor, as well as interval ultrasound examinations to 645 watch for the development of pulmonary hypoplasia. If this occurs, antenatal antibiotics and corticosteroids should be considered if more aggressive management is desired. PRENATAL TREATMENT Treatment of oligohydramnios with amnioinfusion with saline has been proposed to improve fetal survival by increasing the latency period (the interval between PROM and delivery) and thus the gestational age at delivery. In a small study, it appeared to reduce the incidence of pulmonary hypertension from about 50% to 10% with PPROM in the 20th week of gestation (De Carolis et al, 2004). However, the procedure restores adequate amniotic fluid volume in only a minority of patients in which it is performed, and there are procedure-related complications such as chorioamnionitis and placental abruption (Tan et al, 2003). When it is successful, the benefit may come from restoration of back-pressure from the amniotic sac fluid to the lungs, stimulating fetal lung growth and development during the critical canalicular stage between 16 and 26 weeks’ gestation. Based on this same logic, there have been efforts to reverse pulmonary hypoplasia in utero by fetal tracheal occlusion. In a case report of PPROM at 16 weeks’ gestation, fetoscopic tracheal balloon occlusion was performed at 27 6⁄7 weeks (Kohl et al, 2009). Within 6 days, the fetal lung volume as measured by MRI increased from 13 to 70 mL (Figure 46-12, C) and the lung blood flow normalized. The fetus was delivered at 28 6⁄7 weeks using the EXIT procedure (ex utero–intrapartum treatment), had no signs of pulmonary hypertension, and was discharged home at 8 weeks of age. This approach has been used for fetuses with congenital diaphragmatic hernia (Done et al, 2008), but with mixed results. Whether this technique proves to be valuable for PPROM is uncertain, but the restoration of lung size in only 6 days suggests that pulmonary hypoplasia may be more reversible than previously assumed. PNEUMONIA INCIDENCE The incidence of pneumonia/sepsis in preterm infants with birthweight 1500 to 2500 g is only 0.28%, whereas in patients with birthweight <1000 g the incidence is severalfold higher at 1.9% (see Figure 46-1) (Rubaltelli et al, 1998). If only newborns with respiratory distress are considered, the overall incidence (in mostly term infants) is 5% (see Table 46-1)—the third most likely cause after RDS (46%) and TTNB (37%). The data in Figure 46-1 indicate that the incidence of pneumonia in newborns with respiratory distress who are <1000 g, 1000 to 1500 g, and 1500 to 2500 g is 4%, 2%, and 1%, respectively. The NICHD Neonatal Research Network reported an incidence of blood culture– proven early-onset sepsis (<72 hours after birth) in newborns of birthweight <1500 g of 1.9% during 1991–1993 (Stoll et al, 1996) and 1.5% during 1998–2000 (Stoll et al, 2003). In the earlier period, group B streptococcus was the most frequent pathogen (31%) followed by Escherichia coli (16%) and Haemophilus influenzae (12%). In the more recent period, the most common pathogens were E. coli 646 PART X Respiratory System (44%), group B streptococcus (11%), coagulase-negative staphylococcus (11%), viridans streptococci and other streptococci (8%), H. influenzae (8%), Citrobacter (2%), Listeria monocytogenes (2%), and Candida albicans (2%). CLINICAL SIGNS The clinical signs of pneumonia after preterm birth are often indistinguishable from the more common problem of RDS. Bacterial pneumonia is usually accompanied by sepsis because newborns are frequently unable to confine bacteria to the lung, and therefore some infants will exhibit clinical signs of sepsis or shock, including poor perfusion and hypotension, in addition to respiratory failure. Many premature infants with pneumonia also have surfactant deficiency from RDS, further obscuring the diagnosis of pneumonia. LABORATORY AND RADIOGRAPHIC SIGNS Blood cultures will be positive in some premature newborns with pneumonia, but the presence of maternal antibiotics in the blood of the newborn reduces confidence in a negative result. Leukopenia, increased percentage of immature granulocytes, and elevated inflammatory markers such as C-reactive protein increase the likelihood of sepsis/pneumonia, but with poor positive predictive value. Tracheal aspirate culture (but not Gram stain) obtained immediately after placement of an endotracheal tube may help with diagnosis and guide therapy, especially when the blood culture is negative (Booth et al, 2009). Premature newborns with pneumonia will have pulmonary infiltrates on chest radiograph, but the radiographic appearance is difficult to distinguish from RDS (although classically in term newborns it has a more coarse and wet appearance). Because newborns are unable to localize pulmonary infection, lobar infiltrates are rarely an indication of pneumonia—plugging of airways with secretions is more likely. TREATMENT Antibiotics directed at the most common organisms (see earlier discussion) should be started immediately when pneumonia in the preterm infant is suspected. Ampicillin and gentamicin are reasonable choices, to be administered for 48 hours pending culture results. Empiric vancomycin has the disadvantage of promoting the emergence of vancomycin-resistant organisms, and a delay in treating coagulase-negative staphylococcal bacteremia until cultures are positive is rarely of consequence to the patient. If the blood culture is negative, and the mother has been pretreated with antibiotics, a longer course of antibiotics (e.g., 5 to 7 days) may be prudent, especially if there are laboratory abnormalities such as elevated C-reactive protein and a clinical course suggestive of sepsis or pneumonia. Because it is so difficult to tell which preterm infants with respiratory distress have pneumonia and which have the more common problems of RDS or TTNB, one approach is to empirically treat with antibiotics all premature newborns with respiratory symptoms. The likelihood of infection in newborns <1000 g with respiratory distress is about 4% (see Incidence, earlier), and so 24 extremely premature newborns will be needlessly treated for every one who will benefit. This number can be reduced by avoiding empiric antibiotics at birth for newborns who are prematurely delivered for maternal indications, such as hypertension. Both PROM and premature labor increase the risk of infection. Among the possible adverse consequences of unnecessary empiric antibiotics are interference with the colonization of the intestinal tract with nonpathogenic bacteria, selection of antibiotic-resistant bacteria, and fungal infection. On the other hand, the development of bronchopulmonary dysplasia is associated with inflammation from chorioamnionitis (Speer, 2009), and so antibiotics even in the absence of frank pneumonia may be beneficial. SUGGESTED READINGS Engle WA: American Academy of Pediatrics Committee of Fetus and Newborn: Surfactant-replacement therapy for respiratory distress in the preterm and term neonate, Pediatrics 121:419-432, 2008. Halliday HL: Surfactants: past present and future, J Perinatol 28:S47-S56, 2008. Lampland AL, Plumm B, Meyers PA, et al: Observational study of humidified highflow nasal cannula compared with nasal continuous positive airway pressure, J Pediatr 154:177-182, 2009. Rojas MA, Lozano JM, Rojas MX, et al: Colombian Neonatal Research Network: Very early surfactant without mandatory ventilation in premature infants treated with early continuous positive airway pressure: a randomized, controlled trial, Pediatrics 123:137-142, 2009:2009. Seger N, Soll R: Animal derived surfactant extract for treatment of respiratory distress syndrome, Cochrane Database Syst Rev 2:CD007836, 2009. Sinha S, Moya F, Donn SM: Surfactant for respiratory distress syndrome: are there important clinical differences among preparations, Curr Opin Pediatr 19:150154, 2007. Soll R, Ozek E: Multiple versus single doses of exogenous surfactant for the prevention and treatment of neonatal respiratory distress syndrome, Cochrane Database Syst Rev 1:CD000141, 2009. Sweet DG, Halliday HL: The use of surfactants in 2009, Arch Dis Child Educ Pract Ed 94:78-83, 2009. Waters TP, Mercer BM: The management of preterm premature rupture of the membranes near the limit of fetal viability, Am J Obstet Gynecol 201:230-240, 2009. Complete references used in this text can be found online at www.expertconsult.com C H A P T E R 47 Respiratory Failure in the Term Newborn Thomas A. Parker and John P. Kinsella The evaluation and management of respiratory failure in the term newborn poses unique challenges and remains one of the most vexing problems facing clinicians in the newborn intensive care unit. Although some of the pathophysiologic features of respiratory failure in the term infant are similar to the premature newborn condition, several disorders occur more commonly in the term newborn (e.g., meconium aspiration) and are often made more difficult to evaluate and manage because of cardiac and pulmonary vascular abnormalities that complicate the clinical course. Indeed, the traditional perspective of categorizing hypoxemia and respiratory failure in the term newborn as cardiac, pulmonary vascular, or due to air-space (lung) disease is insufficient. For example, the syndrome of persistent pulmonary hypertension of the newborn (PPHN) is defined by severe pulmonary vasoconstriction leading to suprasystemic pulmonary artery pressure with extrapulmonary right-to-left venoarterial admixture across the fetal channels of the oval foramen and the arterial duct. However, PPHN rarely occurs without concomitant parenchymal lung disease and disturbances in cardiac performance. In this chapter we present an algorithm for evaluation of the term newborn with hypoxemia and respiratory failure, review the syndrome of PPHN, and discuss common causes of respiratory failure in the term newborn. EVALUATION OF THE TERM NEWBORN WITH HYPOXEMIA/ RESPIRATORY DISTRESS One of the most anxiety-provoking experiences for many clinicians (particularly those in training) is the initial evaluation and management of a term newborn with hypoxemia/respiratory distress. Traditional textbooks provided a wealth of information about individual conditions once identified. However, there are few sources designed to guide the clinician in an ordered fashion through a comprehensive diagnostic evaluation. In this section, we propose an approach to the evaluation of the hypoxemic newborn that may be useful in clarifying the etiology of hypoxemia/respiratory distress and in determining the proper sequence of diagnostic and therapeutic interventions. HISTORY Marked hypoxemia in the newborn can be caused by parenchymal lung disease with V/Q mismatch or intrapulmonary shunting, pulmonary vascular disease causing extrapulmonary right-to-left shunting (PPHN), or anatomic right-to-left shunting associated with congenital heart disease. Evaluation should begin with the history and assessment of risk factors for hypoxemic respiratory failure. Relevant history may include the results of prenatal ultrasound studies. Lesions such as congenital diaphragmatic hernia (CDH) and congenital cystic adenomatoid malformation are diagnosed prenatally with increasing frequency. Although many anatomic congenital heart defects can be diagnosed prenatally, vascular abnormalities (e.g., coarctation of the aorta, total anomalous pulmonary venous return) are more difficult to diagnose with prenatal ultrasound. A history of a structurally normal heart by fetal ultrasonography should be confirmed by echocardiography in the newborn with cyanosis (see later). Other historical information that may be important in the evaluation of the cyanotic newborn includes a history of severe and prolonged oligohydramnios causing pulmonary hypoplasia. Also important is a history of prolonged fetal bradyarrhythmia and/or tachyarrhythmia and marked anemia (caused by hemolysis, twin-twin transfusion, or chronic hemorrhage) that may cause congestive heart failure, pulmonary edema, and respiratory distress. Maternal illness (e.g., diabetes mellitus), medication use (e.g., aspirin or medications containing nonsteroidal antiinflammatory drugs causing premature constriction of the ductus arteriosus, association of Ebstein’s malformation with maternal lithium use), and illicit drug use may contribute to acute cardiopulmonary distress in the newborn. Risk factors for infection that cause sepsis/pneumonia should be considered, including premature or prolonged rupture of membranes, fetal tachycardia, maternal leukocytosis, uterine tenderness, and other signs of intraamniotic infection. Events at delivery may provide clues to the etiology of hypoxemic respiratory failure in the newborn. For example, if positive-pressure ventilation is required in the delivery room, the risk of pneumothorax increases. A history of meconium-stained amniotic fluid, particularly if meconium is present below the cords, is the sine qua non of meconium aspiration syndrome. Birth trauma (e.g., clavicular fracture, phrenic nerve injury) or acute fetomaternal or fetoplacental hemorrhage may cause respiratory distress in the newborn (Box 47-1). PHYSICAL EXAMINATION The initial physical examination provides important clues to the etiology of cyanosis. Marked respiratory distress in the newborn (retractions, grunting, nasal flaring) suggests the presence of pulmonary parenchymal disease with decreased lung compliance. However, it is important to recognize that upper airway obstruction (e.g., Pierre Robin sequence or choanal atresia) and metabolic acidemia also can cause severe respiratory distress. In contrast, the newborn with cyanosis alone or cyanosis plus tachypnea (i.e., nondistressed tachypnea) typically has cyanotic congenital heart disease, most commonly transposition of the great vessels (TGV) or idiopathic PPHN. 647 648 PART X Respiratory System BOX 47-1 N  eonatal Respiratory Failure: History and Risk Factor Assessment PRENATAL Prenatal ultrasound study results History of oligohydramnios and duration History of fetal brady/tachyarrhythmia Maternal illnesses, drugs , medications History of fetal distress Risk factors for infection DELIVERY History of positive pressure ventilation in DR Meconium stained amniotic fluid Hemorrhage Birth trauma Low Apgar score The presence of a heart murmur in the first hours of life is an important sign in the newborn with cyanosis or respiratory distress. In this setting, it is unusual for the common left-to-right shunt lesions (patent ductus arteriosus, atrial septal defect, ventricular septal defect) to produce an audible murmur because pulmonary vascular resistance (PVR) remains high and little turbulence is created across the defect. A murmur that sounds like a ventricular septal defect in the first hours of life is most commonly caused by tricuspid regurgitation (associated with PPHN or an ischemic myocardium). The response to supplemental oxygen can also provide important clues to the pathophysiology of hypoxemic respiratory failure in the term newborn (Boxes 47-2 and 47-3). BOX 47-2 Physical Examination RESPIRATORY DISTRESS (RETRACTIONS, GRUNTING, NASAL FLARING) Suggests lung parenchymal disease (compliance), upper airway disease or metabolic acidemia NO SIGNIFICANT RESPIRATORY DISTRESS (TACHYPNEA ALONE) Suggests hypoxemia caused by cyanotic heart disease without lung disease BOX 47-3 A  cute Response to Supplemental Oxygen (High Fio2 by Hood, Mask) MINIMAL OR TRANSIENT CHANGE IN SA o2 Cyanotic heart disease, PPHN MARKED IMPROVEMENT IN SA o2 Parenchymal lung disease, CHD with ductal-dependent systemic blood flow TABLE 47-1 Role of Pulse Oximetry in Evaluation of Neonatal Hypoxemic Respiratory Failure Preductal Sao2 = postductal Sao2 1. Intrapulmonary shunt: PVR < SVR 2. Cyanotic congenital heart disease with L→R PDA: Ductal-dependent Qp: pulmonary atresia/stenosis, tricuspid atresia, Ebstein’s anomaly 3. PPHN:R→L shunt at PFO: PVR > SVR, ductus closed Preductal Sao2 > postductal Sao2 1. PVR > SVR with R→L PDA: PPHN: MAS, RDS, CDH 2. Ductal-dependent Qs: HLHS, IAA, coarctation 3. Anatomic PV disease: alveolarcapillary dysplasia