© 2003 Wiley-Liss, Inc.

Birth Defects Research (Part A) 67:193–201 (2003)

Guidelines

Guidelines for Case Classification for the National Birth

Defects Prevention Study
Sonja A. Rasmussen,1* Richard S. Olney,1 Lewis B. Holmes,2 Angela E. Lin,2 Kim M. Keppler-Noreuil,3
Cynthia A. Moore,1 and the National Birth Defects Prevention Study
1
National Center on Birth Defects and Developmental Disabilities, Centers for Disease Control and Prevention, Atlanta, Georgia
2
Genetics and Teratology Unit, Pediatric Service, Massachusetts General Hospital, Boston, Massachusetts
3
Department of Pediatrics, Division of Medical Genetics, University of Iowa Hospitals and Clinics, Iowa City, Iowa
Received 13 May 2002; Accepted 29 October 2002

BACKGROUND: Previous studies have suggested that etiologic heterogeneity may complicate epidemiologic analyses designed to identify
risk factors for birth defects. Case classification uses knowledge of embryologic and pathogenetic mechanisms to make case groups more
homogeneous and is important to the success of birth defects studies. METHODS: The goal of the National Birth Defects Prevention Study
(NBDPS), an ongoing multi-site case– control study, is to identify environmental and genetic risk factors for birth defects. Information on
environmental risk factors is collected through an hour-long maternal interview, and DNA is collected from the infant and both parents for
evaluation of genetic risk factors. Clinical data on infants are reviewed by clinical geneticists to ensure they meet the detailed case
definitions developed specifically for the study. To standardize the methods of case classification for the study, an algorithm has been
developed to guide NBDPS clinical geneticists in this process. RESULTS: Methods for case classification into isolated, multiple, and
syndrome categories are described. Defects considered minor for the purposes of case classification are defined. Differences in the approach
to case classification for studies of specific defects and of specific exposures are noted. CONCLUSIONS: The case classification schema
developed for the NBDPS may be of value to other clinicians working on epidemiologic studies of birth defects etiology. Consideration of
these guidelines will lead to more comparable case groups, an important element of careful studies aimed at identifying risk factors for birth
defects. Birth Defects Research (Part A) 67:193–201, 2003. © 2003 Wiley-Liss, Inc.

INTRODUCTION coded diagnoses), methods of diagnosis, laboratory results,

and relevant exposures or family history, as well as the
Birth defects are a leading cause of infant mortality in the study clinical geneticist’s assessment of whether these
United States (Hoyert et al., 2001), yet the causes of most findings represent a recognized pattern of malformation, is
birth defects are unknown (Nelson and Holmes, 1989). The entered into a centralized clinical database.
National Birth Defects Prevention Study (NBDPS) is a The etiologic heterogeneity of birth defects has long been
large, ongoing case– control study, sponsored by the Cen- recognized (Holmes et al., 1976; Khoury et al., 1982; Martin
ters for Disease Control and Prevention (CDC) and de- et al., 1983; Murray et al., 1985; Jones, 1988; Cunniff et al.,
signed to identify genetic and environmental factors im- 1990; Ferencz, 1993). A single defect type, such as spina
portant in the etiology of birth defects (Yoon et al., 2001).
bifida, may be caused by a chromosome abnormality, a
This study, based in eight birth defects surveillance sys-
single-gene condition, or a teratogenic exposure, or may be
tems located in Arkansas, California, Iowa, Massachusetts,
of unknown cause. Etiologic heterogeneity may complicate
New Jersey, New York, Texas, and metropolitan Atlanta,
epidemiologic studies designed to identify causes of birth
Georgia (CDC), includes collection of data on many poten-
defects (Friedman, 1992; Khoury et al., 1992a,b). Isolated
tial exposures through maternal interview and collection
birth defects have been shown to be epidemiologically and
of biological specimens for study of possible genetic sus-
most likely etiologically distinct from defects associated
ceptibility and gene– environment interaction. Infants with
over 30 types of major congenital defects are included in
the study (Table 1). Since the commencement of the study †
This article is a US government work and, as such, is in the public domain in
in October 1997, each site has contributed approximately the United States of America.
300 cases and 100 controls to the study per year. As of *Correspondence to: Sonja A. Rasmussen, MD, MS, 4770 Buford Highway, NE,
August 15, 2002, 12,190 cases and 5034 controls have been MS-F45, Centers for Disease Control and Prevention, Atlanta, GA 30341.
E-mail: [email protected]
entered into the study. Clinical information on each infant, Published online in Wiley InterScience (www.interscience.wiley.com).
including all major and minor defects (both verbatim and DOI: 10.1002/bdra.10012
194 RASMUSSEN ET AL.
Table 1 analysis into separate groups, based on whether the defect
Birth Defects Eligible for Inclusion in the National is isolated, one of multiple congenital anomalies, or asso-
Birth Defects Prevention Study ciated with a syndrome, to make them presumably more
homogeneous. In other studies, case classification may al-
Birth defect low for infants with anatomically different, but presumed
Anencephaly, craniorachischisis pathogenetically similar, defects to be combined to in-
Spina bifida crease the power of a study. An example of this is com-
Encephalocele, cranial meningocele, encephalomyelocele bining infants with defects of presumed vascular etiology
Holoprosencephaly for study (Martin et al., 1992; Van Allen, 1992). NBDPS
Hydrocephalus
clinical geneticists have developed a system of terminology
Dandy-Walker malformation
Anophthalmia, microphthalmia and case classification guidelines, adapted from the work
Cataracts, glaucoma and related eye defectsa of others (Spranger et al., 1982; Khoury et al., 1994a; Jones,
Anotia, microtia 1997), to standardize the methods of case classification for
Conotruncal heart defects the study. It should be acknowledged that some of the
Single ventricle decisions made in developing these case classification
Septal heart defects (atrial septal defects, ventricular septal guidelines were arbitrary; however, we believe that it is
defects) important that methods of case classification be as well-
Atrioventricular septal heart defects defined as possible, so that the process is uniform among
Ebstein malformation
different clinical geneticists. It is inevitable that some may
Obstructive heart defects (right and left ventricular outflow tract
defects) wish to classify cases differently; we believe this is appro-
Anomalous pulmonary venous return priate as long as details on how the case classification was
Heterotaxia done are provided and the process is consistent within the
Choanal atresia particular study. We are hopeful that the approach delineated
Cleft lip ⫹/⫺ palate here may be helpful to other clinicians involved in case clas-
Cleft palate sification for epidemiologic studies of birth defects.
Esophageal atresia ⫹/⫺ tracheoesophageal fistula
Intestinal atresia/stenosis
Biliary atresia CLASSIFICATION FOR STUDIES OF
Hypospadias, second or third degree SPECIFIC DEFECTS
Renal agenesis/hypoplasia
Exstrophy, bladder In the epidemiologic study of specific birth defects for
Exstrophy, cloacal possible risk factors, classification of infants involves two
Limb deficiency, intercalary issues: 1) Does the infant have the defect of interest as an
Limb deficiency, longitudinal isolated defect, as one of multiple congenital anomalies, or
Limb deficiency, transverse as a component of a syndrome? We use the term “syn-
Limb deficiency, not elsewhere classified drome” here to refer to a recognizable pattern of multiple
Craniosynostosis
Diaphragmatic hernia
malformation that is known or presumed to have a specific
Sacral agenesis cause (e.g., a single-gene condition, chromosome abnor-
Omphalocele mality, or teratogenic exposure) (Khoury et al., 1994b); and
Gastroschisis 2) based on what is known about the pathogenesis of the
Amnion rupture sequence defect of interest, is further classification warranted?
a Given the complexity of the process of determining
Included in the study as of January 1, 2000.
whether an infant has the defect of interest as an isolated
defect, as one of multiple congenital anomalies or as a
component of a syndrome, a stepwise approach may be
advantageous and is summarized in Figure 1. This process
with additional major defects. For example, isolated neural requires that the reviewer have specific training in a mech-
tube defects were more often observed in females and anistic approach to birth defects and be familiar and up-
Caucasians, but these associations were not seen for neural to-date with the birth defects, genetics and dysmorphology
tube defects associated with other major defects (Khoury et literature; thus, we suggest that case classification is best
al., 1982). In addition, different risk factor associations carried out by a clinical geneticist/dysmorphologist when
have been noted in isolated and multiple cases (Khoury et possible. If unavailable, a clinician with experience in birth
al., 1989). For example, a protective effect of periconcep- defects and the availability of a clinical geneticist/dysmor-
tional multivitamin use was found for isolated conotruncal phologist for consultation on complicated cases, especially
heart defects, but not for those associated with other non- those with multiple defects, will be adequate for analyses
cardiac defects or with a recognized syndrome (Botto et al., of certain defects.
1996). Inclusion of infants with different causes in the
study of a birth defect may dilute the magnitude of an
observed association toward the null (Khoury et al., 1992a).
Does the Infant Have at Least One Defect that
Thus, the process of case classification is important to the Meets Case Definition Criteria for the NBDPS?
success of epidemiologic studies of birth defects. To maximize the usefulness of the data, case definitions
The goal of case classification is to use knowledge of have been standardized for the study, and clinical infor-
embryologic and pathogenetic mechanisms to make case mation on each infant is evaluated by a clinical geneticist
groups used for analysis more comparable (Khoury et al., located at each site before inclusion in the study. These
1994a; Martinez-Frias et al., 1990, 1991, 2000). In some case definitions include information on eligibility criteria
studies, infants with the same defect will be classified for (e.g., infants must have Type II , III, or IV microtia [Meur-

Birth Defects Research (Part A) 67:193–201 (2003)

 CASE CLASSIFICATION IN BIRTH DEFECTS STUDIES 195

Figure 1. A summary of the process of deter-

mining whether an infant has the defect of in-
terest as an isolated defect, as one of multiple
congenital anomalies, or as component of a
syndrome. Please refer to text for details on
decision points.

man, 1957] to be included in the study as having anotia/ Has a Single-Gene Condition or Chromosome
microtia), methods of diagnosis (e.g., cardiac defects must Abnormality Been Diagnosed?
be diagnosed by echocardiography, catheterization, sur-
gery, or autopsy to be included in the study), and essential Because the focus of the NBDPS is on cases of unknown
clinical information to be abstracted verbatim from medi- etiology, infants with genetic syndromes (single-gene con-
cal records (e.g., information on other birth defects that ditions or chromosome abnormalities) are excluded from
frequently accompany the birth defect of interest). Al- the study. In the case of chromosome abnormalities, results
though the specific case definitions developed for the of chromosome analysis (karyotype or fluorescence in situ
NBDPS may not be appropriate for other birth defects hybridization [FISH] analysis) to support the diagnosis
studies, the importance of a careful, well-characterized must be available. For single-gene conditions, only infants
case definition to studies of birth defects should be empha- with single-gene conditions documented in the medical
sized. record are excluded. The clinical reviewer must determine

Birth Defects Research (Part A) 67:193–201 (2003)

196 RASMUSSEN ET AL.
if the stated diagnosis is consistent with the defects de- Table 2
scribed and was made by a qualified professional, based on Examples of Case Classification
the medical record data available. These genetic syn-
dromes must be related to the defect, as opposed to being Example Case classification
“additive.” A defect can be described as additive to a Isolated diaphragmatic hernia, Isolated (isolated major defect)
syndrome if the defect has not been described previously no other major or minor
in association with the syndrome, and has no known or defects
plausible connection with the phenotype (e.g., galac- Cleft lip, left ear pit, mild Isolated (one major defect ⫹
tosemia with limb deficiency). hydronephrosis two minors)
Tetralogy of Fallot, atrial Isolated (two defects involving
septal defect the same organ system)
Is an Exposure to a Known Teratogen Present and Spina bifida, talipes Isolated (several defects
Is/Are Observed Defect(s) Strongly Associated equinovarus, hip constituting a sequence
dislocation, hydrocephalus [spina bifida is primary
with this Exposure? defect]- -infant should be
Infants with defects believed to be related to a terato- excluded from studies of
genic exposure (e.g., sacral agenesis in a baby whose other defects)
mother had diabetes mellitus) are included in the NBDPS. Cleft lip, transposition of Multiple (two major, unrelated,
great vessels specific defects)
One reason for including these infants is that they offer an
Esophageal atresia with Multiple (three major defects
opportunity to study genetic factors that may contribute to tracheoesophageal fistula, consistent with VACTERL
the observed outcome (Buehler et al., 1994). In some anal- hemivertebrae, imperforate association; may wish to
yses, an infant with defects that are associated strongly anus analyze separately from
with a specific teratogenic exposure (e.g., an infant with other multiple cases)
anotia or microtia with maternal retinoic acid [Accutane] Holoprosencephaly associated Syndrome (chromosome
exposure) (Lammer et al., 1985) could be classified as hav- with Trisomy 13 abnormality)
ing a teratogenic syndrome and excluded from specific Interrupted aortic arch Syndrome (chromosome
investigations, depending on the analysis being carried associated with 22q11.2 abnormality)
deletion
out. We recommend that infants with defects that have a
Cleft palate associated with Syndrome (autosomal
weaker association with a specific exposure (e.g., an infant Stickler syndrome dominant)
with cleft lip with maternal phenobarbital exposure)
(Arpino et al., 2000) not be excluded. Instead, they should
be classified as having isolated or multiple defects, de-
pending on additional defects present in that infant. To define minor defects for NBDPS case classification
purposes, lists of minor defects collected from previous
How Many Major Defects Are Present? sources (Marden et al., 1964; Hook et al., 1976; Leppig et al.,
1987; Cohen, 1997; Chambers et al., 2001) were reviewed.
Several types of cases should be classified as “isolated.”
Table 3 delineates the minor defects agreed upon by
Infants who have only a single major defect should be
NBDPS clinical geneticists. Although an attempt has been
classified as having an isolated defect; however, the con-
made to make this list as complete as possible, clinical
verse is not true. Classification of an infant with more than
judgment will be necessary for its use. This list is also
one major defect must be based on information of known
somewhat arbitrary, because some of the defects included
embryologic and pathogenetic mechanisms. Infants that
as minor may, at times, be of surgical, medical, or serious
should be classified as having an isolated defect include
cosmetic importance, or may be believed to be mild man-
those with a single major defect with additional minor
ifestations of a major defect (e.g., cleft uvula and cleft
defects in the absence of a defined syndrome; with a major
palate) (Frias and Carey, 1996). It is important, however, to
defect accompanied by other major defects in the same
designate a standard group of minor defects for an epide-
organ, organ system or body part; and with a major defect
miologic study; deviations from this list are acceptable but
accompanied by other pathogenetically related defects (Ta-
should be noted.
ble 2) (Khoury et al., 1994a).
Most epidemiologic studies of birth defects have concen- Are All Major Defects of the Same Organ, Organ
trated on major defects, that is, those that have surgical, System or Body Part?
medical, or serious cosmetic importance. One reason for
this is that ascertainment of minor anomalies has not been Often, a major defect is accompanied by other related
standardized (Lechat and Dolk, 1993) for birth defects major defects. In some infants, these defects affect the same
surveillance programs that focus on abstraction of inpa- organ, organ system, or body part. Some examples include
tient records. Minor defects are known to be important in syndactyly and split hand deficiency of the same limb,
the study of birth defects, however, because they often multiple cardiac defects, esophageal atresia and tracheo-
may accompany, and serve as an indication of, a syndrome esophageal fistula, and multiple neural tube defects, with
of known etiology (Frias and Carey, 1996). In addition, the other examples listed elsewhere (Khoury et al., 1994a).
presence of three or more minor anomalies has been shown Because these defects are believed to be embryologically
predictive of the presence of major malformations (Leppig and pathogenetically related, we classify infants with these
et al., 1987). Because of their frequent occurrence in babies defects as having an isolated defect.
with major defects, we classify infants with a single major
defect accompanied by any number of minor defects as Are All Major Defects Related Pathogenetically?
having an isolated defect, assuming that a recognized syn- Sometimes a major defect is accompanied by other major
drome is not present. defects of a different organ, organ system, or body part,

Birth Defects Research (Part A) 67:193–201 (2003)

 CASE CLASSIFICATION IN BIRTH DEFECTS STUDIES 197
but the pattern of structural defects can be attributed to a tified by MRI in individuals presumed to have isolated,
primary problem in morphogenesis that leads to a cascade nonsyndromic cleft lip or palate (Nopoulos et al., 2001,
of consequent defects. This pattern of defects is termed a 2002).
“sequence” (Spranger et al., 1982; Jones, 1997). In many When a defect of interest is accompanied by at least one
instances, the occurrence of one defect is thought to pre- additional unrelated, major and specified defect and the
cede and directly influence the occurrence of one or more etiology of the defects is unknown, we recommend that the
additional defects. Examples include spina bifida leading infant should be classified as having multiple defects
to the sequence defects talipes, hydrocephalus and axial (Khoury et al., 1994a) (Table 2). The term “unrelated”
skeleton malformations, and severe micrognathia leading refers to defects in different body parts or systems and not
to the sequence defects glossoptosis and cleft palate. In part of a sequence, as discussed previously. The term
other instances, the error in morphogenesis seems to have “major” refers to the exclusion of minor defects, discussed
been earlier, involving cells or tissues that will ultimately above and listed in Table 3. The defect also must be “spec-
form more than one, often contiguous, body structure. ified” or adequately described. This excludes defects that
Examples include hemifacial microsomia with defects of are not well-delineated (e.g., ear defect, malformed limbs)
ear, jaw, and oral structures, and holoprosencephaly with and often coded as “not otherwise specified” (NOS). In-
defects of the brain, midface, and oral structures. In both of fants with genetic syndromes of known etiology should be
these situations, we classify infants with these combina- excluded from this group (see below).
tions of defects as having an isolated defect because there
is one “primary” defect (primary refers to the earliest Does the Reviewer Strongly Suspect a Genetic
defect in morphogenesis) (Jones, 1997).
Syndrome of Known Etiology?
Infants in whom a chromosome abnormality or single-
Is the Defect of Interest Primary or Secondary? gene disorder is suspected by the study clinical geneticist,
Another important issue raised by these situations is but not identified by clinicians who examined the infant,
identification of the group in which the defect should be have been included in the study. Depending on the ana-
analyzed. Clinical information should be evaluated to de- lytic study planned, these infants may be excluded by the
termine if the defect under study is primary, or whether clinical geneticist involved in an analysis. For example, a
the defect of interest is presumed to be secondary to an- stillborn infant with holoprosencephaly and polydactyly in
other defect. For example, infants with meningomyelocele whom chromosome analysis was not carried out may be
often also have clubfoot secondary to the neural deficit excluded from a study of risk factors of holoprosen-
related to the lesion (Jones, 1997). We believe these infants cephaly, given the suspicion that the infant may have
would be more appropriately analyzed for etiologic risk Trisomy 13 or Pseudotrisomy 13 (postulated to be autoso-
factors with other infants with meningomyelocele, rather mal recessive) (Lurie and Wulfsberg, 1993). Exclusions
than with infants with clubfoot, because the clubfoot is such as these, however, should be specifically noted in the
believed to be secondary to the meningomyelocele. Some- study methods.
times, the selection of an appropriate analysis group is A syndrome classification might also be considered in
more apparent (e.g., an infant with holoprosencephaly and the absence of a definitive diagnosis in the case of a posi-
midline cleft lip should be analyzed with other infants with tive family history. Because the etiology of most isolated
holoprosencephaly, and not with infants with cleft lip), but birth defects is believed to be multifactorial, increased risk
other times determining the appropriate analysis group among relatives is expected, but the magnitude of recur-
can be challenging. For example, the appropriate analysis rence risk does not approach that of a single-gene disorder.
group for an infant with hemifacial microsomia consisting For some defects, however, the relative contribution of
of microtia, mandibular hypoplasia, and cleft lip and pal- single-gene disorders may be notably higher. For example,
ate is not as clear. These infants may be excluded from the congenital cataracts are often inherited in an autosomal
analysis or analyzed separately, if sufficient numbers of dominant manner (Francis et al., 2000); thus, an infant with
infants with these phenotypes are available. congenital cataracts whose parent also had congenital cat-
We suggest that sequence designation should be limited aracts could be classified as having a syndrome (presumed
to those defects that occur as a consistent, frequent finding autosomal dominant single-gene condition), even if the
with the primary defect (e.g., spina bifida and clubfoot). In particular single-gene condition had not been identified. In
some instances, a sequence may be suspected, but the contrast, a family in which both a child and his parent have
finding is not consistent or frequent and could represent a cleft lip (not an infrequent occurrence) would not be
unrelated malformations. For example, in an infant with a classified as having a syndrome because the relative con-
large omphalocele and clubfeet, one could postulate that tribution of single-gene disorders to clefting is low. Several
the clubfeet are part of a sequence, related to constricted issues should be taken into account when considering a
movement as the result of the space-occupying lesion. positive family history, including whether the family his-
Because clubfeet rarely accompany omphalocele, however, tory is consistent with a specific type of inheritance (e.g.,
the clinical geneticist should not presume that this is a autosomal dominant), the degree of relationship between
sequence; instead, the infant should be classified as having the proband and the affected family members, and the
multiple defects. relative contribution of single-gene disorders to the defect.
In considering an infant as having an isolated defect, it Positive family history does not necessarily imply genetic
should be noted that the defects identified in a child may etiology. Other causes of a positive family history include
be time-dependent, because some defects may not be rec- shared environmental exposure and, for common defects,
ognized until later in life or may be dependent on addi- coincidence in large families. Information about how in-
tional studies, such as echocardiography or brain imaging fants with positive family history were classified should be
studies. For example, brain abnormalities have been iden- provided in the study methods. Improved understanding

Birth Defects Research (Part A) 67:193–201 (2003)

 Table 3
Defects Considered to be Minor, Prematurity-Related, or Nonstructural for Case Classification Purposes of the
National Birth Defects Prevention Study
Anomaly Case* Anomaly Case*
Congenital anomalies of the brain and Ventricular hypertrophy (right or left) 746.886
nervous system Thickened cardiac valve 746.900
Absent septum pellucidum 742.280 Heart murmur 746.990
Hydrocephalus secondary to 742.385 Other congenital anomalies of the circulatory
intraventricular hemorrhage (IVH) system
Macrocephaly 742.400 Patent ductus arteriosus 747.000
Congenital anomalies of the eye Peripheral pulmonic stenosis (PPS) 747.325
Iris coloboma 743.430 Single umbilical artery 747.500
Iris freckles 743.440 Congenital anomalies of the respiratory system
Blue sclera 743.450 Small nares 748.180
Ptosis 743.600 Anteverted nares 748.180
Absence of eyelashes 743.630 Notched or hypoplastic alae nasi 748.180
Long eyelashes 743.630 Flat or wide nasal bridge 748.180
Weakness of eyelids 743.630 Congenital laryngeal stridor 748.360
Fused eyelids 743.630 Hypoplasia of lung (in premature infants) 748.510
Short palpebral fissures 743.635 Cleft palate and cleft lip
Stenosis or stricture of lacrimal duct 743.650 Bifid uvula 749.080
Exophthalmos 743.800 Cleft gum 749.190
Epicanthal folds 743.800 Other congenital anomalies of the upper
Epicanthus inversus 743.800 alimentary tract
Upward or downward slanting 743.800 Tongue-tie 750.000
palpebral fissures Microglossia 750.110
Brushfield spots 743.800 Macroglossia 750.120
Epibulbar dermoid cyst 743.810 Aberrant frenula 750.180
Congenital anomalies of the ear, face, and High arched palate 750.240
neck Angular lip pits 750.260
Preauricular appendage, tag, or lobule 744.110 Short/long columella 750.270
Ear tag 744.120 Thin vermilion border 750.270
Large ears 744.200 Smooth philtrum 750.270
Misshapen ears 744.220, 744.280 Broad alveolar ridge 750.280
Crumpled ears 744.230 Pylorospasm 750.500
Protruding ears 744.230 Other congenital anomalies of the digestive
Small ears (excludes true microtia) 744.230 system
Lack of helical fold 744.230 Meckel diverticulum 751.010
Thickened or overfolded helix 744.230 Rectal fissure 751.580
Absent tragus 744.230 Hepatomegaly 751.620
Asymmetric sized ears 744.230 Congenital anomalies of the genital organs
Darwinian tubercle 744.230 Imperforate hymen 752.430
Double lobule 744.230 Fusion of vulva 752.440
Bridged concha 744.230 Prominent clitoris 752.450
Ear lobe crease 744.230 Embryonal cyst of vagina 752.460
Low-set ears 744.245 Cyst of vagina, vulva, or canal of Nuck 752.470
Posteriorly rotated ears 744.246 Vaginal or hymenal tags 752.480
Narrow external auditory meatus 744.280 Hypoplastic labia majora 752.480
Ear pit 744.410 Hypoplastic labia minora 752.480
Webbed neck 744.500 Median raphe present (female) 752.480
Redundant neck folds 744.500 Undescended testicles 752.500–752.520
Large, wide lips 744.820 First degree hypospadias ⫹/⫺ chordee 752.605, 752.625
Small lips 744.830 Chordee 752.621
Short neck 744.900 Hypoplasia of testis and scrotum 752.810
Congenital anomaly of face, not 744.910 Shawl scrotum 752.820
otherwise specified (NOS); abnormal Absent or hooded foreskin of penis 752.860
facies Redundant foreskin 752.860
Bulbus cordis anomalies and anomalies of Small penis (unless documented as
cardiac septal closure micropenis) 752.865
Patent foramen ovale 745.500 Scrotalization of phallus 752.880
Other congenital anomalies of the heart Absent median raphe (male) 752.880
Pulmonary valve insufficiency 746.020 Congenital anomalies of the urinary system
Thickened pulmonary valve 746.080 Mild, minimal hydronephrosis 753.200
Tricuspid valve insufficiency 746.105 Ectopic kidney (unless documented as pelvic
Aortic valve insufficiency 746.400 kidney) 753.330
Bicuspid aortic valve 746.400 Patent urachus 753.700
Thickened aortic valve 746.480 Urachal cyst 753.710
Mitral valve insufficiency 746.600 Certain congenital musculoskeletal deformities
Dextrocardia without congenital heart 746.800 Facial asymmetry 754.000
defects Deviation of nasal septum 754.020
Anomalies of myocardium 746.860 Dolichocephaly 754.030
a
Refers to ICD-9-based six-digit coding scheme for birth defects developed by CDC from the BPA-modification of ICD-9 (Rasmussen and
Moore, 2001). Not all defects included in these codes should be considered minor.

Birth Defects Research (Part A) 67:193–201 (2003)

 CASE CLASSIFICATION IN BIRTH DEFECTS STUDIES 199
Table 3
Defects Considered to be Minor, Prematurity-Related, or Nonstructural for Case Classification Purposes of the
National Birth Defects Prevention Study (continued)
Anomaly Code* Anomaly Code*
Third fontanelle 754.040 Spina bifida occulta 756.100
Large or small fontanelles 754.040 Cervical rib 756.200
Metopic suture open to bregma 754.040 Diastasis recti 756.790
Plagiocephaly 754.050 Torticollis 756.860
Head asymmetry 754.055 Congenital anomalies of the integument
Trigonocephaly, other head Single transverse palmar crease 757.200
deformations without synostosis 754.070 Extra or absent hand/interphalangeal 757.200
Hip click 754.310 creases
Hip subluxation 754.310 Unusual dermatoglyphics 757.200
Genu recurvatum 754.430 Skin tag 757.310
Metatarsus varus or metatarsus Anal tag 757.310
adductus 754.520 Nevus flammeus 757.380
Pectus carinatum 754.800 Port-wine stain 757.380
Pectus excavatum 754.810 Birthmark 757.385
Deformed chest 754.820 Mongolian spot 757.386
Barrel chest 754.820 Depigmented or hyperpigmented spot 757.390
Bifid xiphoid 754.820 Café-au-lait spot 757.390
Shieldlike chest 754.825 Cutis marmorata 757.390
Other congenital anomalies of limbs Skin cyst 757.390
Postaxial minimus polydactyly in Excessive or persistent lanugo 757.450
African-Americans 755.006 Aberrant scalp hair patterning 757.480
Syndactyly (involving 2nd and 3rd Hair upsweep 757.480
toes) 755.130 Low posterior hairline 757.480
Broad, triphalangeal thumb 755.500 Depigmentary hair changes 757.480
Tapered fingers 755.500 Synophrys, confluent or medial flare 757.480
Overlapping fingers 755.500 eyebrows
Short fingers 755.500 Thickened toenails 757.510
Long fingers 755.500 Hyperconvex fingernails 757.580
Clinodactyly 755.500 Hyperconvex toenails 757.580
Camptodactyly 755.500 Hypoplastic fingernails 757.585
Dimple—hand 755.510 Hypoplastic toenails 757.585
Short 4th metacarpal 755.510 Absent nipple 757.630
Cubitus valgus 755.540 Small nipple (hypoplastic) 757.640
Dimple shoulder 755.550 Accessory nipple 757.650
Recessed 4th and 5th toes 755.600 Widely spaced nipples 757.680
Widely spaced 1st and 2nd toes 755.600 Inverted nipples 757.680
Overlapping toes 755.600 Other specified and unspecified
Short or broad great toe 755.600 congenital anomalies
Long toes 755.600 Splenomegaly 759.020
Hallux valgus 755.605 Anomalies of thymus, thymic 759.240
Hallux varus 755.606 hypertrophy (absent thymus should
Short 4th metatarsus 755.610 be considered a major defect)
Plantar furrow 755.610 Anomalies of umbilicus, low-lying 759.900
Sole crease 755.610 umbilicus
Prominent heel 755.610 Other anomalies (not coded in congenital
Rocker-bottom feet 755.616 anomalies section)
Tibial torsion 755.630 Benign skin neoplasm, pigmented 216.000–216.900
Hyperextended knee 755.640 nevus
Genu valgum 755.645 Noncavernous, single, small 228.000; 228.010
Genu varum 755.646 hemangioma (⬍4” diameter)
Anteversion of femur 755.650 Facial palsy 351.000
Coxa valga 755.660 Esotropia 368.000
Coxa vara 755.660 Exotropia 378.000
Hyperextended joints, NOS 755.880 Strabismus 378.900
Overlapping digits, NOS 755.880 Nystagmus 379.500
Other musculoskeletal anomalies Natal teeth 520.600
Flat occiput 756.080 Micrognathia 524.000
Prominent occiput 756.080 Prognathia 524.000
Bony occipital spur 756.080 Inguinal hernia 550.000–550.900
Narrow bifrontal diameter 756.080 Umbilical hernia 553.100
Prominent or hypoplastic supraorbital Testicular torsion 608.200
ridges 756.080 Pilonidal or sacral dimple 685.100
Frontal bossing 756.080 Erb’s palsy 767.600
Minor hypotelorism 756.080 Meconium plug syndrome 777.100
Maxillary hypoplasia/prominence 756.080 Meconium peritonitis 777.600
Dystopia canthorum 756.085 Ascites, congenital 778.000
Minor hypertelorism 756.085 Hydrocele 778.600

Birth Defects Research (Part A) 67:193–201 (2003)

200 RASMUSSEN ET AL.
of genes and their contribution to individual birth defects because many human teratogens have been recognized
may assist in deciding which infants with a positive family because of similar patterns of multiple congenital anoma-
history should be classified as having syndromes (Schott et lies (Friedman, 1992). The clinical geneticist should scruti-
al., 1998). nize infants with multiple congenital anomalies for possi-
ble new patterns of malformation that may be associated
Is a Previously Described Pattern Present? with the exposure. In addition to the use of the clinical
geneticist’s expertise to recognize new phenotypes, statis-
Some infants have a recognized phenotype, but of un- tical associations may also be explored using defined meth-
known etiology. In some cases, these constitute “associa- ods (Kallen et al., 1999).
tions”, nonrandom occurrences of certain defects of un- If information is available on the potential action of the
known etiology, such as the VACTERL association exposure of interest, this can be applied to case classifica-
(Khoury et al., 1983) or CHARGE association (Blake et al.,
tion. As an example, cocaine exposure has been hypothe-
1998). Other infants with recognized phenotypes may have
sized to be associated with vascular disruptive defects
“recurrent pattern syndromes” (Cohen, 1997), defined as a
(Hoyme et al., 1990); therefore, lumping of defects believed
similar set of anomalies in two or more unrelated patients
to be secondary to vascular disruption (gastroschisis,
of unknown etiology. Although these recognized pheno-
transverse limb deficiency, and small intestinal atresia)
types of unknown etiology are often referred to as syn-
may be appropriate in a study of cocaine teratogenesis
dromes, the use of this terminology has been questioned
(Khoury et al., 1992b; Martin et al., 1992).
(Khoury et al., 1994b), given that their etiology remains
unknown and may be heterogeneous (Khoury et al., 1983). An issue separate from case classification, but related
Recognized phenotypes of unknown etiology should be and important to studies of specific exposures, is whether
noted by the clinical geneticist. Depending on the study, the defects observed are consistent with the known timing
these infants may be analyzed separately from other in- of the exposure. For example, transposition of the great
fants classified as having multiple defects (Lammer et al., arteries could not have been caused by a third-trimester
1986). exposure, and limb deficiency and ring constriction of
Sometimes, based on what is known about the patho- digits related to amniotic band sequence is unlikely to be
genesis of the defect, further case classification may be due to a periconceptional exposure. The pathogenesis of
appropriate. For example, because neural tube defects may the defect (malformation, deformation, disruption, or dys-
be due to different embryologic mechanisms, depending plasia) also needs to be assessed in light of what is known
on the level of the defect, classifying infants with neural about the action of the exposure. Information about the
tube defects based on the site of their lesion may be useful pathogenesis of the defects observed and the timing of
(Park et al., 1992). Another possible scenario is that indi- exposure reported needs to be consistent for an association
vidual defects that are believed to be embryologically or to be plausible. These are all areas where the contribution
pathogenetically similar can be combined to maximize the of the clinical geneticist to the study of exposure is critical.
number of cases analyzed. For example, grouping of con- We have summarized here the guidelines for case clas-
genital heart defects according to their presumed underly- sification in birth defects epidemiology used by the
ing pathogenetic mechanism (Clark, 1996) may be a rea- NBDPS. We believe adoption of these guidelines will lead
sonable approach to their study. A recent study of risk to more comparable and etiologically homogeneous case
factors in different individual conotruncal defects showed groups for the study of birth defects, an important element
little evidence of risk factor heterogeneity, providing sup- of careful studies aimed at identifying risk factors for birth
port for analyzing these defects as a single category defects. To ensure that NBDPS clinical reviewers consis-
(O’Malley et al., 1996); however, other studies have shown tently review and classify cases, inter-reviewer reliability
more heterogeneity within this category (Ferencz et al., studies are carried out periodically. Sufficient numbers of
1997). cases with specific defect types have become available in
It is important to recognize that improved understand- the NBDPS only recently; thus, evidence of the utility of
ing of the pathogenesis of birth defects may result in the case classification process in the NBDPS remains to be
changes in case classification. This issue needs to be con- demonstrated. Previous studies have shown, however, that
sidered in the planning of epidemiologic studies of birth the case classification process can help to define risk factors
defects, because it is essential that clinical information that might otherwise be missed (Khoury et al., 1989; Botto
continue to be available so that case classification can be et al., 1996). Other clinicians may find consideration of
modified as advances in the understanding of birth defects these guidelines beneficial in their work on epidemiologi-
occur. cal studies of birth defects.

CLASSIFICATION FOR STUDIES OF ACKNOWLEDGMENTS

SPECIFIC EXPOSURES We thank other members of the Clinicians Committee of
Case classification for studies of specific exposures (e.g., the National Birth Defects Prevention Study, especially the
case– control study of maternal use of a specific prescrip- study clinical geneticists (Drs. M. Curtis, S. Fallet, E. Lam-
tion drug) differs somewhat from the approach to case mer, L. Robinson, A. Scheuerle, and L. Shapiro), as well as
classification for studies of specific defects (e.g., case– con- the other study investigators, for their many contributions
trol study examining several risk factors for gastroschisis). to the study. We also thank Drs. J. Frias and E. Lammer for
The focus of case classification, however, continues to be insightful comments on the manuscript, and J. Yu for her
on what is known about embryogenesis and pathogenesis assistance with the list of minor anomalies. We acknowl-
of the defects and on the exposure of interest. Special edge the generous participation of the many study families
interest may be given to infants with multiple defects, who made this work possible.

Birth Defects Research (Part A) 67:193–201 (2003)

 CASE CLASSIFICATION IN BIRTH DEFECTS STUDIES 201
REFERENCES Khoury MJ, Moore CA, Evans JA. 1994b. On the use of the term “syndrome”
in clinical genetics and birth defects epidemiology. Am J Med Genet
Arpino C, Brescianini S, Robert E, et al. 2000. Teratogenic effects of antiepi- 49:26 –28.
leptic drugs: use of an International Database on Malformations and Khoury MJ, Moore CA, James LM, Cordero JF. 1992b. The interaction
Drug Exposure (MADRE). Epilepsia 41:1436 –1443. between dysmorphology and epidemiology: methodologic issues of
Blake KD, Davenport SL, Hall BD, et al. 1998. CHARGE association: an lumping and splitting. Teratology 45:133–138.
update and review for the primary pediatrician. Clin Pediatr 37:159 – Lammer EJ, Chen DT, Hoar RM, et al. 1985. Retinoic acid embryopathy.
173. N Engl J Med 313:837– 841.
Botto LD, Khoury MJ, Mulinare J, Erickson JD. 1996. Periconceptional Lammer EJ, Cordero JF, Khoury MJ. 1986. Exogenous sex hormone expo-
multivitamin use and the occurrence of conotruncal heart defects: sure and the risk for VACTERL association. Teratology 34:165–169.
results from a population-based, case-control study. Pediatrics 98:911– Lechat MF, Dolk H. 1993. Registries of congenital anomalies: EUROCAT.
917. Environ Health Perspect 101(Suppl):153–157.
Buehler BA, Rao V, Finnell RH. 1994. Biochemical and molecular teratology Leppig KA, Werler MM, Cann CI, Cook CA, Holmes LB. 1987. Predictive
of fetal hydantoin syndrome. Neurol Clin 12:741–748. value of minor anomalies. I. Association with major malformations.
Chambers CD, Braddock SR, Briggs GG, Einarson A, Johnson YR, Miller J Pediatr 110:531–537.
RK, et al. 2001. Postmarketing surveillance for human teratogenicity: a Lurie IW, Wulfsberg EA. 1993. “Holoprosencephaly-polydactyly” (pseudo-
model approach. Teratology 64:252–261. trisomy 13) syndrome: expansion of the phenotypic spectrum. Am J
Clark EB. 1996. Pathogenetic mechanisms of congenital cardiovascular mal- Med Genet 47:405– 409.
formations revisited. Semin Perinatol 20:465– 472. Marden PM, Smith DW, McDonald MJ. 1964. Congenital anomalies in the
Cohen MM Jr. 1997. The child with multiple birth defects. New York: newborn infant, including minor variations: a study of 4,412 babies by
Oxford University Press.
surface examination for anomalies and buccal smear for sex chromatin.
Cunniff C, Jones KL, Jones MC. 1990. Patterns of malformation in children
J Pediatr 64:357–371.
with congenital diaphragmatic defects. J Pediatr 116:258 –261.
Martin ML, Khoury MJ, Cordero JF, Waters GD. 1992. Trends in rates of
Ferencz C. 1993. Congenital heart disease: an epidemiologic and teratologic
multiple vascular disruption defects, Atlanta, 1968 – 1989: is there evi-
challenge. In: Ferencz C, Rubin, JD, Loffredo CA, Magee CA, editors.
dence of a cocaine teratogenic epidemic? Teratology 45:647– 653.
Epidemiology of congenital heart disease: the Baltimore-Washington
Martin RA, Fineman RM, Jorde LB. 1983. Phenotypic heterogeneity in
Infant Study 1981–1989. Vol. 4. Mount Kisco, NY: Future. pp. 1–15.
neural tube defects: a clue to causal heterogeneity. Am J Med Genet
Ferencz C, Loffredo CA, Correa-Villasenor A, Wilson PD. 1997. Genetic and
16:519 –525.
environmental risk factors of major cardiovascular malformations: the
Martinez-Frias ML, Bermejo E, Frias JL. 2000. Pathogenetic classification of
Baltimore-Washington Infant Study 1981–1989. Armonk, NY: Futura
a series of 27,145 consecutive infants with congenital defects. Am J Med
Publishing.
Francis PJ, Berry V, Bhattacharya SS, Moore AT. 2000. The genetics of Genet 90:246 –249.
childhood cataract. J Med Genet 37:481– 488. Martinez-Frias ML, Frias JL, Rodriguez-Pinilla E, et al. 1991. Value of
Frias JL, Carey JC. 1996. Mild errors of morphogenesis. Adv Pediatr 43:27– clinical analysis in epidemiological research: the Spanish registry ex-
75. perience. Am J Med Genet 41:192–195.
Friedman JM. 1992. The use of dysmorphology in birth defects epidemiol- Martinez-Frias ML, Frias JL, Salvador J. 1990. Clinical/epidemiological
ogy. Teratology 45:187–193. analysis of malformations. Am J Med Genet 35:121–125.
Holmes LB, Driscoll SG, Atkins L. 1976. Etiologic heterogeneity of neural- Meurman Y. 1957. Congenital microtia and meatal atresia. Arch Otolaryn-
tube defects. N Engl J Med 294:365–369. gol 66:443.
Hook EB, Marden PM, Reiss NP, Smith DW. 1976. Some aspects of the Murray JC, Johnson JA, Bird TD. 1985. Dandy-Walker malformation: etio-
epidemiology of human minor birth defects and morphological vari- logic heterogeneity and empiric recurrence risks. Clin Genet 28:272–
ants in a completely ascertained newborn population (Madison study). 283.
Teratology 13:47–55. Nelson K, Holmes LB. 1989. Malformations due to presumed spontaneous
Hoyert DL, Arias E, Smith BL, Murphy SL, Kochanek KD. 2001. Deaths: mutations in newborn infants. N Engl J Med 320:19 –23.
final data for 1999. Natl Vital Stat Rep 49:1–113. Nopoulos P, Berg S, Canady J, Richman L, Van Demark D, Andreasen NC.
Hoyme HE, Jones KL, Dixon SD, Jewett T, Hanson JW, Robinson LK, et al. 2002. Structural brain abnormalities in adult males with clefts of the lip
1990. Prenatal cocaine exposure and fetal vascular disruption. Pediat- and/or palate. Genet Med 4:1–9.
rics 85:743–747. Nopoulos P, Berg S, VanDemark D, Richman L, Canady J, Andreasen NC.
Jones KL. 1997. Smith’s recognizable patterns of human malformation. 2001. Increased incidence of a midline brain anomaly in patients with
Philadelphia: W.B. Saunders Company. nonsyndromic clefts of the lip and/or palate. J Neuroimaging 11:418 –
Jones MC. 1988. Etiology of facial clefts: prospective evaluation of 428 424.
patients. Cleft Palate J 25:16 –20. O’Malley CD, Shaw GM, Wasserman CR, Lammer EJ. 1996. Epidemiologic
Kallen K, Robert E, Mastroiacovo P, Castilla EE, Kallen B. 1999. CHARGE characteristics of conotruncal heart defects in California, 1987–1988.
Association in newborns: a registry-based study. Teratology 60:334 – Teratology 53:374 –377.
343. Park CH, Stewart W, Khoury MJ, Mulinare J. 1992. Is there etiologic heter-
Khoury MJ, Botto L, Mastroiacovo P, Skjaerven R, Castilla E, Erickson JD. ogeneity between upper and lower neural tube defects? Am J Epide-
1994a. Monitoring for multiple congenital anomalies: an international miol 136:1493–1501.
perspective. Epidemiol Rev 16:335–350. Rasmussen SA, Moore CA. 2001. Effective coding in birth defects surveil-
Khoury MJ, Cordero JF, Greenberg F, James LM, Erickson JD. 1983. A lance. Teratology 64(Suppl):3–7.
population study of the VACTERL association: evidence for its etio- Schott JJ, Benson DW, Basson CT, et al. 1998. Congenital heart disease
logic heterogeneity. Pediatrics 71:815– 820. caused by mutations in the transcription factor NKX2-5. Science 281:
Khoury MJ, Erickson JD, James LM. 1982. Etiologic heterogeneity of neural 108 –111.
tube defects: clues from epidemiology. Am J Epidemiol 115:538 –548. Spranger J, Benirschke K, Hall JG, et al. 1982. Errors of morphogenesis:
Khoury MJ, Gomez-Farias M, Mulinare J. 1989. Does maternal cigarette concepts and terms. Recommendations of an international working
smoking during pregnancy cause cleft lip and palate in offspring? Am J group. J Pediatr 100:160 –165.
Dis Child 143:333–337. Van Allen MI. 1992. Structural anomalies resulting from vascular disrup-
Khoury MJ, James LM, Flanders WD, Erickson JD. 1992a. Interpretation of tion. Pediatr Clin North Am 39:255–77.
recurring weak associations obtained from epidemiologic studies of Yoon PW, Rasmussen SA, Moore CA, et al. 2001. The National Birth Defects
suspected human teratogens. Teratology 46:69 –77. Prevention Study. Public Health Rep 116:32– 40.

Birth Defects Research (Part A) 67:193–201 (2003)