Handbook19 Book Final

IARC HANDBOOKS
ORAL CANCER
PREVENTION
VOLUME 19
IARC HANDBOOKS OF
CANCER PREVENTION
IARC HANDBOOKS
ORAL CANCER
PREVENTION
VOLUME 19
This publication represents the views and

expert opinions of an IARC Working Group on
the Evaluation of Cancer-Preventive Interventions,
which met remotely, 4–11 December 2021
LYON, FRANCE - 2023
IARC HANDBOOKS OF
CANCER PREVENTION
Published by the International Agency for Research on Cancer, 25 avenue Tony Garnier, CS 90627, 69366
Lyon CEDEX 07, France
©International Agency for Research on Cancer, 2023
Distributed by WHO Press, World Health Organization, 20 Avenue Appia, 1211 Geneva 27, Switzerland
(tel.: +41 22 791 3264; fax: +41 22 791 4857; website: https://www.who.int/publications/book-orders;
email: [email protected]).
Permissions and rights: Some rights reserved. This work is available under the Creative Commons Attribution-
NonCommercial-NoDerivs 3.0 IGO licence (CC BY-NC-ND 3.0 IGO; https://creativecommons.org/licenses/by-nc-nd/3.0/igo/).
Under the terms of this licence, you may copy and redistribute the work for non-commercial purposes, provided the
work is appropriately cited, as indicated below. In any use of this work, there should be no suggestion that WHO endorses any
specific organization, products, or services. The use of the WHO logo is not permitted.
Any mediation relating to disputes arising under the licence shall be conducted in accordance with the mediation rules of
the World Intellectual Property Organization.
To submit requests for adaptations or commercial use and queries on rights and licensing, see the IARC Publications
website (https://publications.iarc.fr/Rights-And-Permissions).
Third-party materials: If you wish to reuse material from this work that is attributed to a third party, such as tables, figures
or images, it is your responsibility to determine whether permission is needed for that reuse and to obtain permission from the
copyright holder. The risk of claims resulting from infringement of any third-party-owned component in the work rests solely
with the user.
General disclaimers: The designations employed and the presentation of the material in this publication do not imply
the expression of any opinion whatsoever on the part of the Secretariat of the World Health Organization concerning the legal
status of any country, territory, city, or area or of its authorities, or concerning the delimitation of its frontiers or boundaries.
The mention of specific companies or of certain manufacturers’ products does not imply that they are endorsed or
recommended by the World Health Organization in preference to others of a similar nature that are not mentioned. Errors and
omissions excepted, the names of proprietary products are distinguished by initial capital letters.
All reasonable precautions have been taken by WHO to verify the information contained in this publication. However,
the published material is being distributed without warranty of any kind, either expressed or implied. The responsibility for the
interpretation and use of the material lies with the reader. In no event shall WHO or contributing agencies be liable for damages
arising from its use.
The IARC Handbooks Working Group alone is responsible for the views expressed in this publication.
IARC Handbooks (and Corrigenda) are published online at https://publications.iarc.fr/.
To report an error, please contact [email protected].
How to cite this publication:

IARC (2023). Oral cancer prevention. IARC Handb Cancer Prev. 19:1–358. Available from: https://publications.iarc.fr/617.
IARC Library Cataloging-in-Publication Data

Names: IARC Working Group on the Evaluation of Cancer-Preventive Interventions.
Title: Oral cancer prevention.
Description: Lyon: International Agency for Research on Cancer, 2023. | Series: IARC handbooks of cancer prevention,
ISSN 1027-5622; v. 19. | “This publication represents the views and expert opinions of an IARC Working Group on the
Evaluation of Cancer-Preventive Interventions, which met remotely, 4–11 December 2021.” | Includes bibliographical
references.
Identifiers: ISBN 9789283230274 (pbk.) | ISBN 9789283230267 (ebook)
Subjects: MESH: Mouth Neoplasms. | Primary Prevention. | Early Detection of Cancer.
Classification: NLM W1
International Agency for Research on Cancer
The International Agency for Research on Cancer (IARC) was established in 1965 by the World
Health Assembly, as an independently funded organization within the framework of the World Health
Organization. The headquarters of the Agency are in Lyon, France.
The Agency has as its mission to reduce the cancer burden worldwide through promoting international
collaboration in research. The Agency addresses this mission through conducting cancer research for
cancer prevention in three main areas: describing the occurrence of cancer, identifying the causes of
cancer, and evaluating preventive interventions and their implementation. Each of these areas is a vital
contribution to the spectrum of cancer prevention.
The publications of the Agency contribute to the dissemination of authoritative information on
different aspects of cancer research. Information about IARC publications, and how to order them, is
available at https://publications.iarc.fr/.
IARC Handbooks of Cancer Prevention
In 1969, the International Agency for Research on Cancer (IARC) initiated a programme on the
evaluation of the carcinogenic risk of chemicals to humans involving the production of monographs of
critical reviews and evaluations of individual chemicals.
The IARC Handbooks of Cancer Prevention complement the IARC Monographs’ identifications of
carcinogenic hazards. The objective of the programme is to coordinate and publish critical reviews of data
on the cancer-preventive effects of primary or secondary interventions, and to evaluate these data in terms
of cancer prevention with the help of international working groups of experts in prevention and related
fields. The lists of evaluations are regularly updated and are available at https://handbooks.iarc.fr/.
This IARC Handbook of Cancer Prevention is partly funded by the French National Cancer Institute
(INCa) (grant number INCA_14774). Its contents are solely the responsibility of the authors and do not
necessarily represent the official views of INCa.
Cover image: Clinical photograph of the mouth of a 72-year-old woman presenting with an extensive,
non-homogeneous leukoplakia (with ulcerated areas) on the left lateral border of the tongue, diagnosed by
biopsy as a squamous cell carcinoma. © Oral Medicine (Stomatology) Service, OROCENTRO, Piracicaba
Dental School, University of Campinas (UNICAMP), São Paulo.
CONTENTS
NOTE TO THE READER. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
LIST OF PARTICIPANTS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
PREAMBLE – PRIMARY PREVENTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

A. GENERAL PRINCIPLES AND PROCEDURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1. Background. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2. Objectives, scope, and definitions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3. Identification and selection of interventions and outcomes for review . . . . . . . . . . . . . . . . . . . . . . . . . 10
4. The Working Group and other meeting participants. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
5. Development of a volume of the IARC Handbooks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
6. Overview of the scientific review and evaluation process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
7. Responsibilities of the Working Group. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
B. SCIENTIFIC REVIEW AND EVALUATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1. Intervention and outcome characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2. Studies of cancer prevention in humans. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3. Studies of cancer prevention in experimental animals. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4. Mechanistic evidence and other relevant biological data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
5. Summary of data reported. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
6. Evaluation and rationale. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
PREAMBLE − SECONDARY PREVENTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

A. GENERAL PRINCIPLES AND PROCEDURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
1. Background. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
2. Objectives, scope, and definitions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3. Identification and selection of interventions and outcomes for review . . . . . . . . . . . . . . . . . . . . . . . . . 44
4. The Working Group and other meeting participants. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
5. Development of a volume of the IARC Handbooks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
6. Overview of the scientific review and evaluation process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
7. Responsibilities of the Working Group. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
B. SCIENTIFIC REVIEW AND EVALUATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
1. Definitions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
2. Characterization of the disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
3. Screening methods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
V
IARC HANDBOOKS OF CANCER PREVENTION – 18
4. Current global screening practices. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

5. Epidemiological studies of each screening method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
6. Summary of data reported. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
7. Evaluation and rationale. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
GENERAL REMARKS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
LIST OF ABBREVIATIONS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
1. ORAL CANCER AND ORAL POTENTIALLY MALIGNANT DISORDERS. . . . . . . . . . . . . . . . . . . . . . 77

1.1 Anatomy of the oral cavity and the oropharynx . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
1.1.1 Anatomy of the oral cavity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
1.1.2 Anatomy of the oropharynx and the soft palate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
1.2 Global burden of oral cancer, oropharyngeal cancer, and oral potentially malignant disorders. . . . . . . 81
1.2.1 Oral cancer and oropharyngeal cancer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
1.2.2 Oral potentially malignant disorders. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
1.3 Oral neoplasia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
1.3.1 Classification and natural history of OPMDs and oral cancer. . . . . . . . . . . . . . . . . . . . . . . . . . . 88
1.3.2 Stage at diagnosis and stage-related survival. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
1.3.3 Treatment and management of OPMDs and oral cancer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
2. REDUCING INCIDENCE OF CANCER OR PRECANCER. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105

2.1 Established risk factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
2.1.1 Tobacco smoking. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
2.1.2 Alcohol consumption. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
2.1.3 Smokeless tobacco use. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
2.1.4 Chewing areca nut products (including betel quid) with added tobacco. . . . . . . . . . . . . . . . . . 109
2.1.5 Chewing areca nut products (including betel quid) without tobacco. . . . . . . . . . . . . . . . . . . . . 110
2.1.6 HPV16 infection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
2.1.7 Combined effects of established risk factors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
2.2 Additional potential risk factors for oral cancer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
2.2.1 Environmental factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
2.2.2 Lifestyle factors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
2.2.3 Demographic factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
2.2.4 Orodental factors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
2.2.5 Systemic factors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
2.2.6 Familial or genetic predisposition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
2.3 Impact upon quitting. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
2.3.1 Tobacco smoking. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
2.3.4 Chewing areca nut products (including betel quid) with added tobacco. . . . . . . . . . . . . . . . . . 158
2.3.5 Chewing areca nut products (including betel quid) without tobacco. . . . . . . . . . . . . . . . . . . . . 169
2.3.6 HPV16 infection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
VI
Contents
2.4 Preventive dietary agents. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191

2.4.1 Preventive dietary agents for the development of oral cancer. . . . . . . . . . . . . . . . . . . . . . . . . . 191
2.4.2 Preventive dietary agents for the development or progression of OPMDs. . . . . . . . . . . . . . . . 193
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
3. CESSATION OF SMOKELESS TOBACCO AND/OR ARECA NUT USE. . . . . . . . . . . . . . . . . . . . 209

3.1 Product definition and description. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
3.2 Prevalence of consumption. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
3.2.1 WHO South-East Asia Region. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
3.2.2 WHO Western Pacific Region. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
3.2.3 WHO European Region. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
3.2.4 WHO Region of the Americas. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
3.2.5 WHO African Region. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
3.2.6 WHO Eastern Mediterranean Region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
3.2.7 Determinants of use. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
3.3 Interventions for cessation of use. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
3.3.1 Behavioural interventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
3.3.2 Pharmacological interventions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
3.3.3 Combined pharmacological and behavioural interventions. . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
3.4 Policies and their impacts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264
3.4.1 Control policies for smokeless tobacco. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264
3.4.2 Control policies for areca nut products (including betel quid) . . . . . . . . . . . . . . . . . . . . . . . . . 280
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
4. SCREENING AND EARLY DIAGNOSIS OF ORAL CANCER. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295

4.1 Screening methods and technologies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295
4.1.1 Clinical oral examination. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295
4.1.2 Mouth self-examination. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298
4.1.3 Adjunctive techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299
4.1.4 Cytology and quantitative DNA cytometry. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306
4.1.5 Liquid biopsy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308
4.1.6 Use of emerging technologies in the primary screening setting . . . . . . . . . . . . . . . . . . . . . . . . 310
4.2 Organized and opportunistic oral cancer screening activities. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
4.3 Determinants of participation in screening for oral cancer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314
4.3.1 Individual level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316
4.3.2 Health-care provider level. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316
4.3.3 Health-care system level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317
4.3.4 Health-care policies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317
4.3.5 Strategies to increase participation in oral cancer screening. . . . . . . . . . . . . . . . . . . . . . . . . . . 317
4.4 Effectiveness of screening. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317
4.4.1 Preventive effects of screening. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317
4.4.2 Harms of screening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323
4.5 Risk-based model for screening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325
VII
5. SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
5.1 Oral cancer and oral potentially malignant disorders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
5.1.1 Anatomy of the oral cavity and the oropharynx. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
5.1.2 Global burden. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
5.1.3 Oral neoplasia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338
5.2 Reducing incidence of cancer or precancer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
5.2.1 Established risk factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
5.2.2 Impact upon quitting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340
5.2.3 Preventive dietary agents. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
5.3 Cessation of smokeless tobacco and/or areca nut use. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
5.3.1 Product definition and description. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
5.3.2 Prevalence of consumption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344
5.3.3 Interventions for cessation of use. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346
5.3.4 Policies and their impacts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347
5.4 Screening and early diagnosis of oral cancer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348
5.4.1 Screening methods and technologies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348
5.4.2 Organized and opportunistic oral cancer screening. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 350
5.4.3 Determinants of participation in screening for oral cancer. . . . . . . . . . . . . . . . . . . . . . . . . . . . 350
5.4.4 Effectiveness of screening. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 350
5.4.5 Risk-based model for screening. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 351
6. EVALUATIONS, STATEMENTS, AND CONSIDERATIONS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353

6.1 Impact of quitting exposure to a risk factor on incidence of or mortality from oral cancer. . . . . . . . . 353
6.1.1 Tobacco smoking. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353
6.1.4 Chewing areca nut products (including betel quid) with or without tobacco . . . . . . . . . . . . . . 354
6.1.5 HPV16 infection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355
6.2 Interventions for cessation of smokeless tobacco or areca nut use. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355
6.2.1 Behavioural interventions in adults. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355
6.2.2 Behavioural interventions in youth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356
6.2.3 Pharmacological interventions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356
6.2.4 Combined pharmacological and behavioural interventions. . . . . . . . . . . . . . . . . . . . . . . . . . . . 356
6.2.5 Policies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357
6.3 Screening for oral cancer and OPMDs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357
6.3.1 Effectiveness of screening by clinical oral examination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357
6.3.2 Additional considerations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 358
VIII
NOTE TO THE READER
The IARC Handbooks of Cancer Prevention series was launched in 1995 to complement the IARC
Monographs’ evaluations of carcinogenic hazards. The IARC Handbooks of Cancer Prevention evalu-
ate the published scientific evidence of cancer-preventive interventions.
Inclusion of an intervention in the Handbooks does not imply that it is cancer-preventive, only
that the published data have been examined. Equally, the fact that an intervention has not yet been
evaluated in a Handbook does not mean that it may not prevent cancer. Similarly, identification of
organ sites with sufficient evidence or limited evidence that the intervention has a cancer-preventive
activity in humans should not be viewed as precluding the possibility that an intervention may pre-
vent cancer at other sites.
The evaluations of cancer-preventive interventions are made by international Working Groups
of independent scientists and are qualitative in nature. No recommendation is given for regulation
or legislation.
Anyone who is aware of published data that may alter the evaluation of cancer-preventive inter-
ventions is encouraged to make this information available to the IARC Handbooks programme,
International Agency for Research on Cancer, 25 avenue Tony Garnier, CS 90627, 69366 Lyon
CEDEX 07, France, or by email to [email protected], in order that these data may be considered for
re-evaluation by a future Working Group.
Although every effort is made to prepare the Handbooks as accurately as possible, mistakes may
occur. Readers are requested to communicate any errors to the IARC Handbooks programme at ihb@
iarc.who.int. Corrigenda are published online on the relevant webpage for the volume concerned (IARC
Publications: https://publications.iarc.fr/).
1
LIST OF PARTICIPANTS
Members 1
Devasena Anantharaman Anil Chaturvedi 2 (Subgroup Co-Chair,

Rajiv Gandhi Centre for Biotechnology Impact of Cessation)
Thiruvananthapuram Division of Cancer Epidemiology and
Genetics
National Cancer Institute
Olalekan A. Ayo-Yusuf (Subgroup Chair,
National Institutes of Health
Interventions for Cessation) Bethesda
Sefako Makgatho Health Sciences University
and
Tony Hsiu-Hsi Chen (Subgroup Co-Chair,
School of Health Systems and Public Health
University of Pretoria Impact of Cessation)
Pretoria Institute of Epidemiology and Preventive
Medicine
NTU
Taipei
1
Working Group Members and Invited Specialists serve in their individual capacities as scientists and not as represen-
tatives of their government or any organization with which they are affiliated. Affiliations are provided for identifica-
tion purposes only. Invited Specialists do not serve as Meeting Chair or Subgroup Chair, draft text that pertains to the
description or interpretation of cancer data, or participate in the evaluations. Each participant was asked to declare
potentially relevant research, employment, and financial interests that are current or that have occurred during the past
4 years. Minimal interests are not disclosed here and include stock valued at no more than US$ 1000 overall, grants that
provide no more than 5% of the research budget of the expert’s organization and that do not support the expert’s re-
search or position, and consulting or speaking on matters not before a court or government agency that does not exceed
2% of total professional time or compensation. All other non-publicly funded grants that support the expert’s research
or position and all consulting or speaking on behalf of an interested party on matters before a court or government
agency are disclosed as potentially significant conflicts of interests.
2
Anil Chaturvedi reports that his unit at NCI/NIH benefits from VELscope equipment for oral cancer screening from
LED Dental Ltd.
3
David I. Conway Ravi Mehrotra (Meeting Co-Chair)

School of Medicine, Dentistry, and Nursing Centre for Health Innovation and Policy
University of Glasgow Foundation
Glasgow Noida
and
Rollins School of Public Health
Prakash C. Gupta Emory University
Healis Sekhsaria Institute for Public Health Atlanta
Navi Mumbai
Toru Nagao
Newell W. Johnson Department of Maxillofacial Surgery
Griffith University School of Dentistry
Gold Coast Aichi Gakuin University
and Nagoya
Faculty of Dentistry, Oral, and Craniofacial
Sciences
Vinayak M. Prasad
King’s College London
London World Health Organization
Geneva
Alexander Ross Kerr (Subgroup Chair,
Screening) Kunnambath Ramadas
New York University College of Dentistry Regional Cancer Centre
New York Thiruvananthapuram
Luiz P. Kowalski Felipe Roitberg

University of São Paulo Medical School World Health Organization
and Geneva
A.C. Camargo Cancer Center
São Paulo
Pierre Saintigny
Univ Lyon, Université Claude Bernard
Maria Elena Leon Lyon 1, INSERM 1052, CNRS 5286,
Retired Centre Léon Bérard, Centre de recherche
Panama City en cancérologie de Lyon
and
Department of Medical Oncology,
Olena Mandrik Centre Léon Bérard
The University of Sheffield School of Health Lyon
and Related Research
Sheffield
4
Participants
Alan Roger Santos-Silva Invited Specialists

Piracicaba Dental School
University of Campinas Ann Gillenwater 3 (unable to attend)
Piracicaba, São Paulo UT MD Anderson Cancer Center
Houston
Dhirendra Narain Sinha
School of Preventive Oncology Rengaswamy Sankaranarayanan 4
Patna Preventive Oncology
Karkinos Healthcare Pvt Ltd
Wanninayake M. Tilakaratne Coimbatore
and
University of Peradeniya International Agency for Research on Cancer
Peradeniya Lyon
and
Faculty of Dentistry, University of Malaya
Kuala Lumpur Representatives
Patravoot Vatanasapt Frédéric de Bels

Department of Otorhinolaryngology French National Cancer Institute
Faculty of Medicine Boulogne-Billancourt
Khon Kaen University
Khon Kaen
Luciana Neamtiu
European Commission
Saman Warnakulasuriya (Meeting Co-Chair)
Ispra
King’s College London
WHO Collaborating Centre for Oral Cancer
London Carolina Wiesner
Colombian National Cancer Institute
Bogotá
Rosnah B. Zain
MAHSA University
Selangor
and
Faculty of Dentistry, University of Malaya
Kuala Lumpur
3
Ann Gillenwater reports receiving personal consultancy fees from Rakuten Medical, that her unit at the University of
Texas benefits from research funding from NRG Oncology, and holding United States Patents no. 7365844 B2 and no.
7499161 B2.
4
Rengaswamy Sankaranarayanan reports being a salaried advisor for Karkinos Healthcare Pvt Ltd.
5
Observers 5 Rok Ho Kim, WHO headquarters

Béatrice Lauby-Secretan, Head of the IARC
Handbooks Programme, Deputy Head of
Pankaj Chaturvedi
the Evidence Synthesis and Classification
Tata Memorial Centre Branch
Department of Atomic Energy, Government Filip Meheus, Cancer Surveillance Branch
of India Karen Müller, Services to Science and
Mumbai Research Branch (Editor)
Richard Muwonge, Early Detection,
Prevention, and Infections Branch
Moni Abraham Kuriakose Suzanne Nethan, Evidence Synthesis and
Roswell Park Comprehensive Cancer Center Classification Branch (Rapporteur)
Buffalo Sandra Perdomo, Genomic Epidemiology
Branch
Deependra Singh, Cancer Surveillance
Andrey Moreno Torres Branch (Rapporteur)
Colombian National Cancer Institute Maria Pura Solon, WHO headquarters
Bogotá Benoît Varenne, WHO headquarters
Shama Virani, Genomic Epidemiology
Branch
IARC/WHO Secretariat
Véronique Bouvard, Evidence Synthesis and Administrative Assistance

Classification Branch (Responsible Officer,
Rapporteur) Marieke Dusenberg
Andre Carvalho, Early Detection,
Prevention, and Infections Branch
Chrysovalantou Chatziioannou, Nutrition Production Team
and Metabolism Branch
Ana Carolina de Carvalho, Genomic Niree Kraushaar
Epidemiology Branch Solène Quennehen
Gampo Dorji, WHO Regional Office for
South-East Asia
Thushara Eraj Indranath Fernando, WHO Scientific Assistance
Regional Office for South-East Asia
Andre Ilbawi, WHO headquarters Irena Duś-Ilnicka
Jagdish Kaur, WHO Regional Office for Ana Carolina de Carvalho, IARC
South-East Asia Chrysovalantou Chatziioannou, IARC
5
Each Observer agreed to respect the Guidelines for Observers at IARC Handbooks meetings. Observers did not serve as
Meeting Chair or Subgroup Chair, draft or revise any part of the Handbook, or participate in the evaluations. They also
agreed not to contact participants before or after the meeting, not to lobby them at any time, not to send them written
materials, and not to offer them meals or other favours. IARC asked and reminded Working Group Members to report
any contact or attempt to influence that they may have encountered, either before or during the meeting.
6
PREAMBLE – PRIMARY PREVENTION
The Preamble to the IARC Handbooks of Cancer Prevention describes the objectives and
scope of the programme, general principles and procedures, and scientific review and
evaluations. The IARC Handbooks embody the principles of scientific rigour, impartial eval-
uation, transparency, and consistency. The Preamble should be consulted when reading
an IARC Handbook or a summary of an IARC Handbook’s evaluations. Separate Instructions
for Authors describe the operational procedures for the preparation and publication of a
volume of the IARC Handbooks.
A. GENERAL PRINCIPLES AND developing cancer. Secondary prevention entails

PROCEDURES methods that can find and ameliorate precan-
cerous conditions or find cancers in the early
stages, when they can be treated more success-
1. Background fully. Tertiary prevention is the application of
Prevention of cancer is the mission of the measures aimed at reducing the impact of long-
International Agency for Research on Cancer term disease and disability caused by cancer or
(IARC). Cancer prevention is needed even more its treatment.
today than when IARC was established, in 1965, The IARC Handbooks of Cancer Prevention
because the global burden of cancer is high and provide critical reviews and evaluations of the
continues to increase, as a result of population scientific evidence on the preventive effects
growth and ageing and increases in cancer- of primary or secondary cancer preven-
causing exposures and behaviours, especially in tion measures. The evaluations of the IARC
low- and middle-income countries (Stewart & Handbooks are used by national and interna-
Kleihues, 2003; Boyle & Levin, 2008; Stewart & tional health agencies to develop evidence-based
Wild, 2014). interventions or recommendations for reducing
Broadly defined, prevention is “actions aimed cancer risk.
at eradicating, eliminating, or minimizing the The IARC Handbooks of Cancer Prevention
impact of disease and disability, or if none of series was launched in 1995 by Dr Paul Kleihues,
these is feasible, retarding the progress of disease then Director of IARC, in recognition of the
and disability” (Porta, 2014). Cancer prevention need for a series of publications that would criti-
encompasses primary, secondary, and tertiary cally review and evaluate the evidence on a wide
prevention. Primary prevention consists of range of cancer-preventive interventions. The
actions that can be taken to lower the risk of first volume of the IARC Handbooks (IARC,
7
1997) reviewed the evidence on cancer-preven- 2. Objectives, scope, and

tive effects of non-steroidal anti-inflammatory definitions
drugs, specifically aspirin, sulindac, piroxicam,
and indomethacin. Handbooks Volume 6 (IARC, 2.1 Objectives and scope
2002a) was the first that evaluated behavioural
interventions (weight control and physical The scope of the IARC Handbooks of Cancer
activity), and Handbooks Volume 7 (IARC, Prevention series is to contribute to reducing the
2002b) was the first that evaluated cancer incidence of or mortality from cancer worldwide.
screening (breast cancer screening). Handbooks To this end, the IARC Handbooks programme
Volumes 11–14 (IARC, 2007, 2008, 2009, 2011) prepares and publishes, in the form of volumes
focused on tobacco control. After a 3-year hiatus, of Handbooks, critical scientific reviews and
the IARC Handbooks series was relaunched in evaluations of the available evidence on the effi-
2014 with the preparation of Handbooks Volume cacy, effectiveness, and harms of a wide range
15 (IARC, 2016), which re-evaluated breast of cancer-preventive interventions. The primary
cancer screening. target audiences for the Handbooks are national
IARC’s process for developing Handbooks and international agencies with responsibility
engages international, expert scientific Working for, or advocating for, public health. The IARC
Groups in a transparent synthesis of different Handbooks are an important part of the body
streams of evidence, which is then translated of information on which public health decisions
into an overall evaluation according to criteria for cancer prevention may be based. However,
that IARC has developed and refined (see Part A, public health options to prevent cancer vary
Section 6). Scientific advances are periodically from one setting to another and from country
incorporated into the evaluation methodology, to country, and relate to many factors, including
which must be sufficiently robust to encompass socioeconomic conditions and national prior-
a wide variety of interventions, ranging from ities. Therefore, no recommendations are given
broad societal measures to individual behaviour in the Handbooks with regard to regulations
and to chemoprevention. or legislation, which are the responsibility of
This Preamble, first prepared as the individual governments or other international
Handbooks Working Procedures in 1995 and authorities. However, the IARC Handbooks may
later adapted to the topics of cancer screening aid national and international authorities in
and tobacco control, is primarily a statement of devising programmes of health promotion and
the general principles and procedures used in cancer prevention, understanding important
developing a Handbook, to promote transpar- benefits and harms, and considering cost–effec-
ency and consistency across Handbooks evalu- tiveness evaluations.
ations. In addition, IARC provides Instructions The IARC Handbooks programme also
for Authors to specify more detailed operating does not make formal research recommenda-
procedures. tions. However, because Handbooks synthesize
and integrate streams of evidence on cancer
prevention, critical gaps in knowledge that merit
research may be identified.
8
Preamble – Primary Prevention
2.2 Definition of interventions for primary which an evaluation of the association between
prevention exposure to an agent and cancer incidence, i.e.
a static comparison, is appropriate. In prin-
The current IARC Handbook addresses a ciple, the approaches to scientific review of the
specific intervention or class of interventions for relevant studies in this section will not differ
primary prevention. Primary prevention “aims between those entailing dynamic interventions
to reduce the incidence of disease by personal and those entailing static exposures. Therefore,
and communal efforts” (Porta, 2014). The term in this Preamble the term “intervention” applies
“intervention” in this Handbook refers to any to studies of both types, unless specifically stated
action aimed at reducing the incidence of cancer otherwise.
in humans. Primary prevention interventions
include increasing human exposure to known
2.3 Definitions of efficacy, effectiveness,
cancer-preventive agents, reducing human expo-
sure to known cancer hazards, providing means
and harms
to reduce the effects of exposure to cancer hazards, Efficacy and effectiveness are two funda-
or otherwise intervening on human pathological mental concepts underlying the evaluation
states that cause cancer. In broad terms, such of preventive interventions (Cochrane, 1972).
interventions include, for example, regulating Efficacy was defined by Porta (2008) as “the
exposure to carcinogens, administering chemo- extent to which a specific intervention, proce-
preventive pharmaceuticals or other agents, dure, regimen or service produces a beneficial
vaccinating against cancer-causing infections, result under ideal conditions … Ideally, the
modifying the environment (e.g. planting trees determination of efficacy is based on the results
or constructing shade structures in areas of high of a randomized controlled trial”. Effectiveness
ambient levels of solar ultraviolet radiation), or was defined by Porta (2008) as “a measure of the
promoting personal or societal action to increase extent to which a specific intervention, proce-
the prevalence of healthy lifestyles or behaviours dure, regimen or service, when deployed in the
or decrease the prevalence of unhealthy lifestyles field in routine circumstances, does what it is
or behaviours. intended to do for a specific population”.
Primary preventive interventions can be The distinction between efficacy and effec-
applied across a continuum of: tiveness of an intervention at the population level
(i) the general population (often circum- is an important one to make when evaluating
scribed by age and sex); preventive interventions. Efficacy is a necessary,
but not sufficient, basis for recommending an
(ii) subgroups with particular predisposing
intervention. Whereas efficacy of an interven-
host characteristics, such as genetic suscepti-
tion can be inferred if effectiveness is estab-
bility, precursor lesions, or particular diseases
lished, efficacy does not guarantee effectiveness
other than cancer, or with high exposure to
because of the number of implementation steps,
environmental, occupational, or behavioural
each with uncertainty, required to deliver an
risk factors; and
efficacious prevention intervention as an effec-
(iii) people with a history of cancer who are at tive programme in a target population. Ideally,
high risk of a further primary cancer. efficacy is established before a preventive inter-
Although the intent of the IARC Handbooks vention is implemented in a whole community or
is to evaluate interventions, i.e. a dynamic population, so as to determine whether a case for
comparison, there will be circumstances under population-wide implementation can be made
9
on the basis of the balance of the benefits and a framework depicts the relationships among
harms and the financial costs of the intervention. the study population, intervention, compar-
However, it has not been unusual for preventive ator, and intermediate outcomes or changes in
interventions to be implemented in the absence health status as relevant. The analytical frame-
of evidence of efficacy. Should that occur, eval- work includes both benefits and harms, and
uation of effectiveness may be the only way to key contextual issues related to participation
determine whether the case for the intervention and implementation of the intervention and its
is strong enough to justify its continuation or impact on population health. The framework
implementation elsewhere. defines the intervention in its broadest context
In addition to being shown to be efficacious and specifies the aspects for which the Handbook
or effective, preventive interventions must satisfy will review and evaluate the evidence.
other requirements if they are to be considered In this framework, IARC defines the interven-
for implementation in practice, including an tion and the outcome to be evaluated, according
acceptable balance of benefits and harms. In the to one of two scenarios:
present context, harm is defined as any impair- Scenario 1: evaluation of the effect of a speci-
ment or increase in risk of impairment as a result fied intervention, that is, an action that results in
of exposure to or participation in a preventive a change in a potentially preventive exposure, in
intervention. Harms include physical, psycho- producing a specified change in cancer incidence.
logical, social, and economic consequences of a Scenario 2: a two-step evaluative framework
preventive intervention. Adverse events in health from which, for scientific reasons, the level of
care are a subset of harms. Evaluation of these evidence that an intervention prevents cancer is
potential harms is an important component of established by way of an intermediate outcome.
the summary of the evidence. • In Step 1, the effect of a specified intervention
Other issues to be considered include the on an intermediate outcome, such as expo-
cost, cost–effectiveness, affordability, economic sure to a particular risk factor or preventive
efficiency, health equity impact, feasibility, factor for cancer in humans, is evaluated
acceptability, relative value, and human rights (Jonas et al., 2018). Step 1 alone might be
impact of the intervention. Depending on the taken if it has been established in author-
specific intervention, some of these issues may itative sources (e.g. the IARC Monographs
be of sufficiently high interest to be reviewed in programme) that a change in the interme-
the IARC Handbook. diate outcome (decreasing exposure to a risk
factor or increasing exposure to a preventive
3. Identification and selection of factor) reduces the risk of cancer in humans.
interventions and outcomes for • In Step 2, the effect of the change in the
review intermediate outcome (decrease in exposure
to the risk factor or increase in exposure to
3.1 Development of an analytical the preventive factor) on cancer incidence
framework in humans is evaluated. Evaluation of data
streams to support Step 2 alone might be done
As one of the first steps in the review and eval- in preparation for a subsequent evaluation of
uation process of the IARC Handbooks, the IARC data to support Step 1 if it has not yet been
Secretariat, with the support of the Working established in authoritative sources that a
Group, drafts an analytical framework. Such
10
change in the intermediate outcome reduces have a significant impact on an interme-

the risk of cancer in humans. diate outcome or outcomes known or highly
The analytical framework determines suspected to be linked to cancer (see Section
whether evidence is reviewed for Step 1 only, 3.1; see also Part A, Section 6, Step 2).
Step 2 only, or both Steps 1 and 2. A Handbook In addition, an intervention previously evalu-
might, for example, include both Steps 1 and 2 ated in a Handbook may be re-evaluated if impor-
when a systematic review and evaluation of Step tant new data become available about its effects or
2 is necessary (e.g. is not yet available from other if its technology or implementation has changed
authoritative sources) and the number of studies enough for there to be substantial changes in
to be reviewed for Steps 1 and 2 is manageable. its effects. Occasionally, a re-evaluation may be
Taking Steps 1 and 2 together is equivalent to limited to one or several specific cancer sites or
Scenario 1 with inclusion of one or more inter- to specific aspects of the preventive interven-
mediate outcomes in the evaluation scheme. The tion (e.g. reduction in excess body fatness) to
sections below provide additional details on the which the new evidence predominantly relates.
selection of the interventions and outcomes for For re-evaluations, the full body of evidence
review. relevant to the intervention of interest is consid-
ered, either by de novo review of all evidence or
3.2 Selection of the interventions by accepting as accurate the evidence review of
the previously published Handbook and under-
For each new volume of the Handbooks, IARC taking a de novo review of evidence published
selects one or more interventions for review by since the previous review. Both approaches lead
considering the availability of pertinent research to an evaluation based on all relevant evidence
studies, the need to evaluate an important devel- (see Part A, Section 6, Steps 4 and 5). The choice
opment in cancer prevention, or the need to of the approach is subject to the judgement of the
re-evaluate a previously evaluated intervention. Working Group.
IARC will also consider current public health
priorities in specific geographical regions, for
3.3 Selection of the outcomes
example the concerns of countries or regions
with a high risk of specific cancer types (see Part In primary prevention of cancer, the outcome
A, Section 6, Step 1). IARC will also pay atten- targeted by the preventive intervention or inter-
tion to topics that extend beyond those covered ventions is reduction in the incidence of cancer
by other agencies. (Scenario 1; see Part A, Section 3.1).
Interventions not previously evaluated in the As described above, an intermediate outcome
IARC Handbooks series are selected for evalua- may be chosen as the evaluation outcome for a
tion, where the body of evidence is large enough Handbook when there is evidence that a change
to warrant evaluation, on the basis of one or both in the intermediate outcome (decreasing expo-
of the following criteria: sure to the risk factor or increasing exposure to
• The intervention is of putative preventive the preventive factor) can lead to a reduction in
value, but its effects have not been established the incidence of one or more types of cancer.
formally; An example of such a target is an increase in the
smoking cessation rate, which is a commonly
• The available evidence suggests that the used outcome for studies designed to deter-
intervention has the potential to signifi- mine the preventive effects of new methods of
cantly reduce the incidence of cancer, or to reducing the incidence of tobacco-caused cancer
11
Table 1 Roles of participants at IARC Handbooks meetings
Category of participant Role

Prepare text, Participate in Participate in Eligible to serve as
tables, and discussions evaluations Meeting Chair or
analyses Subgroup Chair
Working Group members ✓ ✓ ✓ ✓
Invited Specialists ✓a ✓
Representatives of health agencies ✓b
Observers ✓b
IARC Secretariat ✓c ✓ ✓d
a Only for sections not directly relevant to the evaluation
b Only at times designated by the Meeting Chair and/or Subgroup Chair
c Only when needed or requested by the Meeting Chair and/or Subgroup Chair
d Only for supporting Working Group members and for clarifying or interpreting the Preamble
by way of reducing the prevalence of tobacco (i) Working Group members have ultimate
smoking. Other examples of changes in inter- responsibility for determining the final list
mediate outcomes include a decrease in excess of studies that contribute evidence to the
body fatness, a decrease in the levels of diesel evaluation, performing the scientific review
engine emissions in urban environments, and of the evidence, and making the final, formal
an increase in the population coverage of human evaluation of the strength of evidence for the
papillomavirus (HPV) vaccination. capacity of the screening interventions to
Alternatively, a Handbook could, as a first reduce cancer incidence or cancer mortality.
step, evaluate the evidence that changing the The Working Group is multidisciplinary and
intermediate outcome can lead to a reduction is organized into Subgroups of experts in the
in the incidence of one or more types of cancer fields that the Handbook covers.
if such evidence is not already available from IARC selects the Working Group members
authoritative sources, followed by an evaluation on the basis of relevant expertise and an
of the effect of an intervention on the interme- assessment of declared interests (see Part A,
diate outcome (Scenario 2, Step 2 followed by Section 5). Consideration is also given to diver-
Step 1; see Part A, Section 3.1). An example of sity in scientific approaches, in stated positions
such a scenario is evaluation of the evidence on the strength of the evidence supporting
that reducing consumption of alcoholic bever- the intervention, and in demographic char-
ages reduces incidence of alcohol-related cancer acteristics. Working Group members gener-
or precancer, followed by evaluation of the ally have published research related to the
efficacy or effectiveness of a specific interven- interventions being reviewed or to the cancer
tion in reducing the consumption of alcoholic types or intermediate outcomes that the
beverages. interventions being reviewed are thought
to prevent or affect; IARC uses literature
4. The Working Group and other searches to identify most experts. IARC also
meeting participants encourages public nominations through its
Call for Experts. IARC’s reliance on Working
Five categories of participants can be present Group members with expertise on the subject
at IARC Handbooks meetings (Table 1): matter or relevant methodologies is supported
12
by decades of experience documenting that Specialists are invited in limited numbers,

there is value in specialized expertise and when necessary, to assist the Working Group
that the overwhelming majority of Working by contributing their unique knowledge and
Group members are committed to the objec- experience to the discussions.
tive evaluation of scientific evidence and not (iii) Representatives of national and interna-
to the narrow advancement of their own tional health agencies may attend because
research results or a predetermined outcome their agencies are interested in the subject
(Wild & Cogliano, 2011). Working Group of the Handbook. The Representatives of
members are expected to serve the public national and international health agencies
health mission of IARC and to refrain from do not draft any section of the Handbook or
using inside information from the meeting or participate in the evaluations. Representatives
meeting drafts for financial gain until the full can participate in discussions at times desig-
volume of the Handbooks is published (see nated by the Meeting Chair or a Subgroup
also Part A, Section 7). Chair. Relevant World Health Organization
IARC selects, from among the Working (WHO) staff members attend as members of
Group members, individuals to serve as the IARC Secretariat (see below).
Meeting Chair and Subgroup Chairs. (iv) Observers with relevant scientific creden-
Subgroup Chairs have preferably served in tials are admitted in limited numbers.
previous Handbooks meetings as Working Attention is given to the balance of Observers
Group members or in similar review processes. from entities with differing perspectives on
At the opening of the meeting, the Working the interventions under review. Observers
Group is asked to endorse the Meeting Chair are invited only to observe the meeting, do
selected by IARC or to propose an alterna- not draft any section of the Handbook or
tive. The Meeting Chair and Subgroup Chairs participate in the evaluations, must agree to
take a leading role at all stages of the review respect the Guidelines for Observers at IARC
process (see Part A, Section 7) to promote Handbooks meetings (IARC, 2018), and must
open scientific discussions that involve all not attempt to influence the outcomes of the
Working Group members in accordance meeting. Observers may speak at Working
with committee procedures and to ensure Group or Subgroup sessions at the discretion
adherence to the processes described in this of the Chair.
Preamble. (v) The IARC Secretariat consists of scien-
(ii) Invited Specialists are experts with critical tists who are designated by IARC or WHO
knowledge and experience on the interven- and who have relevant expertise. The IARC
tions being reviewed, the cancer types that Secretariat coordinates and facilitates all
the interventions being reviewed are thought aspects of the review and evaluation process
to prevent, or relevant methodologies, but and ensures adherence to the processes
who have a declared conflict of interest that described in this Preamble throughout the
warrants exclusion from developing or influ- development of the scientific reviews and
encing the evaluations. The Invited Specialists evaluations (see Part A, Sections 5 and 6).
do not draft any section of the Handbook that The IARC Secretariat announces and orga-
pertains to the description or interpretation nizes the meeting, identifies and invites the
of the data on which the evaluation is based, Working Group members, and assesses the
or participate in the evaluations. Invited declared interests of all meeting participants
13
in accordance with WHO requirements (see Data and a Call for Experts, is announced on
Part A, Section 5). The IARC Secretariat the Handbooks programme website (https://
supports the activities of the Working Group handbooks.iarc.fr/).
(see Part A, Section 7) by performing system- The IARC Secretariat selects potential
atic literature searches, performing title Working Group members based on the criteria
and abstract screening, organizing confer- described in Part A, Section 4. Before a meeting
ence calls to coordinate the development of invitation is extended, each potential partici-
drafts and to discuss cross-cutting issues, pant, including the IARC Secretariat, completes
and reviewing drafts before and during the the WHO Declaration of Interests form to report
meeting. Members of the IARC Secretariat financial interests, employment and consulting
serve as meeting rapporteurs, assist the (including remuneration for serving as an
Meeting Chair and Subgroup Chairs in facil- expert witness), individual and institutional
itating all discussions, and may draft text or research support, and non-financial interests,
tables or assist a Subgroup in the conduct of such as public statements and positions related
additional analyses when designated by the to the subject of the meeting. IARC assesses the
Meeting Chair or a Subgroup Chair. After declared interests to determine whether there is
the meeting, the IARC Secretariat reviews a conflict that warrants any limitation on partic-
the drafts for factual accuracy of research ipation (see Table 1).
results cited. The participation of the IARC Approximately 2 months before a meeting,
Secretariat in the evaluations is restricted to IARC publishes on the Handbooks programme
clarifying or interpreting the Preamble. website the names and principal affiliations of
all participants and discloses any pertinent and
All meeting participants are listed, with their significant conflicts of interest, for transparency
principal affiliations, in the front matter of the and to provide an opportunity for undeclared
published volume of the Handbooks. Pertinent conflicts of interest to be brought to IARC’s
interests, if any, are listed in a footnote to the attention. It is not acceptable for Observers or
participant’s name. Working Group members third parties to contact other participants before
and Invited Specialists serve as individual scien- a meeting or to lobby them at any time. Meeting
tists and not as representatives of any organiza- participants are asked to report all such contacts
tion, government, or industry (Cogliano et al., to IARC (Cogliano et al., 2005).
2004). The Working Group meets at IARC to discuss
The roles of the participants are summarized and finalize the scientific review and to develop
in Table 1. summaries and evaluations. At the opening of
the meeting, all meeting participants update
their Declarations of Interests forms, which are
5. Development of a volume of the then reviewed for conflicts of interest by IARC.
IARC Handbooks Declared interests related to the subject of the
Each volume of the Handbooks is developed meeting are disclosed to the meeting partici-
by an ad hoc, specifically convened Working pants during the meeting and in the published
Group of international experts. Approximately 1 volume of the Handbooks (Cogliano et al., 2004).
year before the meeting of a Working Group, a The objectives of the meeting are twofold:
preliminary list of interventions to be reviewed peer review of the drafts and consensus on the
(see Part A, Section 3), together with a Call for evaluations. During the first part of the meeting,
Working Group members work in Subgroups to
14
Table 2 Public engagement during the development of a volume of the IARC Handbooks
Approximate time frame Milestones

~1 year before a Handbooks meeting IARC posts on the Handbooks programme website:
Preliminary List of Interventions to be reviewed
Call for Data and Call for Experts open
Requests for Observer Status open
WHO Declarations of Interests form
~8 months before a Handbooks meeting Call for Experts closes
~4 months before a Handbooks meeting Requests for Observer Status close
~2 months before a Handbooks meeting IARC publishes the names, principal affiliations, and declared conflicts of
interest of all meeting participants, and a statement discouraging contact
of Working Group members by outside parties
~1 month before a Handbooks meeting Call for Data closes
Handbooks meeting
~2–4 months after a Handbooks meeting IARC publishes a summary of evaluations and key supporting evidence
as a scientific article in a high-impact journal or on the Handbooks
programme website
~9–12 months after a Handbooks meeting IARC Secretariat publishes the verified and edited master copy of the
plenary drafts as a Handbooks volume
review the pre-meeting drafts, develop a joint 6. Overview of the scientific review
Subgroup draft, and draft Subgroup summaries. and evaluation process
During the last part of the meeting, the Working
Group meets in plenary sessions to review the Principles of systematic review are applied
Subgroup drafts and summaries and to develop to the identification, screening, synthesis, and
the consensus evaluations. As a result, the entire evaluation of the evidence (as described in Part
volume is the joint product of the Working Group B, Sections 2–6 and detailed in the Instructions
and there are no individually authored sections. for Authors). For each volume of the Handbooks,
After the meeting, the master copy is verified by the information on the conduct of the literature
the IARC Secretariat (see Part A, Section 4(v)), searches, including search terms and the inclu-
edited, and prepared for publication. The aim sion and exclusion criteria that were used for
is to publish the volume of the Handbooks each relevant stream of evidence, is recorded.
within approximately 12 months of the Working The Working Group considers all relevant
Group meeting. The IARC Secretariat prepares studies, including pertinent reports and reviews
a summary of the outcome for publication on: use of the intervention targeted directly to
in a scientific journal or on the Handbooks cancer or to a relevant intermediate outcome or
programme website soon after the meeting. outcomes; all experimental and observational
The time frame and milestones for public studies in humans (including systematic reviews
engagement during the development of a volume and meta-analyses) of the putative effect of the
of the IARC Handbooks are summarized in intervention or interventions on cancer inci-
Table 2. dence or a relevant intermediate outcome, and
any related harms; all relevant experimental
studies in animals; and all relevant mechanistic
studies.
15
In general, only studies that have been the searches rerun, depending on the amount,
published or accepted for publication in the relevance, and perceived completeness of the
openly available scientific literature are reviewed. articles they identify. The IARC Secretariat may
Materials that are publicly available and whose also identify relevant studies from reference lists
content is final may be reviewed if there is suffi- of past Handbooks, retrieved articles, or author-
cient information to enable peer evaluation of the itative reviews, and through the Call for Data
quality of the methods and results of the studies (see Table 2). The Working Group provides input
(see Step 1, below). Such material may include and advice to the IARC Secretariat to refine the
reports from government agencies, disserta- search strategies, and identifies additional arti-
tions for higher degrees, and other apparently cles through other searches and personal expert
reputable scientific sources. Systematic Internet knowledge.
searches for potentially relevant “grey literature” For certain types of interventions (e.g. admin-
are not usually done. The reliance on published istration of regulated pharmaceuticals), IARC
and publicly available studies promotes trans- also gives relevant regulatory authorities, and
parency and protects against citation of infor- parties regulated by such authorities, an oppor-
mation that, although purportedly final, may tunity to make pertinent unpublished studies
change before it is published. publicly available by the date specified in the
The steps of the review process are as follows: Call for Data. Consideration of such studies by
Step 1. Identification of the review question: the Working Group is dependent on the public
After the intervention (or interventions) and availability of sufficient information to enable an
outcome (or outcomes) to be reviewed have been independent peer evaluation of: (i) completeness
specified, the IARC Secretariat, in consulta- of reporting of pertinent data; (ii) study quality;
tion with the Working Group, drafts the review and (iii) study results.
question (or questions) in PICO form (popula- Step 3. Screening, selection, and organiza-
tion, intervention/exposure, comparator, and tion of the studies: The IARC Secretariat screens
outcome) as required to determine the inclusion the retrieved articles by reviewing the title and
and exclusion criteria for the studies. An analyt- abstract against the inclusion and exclusion
ical framework is developed to assist in identi- criteria agreed upon by the Working Group
fying and formulating the review questions, and and technical experts in the review process.
encompasses the inclusion of studies in humans, Potentially relevant studies are then made avail-
studies in experimental animals, and mecha- able to Working Group members for full-text
nistic studies when relevant, with the aim of screening and inclusion in or exclusion from the
making as large a contribution as possible to the evidence base using agreed criteria specific to
global prevention of cancer. this task.
Step 2. Comprehensive and transparent iden- Step 4. Extraction of information from included
tification of the relevant information: The IARC studies, including characteristics relevant to study
Secretariat specifies search terms for the key quality: Working Group members, working indi-
PICO components of each question and identifies vidually as members of defined Subgroups before
relevant studies through initial comprehensive the Handbooks meeting, review and succinctly
literature searches in authoritative biomedical describe pertinent characteristics and results of
databases (e.g. PubMed). The literature searches included studies as detailed in Part B, Sections
are designed in consultation with a librarian and 2–4. Study design and results are tabulated
other technical experts. The scope and speci- systematically in a standard format. This step
fications of the searches may be modified, and may be iterative with Step 5.
16
Step 5. Assessment of study quality: Also Step 8. Interpretation of study results and
before the Handbooks meeting, Working Group evaluation of strength of evidence: The whole
members evaluate the quality and informative- Working Group reviews the study descriptions
ness of each study they included based on the and the summaries of the body of evidence for
considerations (e.g. design, conduct, analysis, each outcome or end-point, discusses the overall
and reporting of results) described in Part B, strengths and limitations of the evidence in
Sections 2–4. Evaluation of study quality can be each stream of data, and evaluates the strength
done either narratively or by use of a risk of bias of evidence for a preventive effect on cancer or
assessment tool when a relevant one is available an intermediate outcome in each stream using
and can add value to the process. Interpretations transparent methods, which may include the use
of the results, and the strengths and limitations of established specific tools. The preventive effect
of each study, are clearly outlined in square is described in terms given in Part B, Sections
brackets as part of the description of that study 6a–c for each stream of evidence. The Working
(see Part B). Group then integrates the strength-of-evidence
Step 6. Peer review: Several months before conclusions from all streams of evidence (see Part
the meeting, the pre-meeting drafts produced B, Section 6d) and develops the rationale for its
from Steps 4 and 5 are peer-reviewed by other overall consensus evaluation of the cancer-pre-
members of the Working Group (usually within ventive effect of the intervention (see Part B,
the same Subgroup). The IARC Secretariat also Sections 6d–e).
reviews the drafts for completeness, consistency
between drafts, and adherence to the Handbooks
Instructions for Authors. The peer-review
7. Responsibilities of the Working
comments are sent to the Working Group Group
members, who produce a revised pre-meeting The Working Group is responsible for the
draft. The revised drafts are reviewed and revised final list of studies included in the evaluation
in Subgroup sessions during the Handbooks and the review and evaluation of the evidence
meeting. for a Handbook, as described above. The IARC
Step 7. Synthesis of results and quality of the Secretariat supports these activities (see Part A,
studies: The results and quality of the included Section 4). To ensure that the process is rigorous,
studies are synthesized by the Working Group independent, and free from individual conflicts
to provide a summary of the evidence and its of interest, Working Group members must accept
quality for each outcome. This synthesis can the following responsibilities:
be narrative or quantitative (for details, see
the Instructions for Authors), and the quality (i) Before the meeting, Working Group
synthesis may include use of an overall quality members:
of evidence assessment tool, such as GRADE • help in developing the analytical frame-
(Siemieniuk & Guyatt, 2019). work;
Meta-analyses of large bodies of evidence • ascertain that all appropriate studies have
may be performed by the Working Group and/ been identified and selected;
or by the IARC Secretariat before the meeting
if such meta-analyses would assist in evidence • assess the methods and quality of each
synthesis and evaluation. For more information included study;
on the conduct and use of such meta-analyses, • prepare pre-meeting drafts that present
see Part B, Section 2.1d. an accurate quantitative and/or textual
17
synthesis of the body of evidence, with key proposed by the Meeting Chair using the
elements of study design and results and guidance provided in Part B, Section 6d.
notable strengths and limitations; The Working Group strives to achieve
• participate in conference calls organized consensus evaluations. Consensus reflects broad
by the IARC Secretariat to coordinate the agreement among the Working Group members,
development of pre-meeting drafts and to but not necessarily unanimity. If unanimity has
discuss cross-cutting issues; and not been reached when the interpretations of the
• review and provide comments on evidence by all Working Group members have
pre-meeting drafts prepared by other been expressed and debated, the judgement of
members of their Subgroup or of the the majority of the Working Group members
Working Group. is taken as the consensus. When consensus
is reached in this way, the Meeting Chair may
(ii) At the meeting, Working Group members poll Working Group members to determine and
work in Subgroups to: record the diversity of scientific opinion on the
• critically review, discuss, and revise the overall evaluation.
pre-meeting drafts and adopt the revised Only the final product of the plenary sessions
versions as consensus Subgroup drafts; represents the views and expert opinions of the
and Working Group. The Handbook is the joint
• develop and propose an evaluation of the product of the Working Group and represents
strength of the evidence summarized in an extensive and thorough peer review of the
the consensus Subgroup drafts (see Part B, body of evidence (review of individual studies,
Section 5), using the IARC Handbooks synthesis, and evaluation) by a multidisciplinary
criteria (see Part B, Section 6a–c). group of experts. Initial pre-meeting drafts and
subsequent revisions are temporarily archived
(iii) At the meeting, Working Group members but are not released, because they would give
work in plenary sessions to: an incomplete and possibly misleading impres-
• present their Subgroup drafts for scientific sion of the consensus developed by the Working
review by and discussion with the other Group over its complete deliberation.
Working Group members, and subsequent
revisions, as needed;
• participate in review and discussion of B. SCIENTIFIC REVIEW AND
other Subgroup drafts and in their adop- EVALUATION
tion as a consensus Working Group draft;
• participate in review and discussion of the This part of the Preamble discusses the types
summaries and evaluations of the strength of evidence that are considered and summarized
of the evidence developed in Subgroups in each section of a Handbook, followed by the
(see Part B, Sections 6a–c), and contribute scientific criteria that guide the evaluations. In
to their revision, as needed, and their addition, a section of General Remarks at the
adoption by consensus of the full Working front of the volume discusses the reasons the
Group; and interventions were scheduled for evaluation and
• contribute to the discussion of and adop- any key issues encountered during the meeting.
tion by consensus of an overall evaluation
18
1. Intervention and outcome experimental and observational epidemiological

characterization studies are described and evaluated. This section
also reports on validated biomarkers of internal
An intervention for primary cancer preven- exposure, metabolites, or other intermediate
tion has been defined in this Preamble to be any outcomes that are routinely used for exposure
action aimed at reducing the incidence of cancer assessment. Concepts of absorption, distribu-
in humans (Part A, Section 2). Given this defi- tion, metabolism, and excretion, where relevant,
nition, the efficacy or effectiveness of an inter- are considered in the section on mechanistic
vention would be most directly approached by evidence (see Part B, Section 4b).
research that examines whether the delivery of
the intervention results in a measurable change (a) Identification of the intervention
in a cancer-related exposure that leads to a The intervention being evaluated is unam-
reduction in the incidence of cancer. However, biguously identified. The information provided
such research is often lacking, and therefore the will vary widely depending on the type of inter-
possibility of cancer-preventive effects has often vention but should be sufficient to enable the
been inferred from static associations of cancer implementation of an intervention in practice
incidence with prevalence of exposure to cancer- with reasonable confidence that its outcomes
causing agents or cancer-preventive agents. For in populations would be similar to those of the
example, all measures that are now taken to intervention from which the bulk of the evidence
minimize environmental exposure to asbestos evaluated in the Handbook originated.
(e.g. regulation of removal of asbestos from Many interventions are multifaceted and
buildings or demolition of buildings known to comprise complex sets of actions. Interventions
contain asbestos) are based on the very strong determined by personal behaviour or circum-
evidence that people who have had identifiable stances may result from, be influenced by, or be
exposure to asbestos have a higher incidence of correlated with a diverse range of behavioural
cancer than people who have not had such expo- and environmental factors, such as smoking,
sure. Similarly, the evaluation of Handbooks alcohol consumption, diet, sleep and physical
Volume 16 that there “is sufficient evidence in activity patterns, remoteness of residence, and
humans for a cancer-preventive effect of absence socioeconomic circumstances. The description
of excess body fatness” is almost exclusively based of such interventions should include their vari-
on the substantial body of evidence that cancer ability across human populations and environ-
incidence is lower in people without excess body ments, and their known relationships with other
fatness than it is in people with excess body health-determining factors.
fatness; this is a static comparison, not a dynamic
comparison as the term “intervention” implies. (b) Global occurrence and use
Geographical patterns and time trends in
1.1 Intervention characterization occurrence are summarized. A concise overview
of quantitative information about sources, prev-
This section provides informative back-
alence, and levels of individual and population
ground on the intervention and the factors that
interventions, whether purposive or incidental,
mediate it. It also summarizes the prevalence
is provided. Representative data from formal
and level of the intervention across geographical
environmental or behavioural monitoring or
areas and across the life-course. Methods used
surveillance data, research studies, government
to assess exposure to the intervention in key
reports and websites, online databases, and other
19
citable, publicly available sources are tabulated. these dimensions. Interpretation of information
Data from low- and middle-income countries are for chemical, biological, or physical interventions
sought and included to the extent that is feasible; may also be informed by consideration of mech-
information gaps for key regions are noted. anistic evidence on absorption, distribution,
If available, data are reported by region and metabolism, and excretion (e.g. as described in
by other relevant characteristics, such as sex, Part B, Section 4b).
age, socioeconomic status, and other variables In experimental epidemiological studies,
considered relevant by the Working Group. the investigators determine, usually by way of
randomization, who will and who will not be
(c) Regulations and guidelines assigned to the intervention; however, in prac-
Regulations or guidelines that have been tice the assignment is not always adhered to.
established for the intervention (e.g. permissible Therefore, a critical assessment of such studies
levels of fortification in food, national dietary requires careful evaluation using appropriate
guidelines) are described and may be tabulated guidelines or assessment frameworks (e.g. fidelity
if they are informative for the interpretation of to intervention implementation and extent of
current or historical levels of the intervention. non-adherence to intervention).
Information on applicable populations, the basis Intervention intensity and timing in obser-
for regulation, and the timing of regulation may vational epidemiological studies can be char-
be noted. acterized by using environmental monitoring
data, records from workplaces or other sources,
(d) Intervention assessment in key and subject or proxy reports collected by way of
epidemiological studies questionnaires or interviews. Both objective and
Epidemiological studies reviewed in the subjective data sources are used, individually
context of the IARC Handbooks programme or in combination, to assign levels or values of
evaluate cancer prevention interventions (or an intervention metric to members of the study
effects on intermediate outcomes) by comparing population.
outcomes across groups differently exposed to Key epidemiological studies with inter-
changes in a putative cancer-preventing interventions on cancer or intermediate outcomes
vention. Therefore, the type and the quality of are identified, and the intervention assessment
intervention assessment methods used are key approach and its strengths and limitations are
considerations when interpreting study findings. summarized in text and tables. The Working
This section summarizes and critically reviews Group identifies concerns about intervention
the intervention assessment methods used in assessment methods and their impacts on the
both experimental and observational epidemi- overall quality of each study reviewed. The
ological studies that contribute data relevant to Working Group notes the studies where the
the Handbooks evaluation. information provided to characterize the inter-
All interventions have two principal dimen- vention properly, the adherence to the intended
sions: (i) dose (sometimes defined as concentra- intervention in each arm of experimental studies,
tion or intensity), and (ii) time considerations, or the assessment of the intervention in observa-
including duration (time from first to last tional studies is inadequate. The Working Group
exposure), pattern or frequency (whether further discusses the likely direction of bias due
continuous or intermittent), and windows of to non-adherence or to error in intervention
susceptibility. This section considers how each assessment in studies where adequate informa-
of the key epidemiological studies characterizes tion is available.
20
1.2 Outcome characterization In what follows, the term “cancer incidence”

refers to the outcome of a Handbooks evalua-
(a) Evaluation of cancer outcomes tion, that is, to the incidence of cancer or of an
The cancers are defined and described in intermediate outcome, as defined in the analyt-
terms of their International Classification of ical framework.
Diseases for Oncology (ICD-O) (IARC, 2019)
or International Classification of Diseases (ICD)
categories, with other relevant morphological or
2. Studies of cancer prevention in
molecular characteristics where relevant. humans
Benign neoplasms, pre-neoplastic lesions, This section includes all pertinent exper-
malignant precursors, and other end-points imental and observational studies in humans
closely related to cancer may also be reviewed that include cancer or a specified intermediate
when they relate to the intervention reviewed outcome (if it is the topic of the Handbook) as
and are known to predict the primary cancer a study outcome. As noted above, only observa-
outcome. These studies can strengthen evidence tional studies in which changes in the exposure
from studies of cancer itself. For example, the (i.e. intervention) in relation to the outcome
results of controlled trials of sun protection have been analysed will be considered, unless
measures in preventing development of cuta- specifically stated otherwise. Among many
neous melanocytic naevi (which are strong risk others, these studies also encompass studies with
factors for development of later cutaneous mela- biomarkers as intervention metrics (Alexandrov
noma) in children provide support for the effi- et al., 2016). As mentioned above, studies that
cacy of sun protection measures in preventing assess biomarkers of early biological effects are
cutaneous melanoma in adults (Thun et al., 2018). reviewed in Part B, Section 4.
(b) Evaluation of intermediate outcomes This section includes specification and
assessment of beneficial effects, as well as poten-
Potentially relevant intermediate outcomes tial harms.
vary widely across human biology, pathology,
and behaviour. (Intermediate outcomes that are
2.1 Assessment of beneficial effects
biomarkers of early biological effects, which are
not topics evaluated in IARC Handbooks, are (a) Types of studies considered
described in Part B, Section 4.) All intermediate Several types of epidemiological study
outcomes are described as precisely as possible, designs contribute to the evaluation of cancer
using an applicable international standard clas- prevention in humans (Table 3). These studies
sification (e.g. ICD classification). When, as with include experimental studies and different
some behavioural or physiological risk factors, types of observational studies (i.e. cohort, case–
they can be defined or measured in a range of control, and ecological). In addition to these
ways, the definitions that are acceptable for the types of studies, innovations in epidemiology
evaluation are clearly defined and acceptable enable other designs that may be considered in
standards for measurement stated. Handbooks evaluations. (b) Identification of
When an intermediate outcome is the eligible studies in humans
outcome being evaluated, the evidence base Relevant studies in humans are identified
establishing that the intermediate outcome has using principles of systematic review as described
an established causal or preventive association in Part A and further detailed in the Instructions
with cancer incidence is briefly summarized.
21
Table 3 Types of epidemiological studies that contribute to the evaluation of cancer prevention
Experimental studies
• High level of investigator control over assignment to the intervention and non-intervention
group
• Ideally random assignment, either of individuals or of groups, to the intervention and non-
intervention group
• Provides evidence for the efficacy or effectiveness of a preventive intervention
• Includes a range of quasi-experimental designs in which there is lack of random assignment to
the intervention and non-intervention; quasi-experimental studies are often at high risk of bias
Observational (non-experimental) studies
Cohort • In a prospective cohort study, information on the intervention and non-intervention is collected
from individuals who are then followed up over time to assess subsequent outcomes. Further
intervention information may be collected at intervals during follow-up.
• In a retrospective cohort study, information on intervention and subsequent outcomes in a
defined group of individuals, which was usually recorded for purposes other than research, is
accessed after the outcomes have occurred.
• Nested within these studies, case–control and case–cohort studies provide efficiency and an
opportunity to collect additional intervention information.
Case–control • In a case–control study, individuals newly diagnosed with the outcome in a defined population
and a sample of “control” individuals without the outcome from the same source population and
time period are enrolled, and their intervention histories are compared.
• Intervention information collected from cases and controls must refer to time before disease
onset to reasonably infer a temporal association.
Mendelian randomization • Mendelian randomization studies are cohort or case–control studies in which an intervention is
inferred using appropriate genomic surrogate(s) (Yarmolinsky et al., 2018).
• These studies are considered to be less prone to bias than other observational studies because the
genomic variants from which intervention is inferred are randomly allocated at conception.
Ecological • The association between an intervention and an outcome is examined not in individual people
but in units of population defined geographically and/or temporally. Uncontrolled confounding
is a major issue for ecological studies.
• Results from ecological studies can support a hypothesis about an intervention–outcome
association or, when taken together with results of case–control and cohort studies, support
judgements on causal associations.
• Results may be persuasive when population-wide implementation of an intervention leads to
changes in cancer incidence or mortality: (a) in several populations, and there is no similar trend
in similar populations not, or much less, subject to the intervention (e.g. Hakama, 1983); or
(b) in a single population, by use of time series analysis when longitudinal data on both the
intervention and the outcome are available (e.g. Bernal et al., 2017).
for Authors provided to each Working Group. focused on different aspects of an interven-
Eligible studies include all studies in humans of tion–outcome association, or from inclusion of
the association of a putative cancer-preventive overlapping populations. In these situations, the
intervention with the occurrence of cancer, or most recent or most informative report is usually
a specified intermediate outcome if it is a topic reviewed first, with recourse to the other reports
of the Handbook. Multiple publications on the if important information (e.g. methodological
same study population are identified so that the detail) is not included in the most recent or most
number of independent studies is accurately informative report.
represented. Multiple publications may result,
for example, from successive follow-ups of a
single trial population or cohort, from analyses
22
(c) Study quality and informativeness can cause an association between hypothesized
Epidemiological studies are susceptible to cause and effect to appear stronger or weaker
several different sources of error. Study quality is than it really is. Confounding arises when a third
assessed as part of the structured expert review factor is associated with both the intervention and
process undertaken by the Working Group. A the outcome and, because of this, influences the
key aspect of quality assessment is consideration apparent association between them (Rothman
of the possible roles of chance and bias in the et al., 2008). An association between the inter-
interpretation of epidemiological studies. vention and another factor that is associated with
Chance, also called “random variation”, can an increase or a decrease in the incidence of or
produce misleading study results. This vari- mortality from the disease can lead to a spurious
ability in study results is strongly influenced by association or the absence of a real association of
the sample size: smaller studies are more likely the intervention with the outcome. When either
than larger studies to have effect estimates that of these occurs, confounding is present.
are imprecise and, therefore, are more likely In principle, experimental studies are less
to be misleading. Confidence intervals around prone to each of these sources of bias, because
a study’s point estimate of effect are routinely selection for intervention or non-intervention is
used to indicate the range of values of the esti- determined by the investigator (usually by random
mate that could be produced by chance. Both allocation) and not by the study participants or
experimental and observational epidemiological their characteristics. However, bias may still arise
studies are prone to effects of chance, and experi- as a result of lack of concealment, non-random
mental studies are arguably more prone, because allocation, lack of blinding, post-randomization
of their smaller sample sizes, associated with the exclusions, non-acceptance of or non-adher-
greater cost of conducting such studies. ence by the study participants to the interven-
Bias is the effect of factors in study design, tion condition of the study arm to which they
conduct, or reporting that lead an association to are randomized, or study loss to follow-up. One
erroneously appear stronger than, weaker than, potential shortcoming of randomized studies is
or opposite in direction to the association that their potentially limited external validity (rele-
really exists between an intervention and an vance) and consequently limited generalizability
outcome. Biases that require consideration are to non-studied populations.
varied and can be broadly categorized as selec- In assessing the quality of the studies, the
tion bias, information bias, and confounding Working Group considers the following aspects:
bias (Rothman et al., 2008). Selection bias in an • Study description: Clarity in describing the
epidemiological study can occur when the inclu- study design, implementation, and conduct,
sion of participants from the eligible population and the completeness of reporting of all
or their follow-up in the study is influenced by other key information about the study and
their intervention status or their outcome (usually its results.
disease occurrence). Under these conditions, the • Study population: Whether the study popu-
measure of association found or not found in the lation was appropriate for evaluating the
study may not accurately reflect the association association between the intervention and the
or lack thereof that might otherwise have been outcome. Whether the study was designed
found in the eligible population (Hernán et al., and conducted in a manner that would
2004). Information bias results from inaccuracy minimize selection bias and other forms of
in intervention or outcome measurement. Both bias. The designated outcomes in the study
23
population must have been identified in a confounding variables (including co-ex-

way that was independent of the intervention posures, as described in Part B, Section 1d)
of interest, and the intervention must have that could influence the occurrence of the
been assessed in a way that was not related to outcome and may be related to the interven-
outcome status. In these respects, complete- tion of interest. Important sources of poten-
ness of recruitment into the study from the tial confounding by such variables should,
population of interest (which is less of an where possible, have been addressed in the
issue for experimental efficacy studies than study design, such as by randomization,
for effectiveness studies and observational matching, or restriction, or in the analysis
studies) and completeness of follow-up for by statistical adjustment. In some instances,
the outcome (see below) are very important. where direct information on confounders is
• Outcome measurement: The appropriate- unavailable, use of indirect methods to eval-
ness of the outcome measure (incidence of uate the potential impact of confounding
cancer, mortality from cancer, or an interme- on intervention–outcome associations is
diate outcome, as defined in Part B, Section appropriate (e.g. Axelson & Steenland, 1988;
1.2) for the intervention and the cancer type Richardson et al., 2014).
under consideration, the outcome ascertain- • Other potential sources of bias: Each
ment methodology, and the extent to which epidemiological study is unique in its study
outcome misclassification may have led to bias population, its design, its data collection,
in the measure or measures of association. and, consequently, its potential biases. For
• Intervention measurement: This includes: example, repeated assessments of exposure to
(i) the adequacy (including the validity and the intervention over time can be influenced
the reliability) of the methods used to assess by the occurrence of the outcome and thus
the intervention in observational studies, and bias the result and sometimes lead to “reverse
adherence to the intervention condition in causation”. All possible sources of bias are
experimental studies, and (ii) the likelihood considered for their possible impact on the
(and direction) of bias in the measure or results, including the possibility of reporting
measures of association because of interven- bias (selective reporting of some results).
tion measurement error or misclassification • Statistical methodology: The studies are
in observational studies and non-adherence evaluated for the adequacy of the statistical
to the intervention condition in experimental analysis methods used and their ability to
studies (see Part B, Section 1.1. Of particular obtain unbiased estimates of intervention–
relevance is an assessment of the error asso- outcome associations, confidence intervals,
ciated with the measurement of change over and test statistics for the significance of
time in several study designs, including measures of association. Appropriateness
prospective longitudinal studies (e.g. change of methods used to address confounding,
in body weight estimated from contemporary including adjusting for matching when
recall of past body weight and self-reported necessary and avoiding treatment of prob-
or measured current body weight at recruit- able mediating variables as confounders, is
ment into a cohort study). considered. For example, the use of directed
• Assessment of potential confounding: The acyclic graphs can inform about whether
extent to which the authors took into account confounding and selection biases have been
in the study design and analysis potentially specified correctly (Hernán et al., 2004).
24
Detailed analyses of cancer risks in relation (d) Meta-analyses and pooled analyses
to summary measures of intervention, such Independent epidemiological studies of the
as cumulative exposure to the intervention, or same intervention with a comparatively weak
temporal variables, such as age at first inter- effect or small sample size may produce incon-
vention or time since first intervention, are clusive results that are difficult to summarize.
reviewed and summarized when available. Combined analyses of data from multiple studies
For the sake of economy and simplicity, this may increase the precision of estimates. There are
Preamble refers to the list of possible sources two types of combined analysis: (i) meta-analysis,
of error with the phrase “chance, bias, and which involves combining summary statistics,
confounding”, but it should be recognized that such as relative risks from individual studies;
this phrase encompasses a comprehensive set and (ii) pooled analysis, which involves a pooled
of concerns pertaining to study quality. These analysis of the raw data from the individual
elements of study quality do not constitute and studies (Greenland & O’Rourke, 2008). There are
should not be used as a formal checklist of indi- also “umbrella reviews”, systematic reviews of
cators of study quality. Rather, the assessment multiple meta-analyses, which may be evaluated
by the Working Group is reported in a narrative by the Working Group.
way, in the form of comments in square brackets. The strengths of combined analyses are
The judgement of the experts is critical in deter- increased precision due to increased sample size
mining how much weight to assign to different and, in the case of pooled studies, the opportu-
issues when considering how all these potential nity to better control for potential confounders
sources of error should be integrated and how and to explore interactions and modifying effects
to rate the potential for error related to each. that may help to explain heterogeneity between
However, it is important that the process under- studies. A disadvantage of combined analyses is
taken, including the weight given to various the possible lack of comparability of results from
studies, be replicable and be described in a way various studies, because of differences in specifi-
that is transparent to readers. cation of the intervention or the outcome, popu-
• Study informativeness: The informativeness lation characteristics, subject recruitment, data
of a study is its ability to show a true preventive collection procedures, methods of measurement,
effect, if one exists, between the intervention and effects of unmeasured covariates, which may
and the outcome in a relevant population, differ among studies. These differences in study
and not to show an effect if one does not exist. methods and quality can influence the results of
Key determinants of informativeness include both pooled analyses and meta-analyses.
having a study population of sufficient size to Meta-analyses considered by the Working
obtain precise estimates of effect, sufficient Group may include high-quality published
elapsed time from intervention to measure- meta-analyses, updates of such meta-anal-
ment of outcome for an effect, if present, to yses, and new meta-analyses. When published
be observable, presence of at least moderate meta-analyses are considered by the Working
heterogeneity of exposure to the intervention Group, they should comply with basic quality
(intensity, frequency, and/or duration) in the standards for meta-analyses and their under-
study population, and biologically relevant lying systematic reviews (e.g. AMSTAR, 2017):
definitions of the intervention. their risk of bias is carefully evaluated, including
the completeness of the studies included, the
methods used to identify and the criteria used
25
to select eligible studies, and the accuracy of the than a single study in a single population will
data extracted from the individual studies. almost always be needed.
Subject to the judgement of the IARC After the quality of individual epidemio-
Secretariat and in consultation with the logical studies of cancer or of an intermediate
Working Group, the updating of meta-analyses outcome has been assessed and the informa-
or the conduct of ad hoc meta-analyses may be tiveness of the various studies on the associ-
performed by the Working Group and/or by ation between the intervention and cancer or
the IARC Secretariat during preparation for a an intermediate outcome has been evaluated,
Handbooks meeting, when there are sufficient the body of evidence is assessed and a consensus
studies of an intervention–outcome association scientific judgement is made about the strength
to aid the Working Group’s assessment of the of the evidence that the intervention under
association. When results from both experi- review prevents cancer in humans. In making its
mental and observational studies are available, judgement, the Working Group considers several
any combined analyses should be conducted aspects of the body of evidence (e.g. Hill, 1965;
separately for experimental and observational Rothman et al., 2008; Vandenbroucke et al.,
studies, with consideration given to separate 2016).
combined analyses of cohort and case–control A strong association (e.g. a large relative
studies, because of their different propensities to risk or a relative risk that is well below 1.0) is
bias. The results of such ad hoc meta-analyses, more likely to be causal than a weak associ-
which are specified in the text of the Handbook ation, because it is harder for confounding
by presentation in square brackets, may come or other biases to create a false strong associ-
from the addition of the results of more recent ation. However, it is recognized that estimates
studies to those of published meta-analyses or of effect of small magnitude do not imply lack
from de novo meta-analyses. Additional details of causality and may have a substantial impact
on the conduct of such ad hoc meta-analyses are on public health if the outcome is common or if
provided in the Instructions for Authors. the intervention is highly feasible. Estimates of
Irrespective of the source of the informa- effects of small magnitude can also contribute
tion for the meta-analyses and pooled analyses, useful information if the magnitude of the effect
the criteria for information quality applied are correlates with the level of intervention in popu-
the same as those applied to individual studies. lations that are differently exposed.
The sources of heterogeneity among the studies Associations that are consistently observed
contributing to them are carefully considered in several studies of the same design, in studies
and the possibility of publication bias evaluated. that use different epidemiological approaches,
or under different circumstances of intervention
(e) Considerations in assessing the body of are more likely to indicate preventive efficacy or
epidemiological evidence effectiveness than are isolated observations from
The ability of the body of epidemiological single studies. If there are inconsistent results
evidence to inform the Working Group about among investigations, possible reasons for such
the cancer-preventive effect of an intervention is inconsistencies are sought – such as differences in
related to both the quantity and the quality of time since initiation of the intervention (latency),
the evidence. There is no formulaic answer to the intervention levels (e.g. dosage), or assessment
question of how many cancer prevention studies methods – and their implications for the overall
in humans are needed from which to draw infer- findings are assessed.
ences about preventive effect, although more
26
Results of studies that are judged to be of high to be modifiers of the intervention–outcome

quality and highly informative are given more relationship, because evaluation of polymor-
weight than those of studies that are judged to be phisms may increase the ability to detect an
methodologically less sound or less informative. effect in susceptible subpopulations. Identifying
Temporality of the association is also an susceptible subpopulations can also improve the
essential consideration, that is, the intervention specificity of targeting interventions.
must precede the outcome. The likelihood of
reverse causation (i.e. the outcome prompts the 2.2 Harms of the intervention
intervention) is greater in observational studies
of interventions, which often entail self-reported Potential harms to individuals that are
behaviour change, than in studies of static linked to the intervention under review are also
exposures. reviewed. Evidence of harm may come from
An observation that cancer incidence any type of epidemiological study and may also
decreases with increasing exposure to a putative be reported separately from evidence on the
preventive intervention is considered to be an potential beneficial effects of the intervention.
indication of a preventive effect, although the Although the IARC Handbooks do not formally
absence of a graded response is not necessarily evaluate the harms associated with an interven-
evidence against a causal relationship, and there tion in the way that is done for the benefits, the
are several reasons why the shape of the inter- review of the evidence of harms aims to be as
vention–outcome association may be non-mono- complete, rigorous, and informative as it is for
tonic (e.g. Stayner et al., 2003). the evidence of beneficial effects.
Confidence in a causal interpretation of the There are three broad categories of possible
evidence from studies in humans is enhanced if harms associated with interventions: (i) biological
it is coherent with physiological and biological harm (e.g. toxicity of a chemopreventive agent),
knowledge, including information about target (ii) physical harm (e.g. injury associated with
organ exposure to the intervention, characteris- increased physical activity), and (iii) psychoso-
tics of tumour subtypes, and evidence of biolog- cial harm (e.g. community-based interventions
ical mechanisms by which the intervention and social marketing campaigns specifically
could exert a cancer-preventive effect (see Part targeting obesity; Walls et al., 2011). Evidence of
B, Section 4b). occurrence of biological, physical, and psycho-
The Working Group considers whether or social harm (including emerging harms identi-
not there are subpopulations with increased fied using qualitative methods in intervention
susceptibility to the cancer-preventive effects studies) is reviewed and described, and the
of the intervention. For example, studies that potential impacts of the harm are discussed.
identify inter-individual differences in cancer Known financial harms or opportunity costs
susceptibility to the intervention on the basis (Walls et al., 2011), which can apply at the indi-
of sociodemographic characteristics (e.g. age, vidual level (e.g. higher cost of healthy foods,
sex, race, ethnicity), other behavioural factors impacts of increases in tobacco taxes on smokers
(e.g. smoking or alcohol consumption), genetic of lower socioeconomic status, membership of
polymorphisms, or age at first intervention (e.g. a weight-loss plan) or the community level (e.g.
childhood interventions) may contribute to the community-based interventions and campaigns),
identification of cancer-preventive interventions may be noted.
in humans. Such studies may be particularly
informative if genetic polymorphisms are found
27
2.3 Balance of benefits and harms Taking a similar approach to that taken for
the balance of benefits and harms described
Ideally, the benefits and harms of primary above, the Working Group identifies published
prevention interventions are expressed in similar reports of well-conducted cost–effectiveness
terms, such as quality-adjusted life years (QALYs) analyses based on the highest-quality evaluative
gained (benefits) or lost (harms) per 1000 individ- studies of the primary preventive intervention,
uals of the target population. After identification critically assesses each, and summarizes the
of all published estimates of the balance of bene- results, in narrative or tabular format as appro-
fits and harms based on the same combination priate. The results do not contribute to the overall
or combinations of intervention and outcome, evaluation of each intervention, but they may be
the Working Group selects those based on the highlighted in the rationale after the evaluation
highest-quality evaluative studies of the inter- and can be used by governments and health
vention, critically assesses each, and summa- services to aid decisions about implementation
rizes the results, in narrative or tabular format of the intervention for which there is sufficient
as appropriate. The results do not contribute to evidence of a preventive effect. In addition, it is
the overall evaluation of each intervention, but important to note that when the intervention is
they may be highlighted in the rationale after the targeted towards a risk factor for cancer that is
evaluation and can be used to aid decisions about also a risk factor for other chronic diseases, any
implementation of and participation in the rele- estimate of cost–effectiveness that is based solely
vant primary preventive interventions. on cancer is of limited use for policy purposes.
2.4 Cost–effectiveness
3. Studies of cancer prevention in
For a primary preventive intervention that
can deliver a beneficial outcome, cost–effective-
experimental animals
ness is usually expressed as the estimated financial (a) Types of study considered
cost of implementing the intervention per unit of
Animal models are an important component
benefit it delivers, which is most often measured
of research on cancer prevention. Models are
in terms of QALYs gained. The ratio of costs to
available that enable the evaluation of the effects
benefits (i.e. level of cost–effectiveness) needed
of interventions on the development or progres-
to implement a health service programme varies
sion of cancer in most major organ sites. Animal
from country to country, depending principally
models for cancer include: (i) carcinogen-in-
on the wealth of the country and on who pays (e.g.
duced (e.g. chemical, physical, or infectious/
the government or individual citizens). Although
biological); (ii) genetically engineered; (iii) trans-
most primary preventive interventions come at
plantable systems (e.g. xenograft, organoid); and
a net cost to health services, some can deliver a
(iv) spontaneously developing tumours. Most
gain in QALYs and a reduction in health service
cancer-preventive interventions investigated can
cost (Vos et al., 2010). Although assessments of
be categorized at the biological level as those
cost–effectiveness that account for all costs (e.g.
that: (i) prevent molecules from reaching or
that are not restricted to health service costs) are
reacting with critical target sites; (ii) reduce the
less frequently done, it is important to note that
sensitivity of target tissues to carcinogens; or
their perspective may differ markedly from one
(iii) interrupt the evolution of the neoplastic
based on health service costs only.
process. There is increasing interest in the use
of combinations of interventions as a means
28
of increasing efficacy and minimizing toxicity; when appropriate, related interventions (see Part
animal models are useful in evaluating such A, Section 7). After a thorough evaluation of the
combinations. The development of optimal strat- pertinent study features (see Part B, Section 3b),
egies for intervention in humans can be facili- studies judged to be irrelevant or inadequate
tated by the use of animal models that mimic according to the criteria determined in consul-
the neoplastic process in humans. The ques- tation with the Working Group may be excluded.
tions posed below (modified from Lewis et al., Guidelines for conducting and reporting studies
2017) may assist in determining the relevance in experimental animals have been published
of individual studies in experimental animals (e.g. OECD, 2018; Percie du Sert et al., 2018).
to the evaluation of cancer-preventive effects in
humans: (b) Study evaluation
• Are the timing, route, level, and frequency Important considerations for assessing study
of exposure comparable with those in quality include: (i) whether the intervention under
humans, after accounting for relevant species review was clearly characterized; (ii) whether
differences? the intervention exposure or dose was charac-
terized and monitored adequately; (iii) whether
• Is the cancer that is induced (i.e. by a biolog-
the control animals, exposure doses, duration of
ical, physical, or chemical agent, or genetic
dosing, timing and frequency of dosing, dura-
manipulation) relevant to the cancer in
tion of observation, and route of exposure to
humans?
the intervention were appropriate; (iv) whether
• Is the time at which the outcome is assessed appropriate experimental animal species and
relevant and justified? strains were evaluated, including appropriate sex
• Does the study explore only mechanisms or and age; (v) whether there were adequate numbers
pathways of cancer development? of animals per group; (vi) whether animals were
• Is the outcome measure cancer incidence or allocated randomly to groups; (vii) whether all
progression rather than surrogate measures experimental conditions, with the exception of
of tumour activity, such as tumour size or the tested intervention, were identical between
number of tumours? the groups; (viii) whether the histopathology
• Do the outcome measures mimic those being review was adequate; and (ix) whether the data
evaluated in humans? More specifically, does were analysed correctly and reported according
the tumour mimic the human disease in to well-accepted standards (e.g. Percie du Sert et
terms of the organs or tissues affected, and at al., 2018).
the histopathological or genetic level? Does Specific factors to be considered in inter-
the progression of the disease mimic the preting the results of cancer prevention experi-
cancer in humans? ments include: (i) the timing of the intervention
over the course of the animals’ lifespan; (ii) the
Relevant studies of cancer in experimental timing and duration of administration of the
animals are identified using principles of intervention in relation to any carcinogen admin-
systematic review as described in Part A and istration; (iii) dose–response effects; (iv) the site
further detailed in the Instructions for Authors specificity of the anticipated cancer-preventive
provided to each Working Group. Consideration outcome; (v) the spectrum and relevance of the
is given to all available long-term (i.e. lifetime or preventive outcome, from pre-neoplastic lesions
near-lifetime) studies of cancer in experimental to invasive cancers; (vi) the incidence, latency, and
animals with the intervention under review and, magnitude of the outcome, and the multiplicity
29
of the relevant neoplasms and/or other lesions; providing critical insight into the biological
and (vii) the number and structural diversity of processes that can mediate the relationship
experimental or environmental exposures, and between an intervention and a cancer outcome.
carcinogenic mechanisms underpinning the Studies of mechanisms provide evidence for
animals’ baseline risk of the cancer to which the biological plausibility, inform causality, and can
intervention was targeted. In addition, because identify biomarkers relevant to the carcinogenic
administration of an intervention may result process. The study of mechanistic biomarkers
in prevention of tumours at one site but unin- can provide insights into human heterogeneity
tended consequences at other sites, it is impor- in response to carcinogens according to age, sex,
tant that multiple organs are examined in animal genetic background, and other variables that are
experiments. important to the application of cancer-preventive
Because certain factors, including diet, food interventions in human populations. This array
or water consumption, infection, and stress, may of possible contributions by mechanistic studies
modulate cancer risk, consideration should be means that outcomes and end-points will vary
given to the potential for interaction between widely depending on the types of intervention
these factors and the intervention being studied. and the specific types of cancer examined in each
Handbook.
(c) Statistical considerations Mechanistic studies and data are identified,
The statistical methods used should be clearly screened, and evaluated for quality and human
stated and should be the generally accepted tech- relevance using principles of systematic review,
niques refined for this purpose (Peto et al., 1980; as described in Part A and further elaborated in
Gart et al., 1986; Portier & Bailer, 1989; Bieler & the Instructions for Authors provided to each
Williams, 1993). An appropriate unit of analysis Working Group, and as detailed below.
should be used (e.g. cage or individual animal
in feed studies). The statistical methods should (a) Types of studies considered
reflect the outcomes of the study (e.g. tumour This section focuses primarily on studies in
incidence or multiplicity, or overall survival of the humans, including intervention trials and longi-
animals). For outcomes other than survival, the tudinal studies with cancer-relevant biomarkers
potential influence of different overall survival that may serve as exposure or intermediate
time between exposed and unexposed animals end-points. Data from relevant experimental
should be considered. models may also be incorporated, especially
when data from studies in humans are limited
or are not practical to obtain.
4. Mechanistic evidence and other
relevant biological data (b) Evidence of cancer prevention
For a rational implementation of cancer-pre- Possible mechanisms of action of interven-
ventive measures, it is important not only to tions aiming at cancer prevention may include,
assess preventive end-points but also to under- but are not limited to: (i) altering the absorption,
stand the mechanisms by which the intervention distribution, metabolism, and excretion of a
exerts its cancer-preventive action. Mechanistic known cancer-promoting or cancer-preventive
studies derived from human research and agent; (ii) reducing endogenous DNA damage
complemented by experimental models support (e.g. by decreasing the oxidative stress and
cancer prevention research in humans by DNA–protein cross-links) or activating DNA
repair or modulating epigenetic mechanisms;
30
(iii) altering host physiology, such as the endo- systems. Based on considerations of the quality
crine environment (e.g. by modulation of exog- of the studies (e.g. design, methods and reporting
enous ligands, including hormones) or the of results, as described in Part B, Section 3b) and
microbiome; (iv) affecting cell biology to reduce a relevance to humans, the Working Group may
cell’s susceptibility to transformation, initiation, give greater weight to some included studies.
and progression of tumorigenesis (e.g. by regu- Evaluation of the results of studies in
lating cell differentiation, proliferation, migra- humans includes consideration of study quality,
tion, invasion, and cell death through apoptosis as discussed in Part B, Section 2. For obser-
and senescence); and (v) modifying the tumour vational and other studies of mechanisms of
microenvironment, including the inflammatory cancer prevention in humans, the quality of the
and immune responses. Inter-individual varia- study design, the intervention exposure assess-
tions in these responses or outcomes associated ment, and the accuracy (validity and precision)
with host factors such as age, sex, race/ethnicity, of the biomarker measurement are considered,
and genetic heterogeneity (e.g. metabolic poly- as are other important factors, including those
morphisms) are also considered. described for the evaluation of studies of cancer
In the case of potentially chemopreven- prevention in humans (Vermeulen et al., 2018).
tive agents, studies of absorption, distribution, Specific guidelines to assess the quality of molec-
metabolism, and excretion in humans and other ular biomarker and genetic studies are given in
mammalian species are summarized. The meta- STROBE-ME (Gallo et al., 2011) and STREGA
bolic fate of the intervention agent is described, (Little et al., 2009), respectively.
noting the metabolites that have been identified In addition to studies in humans, mechanistic
and their reactivity. A metabolic schema may insights may be complemented by studies in
indicate the relevant metabolic pathways and experimental systems, including animal models
products, and whether supporting evidence is (Le Magnen et al., 2016) and in vitro studies.
derived from studies in humans, in experimental Important considerations for in vitro studies
animal systems, or in in vitro models. When include the ability of the system to recapitulate
available, physiologically based pharmacokinetic the carcinogenic process that occurs in humans
models and their parameter values are included. and to model the exposure of the intervention as
would be experienced in vivo (Lewis et al., 2017;
(c) Harms of the preventive intervention Gordon et al., 2018).
Any intervention that has putative beneficial The synthesis is focused on the evidence that
effects must be assessed for potential harms. is most informative for the overall evaluation.
Toxic and other potentially harmful effects of a Evidence from several streams of mechanistic
cancer-preventive intervention that are observed data, especially those from studies in humans,
in studies in humans or studies in experimental can strengthen mechanistic conclusions.
animals and that might predict harmful effects in
humans are reviewed, and the relevant evidence 5. Summary of data reported
about them is summarized.
(a) Intervention characterization
(d) Study quality and evidence synthesis
The nature of the intervention and its char-
The Working Group summarizes the studies, acteristics, common use, and implementation
with an emphasis on characterizing consistencies in different settings, including geographical
or differences in results within and across studies patterns and time trends, are summarized as
of varying experimental designs and model
31
appropriate depending on the intervention (d) Mechanistic and other relevant data
under review. Intervention assessment methods Results pertinent to mechanisms of cancer
used in key epidemiological studies reviewed by prevention are summarized. The summary
the Working Group, their strengths, and their encompasses the informative studies on
limitations are also summarized. cancer-preventive mechanisms with adequate
(b) Cancer prevention in humans evidence for evaluation, and on any other aspects
of sufficient importance to affect the overall eval-
Results of epidemiological studies perti- uation. High-quality studies in humans, when
nent to an evaluation of the cancer-preventive available, are prioritized. In addition, supporting
effects of the interventions and their harms in findings from experimental animal models or in
humans are summarized. The overall strengths vitro systems are summarized, especially when
and limitations of the epidemiological evidence data from studies in humans are limited.
are highlighted to indicate how the evaluation
was reached. The target organ(s) or tissue(s)
in which a decrease in cancer occurrence was 6. Evaluation and rationale
observed are identified. Intervention–outcome
Evaluation of the evidence is guided by
associations and other quantitative data may be
an analytical framework that depicts the rela-
summarized when available. When the available
tionships among the population, intervention,
epidemiological studies pertain to a mixed inter-
comparator, and outcomes (including both bene-
vention (e.g. fruits and vegetables), the Working
fits and harms), and key contextual issues related
Group may seek to identify the specific agent or
to adherence to and implementation of the inter-
group of agents most likely to be responsible for
vention and its impact on population health. The
any cancer-preventive effect. The evaluation is
analytical framework may articulate both direct
focused as narrowly as is appropriate or as the
pathways (the intervention has a direct effect on
available data permit. Summaries of the evidence
cancer outcomes) and indirect pathways (the
on the balance of benefits and harms and on
intervention has an effect on an intermediate
cost–effectiveness are also provided.
outcome that has an established causal or preven-
(c) Cancer prevention in experimental animals tive association with cancer incidence).
Consensus evaluations of the strength of the
Results pertinent to an evaluation of a evidence of cancer-preventive effects of the inter-
cancer-preventive effect in animals are summa- vention in humans, in experimental animals,
rized to indicate how the evaluation was reached. and in mechanistic studies are made using trans-
For each animal species and study design, it is parent criteria and defined descriptive terms
stated whether or not changes in overall survival (see below). The Working Group then develops
or tumour incidence, latency, severity, or multi- a consensus overall evaluation of the strength of
plicity were observed, and the tumour sites the evidence that the intervention under review
are indicated. Dose–response patterns are also prevents cancer and assigns the intervention to
summarized. Possible harms of the intervention one of four categories (see below).
are noted. When the Working Group has reviewed
multiple, closely related interventions (e.g.
different forms of an intervention on the same
presumed cause of cancer), they may be grouped
together for the purpose of a unified evaluation
32
of the strength of the evidence that they prevent prevention in humans is classified into one of the
cancer. following categories:
The framework for these evaluations, Sufficient evidence of cancer prevention
described below, may not encompass all factors in humans: A causal preventive association
relevant to a particular evaluation of preventive between the intervention and cancer in humans
effect. After considering all relevant scientific has been established. That is, a cancer-preventive
findings, the Working Group may, exception- association has been observed consistently in the
ally, assign the intervention to a different cate- body of evidence (including several high-quality
gory from the one that a strict application of the studies) and chance, bias, and confounding as
framework would indicate, while providing a causes of this association were ruled out with
clear rationale for the overall evaluation reached. reasonable confidence.
When there are substantial differences of Limited evidence of cancer prevention
scientific interpretation among the Working in humans: A causal preventive association
Group members, the overall evaluation will be between the intervention and cancer in humans
based on the consensus of the Working Group. is plausible. That is, a cancer-preventive associa-
A summary of the alternative interpretations tion has been observed in the body of evidence,
may be provided, together with their scientific but chance, bias, or confounding as causes of this
rationale and an indication of the degree of association could not be ruled out with reason-
support for each. able confidence.
The evaluation categories refer to the strength Inadequate evidence of cancer prevention
of the evidence that an intervention can prevent in humans: The current body of evidence does
cancer in humans. Consideration may be given not enable a conclusion to be drawn about the
to how strongly or weakly the intervention can presence or absence of a preventive association
prevent cancer. In addition, actual and potential between the intervention and cancer in humans.
harms of the proposed intervention are addressed Common situations that lead to a determina-
qualitatively and quantitatively, as the evidence tion of inadequate evidence of cancer preven-
base permits. tion in humans include: (a) no data are available
In what follows, the term “cancer prevention” in humans; (b) there are studies available in
refers to the outcome of a Handbooks evalua- humans, but of poor quality or informativeness;
tion, that is, to a cancer outcome or an inter- and (c) there are studies available in humans
mediate outcome, as defined in the analytical of sufficient quality, but their results are incon-
framework. Thus, the wording of these evalua- sistent or otherwise do not enable a conclusion
tions is the same when an intermediate outcome, to be drawn.
not cancer itself, is the outcome studied. As noted Evidence suggesting lack of cancer preven-
above, evaluation of an intermediate outcome is tion in humans: There are several high-quality
performed only when the intermediate outcome studies covering, through direct or indirect path-
has an established causal or preventive associa- ways, the full range of levels of the intervention
tion with cancer incidence. that humans are known to encounter that are
mutually consistent in not showing a preventive
(a) Cancer prevention in humans association between the intervention and the
Cancer-preventive effects in humans are eval- studied cancers at any observed level of inter-
uated on the basis of the principles outlined in vention. The results from these studies alone or
Part B, Section 2. The evidence relevant to cancer in combination had narrow confidence intervals
with their upper bounds above or close to the
33
null value (e.g. a relative risk of 1.0). Similarly, Inadequate evidence of cancer prevention
bias and confounding as possible causes of this in experimental animals: The studies cannot be
null result were ruled out with reasonable confi- interpreted as showing the presence or absence
dence, and the studies were considered infor- of a preventive association between the interven-
mative. A conclusion of evidence suggesting lack tion and cancer in experimental animals because
of cancer prevention in humans is limited to the of major qualitative or quantitative limitations
cancer sites, populations, life stages, conditions of the data available, or no data are available on
and levels of intervention, and length of observa- cancer in experimental animals.
tion covered by the pertinent studies. The target Evidence suggesting lack of cancer preven-
organ(s) or tissue(s) where evidence suggesting tion in experimental animals: Evidence from
of lack of cancer prevention was observed in high-quality studies in several experimental
humans are identified. models shows that, within the limits of the
tests used (e.g. tumour site, age at intervention,
(b) Cancer prevention in experimental animals conditions and levels of intervention tested), the
Cancer-preventive effects in experimental intervention has no preventive association with
animals are evaluated on the basis of the princi- cancer in experimental animals.
ples outlined in Part B, Section 3. The evidence
relevant to cancer prevention in experimental (c) Mechanistic evidence
animals is classified into one of the following Mechanistic studies are evaluated on the
categories: basis of the principles outlined in Part B, Section
Sufficient evidence of cancer prevention in 4. The mechanistic evidence is classified into one
experimental animals: A preventive association of the following categories:
has been established between the intervention Strong mechanistic evidence: There are a
and increased cancer-related survival, decreased substantial number of high-quality studies in
incidence, increased latency, and/or decreased humans that consistently link the intervention to
multiplicity of malignant neoplasms or of an a mechanistic pathway by which it could prevent
appropriate combination of benign and malig- cancer.
nant neoplasms in several independent, high- Limited mechanistic evidence: The evidence
quality studies and model systems. from mechanistic data in humans is suggestive
Limited evidence of cancer prevention of a cancer-preventive effect of the intervention,
in experimental animals: The data suggest a but (a) there are a limited number of high-quality
preventive association between the intervention studies, or (b) the studies cover a narrow range of
and cancer in experimental animals. That is, experiments or relevant end-points, or (c) there
an association has been observed but the data are some inconsistencies in studies of similar
are limited for making a definitive evaluation design, or (d) there is unexplained incoherence
because: (a) the evidence of a cancer-preventive across studies of different end-points, or (e) the
association is based on only a few high-quality available data are limited to studies in experi-
studies; (b) the intervention decreases incidence, mental model systems.
increases latency, and/or decreases multiplicity Inadequate mechanistic evidence: The
only of benign neoplasms; or (c) there are unre- evidence from mechanistic data in both humans
solved questions about the adequacy of the and experimental model systems is lacking, or
design, conduct, or interpretation of the available the data are inconsistent in linking the inter-
studies. vention to any mechanistic pathway by which it
could prevent cancer.
34
(d) Overall evaluation intervention as commonly implemented in prac-

Finally, the body of evidence is considered as tice, and whether or not the benefits outweigh
a whole. Overall evaluation of the intervention is the harms.
a matter of scientific judgement that reflects the (ii) The intervention probably prevents
strength of the evidence derived from the studies cancer in humans (Group B1)
reviewed. The levels of evidence from studies in In Scenario 1, this category is used for inter-
humans, mechanistic data, and studies in experi- ventions for which there is limited evidence of
mental animals are weighed into the overall eval- cancer prevention in humans and either strong
uation, and statements are made about cancer mechanistic evidence in humans or sufficient
prevention in humans with the wording of one evidence in experimental animals with all the
of the standard categories as described below. criteria for the relevance to humans being met
One of the two overall evaluation scenarios (see Part B, Section 3a).
(see Part A, Section 3.1) will apply, depending on In Scenario 2, this category is used for inter-
the nature of the evidence that has been reviewed ventions for which there is sufficient evidence in
(Table 4; see also Part A). If, for logistic reasons, humans that the intervention has a cancer-pre-
evidence for Step 1 and Step 2 of Scenario 2 has ventive effect on the intermediate outcome
been reviewed at two separate Handbooks meet- (Step 1), limited evidence that the interme-
ings, no overall evaluation will be made for Step diate outcome has a cancer-preventive effect in
2 alone. humans (Step 2), and either sufficient evidence in
None of these evaluations quantify the frac- experimental animals with all the criteria for the
tion of the burden of a particular cancer that a relevance to humans being met or strong mecha-
specific intervention would prevent; thus, some nistic evidence in humans (see Part B, Section 3a).
interventions may prevent a small fraction of the Alternatively, this category is used when there is
cancer, some may prevent a larger fraction, and limited evidence in humans that the intervention
these fractions may vary across populations, for has a cancer-preventive effect in the intermediate
example as a function of the prevalence of the outcome (Step 1) and sufficient evidence that the
relevant risk factors. intermediate outcome has a cancer-preventive
effect in humans (Step 2).
Overall evaluation categories (iii) The intervention possibly prevents
cancer in humans (Group B2)
(i) The intervention is established to prevent
In Scenario 1, this category is used for inter-
cancer in humans (Group A)
ventions for which there is limited evidence of
This category is used for interventions for
cancer prevention in humans, less than strong
which there is sufficient evidence of cancer preven-
evidence from mechanistic data, and less than
tion in humans, either directly (Scenario 1) or in
sufficient evidence of cancer prevention in exper-
two steps (Scenario 2): from the intervention to
imental animals.
the intermediate outcome (Step 1) and from the
In Scenario 2, this category is used when
intermediate outcome to cancer (Step 2).
(i) there is sufficient evidence in humans that the
The organ sites on which the evidence in
intervention has a cancer-preventive effect on
humans is based are stated here. A statement is
the intermediate outcome (Step 1), and limited
also made of what the Working Group considers
evidence in humans and less than sufficient
to be the magnitudes of the benefits and the
evidence in experimental animals or less than
harms of the intervention, in as nearly compa-
strong evidence from mechanistic data that the
rable terms as possible, for people adhering to the
intermediate outcome has a cancer-preventive
35
Table 4 Summary of the strength of the evidence in each evidence stream contributing to the
overall evaluation
Scenario 1: Direct evidence that the intervention prevents cancer
Strength of the evidence that Strength of the evidence Strength of the evidence that the Overall evaluation
the intervention prevents from mechanistic studies intervention prevents cancer in
cancer in humans that the intervention experimental animals
prevents cancer
Sufficient – – Group A
Limited Strong – Group B1
Limited – Sufficient Group B1
Limited Less than strong Less than sufficient Group B2
Inadequate – – Group C
Evidence suggesting lack of – Evidence suggesting lack of cancer Group D
cancer prevention prevention
Scenario 2: Evidence that the intervention prevents cancer by way of an intermediate outcome
(risk factor or preventive factor)
Step 1 Step 2 a Overall evaluationa
Strength of the evidence that Strength of the evidence Strength of the evidence that
the intervention decreases that decreasing exposure to decreasing exposure to the risk
exposure to the risk factor the risk factor or increasing factor or increasing exposure to
or increases exposure to the exposure to the preventive the preventive factor prevents
preventive factor in humans factor prevents cancer in cancer in experimental animals or
humans mechanistic studiesb
Sufficient Sufficientc – Group A
Sufficient Limited Sufficient Group B1
Sufficient Limited Less than sufficient Group B2
Limited Sufficient – Group B1
Limited Limited – Group B2
Inadequate – – Group C
– Evidence suggesting lack of Evidence suggesting lack of cancer Group D
cancer prevention prevention
Evidence suggesting lack of – – Group D
cancer prevention
a This overall evaluation applies only when evidence from both Step 1 and Step 2 is available. When a Handbook evaluates only Step 2, no overall
evaluation is made.
b Evidence in experimental animals and mechanistic data is considered to be sufficient when there is strong evidence from mechanistic data
(mechanistic studies in humans) or sufficient evidence in experimental animals.

c The evidence in this category may be considered to be sufficient when it is based on observational studies of change in cancer incidence
associated with self-reported or observed (by way of time-separated repeated measures) change in the level of a risk factor or preventive factor
(e.g. smoking cessation; increase in consumption of fruits and vegetables), OR, exceptionally, studies of variation in cancer incidence with the
level of a risk factor or preventive factor measured at one time point.
36
effect; OR (ii) there is limited evidence in humans evidence for each stream, an indication of the
that the intervention has a cancer-preventive body of evidence that was pivotal to these conclu-
effect on the intermediate outcome (Step 1), and sions, and an explanation of the reasoning of the
limited evidence in humans that the intermediate Working Group in making evaluations.
outcome has a cancer-preventive effect, and any In the rationale, the Working Group may
evidence category in experimental animals and draw attention to the fact that actions on the
mechanistic data. evaluations should be taken in the light of
When the evidence is classified in Group country- or setting-specific circumstances that
B1 or Group B2, the evaluation is followed by a influence the public health priority, feasibility,
description of harms, actual and potential. and acceptability of programmes based on the
(iv) The intervention is not classifiable as interventions evaluated.
to its capacity to prevent cancer in humans
(Group C)
In both Scenario 1 and Scenario 2, this cate- References
gory is used for interventions for which there is
inadequate evidence in humans, irrespective of Alexandrov LB, Ju YS, Haase K, Van Loo P, Martincorena
the level of evidence from mechanistic data and I, Nik-Zainal S, et al. (2016). Mutational signatures
associated with tobacco smoking in human cancer.
studies in experimental animals. Interventions Science. 354(6312):618–22. doi:10.1126/science.aag0299
that do not fall into any other category are also PMID:27811275
placed in this category. AMSTAR (2017). AMSTAR checklist. Assessing the
(v) The intervention probably does not methodological quality of systematic reviews.
Available from: https://amstar.ca/Amstar_Checklist.
prevent cancer in humans (Group D) php, accessed 6 March 2019.
In Scenario 1, this category is used for inter- Axelson O, Steenland K (1988). Indirect methods of
ventions for which there is evidence suggesting assessing the effects of tobacco use in occupational
studies. Am J Ind Med. 13(1):105–18. doi:10.1002/
lack of cancer prevention both in humans and in ajim.4700130107 PMID:3344750
experimental animals. In Scenario 2, this cate- Bernal JL, Cummins S, Gasparrini A (2017). Interrupted
gory is used when there is evidence suggesting time series regression for the evaluation of public
lack of cancer prevention both in humans and health interventions: a tutorial. Int J Epidemiol.
46(1):348–55. doi:10.1093/ije/dyw098 PMID:27283160
in experimental animals for the intermediate Bieler GS, Williams RL (1993). Ratio estimates, the
outcome to cancer, irrespective of the level of delta method, and quantal response tests for
evidence for the intervention to the intermediate increased carcinogenicity. Biometrics. 49(3):793–801.
outcome; or there is evidence suggesting lack of doi:10.2307/2532200 PMID:8241374
Boyle P, Levin B, editors (2008). World cancer report
cancer prevention for the intervention to the 2008. Lyon, France: International Agency for Research
intermediate outcome, irrespective of the level of on Cancer. Available from: https://publications.iarc.
evidence for the intermediate outcome to cancer. fr/395.
Cochrane AL (1972). Effectiveness and efficiency:
(e) Rationale random reflections on health services. Nuffield
Trust. Available from: https://www.nuffieldtrust.org.
The reasoning that the Working Group used uk/research/effectiveness-and-efficiency-random-
to reach its evaluation is summarized so that the reflections-on-health-services.
Cogliano V, Baan R, Straif K, Grosse Y, Secretan B,
basis for the evaluation offered is transparent. It El Ghissassi F, et al. (2005). Transparency in IARC
includes concise statements of the principal line Monographs. Lancet Oncol. 6(10):747. doi:10.1016/
or lines of argument that emerged in the delib- S1470-2045(05)70380-6
Cogliano VJ, Baan RA, Straif K, Grosse Y, Secretan MB,
erations of the Working Group, the conclusions El Ghissassi F, et al. (2004). The science and practice
of the Working Group on the strength of the
37
of carcinogen identification and evaluation. Environ IARC (2016). Breast cancer screening. IARC Handb
Health Perspect. 112(13):1269–74. doi:10.1289/ehp.6950 Cancer Prev. 15:1–469. Available from: https://
PMID:15345338 publications.iarc.fr/545.
Gallo V, Egger M, McCormack V, Farmer PB, Ioannidis IARC (2018). Guidelines for Observers at IARC Hand-
JP, Kirsch-Volders M, et al. (2011). STrengthening the books Meetings. Available from: https://handbooks.
Reporting of OBservational studies in Epidemiology - iarc.fr/preambles/guidelines-for-observers/.
Molecular Epidemiology (STROBE-ME): an extension IARC (2019). International Classification of Diseases for
of the STROBE statement. Prev Med. 53(6):377–87. Oncology, ICD-O-3 online. Lyon, France: International
doi:10.1016/j.ypmed.2011.08.007 PMID:22029945 Agency for Research on Cancer. Available from: http://
Gart JJ, Krewski D, Lee PN, Tarone RE, Wahrendorf codes.iarc.fr/usingicdo.php, accessed 8 January 2019.
J (1986). Statistical Methods in Cancer Research, Jonas DE, Ferrari RM, Wines RC, Vuong KT, Cotter A,
Vol. 3. The design and analysis of long-term animal Harris RP (2018). Evaluating evidence on intermediate
experiments. Lyon, France: International Agency for outcomes: considerations for groups making health-
Research on Cancer (IARC Scientific Publications, No. care recommendations. Am J Prev Med. 54(1S1):S38–
79) 52. doi:10.1016/j.amepre.2017.08.033 PMID:29254524
Gordon JL, Brown MA, Reynolds MM (2018). Cell-based Le Magnen C, Dutta A, Abate-Shen C (2016). Optimizing
methods for determination of efficacy for candidate mouse models for precision cancer prevention. Nat
therapeutics in the clinical management of cancer. Rev Cancer. 16(3):187–96. doi:10.1038/nrc.2016.1
Diseases. 6(4):E85. doi:10.3390/diseases6040085 PMID:26893066
PMID:30249005 Lewis SJ, Gardner M, Higgins J, Holly JMP, Gaunt
Greenland S, O’Rourke K (2008). Meta-analysis. In: TR, Perks CM, et al. (2017). Developing the WCRF
Rothman KJ, Greenland S, editors. Modern epide- International/University of Bristol methodology
miology. Philadelphia (PA), USA: Wolters Kluwer for identifying and carrying out systematic reviews
Health/Lippincott Williams & Wilkins; pp. 652–82. of mechanisms of exposure-cancer associations.
Hakama M (1983). Trends in the incidence of cervical Cancer Epidemiol Biomarkers Prev. 26(11):1667–75.
cancer in the Nordic countries. In: Magnus K, editor. doi:10.1158/1055-9965.EPI-17-0232 PMID:28978562
Trends in cancer incidence: causes and practical impli- Little J, Higgins JP, Ioannidis JP, Moher D, Gagnon F, von
cations. Taylor & Francis Inc.; pp. 279–92. Elm E, et al. (2009). STrengthening the REporting of
Hernán MA, Hernández-Díaz S, Robins JM (2004). A Genetic Association studies (STREGA) – an exten-
structural approach to selection bias. Epidemiology. sion of the STROBE statement. Eur J Clin Invest.
15(5):615–25. doi:10.1097/01.ede.0000135174.63482.43 39(4):247–66. doi:10.1111/j.1365-2362.2009.02125.x
PMID:15308962 PMID:19297801
Hill AB (1965). The environment and disease: associ- OECD (2018). Test no. 451: carcinogenicity studies. OECD
ation or causation? Proc R Soc Med. 58(5):295–300. Guidelines for the Testing of Chemicals, Section 4.
doi:10.1177/003591576505800503 PMID:14283879 Paris, France: OECD Publishing.
IARC (1997). Non-steroidal anti-inflammatory drugs. Percie du Sert N, Hurst V, Ahluwalia A, Alam S,
IARC Handb Cancer Prev. 1:1–202. Available from: Altman DG, Avey MT, et al. (2018). Revision of the
https://publications.iarc.fr/371. ARRIVE guidelines: rationale and scope. BMJ Open
IARC (2002a). Weight control and physical activity. IARC Sci. 2(1):e000002. doi:10.1136/bmjos-2018-000002
Handb Cancer Prev. 6:1–315. Available from: https:// PMID:33954268
publications.iarc.fr/376. Peto R, Pike MC, Day NE, Gray RG, Lee PN, Parish S, et
IARC (2002b). Breast cancer screening. IARC Handb al. (1980). Guidelines for simple, sensitive significance
Cancer Prev. 7:1–229. Available from: https:// tests for carcinogenic effects in long-term animal
publications.iarc.fr/377. experiments. IARC Monogr Eval Carcinog Risk Chem
IARC (2007). Tobacco control: reversal of risk after quit- Hum Suppl. (2 Suppl):311–426. PMID:6935185
ting smoking. IARC Handb Cancer Prev. 11:1–358. Porta M, editor (2008). A dictionary of epidemiology. 5th
Available from: https://publications.iarc.fr/381. ed. New York (NY), USA: Oxford University Press.
IARC (2008). Methods for evaluating tobacco control doi: 10.1093/acref/9780195314496.001.0001
policies. IARC Handb Cancer Prev. 12:1–300. Available Porta M, editor (2014). A dictionary of epidemiology. 6th
from: https://publications.iarc.fr/382. ed. New York (NY), USA: Oxford University Press.
IARC (2009). Evaluating the effectiveness of smoke-free doi:10.1093/acref/9780199976720.001.0001
policies. IARC Handb Cancer Prev. 13:1–334. Available Portier CJ, Bailer AJ (1989). Testing for increased carcino-
from: https://publications.iarc.fr/383. genicity using a survival-adjusted quantal response test.
IARC (2011). Effectiveness of tax and price policies for Fundam Appl Toxicol. 12(4):731–7. doi:10.1016/0272-
tobacco control. IARC Handb Cancer Prev. 14:1–366. 0590(89)90004-3 PMID:2744275
Available from: https://publications.iarc.fr/384.
38
Richardson DB, Laurier D, Schubauer-Berigan MK, Vandenbroucke JP, Broadbent A, Pearce N (2016).
Tchetgen Tchetgen E, Cole SR (2014). Assessment Causality and causal inference in epidemiology:
and indirect adjustment for confounding by smoking the need for a pluralistic approach. Int J Epidemiol.
in cohort studies using relative hazards models. Am 45(6):1776–86. doi:10.1093/ije/dyv341 PMID:26800751
J Epidemiol. 180(9):933–40. doi:10.1093/aje/kwu211 Vermeulen R, Bell DA, Jones DP, Garcia-Closas M, Spira
PMID:25245043 A, Wang TW (2018). Application of biomarkers in
Rothman KJ, Greenland S, Lash TL (2008). Causation cancer epidemiology. In: Thun MJ, Linet MS, Cerhan
and causal inference. In: Rothman KJ, Greenland S, JR, Haiman CA, Schottenfeld D, editors. Cancer epide-
editors. Modern epidemiology. Philadelphia (PA), miology and prevention. 4th ed. New York (NY), USA:
USA: Wolters Kluwer Health/Lippincott Williams & Oxford University Press; pp. 77–96.
Wilkins; pp. 5–31. Vos T, Carter R, Barendregt J, Mihalopoulos C, Veerman
Siemieniuk R, Guyatt G (2019). What is GRADE? BMJ JL, Magnus A, et al.; ACE–Prevention Team (2010).
Best Practice. Available from: https://bestpractice.bmj. Assessing Cost-Effectiveness in Prevention (ACE–
com/info/toolkit/learn-ebm/what-is-grade/, accessed 6 Prevention): final report. Brisbane, Australia:
March 2019. University of Queensland and Melbourne, Australia:
Stayner L, Steenland K, Dosemeci M, Hertz-Picciotto I Deakin University. Available from: https://public-
(2003). Attenuation of exposure-response curves in health.uq.edu.au/files/571/ACE-Prevention_final_
occupational cohort studies at high exposure levels. report.pdf.
Scand J Work Environ Health. 29(4):317–24. doi:10.5271/ Walls HL, Peeters A, Proietto J, McNeil JJ (2011). Public
sjweh.737 PMID:12934726 health campaigns and obesity - a critique. BMC
Stewart BW, Kleihues P, editors (2003). World cancer Public Health. 11(1):136. doi:10.1186/1471-2458-11-136
report 2003. Lyon, France: International Agency PMID:21352562
for Research on Cancer. Available from: https:// Wild CP, Cogliano VJ (2011). A plea on behalf of expert
publications.iarc.fr/397. evaluation and the experts involved. Int J Epidemiol.
Stewart BW, Wild CP, editors (2014). World cancer report 40(1):253–61. doi:10.1093/ije/dyq038 PMID:20223796
2014. Lyon, France: International Agency for Research Yarmolinsky J, Wade KH, Richmond RC, Langdon RJ,
on Cancer. Available from: https://publications.iarc. Bull CJ, Tilling KM, et al. (2018). Causal inference in
fr/396. cancer epidemiology: what is the role of Mendelian
Thun MJ, Linet MS, Cerhan JR, Haiman CA, Schottenfeld randomization? Cancer Epidemiol Biomarkers Prev.
D, editors (2018). Cancer epidemiology and prevention. 27(9):995–1010. doi:10.1158/1055-9965.EPI-17-1177
4th ed. New York (NY), USA: Oxford University Press. PMID:29941659
39
40
PREAMBLE − SECONDARY PREVENTION
The Preamble to the IARC Handbooks of Cancer Prevention describes the objectives and
scope of the programme, general principles and procedures, and scientific review and
evaluations. The IARC Handbooks embody the principles of scientific rigour, impartial eval-
uation, transparency, and consistency. The Preamble should be consulted when reading
an IARC Handbook or a summary of an IARC Handbook’s evaluations. Separate Instructions
for Authors describe the operational procedures for the preparation and publication of a
volume of the IARC Handbooks.
A. GENERAL PRINCIPLES AND developing cancer. Secondary prevention entails

PROCEDURES methods that can find and ameliorate precan-
cerous conditions or find cancers in the early
stages, when they can be treated more success-
1. Background fully. Tertiary prevention is the application of
Prevention of cancer is the mission of the measures aimed at reducing the impact of long-
International Agency for Research on Cancer term disease and disability caused by cancer or
(IARC). Cancer prevention is needed even more its treatment.
today than when IARC was established, in 1965, The IARC Handbooks of Cancer Prevention
because the global burden of cancer is high and provide critical reviews and evaluations of the
continues to increase, as a result of population scientific evidence on the preventive effects
growth and ageing and increases in cancer- of primary or secondary cancer preven-
causing exposures and behaviours, especially in tion measures. The evaluations of the IARC
low- and middle-income countries (Stewart & Handbooks are used by national and interna-
Kleihues, 2003; Boyle & Levin, 2008; Stewart & tional health agencies to develop evidence-based
Wild, 2014). interventions or recommendations for reducing
Broadly defined, prevention is “actions aimed cancer risk.
at eradicating, eliminating, or minimizing the The IARC Handbooks of Cancer Prevention
impact of disease and disability, or if none of series was launched in 1995 by Dr Paul Kleihues,
these is feasible, retarding the progress of disease then Director of IARC, in recognition of the
and disability” (Porta, 2014). Cancer prevention need for a series of publications that would criti-
encompasses primary, secondary, and tertiary cally review and evaluate the evidence on a wide
prevention. Primary prevention consists of range of cancer-preventive interventions. The
actions that can be taken to lower the risk of first volume of the IARC Handbooks (IARC,
41
1997) reviewed the evidence on cancer-preven- 2. Objectives, scope, and

tive effects of non-steroidal anti-inflammatory definitions
drugs, specifically aspirin, sulindac, piroxicam,
and indomethacin. Handbooks Volume 6 (IARC, 2.1 Objectives and scope
2002a) was the first that evaluated behavioural
interventions (weight control and physical The scope of the IARC Handbooks of Cancer
activity), and Handbooks Volume 7 (IARC, Prevention series is to contribute to reducing the
2002b) was the first that evaluated cancer incidence of or mortality from cancer worldwide.
screening (breast cancer screening). Handbooks To this end, the IARC Handbooks programme
Volumes 11–14 (IARC, 2007, 2008, 2009, 2011) prepares and publishes, in the form of volumes
focused on tobacco control. After a 3-year hiatus, of Handbooks, critical scientific reviews and
the IARC Handbooks series was relaunched evaluations of the available evidence on the effi-
in 2014 with the preparation of Handbooks cacy, effectiveness, and harms of a wide range
Volume 15 (IARC, 2016a), which re-evaluated of cancer-preventive interventions. The primary
breast cancer screening. target audiences for the Handbooks are national
IARC’s process for developing Handbooks and international agencies with responsibility
engages international, expert scientific Working for, or advocating for, public health. The IARC
Groups in a transparent synthesis of different Handbooks are an important part of the body
streams of evidence, which is then translated of information on which public health decisions
into an overall evaluation according to criteria for cancer prevention may be based. However,
that IARC has developed and refined (see Part A, public health options to prevent cancer vary
Section 6). Scientific advances are periodically from one setting to another and from country
incorporated into the evaluation methodology, to country, and relate to many factors, including
which must enable the evaluation of new genera- socioeconomic conditions and national prior-
tions of existing methods as well as new screening ities. Therefore, no recommendations are given
methodologies. in the Handbooks with regard to regulations
This Preamble, first prepared as the or legislation, which are the responsibility of
Handbooks Working Procedures in 1995 and individual governments or other international
later adapted to the topics of cancer screening authorities. However, the IARC Handbooks may
and tobacco control, is primarily a statement of aid national and international authorities in
the general principles and procedures used in devising programmes of health promotion and
developing a Handbook, to promote transpar- cancer prevention, estimating the balance of
ency and consistency across Handbooks evalu- benefits and harms, and considering cost–effec-
ations. In addition, IARC provides Instructions tiveness evaluations.
for Authors to specify more detailed operating The IARC Handbooks programme also
procedures. does not make formal research recommenda-
tions. However, because Handbooks synthesize
and integrate streams of evidence on cancer
prevention, critical gaps in knowledge that merit
research may be identified.
42
Preamble – Secondary Prevention
2.2 Definition of interventions for (ii) subgroups with particular predisposing

secondary prevention host characteristics, such as genetic suscepti-
bility, precursor lesions, or particular diseases
The current IARC Handbook addresses a other than cancer, or with high exposure to
specific intervention or class of interventions environmental, occupational, or behavioural
for secondary prevention. The principal instru- risk factors; and
ments of secondary prevention of cancer are
(iii) people with a history of cancer who are at
interventions for early detection of precancerous
high risk of a further primary cancer.
lesions (i.e. precancer) or invasive cancer, which
are currently mostly cancer screening inter- Early diagnosis interventions aim at
ventions. However, there is growing evidence detecting cancer in symptomatic patients as
that action campaigns to increase awareness of early as possible. Delays in accessing cancer
cancer among the general public can increase care are common with late-stage presentation,
the number of people who present to health-care particularly in lower-resource settings and in
providers, leading to earlier diagnosis of cancer vulnerable populations. The consequences of
and, generally, to better cancer outcomes. Such delayed or inaccessible cancer care are lower
interventions for early diagnosis are also within likelihood of survival, greater morbidity of
the scope of the Handbooks programme. treatment, and higher costs of care, resulting
Screening is the systematic application of a in avoidable deaths and disability from cancer.
test that “can be applied rapidly in a presum- Early diagnosis improves cancer outcomes by
ably asymptomatic population, aiming at the providing care at the earliest possible stage and
presumptive identification of unrecognized is therefore an important public health strategy
disease or defect” (Porta, 2014). Screening tests in all settings (https://www.who.int/cancer/
sort out apparently-well people who probably prevention/diagnosis-screening/en/). One of the
have a disease from those who probably do not. most commonly used strategies is to raise aware-
A screening test is not intended to be diagnostic, ness among the public and/or health profes-
because people with positive or suspicious find- sionals of early signs and symptoms of cancer
ings must be referred to their physicians for in order to facilitate diagnosis before the disease
diagnosis and necessary treatment (Porta, 2014). becomes advanced. Other possible interventions
Screening may enable diagnosis of cancer suffi- to promote early diagnosis may involve regula-
ciently early that cure and resulting prevention of tion of health care and organization of health
cancer death or a reduction in risk of cancer are services (WHO, 2017).
realistic possibilities. Screening for some cancers,
such as cervical cancer or colorectal cancer, may 2.3 Definitions of efficacy, effectiveness,
also detect precancer, effective treatment of and harms
which can prevent occurrence of invasive cancer.
Screening can also cause harm, and evidence for Efficacy and effectiveness are two funda-
harm must also be considered when evaluating mental concepts underlying the evaluation
the capacity of screening to reduce the incidence of preventive interventions (Cochrane, 1972).
of cancer or death from cancer. Efficacy was defined by Porta (2008) as “the
Screening interventions can be applied across extent to which a specific intervention, proce-
a continuum of: dure, regimen or service produces a beneficial
result under ideal conditions … Ideally, the
(i) the general population (often circum- determination of efficacy is based on the results
scribed by age and sex);
43
of a randomized controlled trial”. Effectiveness potential harms is an important component of

was defined by Porta (2008) as “a measure of the the summary of the evidence.
extent to which a specific intervention, proce- For screening and for early diagnosis, other
dure, regimen or service, when deployed in the issues to be considered include acceptability to
field in routine circumstances, does what it is the target population, impact on health equity,
intended to do for a specific population”. cost, cost–effectiveness, availability of the
The distinction between efficacy and effec- personnel and facilities required to deliver the
tiveness of an intervention at the population level screening intervention, and access to the health
is an important one to make when evaluating services needed to diagnose and treat the disease
preventive interventions. Efficacy is a necessary, detected. Depending on the specific interven-
but not sufficient, basis for formulating recom- tion, some of these issues may be of sufficiently
mendations for an intervention. Whereas efficacy high interest to programme managers that they,
of an intervention can be inferred if effectiveness too, are reviewed in the IARC Handbook.
is established, efficacy does not guarantee effec- Although the distinction between evidence
tiveness because of the number of implemen- of efficacy and effectiveness is an important one
tation steps, each with uncertainty, required to to make when seeking to act on cancer preven-
deliver an efficacious prevention intervention as tion, the Handbooks evaluations are based on
an effective programme in a target population. evidence from all relevant research into efficacy
Ideally, efficacy is established before a preven- and effectiveness.
tive intervention is implemented in a whole
community or population, so as to determine
whether a case for population-wide implementa-
3. Identification and selection of
tion can be made on the basis of the balance of the interventions and outcomes for
benefits and harms and the financial costs of the review
intervention. However, it has not been unusual
for preventive interventions to be implemented 3.1 Development of an analytical
in the absence of evidence of efficacy. Should that framework
occur, evaluation of effectiveness may be the only As one of the first steps in the review and
way to determine whether the case for the inter- evaluation of a selected cancer screening inter-
vention is strong enough to justify its continua- vention, the IARC Secretariat, with the support
tion or implementation elsewhere. of the Working Group, drafts an analytical
In addition to being shown to be efficacious framework. Such a framework depicts the rela-
or effective, screening interventions must satisfy tionships among the study population, interven-
other requirements if they are to be considered tion, comparator, and intermediate outcomes or
for implementation in practice, including an changes in health status as relevant. The analyt-
acceptable balance of benefits and harms. In the ical framework includes both benefits and harms,
present context, harm is defined as any impair- and key contextual issues related to participation
ment or increase in risk of impairment as a result and implementation of the intervention and its
of exposure to or participation in a preventive impact on population health. The framework
intervention. Harms include physical, psycho- defines the intervention in its broadest context
logical, social, and economic consequences of a and specifies the aspects for which the Handbook
preventive intervention. Adverse events in health will review and evaluate the evidence.
care are a subset of harms. Evaluation of these
44
In this framework, it is most commonly • The intervention is of putative preventive

the case that a single cancer type, usually only value, but its effects or balance of benefits and
topographically defined, is the primary target, harms have not been established formally;
and the reduction of the incidence of and/or • The available evidence suggests that the
mortality from that cancer type is the primary intervention has the potential to significantly
outcome. However, it is sometimes the case that reduce the incidence of or mortality from
intermediate outcomes (i.e. outcomes that are cancer, or to have a significant impact on
not invasive cancer or death from cancer) are an intermediate outcome (e.g. precancerous
important targets. For example, detection and lesions; see below) known or highly suspected
ablation of precancerous polyps is the mecha- to be linked to cancer (see Part A, Section 6,
nism whereby some screening methods for colon Step 2).
cancer and rectal cancer reduce the incidence of
colorectal cancer. Moreover, it is plausible that a In addition, an intervention previously eval-
new test with high sensitivity and specificity for a uated in a Handbook may be re-evaluated if
precancerous lesion, such as high-grade cervical important new data become available about its
intraepithelial neoplasia, could be judged on the effects, or if its technology or implementation
grounds of these characteristics to be efficacious has changed enough for there to be substantial
in preventing invasive cervical cancer and death changes in its effects. Occasionally, a re-evalu-
from cervical cancer, provided that there is also ation may be limited to specific aspects of the
strong evidence that ablation of the precancerous screening intervention to which the new evidence
lesion prevents invasive cervical cancer. These predominantly relates (e.g. tomosynthesis for
possibilities are taken into consideration when breast cancer screening). For re-evaluations, the
defining the framework of a Handbook. full body of evidence relevant to the interven-
tion of interest is considered, either by de novo
review of all evidence or by accepting as accurate
3.2 Selection of the interventions
the evidence review of the previously published
For each new volume of the Handbooks, Handbook and undertaking a de novo review of
IARC selects one or more interventions for evidence published since the previous review.
review by considering the availability of perti- Both approaches lead to an evaluation based
nent research studies, the need to evaluate an on all relevant evidence (see Part A, Section 6,
important development in cancer prevention, Steps 4 and 5). The choice of the approach is
or the need to re-evaluate a previously evaluated subject to the judgement of the Working Group.
intervention. IARC will also consider current
public health priorities in specific geographical
regions, for example the concerns of countries or
4. The Working Group and other
regions with a high risk of specific cancer types meeting participants
(see Part A, Section 6, Step 1). Five categories of participants can be present
Interventions not previously evaluated in the at IARC Handbooks meetings (Table 1):
IARC Handbooks series are selected for evalua-
tion, where the body of evidence is large enough (i) Working Group members have ultimate
to warrant evaluation, on the basis of one or both responsibility for determining the final list
of the following criteria: of studies that contribute evidence to the
evaluation, performing the scientific review
of the evidence, and making the final, formal
45
Table 1 Roles of participants at IARC Handbooks meetings
Category of participant Role

Prepare text, Participate in Participate in Eligible to serve as
tables, and discussions evaluations Meeting Chair or
analyses Subgroup Chair
Working Group members ✓ ✓ ✓ ✓
Invited Specialists ✓a ✓
Representatives of health agencies ✓b
Observers ✓b
IARC Secretariat ✓c ✓ ✓d
a Only for sections not directly relevant to the evaluation
b Only at times designated by the Meeting Chair and/or Subgroup Chair
c Only when needed or requested by the Meeting Chair and/or Subgroup Chair
d Only for supporting Working Group members and for clarifying or interpreting the Preamble
evaluation of the strength of evidence for the matter or relevant methodologies is supported
capacity of the screening interventions to by decades of experience documenting that
reduce cancer incidence or cancer mortality. there is value in specialized expertise and
The Working Group is multidisciplinary and that the overwhelming majority of Working
is organized into Subgroups of experts in the Group members are committed to the objec-
fields that the Handbook covers. tive evaluation of scientific evidence and not
IARC selects the Working Group members to the narrow advancement of their own
on the basis of relevant expertise and an research results or a predetermined outcome
assessment of declared interests (see Part A, (Wild & Cogliano, 2011). Working Group
Section 5). For screening, the fields of exper- members are expected to serve the public
tise are: (i) the cancer targeted and its global health mission of IARC and to refrain from
epidemiology; (ii) worldwide use of preven- using inside information from the meeting or
tive interventions for the cancer targeted; meeting drafts for financial gain until the full
and (iii) specific knowledge and experience volume of the Handbooks is published (see
of screening, in general or as practised for the also Part A, Section 7).
targeted cancer. Consideration is also given IARC selects, from among the Working
to diversity in scientific approaches, in stated Group members, individuals to serve as
positions on the strength of the evidence Meeting Chair and Subgroup Chairs.
supporting the intervention, and in demo- Subgroup Chairs have preferably served in
graphic characteristics. Working Group previous Handbooks meetings as Working
members generally have published research Group members or in similar review processes.
related to the interventions being reviewed or At the opening of the meeting, the Working
to the cancer types or intermediate outcomes Group is asked to endorse the Meeting Chair
that the interventions being reviewed are selected by IARC or to propose an alterna-
thought to prevent; IARC uses literature tive. The Meeting Chair and Subgroup Chairs
searches to identify most experts. IARC also take a leading role at all stages of the review
encourages public nominations through its process (see Part A, Section 7) to promote
Call for Experts. IARC’s reliance on Working open scientific discussions that involve all
Group members with expertise on the subject Working Group members in accordance
46
with committee procedures and to ensure Group or Subgroup sessions at the discretion
adherence to the processes described in this of the Chair.
Preamble. (v) The IARC Secretariat consists of scien-
(ii) Invited Specialists are experts with critical tists who are designated by IARC or WHO
knowledge and experience on the interven- and who have relevant expertise. The IARC
tions being reviewed, the cancer types that Secretariat coordinates and facilitates all
the interventions being reviewed are thought aspects of the review and evaluation process
to prevent, or relevant methodologies, but and ensures adherence to the processes
who have a declared conflict of interests that described in this Preamble throughout the
warrants exclusion from developing or influ- development of the scientific reviews and
encing the evaluations. The Invited Specialists evaluations (see Part A, Sections 5 and 6).
do not draft any section of the Handbook that The IARC Secretariat announces and orga
pertains to the description or interpretation nizes the meeting, identifies and invites the
of the data on which the evaluation is based, Working Group members, and assesses the
or participate in the evaluations. Invited declared interests of all meeting participants
Specialists are invited in limited numbers, in accordance with WHO requirements (see
when necessary, to assist the Working Group Part A, Section 5). The IARC Secretariat
by contributing their unique knowledge and supports the activities of the Working Group
experience to the discussions. (see Part A, Section 7) by performing system-
(iii) Representatives of national and interna- atic literature searches, performing title
tional health agencies may attend because and abstract screening, organizing confer-
their agencies are interested in the subject ence calls to coordinate the development of
of the Handbook. The Representatives of drafts and to discuss cross-cutting issues,
national and international health agencies and reviewing drafts before and during the
do not draft any section of the Handbook or meeting. Members of the IARC Secretariat
participate in the evaluations. Representatives serve as meeting rapporteurs, assist the
can participate in discussions at times desig- Meeting Chair and Subgroup Chairs in facil-
nated by the Meeting Chair or a Subgroup itating all discussions, and may draft text or
Chair. Relevant World Health Organization tables or assist a Subgroup in the conduct of
(WHO) staff members attend as members of additional analyses when designated by the
the IARC Secretariat (see below). Meeting Chair or a Subgroup Chair. After
(iv) Observers with relevant scientific creden- the meeting, the IARC Secretariat reviews
tials are admitted in limited numbers. the drafts for factual accuracy of research
Attention is given to the balance of Observers results cited. The participation of the IARC
from entities with differing perspectives on Secretariat in the evaluations is restricted to
the interventions under review. Observers clarifying or interpreting the Preamble.
are invited only to observe the meeting, do All meeting participants are listed, with their
not draft any section of the Handbook or principal affiliations, in the front matter of the
participate in the evaluations, must agree to published volume of the Handbooks. Pertinent
respect the Guidelines for Observers at IARC interests, if any, are listed in a footnote to the
Handbooks meetings (IARC, 2018), and must participant’s name. Working Group members and
not attempt to influence the outcomes of the Invited Specialists serve as individual scientists
meeting. Observers may speak at Working
47
and not as representatives of any organization, participants are asked to report all such contacts
government, or industry (Cogliano et al., 2004). to IARC (Cogliano et al., 2005).
The roles of the participants are summarized The Working Group meets at IARC to discuss
in Table 1. and finalize the scientific review and to develop
summaries and evaluations. At the opening of
the meeting, all meeting participants update
5. Development of a volume of the their Declarations of Interests forms, which are
IARC Handbooks then reviewed for conflicts of interest by IARC.
Each volume of the Handbooks is developed Declared interests related to the subject of the
by an ad hoc, specifically convened Working meeting are disclosed to the meeting partici-
Group of international experts. Approximately pants during the meeting and in the published
1 year before the meeting of a Working Group, volume of the Handbooks (Cogliano et al., 2004).
a preliminary list of interventions to be reviewed The objectives of the meeting are twofold:
(see Part A, Section 3), together with a Call for peer review of the drafts and consensus on the
Data and a Call for Experts, is announced on evaluations. During the first part of the meeting,
the Handbooks programme website (https:// Working Group members work in Subgroups to
handbooks.iarc.fr/). review the pre-meeting drafts, develop a joint
The IARC Secretariat selects potential Subgroup draft, and draft Subgroup summaries.
Working Group members based on the criteria During the last part of the meeting, the Working
described in Part A, Section 4. Before a meeting Group meets in plenary sessions to review the
invitation is extended, each potential partici- Subgroup drafts and summaries and to develop
pant, including the IARC Secretariat, completes the consensus evaluations. As a result, the entire
the WHO Declaration of Interests form to report volume is the joint product of the Working Group
financial interests, employment and consulting and there are no individually authored sections.
(including remuneration for serving as an After the meeting, the master copy is verified by
expert witness), individual and institutional the IARC Secretariat (see Part A, Section 4(v)),
research support, and non-financial interests edited, and prepared for publication. The aim
such as public statements and positions related is to publish the volume of the Handbooks
to the subject of the meeting. IARC assesses the within approximately 12 months of the Working
declared interests to determine whether there is Group meeting. The IARC Secretariat prepares
a conflict that warrants any limitation on partic- a summary of the outcome for publication
ipation (see Table 1). in a scientific journal or on the Handbooks
Approximately 2 months before a meeting, programme website soon after the meeting.
IARC publishes on the Handbooks programme The time frame and milestones for public
website the names and principal affiliations of engagement during the development of a volume
all participants and discloses any pertinent and of the IARC Handbooks are summarized in
significant conflicts of interests, for transparency Table 2.
and to provide an opportunity for undeclared
conflicts of interests to be brought to IARC’s
attention. It is not acceptable for Observers or
third parties to contact other participants before
a meeting or to lobby them at any time. Meeting
48
Table 2 Public engagement during the development of a volume of the IARC Handbooks
Approximate time frame Milestones

~1 year before a Handbooks meeting IARC posts on the Handbooks programme website:
Preliminary List of Interventions to be reviewed
Call for Data and Call for Experts open
Requests for Observer Status open
WHO Declarations of Interests form
~8 months before a Handbooks meeting Call for Experts closes
~4 months before a Handbooks meeting Requests for Observer Status close
~2 months before a Handbooks meeting IARC publishes the names, principal affiliations, and
declared conflicts of interest of all meeting participants, and a
statement discouraging contact of Working Group members
by outside parties
~1 month before a Handbooks meeting Call for Data closes
Handbooks meeting
~2–4 months after a Handbooks meeting IARC publishes a summary of evaluations and key supporting
evidence as a scientific article in a high-impact journal or on
the Handbooks programme website
~9–12 months after a Handbooks meeting IARC Secretariat publishes the verified and edited master
copy of the plenary drafts as a Handbooks volume
6. Overview of the scientific review Materials that are publicly available and whose
and evaluation process content is final may be reviewed if there is suffi-
cient information to enable peer evaluation of the
Principles of systematic review are applied to quality of the methods and results of the studies
the identification, screening, synthesis, and eval- (see Step 1, below). Such material may include
uation of the evidence (as described in Part B, reports from government agencies, disserta-
Sections 2–7 and detailed in the Instructions for tions for higher degrees, and other apparently
Authors). For each volume of the Handbooks, reputable scientific sources. Systematic Internet
the information on the conduct of the literature searches for potentially relevant “grey literature”
searches, including search terms and the inclu- are not usually done. The reliance on published
sion and exclusion criteria that were used for and publicly available studies promotes trans-
each relevant stream of evidence, is recorded. parency and protects against citation of infor-
The Working Group considers all relevant mation that, although purportedly final, may
studies, including experimental and observa- change before it is published.
tional studies of the efficacy and/or effectiveness The steps of the review process are as follows:
of the intervention and related harms (including Step 1. Identification of the review question:
systematic reviews and meta-analyses), pertinent After the intervention (or interventions) and
information on global practices of the screening outcome (or outcomes) to be reviewed have been
methods, and background information on the specified, the IARC Secretariat, in consulta-
global epidemiology and burden of the targeted tion with the Working Group, drafts the review
cancer type. question (or questions) in PICO form (popula-
In general, only studies that have been tion, intervention/exposure, comparator, and
published or accepted for publication in the outcome) as required to determine the inclu-
openly available scientific literature are reviewed. sion and exclusion criteria for the studies. An
49
analytical framework is developed to assist in available to Working Group members for full-
identifying and formulating the review questions, text screening and inclusion in or exclusion from
with the aim of making as large a contribution as the evidence base using agreed criteria specific to
possible to the global prevention of cancer. this task.
Step 2. Comprehensive and transparent iden- Step 4. Extraction of information from included
tification of the relevant information: The IARC studies, including characteristics relevant to study
Secretariat specifies search terms for the key quality: Working Group members, working
PICO components of each question and identifies individually as members of defined Subgroups
relevant studies through initial comprehensive before the Handbooks meeting, review and
literature searches in authoritative biomedical succinctly describe pertinent characteristics and
databases (e.g. PubMed). The literature searches results of included studies as detailed in Part B,
are designed in consultation with a librarian and Sections 2–5. Study design and results are tabu-
other technical experts. The scope and speci- lated systematically in a standard format. This
fications of the searches may be modified, and step may be iterative with Step 5.
the searches rerun, depending on the amount, Step 5. Assessment of study quality: Also
relevance, and perceived completeness of the before the Handbooks meeting, Working Group
articles they identify. The IARC Secretariat may members evaluate the quality and informative-
also identify relevant studies from reference lists ness of each study they included based on the
of past Handbooks, retrieved articles, or author- considerations (e.g. design, conduct, analysis,
itative reviews, and through the Call for Data and reporting of results) described in Part B,
(see Table 2). The Working Group provides input Sections 2–5. Evaluation of study quality can be
and advice to the IARC Secretariat to refine the done either narratively or by use of a risk of bias
search strategies, and identifies additional arti- assessment tool when a relevant one is available
cles through other searches and personal expert and can add value to the process. Interpretations
knowledge. of the results, and the strengths and limitations
For certain types of interventions (e.g. admin- of each study, are clearly outlined in square
istration of regulated imaging agents), IARC brackets as part of the description of that study
also gives relevant regulatory authorities, and (see Part B).
parties regulated by such authorities, an oppor- Step 6. Peer review: Several months before
tunity to make pertinent unpublished studies the meeting, the pre-meeting drafts produced
publicly available by the date specified in the from Steps 4 and 5 are peer-reviewed by other
Call for Data. Consideration of such studies by members of the Working Group (usually within
the Working Group is dependent on the public the same Subgroup). The IARC Secretariat also
availability of sufficient information to enable an reviews the drafts for completeness, consistency
independent peer evaluation of: (i) completeness between drafts, and adherence to the Handbooks
of reporting of pertinent data; (ii) study quality; Instructions for Authors. The peer-review
and (iii) study results. comments are sent to the Working Group
Step 3. Screening, selection, and organiza- members, who produce a revised pre-meeting
tion of the studies: The IARC Secretariat screens draft. The revised drafts are reviewed and revised
the retrieved articles by reviewing the title and in Subgroup sessions during the Handbooks
abstract against the inclusion and exclusion meeting.
criteria agreed upon by the Working Group Step 7. Synthesis of results and quality of the
and technical experts in the review process. studies: The results and quality of the included
Potentially relevant studies are then made studies are synthesized by the Working Group
50
to provide a summary of the evidence and its of interest, Working Group members must accept
quality for each outcome. This synthesis can the following responsibilities:
be narrative or quantitative (for details, see (i) Before the meeting, Working Group
the Instructions for Authors), and the quality members:
synthesis may include use of an overall quality
of evidence assessment tool, such as GRADE • help in developing the analytical frame-
(Siemieniuk & Guyatt, 2019). work;
Meta-analyses of large bodies of evidence • ascertain that all appropriate studies have
may be performed by the Working Group and/ been identified and selected;
or by the IARC Secretariat before the meeting • assess the methods and quality of each
if such meta-analyses would assist in evidence included study;
synthesis and evaluation. For more information • prepare pre-meeting drafts that present
on the conduct and use of such meta-analyses, an accurate quantitative and/or textual
see Part B, Section 5.1c. synthesis of the body of evidence, with key
Step 8. Interpretation of study results and elements of study design and results and
evaluation of strength of evidence: The whole notable strengths and limitations;
Working Group reviews the study descriptions
and the summaries of the body of evidence for • participate in conference calls organized
each outcome or end-point, discusses the overall by the IARC Secretariat to coordinate the
strengths and limitations of the evidence in development of pre-meeting drafts and to
each stream of data, and evaluates the strength discuss cross-cutting issues; and
of evidence for a preventive effect on cancer or • review and provide comments on
an intermediate outcome in each stream using pre-meeting drafts prepared by other
transparent methods, which may include the members of their Subgroup or of the
use of established specific tools. The preventive Working Group.
effect for each stream of evidence is assessed.
(ii) At the meeting, Working Group members
The Working Group then integrates the assess-
work in Subgroups to:
ments from all streams of evidence (see Part B,
Section 7.1) and develops the rationale for its • critically review, discuss, and revise the
consensus evaluation of the preventive effect pre-meeting drafts and adopt the revised
of the screening or early diagnosis method (see versions as consensus Subgroup drafts;
Part B, Section 7.2). and
• develop and propose an evaluation of the
strength of the evidence summarized in
7. Responsibilities of the Working
the consensus Subgroup drafts (see Part B,
Group Section 6), using the IARC Handbooks
The Working Group is responsible for the criteria (see Part B, Section 7.1).
final list of studies included in the evaluation (iii) At the meeting, Working Group members
and the review and evaluation of the evidence work in plenary sessions to:
for a Handbook, as described above. The IARC
Secretariat supports these activities (see Part A, • present their Subgroup drafts for scientific
Section 4). To ensure that the process is rigorous, review by and discussion with the other
independent, and free from individual conflicts Working Group members, and subsequent
revisions, as needed;
51
• participate in review and discussion of B. SCIENTIFIC REVIEW AND

other Subgroup drafts and in their adop- EVALUATION
tion as a consensus Working Group draft;
• participate in review and discussion of the This part of the Preamble discusses the types
summaries and evaluations of the strength of evidence that are considered and summarized
of the evidence developed in Subgroups in each section of a Handbook, followed by the
(see Part B, Section 7.1), and contribute to scientific criteria that guide the evaluations. In
their revision, as needed, and their adop- addition, a section of General Remarks at the
tion by consensus of the full Working front of the volume discusses the reasons the
Group; and interventions were scheduled for evaluation and
• contribute to the discussion of and adop- any key issues encountered during the meeting.
tion by consensus of an overall evaluation
proposed by the Meeting Chair using the
guidance provided in Part B, Section 7.1. 1. Definitions
The Working Group strives to achieve Secondary prevention of cancer is the use of
consensus evaluations. Consensus reflects broad methods that can lead to the detection of asymp-
agreement among the Working Group members, tomatic or early symptomatic precancerous
but not necessarily unanimity. If unanimity has conditions or cancers at a stage when treatment
not been reached when the interpretations of the of a lesion that is found can prevent progression
evidence by all Working Group members have to invasive cancer or, if the cancer is already inva-
been expressed and debated, the judgement of sive, prevent death from cancer. The two corner-
the majority of the Working Group members stones of secondary prevention are screening
is taken as the consensus. When consensus and early diagnosis. WHO defines these terms as
is reached in this way, the Meeting Chair may follows (https://www.who.int/cancer/prevention/
poll Working Group members to determine and diagnosis-screening/en/).
record the diversity of scientific opinion on the Screening is “the systematic application of
overall evaluation. a screening test in a presumably asymptomatic
Only the final product of the plenary sessions population. It aims to identify individuals with
represents the views and expert opinions of the an abnormality suggestive of a specific cancer.
Working Group. The Handbook is the joint These individuals require further investigation.”
product of the Working Group and represents Early diagnosis is “the early identification
an extensive and thorough peer review of the of cancer in patients who have symptoms of
body of evidence (review of individual studies, the disease”. Early diagnosis is most commonly
synthesis, and evaluation) by a multidisciplinary achieved by raising “the awareness (by the
group of experts. Initial pre-meeting drafts and public or health professionals) of early signs and
subsequent revisions are temporarily archived symptoms of cancer in order to facilitate diag-
but are not released, because they would give nosis before the disease becomes advanced. This
an incomplete and possibly misleading impres- enables more effective and simpler therapy.”
sion of the consensus developed by the Working
Group over its complete deliberation.
52
WHO defines a cancer early detection pro- positive (Miles et al., 2004). The IARC Handbooks
gramme as “the organized and systematic imple- assess all available relevant evidence from both
mentation of early diagnosis or screening (or organized programmes and opportunistic set-
both), diagnosis, treatment, and follow-up”, thus tings in their evaluation of the effectiveness of
encompassing both screening and early diag- a screening method or early diagnosis method.
nosis. Early detection programmes, when imple- Whether organized or opportunistic, screen-
mented, usually operate alongside opportunistic ing is a complex public health strategy that
early diagnosis and/or screening. requires substantial health-care resources, infra-
IARC defines an organized screening pro structure, and coordination to be effective. In
gramme as one that has “an explicit policy with addition, screening should be undertaken only
specified age categories, method, and interval when efficacy and, ideally, effectiveness have been
for screening; a defined target population; a established. It should also only be undertaken
management team responsible for implementa- when resources are sufficient to cover a large
tion; a health-care team for decisions and care; proportion of the intended target group, when
a quality assurance structure; and a method facilities exist for follow-up of screen-positive
for identifying cancer occurrence in the target subjects to confirm or exclude disease and ensure
population” (IARC, 2005). In principle, an treatment, and where the disease is a sufficiently
organized screening programme also includes burdensome public health problem to justify the
systematic invitation of the target population effort and costs of screening. In addition, infor-
for quality-assured screening tests and assured mation systems are essential to monitor inputs
follow-up of screen-positive subjects with diag- and evaluate outcomes.
nostic investigations, treatment, and post-treat- Early diagnosis programmes of cancer also
ment care. The former can minimize inequalities have minimum requirements, specifically the
in access to screening by giving every eligible and facilities needed to confirm or exclude a diag-
contactable person access to screening. nosis of cancer in people who present to health-
Opportunistic refers to the fact that the care providers with symptoms suggestive of a
medical examination is requested by a patient potentially curable cancer, and to ensure treat-
or offered by a health practitioner in the context ment when a diagnosis of cancer is confirmed.
of the patient–practitioner relationship and is At present, the tools of early diagnosis are largely
not, or is minimally, subject to any other orga limited to community education about symp-
nizing principle. The proportion of screening for toms that may suggest cancer, and to educating
a particular cancer that is opportunistic varies or enabling primary care practitioners to ask
widely from country to country; in many coun- at-risk patients presenting for any care about
tries screening is exclusively opportunistic, and symptoms they have that may be signs of cancer.
in some countries screening is almost exclusively Evidence of the effectiveness of such measures is
organized (for particular types of cancer). accumulating (Emery et al., 2014). Other possible
Compared with opportunistic screening, interventions to promote early diagnosis may
organized screening focuses much greater atten- involve regulation of health care and organiza-
tion on higher coverage by way of systematic tion of health services.
invitation and on the quality of the screening It is important to note that in low- and
process, and provides greater protection against middle-income countries, depending on soci-
the harms of screening, including overscreening, etal prioritization, early diagnosis programmes
poor-quality screening, adverse events of may be the only affordable option for increasing
screening, and poor follow-up of those who test the detection of cancer when it is potentially
53
curable. Screening (organized or opportunistic) 3. Screening methods

may be unaffordable, although simulation of
realistic cost–effectiveness (taking into account Screening methods for the relevant cancer
all societal costs) might make some programmes site are considered for evaluation if they have
attractive. been subject to one or more well-conducted ran-
Early diagnosis and screening are the early domized controlled trials with cancer incidence
parts of a multistep process. The Handbooks and/or mortality (see Part A, Section 3) as the
consider for evaluation the methods used for trial outcome. Screening methods for which no
early diagnosis and screening, and not the steps randomized controlled trials are available may
that follow in the process. Although the following be evaluated if the body of evidence from obser-
details about the scientific review and evaluation vational studies is sufficiently large to warrant
refer specifically to screening interventions, they evaluation, especially for screening methods that
will also apply for the evaluation of early diag- are already in use in the community.
nosis interventions, with some adaptation as New screening methods and innovations
needed. in existing methods that may offer signifi-
cant improvements in screening performance,
increases in acceptability of screening, or reduc-
2. Characterization of the disease tions in cost of screening but that did not meet
This type of Handbook addresses screening the threshold for detailed review and evaluation
for cancer at one specific site. Information is described above (i.e. are materially different from
presented on the precursor or invasive lesions other methods under consideration and have
that cancer screening aims to detect. Each cancer not been subject to one or more well-conducted
or other lesion is precisely defined as to its loca- randomized controlled trials or are in widespread
tion and morphology, using the appropriate use in some countries), or for which the body of
codes from the latest International Classification evidence was too limited to enable an evaluation
of Diseases for Oncology (IARC, 2019a) and brief to be performed, are also reviewed. The review
pathological criteria for its diagnosis as published includes a description and critical assessment of
by IARC (IARC, 2019b). The global distribu- any studies on the performance or the screening
tion and burden of the cancer are summarized, effect of these new methods or innovations of
including regional differences, time trends, and existing methods.
credible projections of incidence and/or mortality, Emerging methods may be evaluated in the
based on IARC’s data from cancer registries. The absence of studies of efficacy or effectiveness if
natural history of the cancer and its established comparative data with an established screening
risk factors and preventive factors are briefly method are available. Such comparative data
described. The nature and efficacy of evidence- may include data
based, potentially curative therapy is also briefly (i) on performance against validated reference
described, together with geographical variation standards (including those of the International
in its nature and accessibility worldwide. Organization for Standard ization [ISO] when
relevant);
(ii) on other performance characteristics in
populations at average risk; and
(iii) on intermediate outcomes that provide
data on efficacy or effectiveness (e.g. sensitivity,
54
specificity, and interval cancer rate) (Young et al., • the type of screening offered (e.g. opportu
2016). nistic screening, organized population-wide
Ideally, such comparisons will have been programme);
made under conditions in which potential biases • the screening methods most commonly used
have been minimized. Possible differences in or recommended; and
other important characteristics, such as accept • availability of facilities, extent of population
ability and possibility of harm, are also taken coverage, and participation rates.
into account. In addition, demographic, cultural, and
Each method considered for evaluation is behavioural considerations that affect partic-
described, and its state-of-the-art application is ipation in screening are presented in a global
outlined. The description of each method should perspective, with some specific, local character-
include whether the goal of screening is to reduce istics, as appropriate.
cancer-specific mortality by primarily detecting
invasive lesions, or to reduce cancer-specific inci-
dence by primarily detecting precursor lesions. 5. Epidemiological studies of each
The characteristics of the target population, screening method
such as age ranges and sex, should be stated.
Other relevant issues for the method should be The evaluative processes described here are
addressed, including: repeated in full, as far as they apply, for each
screening method reviewed.
• equipment and training required; Relevant studies of cancer in humans are
• technical quality control; identified using systematic review principles,
• the screening protocol and its expected as described in Part A and further detailed in
performance, including sensitivity and the Instructions for Authors provided to each
specificity; Working Group. Eligible studies include: all
• host factors that affect screening perfor- studies in humans of the association of the
mance; screening intervention of interest with its cancer
incidence, mortality, or intermediate outcome
• any assessment protocol for screen-positive
target (studies of benign neoplasms, pre-neo-
subjects; and
plastic lesions, and other outcomes are reviewed
• quality assurance. when they are outcomes sought by, or interme-
diate outcomes related to, the screening interven-
tion reviewed); studies dealing with the accuracy
4. Current global screening (sensitivity, specificity, and predictive values) of
practices the screening intervention; studies examining a
A brief overview of relevant screening prac- putative harm as an outcome of the screening
tices in different regions of the world is presented, intervention; reports on the balance of benefits
limiting the description to those countries and harms of screening; and reports on the cost–
or settings where screening takes place. The effectiveness of screening. Search strategies must
following aspects are summarized if available: take into account the possibility that any of the
above-mentioned outputs from a single study
• policies and guidelines for, and regulation of, may have been published separately from the
screening; other outputs of the study. Multiple publications
may arise from successive follow-ups of a single
55
trial population or cohort, from analyses focused emerging method and its relative impact on
on different aspects of a screening–outcome asso- screening outcomes.
ciation, or from inclusion of overlapping popula- (ii) Observational studies: Typically, obser-
tions. In these situations, only the most recent vational studies include cohort studies
publication or the one that provides the most, or (including variants such as case–cohort and
most relevant, information should be included, nested case–control studies), case–control
unless circumstances warrant otherwise. studies, cross-sectional studies, and ecolog-
ical studies, all with cancer incidence or
5.1 Evaluation of the preventive and mortality as an outcome. In addition to these
harmful effects of the intervention designs, innovations in epidemiology enable
many variant designs that may be considered
(a) Types of studies considered
in Handbooks evaluations. Observational
Several types of epidemiological studies studies generally provide evidence of effec-
contribute to the evaluation of the benefits and tiveness only.
harms of cancer screening. Benefits are the prin-
Cohort and case–control studies of screening
cipal focus of this section.
typically relate individual exposure to the
(i) Experimental studies: Allocation by screening intervention under study to the inci-
the investigator of the participants to the dence of or mortality from the target cancer in
intervention (screening) or control condi- individuals, and provide an estimate of the rela-
tion, ideally by a random and blind process tive incidence of or mortality from cancer as the
(to the investigator and the participant), is main measure of screening effect. In addition,
the defining characteristic of experimental cross-sectional studies may be used to measure
studies. These studies can include classic accuracy, such as sensitivity, specificity, and
individually randomized controlled trials, predictive values.
cluster-randomized controlled trials that In ecological studies, the unit of investigation
include sufficient clusters to minimize prob- is not an individual but a whole population or
ability of bias, and a range of other designs a set of subgroups of a population, and cancer
in which there is non-random allocation of incidence or mortality is related to a summary
participants to the intervention or control measure of the exposure (screening method) of
condition or there are too few randomization the whole population at different times, or aggre-
units to minimize bias. gate measures of the exposure in the subgroups
In principle, experimental studies can at the same time. Time-based ecological studies
provide evidence for efficacy or effectiveness may be of particular interest in evaluating the
of an intervention that is at low risk of bias. In impact of screening methods, because changes
particular, pragmatic trials (trials designed to in cancer incidence or mortality, or harms, over
test the effectiveness of the intervention in a interrupted time periods can be related to expo-
broad routine clinical practice) can provide sure to the screening method within a single
evidence of effectiveness when conducted in population. Nevertheless, results from ecological
settings with populations at average risk. studies should be interpreted with caution for two
Studies with a tandem design (i.e. the same reasons: (i) because they are prone to misclassi-
population is screened with both methods fication of exposure within individual time or
consecutively) can also be useful, to assess an population units, due to the lack of individual
data on exposure or outcome, and (ii) because
56
of the limited ability to adjust for confounders. are varied and can be broadly categorized as
Therefore, ecological studies should generally selection bias, information bias (e.g. screening
be used to raise hypotheses and to support the intervention and outcome measurement error),
evidence of results from experimental or other and confounding bias (Rothman et al., 2008).
observational studies. Selection bias in an epidemiological study can
occur when the inclusion of participants from
(b) Study quality and informativeness the eligible population or their follow-up in the
The following paragraphs outline the general study is influenced by their exposure (screening
principles of description, analysis, and inter- use) or their outcome (usually disease occur-
pretation of epidemiological studies in a cancer rence). Under these conditions, the measure
screening context. It is important to note that the of association found or not found in the study
evaluation of cancer screening studies involves may not accurately reflect the association or lack
complexities that are uncommon to other fields of thereof that might otherwise have been found
epidemiology. Some examples of these complexi- in the eligible population (Hernán et al., 2004).
ties are self-selection for screening, heterogeneity Information bias results from inaccuracy in
of opportunity to be screened, confounding with intervention or outcome measurement. Both can
differential treatment, and the complexities of cause an association between hypothesized cause
lead time, length sampling, and overdiagnosis and effect to appear stronger or weaker than it
(IARC, 2016b). really is. Confounding arises when a third factor
Epidemiological studies are susceptible to is associated with both the intervention and the
several different sources of error. Study quality is outcome and, because of this, influences the
assessed as part of the structured expert review apparent association between them (Rothman
process undertaken by the Working Group. A et al., 2008). An association between the purport-
key aspect of quality assessment is consideration edly preventive intervention and another factor
of the possible roles of chance and bias in the that is associated with an increase or a decrease
interpretation of epidemiological studies. in the incidence of or mortality from the disease
Chance, also called “random variation”, can can lead to a spurious association or the absence
produce misleading study results. This vari of a real association of the purportedly preven-
ability in study results is strongly influenced by tive intervention with the disease. When either
the sample size: smaller studies are more likely of these occurs, confounding is present.
than larger studies to have effect estimates that In principle, experimental studies are less
are imprecise and, therefore, are more likely to prone to each of these sources of bias, because
be misleading. Confidence intervals around a selection for intervention or non-intervention
study’s point estimate of effect are routinely used is determined by the investigator (usually by
to indicate the range of values of the estimate that random allocation) and not by the study partici-
could be produced by chance. Both experimental pants or their characteristics. However, bias may
and observational epidemiological studies are arise because of lack of concealment, non-random
prone to effects of chance. allocation, lack of blinding, post-randomization
Bias is the effect of factors in study design, exclusions, or non-acceptance of or non-adher-
conduct, or reporting that lead an association ence by the study participants to the conditions
to erroneously appear stronger than, weaker of the study arm (screening or not screening) to
than, or opposite in direction to the associa- which they were randomized when, as is usual
tion that really exists between an exposure and in experimental studies of cancer screening, they
an outcome. Biases that require consideration are not blind to their study arm. In addition,
57
even when they are blind to the study arm, a • Study population: Whether the study popula-
high degree of participant non-adherence may tion was appropriate for evaluating the associ-
cause important information bias and potential ation between the screening intervention and
confounding with variables related to the choice cancer. Whether the study was designed and
of whether to adhere or not adhere to the study conducted in a manner that would minimize
conditions. Because of such possibilities for selection bias and other forms of bias. The
confounding, it is common practice to include designated outcomes in the study population
key confounding variables in the data collected must have been identified in a way that was
from or about participants, to enable statistical independent of the screening intervention,
control of confounding. for both experimental studies and observa-
Two other sources of bias may have impor- tional studies, and the screening intervention
tant effects on the estimates of the screening must have been assessed in a way that was not
efficacy: lead-time bias and length bias (Cole related to disease (outcome) status. In these
and Morrison, 1980; IARC, 2016b). Lead time is respects, completeness of recruitment into
the period between screen detection and when the study from the population of interest and
a tumour would have been clinically diagnosed completeness of follow-up for the outcome
in the absence of screening. The survival time, (see below) are very important.
defined as the time from the date of diagnosis • Outcome measurement: The appropriate-
of cancer to the date of death, of screen-detected ness of the outcome measure (incidence of
cases is overestimated because of this lead time, cancer, mortality from cancer, or an interme-
even for individuals who do not benefit from diate outcome, as defined in Part B, Section 1)
screening. Therefore, lead-time bias can produce for the screening intervention and the cancer
data that appear to support a favourable effect of type under consideration, the outcome
screening, if conclusions are based on survival ascertainment methodology, and the extent
analysis. to which outcome misclassification may
The other important bias is length bias (or have led to bias in the measure or measures
length-sampling bias). The probability of a of association (e.g. because of systematic
tumour being detected at screening depends, at differences between exposed and unexposed
least in part, on its growth rate, because slow- people in the way in which the outcome was
growing tumours have a longer preclinical detect- ascertained, and lack of blinding of ascer-
able phase compared with fast-growing tumours. tainment of cancer outcomes, which requires
Thus, tumours detected at screening are a biased the exercise of human judgement).
sample of preclinical lesions, weighted towards • Intervention measurement: This includes
slower-growing tumours, which are generally (i) the adequacy (including the validity and
thought to be associated with a better prognosis the reliability) of the methods used to assess
and therefore longer survival. This again leads to the intervention in observational studies,
bias apparently in favour of screening. and adherence to the intervention condition
In assessing the quality of the studies, the in experimental studies, and (ii) the likeli-
Working Group considers the following aspects: hood (and direction) of bias in the measure
• Study description: Clarity in describing the or measures of association because of inter-
study design and its implementation, and the vention measurement error or misclas-
completeness of reporting of all other key sification in observational studies and
information about the study and its results. non-adherence to the intervention condition
58
and cross-contamination of the non-inter- including adjusting for matching when

vention group in experimental studies (as necessary and avoiding treatment of prob-
described in Part B, Section 5.1). able mediating variables as confounders, is
• Assessment of potential confounding: The considered. Detailed analyses of cancer risks
extent to which the authors took into account in relation to summary measures of inter-
in the study design and analysis potentially vention, such as cumulative exposure to the
confounding variables, including co-expo- intervention, or temporal variables, such as
sures, that could influence the occurrence of age at first intervention or time since first
the outcome and may be related to the inter- intervention, are reviewed and summarized
vention of interest. Particular to screening when available.
interventions is the possibility that for a For the sake of economy and simplicity, this
given stage, people with screen-detected Preamble refers to the list of possible sources
cancers receive better treatment than those of error with the phrase “chance, bias, and
with symptom-detected cancers. Important confounding”, but it should be recognized that
sources of potential confounding by such this phrase encompasses a comprehensive set
variables should, where possible, have been of concerns pertaining to study quality. These
addressed in the study design, such as by elements of study quality do not constitute and
randomization, matching, or restriction, or should not be used as a formal checklist of indi-
in the analysis by statistical adjustment. In cators of study quality. Rather, the assessment by
some instances, where direct information on the Working Group is reported in a narrative
confounders is unavailable, use of indirect way, in the form of comments in square brackets.
methods to evaluate the potential impact The judgement of the experts is critical in deter-
of confounding on intervention–outcome mining how much weight to assign to different
associations is appropriate (e.g. Axelson & issues when considering how all these potential
Steenland, 1988; Richardson et al., 2014). sources of error should be integrated and how
• Other potential sources of bias: Each to rate the potential for error related to each.
epidemiological study is unique in its study However, it is important that the process under-
population, its design, its data collection, taken, including the weight given to various
and, consequently, its potential biases. All studies, be replicable and be described in a way
possible sources of bias are considered for that is transparent to readers.
their possible impact on the results. Several • Study informativeness: The informativeness
sources of bias have important effects on the of a study is its ability to show a true preven-
estimation of screening efficacy. The possi- tive effect, if one exists, of the intervention on
bility of reporting bias (selective reporting of the outcome, and not to show an effect if one
some results) should also be explored. does not exist. Key determinants of informa-
• Statistical methodology: The studies are tiveness include having a study population
evaluated for the adequacy of the statistical of sufficient size to obtain precise estimates
analysis methods used and their ability to of effect, sufficient elapsed time from inter-
obtain unbiased estimates of intervention– vention to measurement of outcome for an
outcome associations, confidence intervals, effect, if present, to be observable, presence of
and test statistics for the significance of adequate intervention contrast, and relevant
measures of association. Appropriateness and well-defined time windows for interven-
of methods used to address confounding, tion and outcome.
59
(c) Meta-analyses and pooled analyses additional checks made of the alignment of
Independent epidemiological studies of the the systematic review specifications with those
same intervention with a comparatively weak required for the Handbooks evaluation, the
effect or small sample size may produce incon- completeness of coverage of articles relevant to
clusive results that are difficult to summarize. the evaluation compared with those ultimately
Combined analyses of data from multiple studies included in the meta-analysis, and the accuracy
may increase the precision of estimates. There of extraction of required data from the results of
are two types of combined analysis: (i) meta- the individual studies.
analysis, which involves combining summary Subject to the judgement of the IARC Sec-
statistics, such as relative risks from individual retariat and in consultation with the Working
studies, and (ii) pooled analysis, which involves Group, the updating of meta-analyses or
a pooled analysis of the raw data from the indi- the conduct of ad hoc meta-analyses may be
vidual studies (Greenland & O’Rourke, 2008). performed by the Working Group and/or by
There are also “umbrella reviews”, systematic the IARC Secretariat during preparation for a
reviews of multiple meta-analyses, which may Handbooks meeting, when there are sufficient
be evaluated by the Working Group. studies of an intervention–outcome association
The strengths of combined analyses are to aid the Working Group’s assessment of the
increased precision due to increased sample size association. When results from both experi-
and, in the case of pooled studies, the opportu- mental and observational studies are available,
nity to better control for potential confounders any combined analyses should be conducted
and to explore interactions and modifying effects separately for experimental efficacy studies,
that may help to explain heterogeneity between experimental effectiveness studies, and obser-
studies. A disadvantage of combined analyses is vational studies, with consideration given to
the possible lack of comparability of results from separate combined analyses of cohort and
various studies, because of differences in specifi- case–control studies, because of their different
cation of the intervention or the outcome, popu- propensities to bias. The results of such ad hoc
lation characteristics, subject recruitment, data meta-analyses, which are specified in the text of
collection procedures, methods of measurement, the Handbook by presentation in square brackets,
and effects of unmeasured covariates, which may may come from the addition of the results
differ among studies. These differences in study of more recent studies to those of published
methods and quality can influence the results of meta-analyses or from de novo meta-analyses.
both pooled analyses and meta-analyses. Additional details on the conduct of such ad hoc
Meta-analyses considered by the Working meta-analyses are provided in the Instructions
Group may include high-quality published for Authors.
meta-analyses, updates of such meta-ana- Irrespective of the source of the informa-
lyses, and new meta-analyses. When published tion for the meta-analyses and pooled analyses,
meta-analyses are considered by the Working the criteria for information quality applied are
Group, the conduct and reporting quality of the the same as those applied to individual studies.
meta-analyses will be carefully assessed against The sources of heterogeneity among the studies
prior expectations set with reference to items in contributing to them are carefully considered
checklists for published systematic reviews and and the possibility of publication bias evaluated.
meta-analyses, such as AMSTAR (AMSTAR,
2017) and/or PRISMA (Moher et al., 2009), with
60
(d) Evaluation of new technologies (e) Considerations in assessing the body of

It is important that a new screening test or epidemiological evidence
method is evaluated before it replaces existing The ability of the body of epidemiological
technology. New technology need not be subject evidence to inform the Working Group about
to a full controlled trial of efficacy if it is similar the efficacy or effectiveness of a screening inter-
enough to the old technology and if the old tech- vention is related to both the quantity and the
nology has been shown to reduce cancer incidence quality of the evidence. There is no formulaic
or cancer mortality. A new technology is consid- answer to the question of how many studies are
ered similar enough if the method of screening needed from which to draw inferences about
is based on the same principles as the old tech- the efficacy or effectiveness of a screening inter-
nology and targets lesions with the same biology. vention, although more than a single study in a
In such instances, instead of a full controlled single population will almost always be needed.
trial of efficacy, the following are required: Experimental and observational studies are
(i) adequate analytical and clinical validity of to be considered. Randomized controlled trials
the test in human subjects; (ii) cross-sectional typically provide the strongest evidence, but
evaluation of diagnostic accuracy of the new observational studies also provide valuable and
method for intermediate outcomes validated timely information. For example, observational
in randomized controlled trials or in tandem studies can be done for initial evaluation of
studies in a screening population at average proposed screening methods and for evaluation
risk (Young et al., 2016); and (iii) a prospec- of their effectiveness after dissemination has
tive evaluation over more than one screening occurred.
round of the comparative performance of the After the quality of individual epidemiolog-
two methods, including participation, detection ical studies has been assessed and the informa-
rates, false-positive rates, interval cancer rates, tiveness of the various studies on the association
and the burden and harms of screening (Irwig between screening and cancer or an interme-
et al., 2006; Young et al., 2016). In the absence of a diate outcome has been evaluated, the body of
reduction in risk of interval cancer, any increase evidence is assessed and a consensus scientific
in test sensitivity is probably due to an increase judgement is made about the strength of the
in overdiagnosis (see Section 5.2), which could evidence that the screening method under review
make the new technology more harmful, rather reduces the incidence of cancer or death from
than more beneficial, than the old technology. If cancer. In making its judgement, the Working
the Working Group decides to make a full eval- Group considers several aspects of the body of
uation of a new screening method in comparison evidence (e.g. Hill, 1965; Rothman et al., 2008;
with an existing screening method that has been Vandenbroucke et al., 2016).
established to reduce the incidence of cancer or A strong association (e.g. a large relative risk or
death from cancer, it does a full systematic review a relative risk that is well below 1.0) is more likely
of research evidence relevant to this question, as to be causal than a weak association, because it is
described in Part A, Section 6. harder for confounding or other biases to create a
greater association than the one that is observed.
However, it is recognized that estimates of
effect of small magnitude do not imply lack of
causality and may have a substantial impact on
public health if the disease is common or if the
61
screening intervention is highly feasible and/or rigorous, and informative as it is for the evidence
widely applicable. Estimates of effects of small of beneficial effects.
magnitude can also contribute useful informa- Occurrence of screening harms is reviewed
tion to the assessment of screening efficacy or and described, and their potential impacts are
effectiveness if the magnitude of the effect corre- discussed. The evaluation of harms includes:
lates with the level of screening intervention in (i) estimates of rates of false-positive and
populations that are differently exposed. false-negative findings, overdiagnosis, and
Associations that are consistently observed overtreatment, which are harms shared by all
in several studies of the same design, in studies screening methods; and (ii) estimates of risks
that use different epidemiological approaches, of harm intrinsic to the screening method, and
or under different circumstances of intervention not necessarily shared by other methods (e.g.
are more likely to indicate screening efficacy or radiation-induced cancer due to radiographic
effectiveness than are isolated observations from screening). Interval cancers are not considered
single studies. If there are inconsistent results to be a harm, because they are, in essence, a
among investigations, possible reasons for such planned outcome of the frequency with which
inconsistencies are sought (e.g. populations screening is offered to members of the target
studied, intervention characteristics, measure- population and are balanced against harms that
ments of outcomes, differences in study informa- would increase in probability with increasing
tiveness because of time since initiation of the frequency of screening. However, it is recognized
intervention, screening participation), and their that some interval cancers are a consequence of a
implications for the overall findings are assessed. false-negative test.
Results of studies that are judged to be of high The actual harms of the screening test itself or
quality and highly informative are given more mediated by the screening-related events listed
weight than those of studies that are judged to be above include: (i) physical and psychological
methodologically less sound or less informative. discomfort due to, and medical complications
Temporality of the association is also an of, the screening method or further investigation
essential consideration, that is, the interven- of positive findings and subsequent treatment;
tion must precede the outcome, and by a time (ii) all harmful consequences of overdiag-
period that is sufficiently long for observation of nosis and/or overtreatment of screen-detected
a screening effect to be plausible. cancers, including preclinical cancers, and of
precancerous lesions; (iii) unnecessary diagnosis
5.2 Harms of screening and treatment of overdiagnosed cancers; and
(iv) delay in diagnosis, a possibly poorer outcome
Potential harms to individuals that are linked of the targeted cancer, and feelings of betrayal
to the screening method under review are also due to the false reassurance of a false-negative
reviewed. Evidence of harm may come from any finding.
type of epidemiological study (see Section 5.1a) Overdiagnosis is defined in the Handbooks as
and may also be reported in studies separately the diagnosis of a cancer as a result of screening
from evidence on the benefits of screening that would never have caused any symptoms
using the same criteria as for preventive effects. or problems if it had not been detected by
Although the IARC Handbooks do not formally screening. Screening may also detect a large
evaluate the harms associated with screening in number of precursors of cancer that would
the way that is done for the benefits, the review not have progressed to clinical cancer in the
of the evidence of harms aims to be as complete, person’s lifetime. The main concern in such
62
cases is overtreatment. There are challenges to in absolute terms (e.g. numbers of beneficial and
estimating overdiagnosis, and there are several harmful outcomes per 1000 screened individ-
ways in which it can be estimated, including the uals), the Working Group selects those based
excess-incidence approach and the mean-lead- on the highest-quality evaluative studies of the
time approach. Estimates can be made from commonly implemented screening regimens,
“well-conducted, population-based random- critically assesses each study, summarizes
ized controlled trials with long follow-up and the results in narrative or tabular format as
minimal to no screening in the control group” appropriate, and critically assesses the body of
(Davies et al., 2018), as well as from statistical evidence. The Working Group may also propose
modelling and from ecological studies. When one or more “best” estimates of the balance of
there are several plausible estimates of overdiag- benefits and harms, while noting the limits of
nosis, results of any combined analyses of these applicability of those estimates to settings other
estimates are also reviewed. than the populations and screening experience
The IARC Secretariat, in consultation with from which they were derived.
the Working Group, may also commission or As noted in Part B, Section 1, the balance
conduct a meta-analysis of such studies. of benefits and harms of screening is expected
to be more favourable in organized screening
5.3 Balance of benefits and harms programmes than in the case of opportunistic
screening. The balance may also differ substan-
A sound estimate of the balance of benefits tially between specific population subgroups,
and harms of a screening programme is impor- for example human papillomavirus (HPV)-
tant to aid decisions about whether to offer the vaccinated and non-vaccinated women for
programme and is most important for people cervical cancer screening. Major factors that
who are deciding whether to participate in the influence the balance of benefits and harms
programme. Estimates of the balance of bene- include background cancer risk, life expectancy,
fits and harms for a particular cancer screening sex, and age. Where possible, the Working Group
programme usually comprise one estimate of will acknowledge these factors and consider
benefit (e.g. number of cancer deaths prevented comparing benefits and harms for different
per 1000 eligible people fully participating in population subgroups.
the programme) and several estimates of harm In addition to the balance of benefits and
(e.g. number of false-positive screening tests, harms, the net benefit of screening (which can
and number of overdiagnosed cancers, per be positive or negative) may be estimated in an
1000 eligible people fully participating in the aggregate manner, for example by calculating
programme). These estimates are usually based the average number of quality-adjusted life years
on experimental or high-quality observational (QALYs) gained or disability-adjusted life years
evaluations (e.g. incidence-based mortality (DALYs) averted as a result of screening. QALYs
analyses done under optimal circumstances) and DALYs are generic measure of disease burden
of the performance of screening methods or that include quality and quantity of life in their
programmes. To project estimates of benefits and estimation. Because both are based on estima-
harms to a steady-state programme operating in tion of lifetime outcomes and are estimated by
a particular general population, modelling is modelling, they cannot be estimated directly
required. from trials.
After identification of all published estimates In consultation with the Working Group and
of the balance of benefits and harms expressed when it is feasible and potentially contributory,
63
the IARC Secretariat may commission or conduct 5.4 Cost–effectiveness

a systematic review of modelling studies that
have estimated QALYs gained or DALYs averted For a screening method or programme that is
from screening, and also modelling studies that capable of delivering a beneficial outcome, cost–
have estimated disaggregated measures of bene- effectiveness is usually expressed as the estim-
fits or harms. The Working Group will critically ated financial cost of implementing the method
appraise the quality of the studies using inter- or programme per unit of the benefit it delivers,
nationally accepted criteria for good model- which is most often measured in terms of life
ling conduct (Caro et al., 2012) and applicable years, as QALYs gained or DALYs averted. The
subject-specific quality frameworks for models. ratio of costs to benefits (i.e. level of cost–effec-
High-quality collaborative modelling studies tiveness) needed to implement a health service
(i.e. studies in which different modelling groups programme varies from country to country,
work together using standardized assump- depending principally on the wealth of the
tions) will be favourably viewed in considering country and on who pays (e.g. the government
the overall quality of a particular evaluation. or individual citizens). Therefore, the specific
Petitti et al. (2018) provided a checklist for the ratio derived from cost–effectiveness analyses
critical appraisal of collaborative modelling from a certain country is usually not general-
reports specific to cancer screening, which can izable to other countries and settings. However,
also be used for the appraisal of single model- if there are sufficient (high-quality) analyses
ling studies. Baseline parameters used and their from different parts of the world with consistent
sources, most particularly the sources of calibra- results on the cost–effectiveness of the screening
tion data, and other assumptions made in the intervention of interest within their respective
absence of relevant baseline data require careful settings, qualitative statements can be made
scrutiny. Special attention needs to be paid to the about the cost–effectiveness of the screening
extent to which weights for quality and disability intervention. Although assessments of cost–
have been incorporated for all relevant phases of effectiveness that account for all costs (e.g. that
screening and management of cancer, and also are not restricted to health service costs) are less
whether disutility is available for all downstream frequently done, it is important to note that their
management pathways after the screening test, perspective may differ markedly from one based
and whether these have been modelled in detail or on health service costs only. Like the balance of
as a single aggregate disutility. Currently, there is benefits and harms, cost–effectiveness estimates
a general paucity of evidence to support detailed can be markedly different in different population
modelling of disutility for each step involved in subgroups, depending on background cancer
screening, triage, diagnosis, surveillance, and risk, life expectancy, sex, and age, among others.
treatment (all of which are required to model the Ideally, the cost–effectiveness analysis should be
detailed impact of a screening programme on based on the primary population targeted for
QALYs or DALYs). As a result, primary studies screening; incremental analyses can consider the
may judiciously choose to present aggregate inclusion of additional populations (e.g. extended
benefits information summarized as life years age range for screening).
saved, and these data should be considered Taking a similar approach to that taken for the
very carefully as less prone to issues around the balance of benefits and harms described above,
uncertainty inherent in estimation of QALYs or the IARC Secretariat may commission or conduct
DALYs. a systematic review of published reports of cost–
effectiveness analyses. Studies to be included
64
report on net costs (including upfront costs of such as screening protocol, acceptability, harms,
screening and downstream costs and savings for costs, and equity of access, that can influence the
follow-up and management of cancers) as well as population impact of a screening method.
net benefits, preferably in the form of life years In the absence of such evidence, the Working
gained, QALYs, or DALYs. Methods for all such Group may critically appraise the commonly
studies will include modelling. Where appli- advanced reasons for choosing one method over
cable, study quality will be appraised in ways another and the justifications given for them,
similar to those described in Section 5.1b, with taking into account all the dimensions listed
the addition of appraisal against internationally above.
accepted criteria for good conduct of cost–effec-
tiveness analysis, such as the Recommendations 5.6 Surveillance in populations at increased
for Conduct, Methodological Practices, and risk
Reporting of Cost-effectiveness Analyses by the
Second Panel on Cost-Effectiveness in Health Screening in people with a personal history of
and Medicine (Sanders et al., 2016). Methods, the cancer type subject to screening is not evalu-
assessment against quality criteria, and results ated in the Handbooks.
will be tabulated for high-quality studies of Population subgroups at substantially in -
commonly implemented screening regimens. To creased risk of the target cancer(s) are briefly
ensure sufficient regional variation in the reports, described. Available evidence relating to the
low-quality cost–effectiveness analyses may effect of screening in any of these populations
also be reported and considered in the overall using any of the separately considered screening
assessment of cost–effectiveness for regions methods is systematically reviewed and analysed
without high-quality reports. The results do with the same rigour as evidence in whole popu-
not contribute to the overall evaluation of each lations or populations at average risk, and, where
screening method but can be used by govern- possible, a statement is made as to the strength
ments and health services to aid decisions about of the evidence that use of any screening method
implementation of screening for which there is or particular screening method regimen in the
sufficient evidence of a screening effect. group at high risk is more efficacious or effec-
tive than use of any other screening method or
regimen. Where possible, the magnitudes of the
5.5 Comparison of effects of separately
benefits and the harms of the screening method
reviewed screening methods
or regimen in these populations are given.
When two screening methods have been In the absence of such evidence, the Working
established to reduce cancer incidence or cancer Group may critically appraise approaches
mortality, an evaluation may be conducted commonly taken to screening in defined groups
of the comparative efficacy or effectiveness of at high risk and the justifications that have been
these methods. Studies that compare the effects given for them.
of screening of two or more different screening
methods are reviewed and rigorously assessed. 5.7 Other topics reviewed
Where possible, a statement is made as to the
strength of the evidence that use of one screening Some other topics important to the practice
method is more efficacious or effective than use of screening may be reviewed in a Handbook
of another, together with an evaluation of any by summarizing a representative set of studies.
comparative data about additional dimensions, These topics do not contribute to the overall
65
evaluations of the screening methods. They may reductions in incidence and/or mortality in
include, among others: populations adhering to the screening regimen
evaluated are presented. Harms of the screening
(a) Determinants of participation in screening intervention are described, both qualitatively
Given an often large and complex literature, and quantitatively, as the evidence base permits.
a review of reviews of studies in high-income Depending on the amount and relevance of
populations and of individual studies from low- the data, the Working Group may also summa-
and middle-income countries is performed. rize the reviewed evidence for cost–effectiveness,
Special attention is given to the impact on equity and for any other item that the Working Group
of access to effective screening when assessing considers sufficiently important to note.
the role of barriers and the effectiveness of inter-
ventions aimed at promoting participation. 7. Evaluation and rationale
(b) Quality of life Although the following details about the
The results of studies on gain or loss in quality evaluation and rationale refer specifically to
of life of participants in screening programmes screening interventions, they will also apply for
that add useful information on the value of the evaluation of early diagnosis interventions,
screening are reviewed. Only a few studies have with some adaptation as needed.
directly investigated change in quality of life as Consensus evaluations of the strength of
an outcome of screening programmes. These the evidence of a reduction of cancer incidence
estimates can be used in health (economic) and/or cancer mortality (preventive effects) in
assessments as disability weights when esti- humans of each screening method reviewed
mating DALYs, QALYs, and cost–effectiveness. are made using transparent criteria and defined
Although the available quality-of-life studies descriptive terms (see below). Statements should
usually address physical, social, and emotional also be made about the evidence for harms and
functional abilities and general satisfaction, the for the balance of benefits and harms.
assessment of health-related quality of life gained Where the evaluation of several cancer
or lost through screening programmes is chal- screening methods indicates that they can
lenging and is heavily context-dependent. reduce cancer incidence and/or cancer mortality
(Group A; see below), the Working Group may
also choose to indicate whether the efficacy or
6. Summary of data reported effectiveness in reducing cancer incidence and/
Each section or subsection of the Handbook is or cancer mortality and the balance of benefits
summarized. The cancer type subject to screening and harms of one screening method are superior
and its global burden are described, the screening to those of another screening method.
methods evaluated are identified, and their global Similarly, the Working Group may choose
use is briefly presented. The results of epidemio- to evaluate the efficacy or effectiveness of one
logical studies addressing the efficacy, effective- screening method or protocol implemented
ness, and harms of each screening method are in a population at increased risk of the cancer,
also summarized. The overall strengths and depending on whether relevant evidence is
limitations of the epidemiological evidence base available.
are highlighted to indicate how the evaluation The framework for these evaluations, de -
was reached. Typically, the relative and absolute scribed below, may not encompass all factors
relevant to a particular evaluation of preventive
66
efficacy or effectiveness. After considering all (i) The cancer screening method is estab-
relevant scientific findings, the Working Group lished to reduce the incidence of cancer of the
may exceptionally assign the intervention to a [target organ] OR is established to reduce
different category than a strict application of the mortality from cancer of the [target organ]
framework would indicate, while providing a (Group A)
clear rationale for such an evaluation. A causal preventive association between use
The wording of these evaluations is the same of the screening method or screening methods
when inferences about preventive effects are and cancer incidence or mortality has been estab-
made from the results of studies in which an lished. That is, a preventive association has been
intermediate outcome, not cancer incidence and/ observed consistently in the body of evidence
or cancer mortality, was the outcome studied. on use of the screening method or methods and
Such evaluations are made only when a causal cancer incidence or mortality, and chance, bias,
association has been established between the and confounding as explanations for the associ-
intermediate outcome and cancer. A statement ation were ruled out with reasonable confidence.
to this effect is added. When the evidence is classified in Group A,
The evaluation is followed by a description or the evaluation is followed by separate sentences
discussion of harms, with a qualitative and quan- to:
titative overall evaluation considered in the light
of potential and actual harms. • make a statement as to the screening regimen
When there are substantial differences of to which the Working Group considers each
scientific interpretation among the Working evaluation of a screening method applies or
Group members, the overall evaluation will be applies most strongly, and as to whether or
based on the consensus of the Working Group. not the effectiveness of that screening method
A summary of the alternative interpretations has been established;
may be provided, together with their scientific • make a statement of what the Working Group
rationale and an indication of the degree of considers to be the magnitudes of the bene-
support for each. fits and the harms of the screening method,
The evaluation categories refer to the strength in as nearly comparable terms as possible,
of the evidence that an intervention can reduce for people adhering fully to the screening
the incidence of cancer or death from cancer; approach most commonly implemented in
they do not address how strongly or weakly the practice, and whether or not the benefits
intervention reduces cancer incidence and/or outweigh the harms.
cancer mortality, if it can. Put another way, they (ii) The cancer screening method may
do not address the question “By how much might reduce the incidence of cancer of the [target
or does this intervention reduce cancer incidence organ] OR may reduce mortality from cancer
or cancer mortality in exposed people?” of the [target organ] (Group B)
7.1 Evaluation A causal preventive association between use

of the screening method or methods and cancer
On the basis of the principles outlined in incidence or mortality is credible, but chance,
Part B, Section 5, the evidence relevant to cancer bias, or confounding as explanations for the
prevention is classified into one of the following association could not be ruled out with reason-
categories: able confidence; OR a causal preventive associ-
ation between use of the screening method and
67
incidence of precancer or clinically advanced bias, and confounding as explanations for the
cancer has been established in the absence of an null results were ruled out with reasonable confi-
established association for cancer incidence or dence, and the studies were considered informa-
mortality, respectively. tive. Consistent and substantial evidence that the
When the evidence is classified in Group B, screening method does not result in diagnosis
a sentence makes a statement as to the screening that is earlier in the natural history of cancer than
regimen to which the Working Group considers is observed in the absence of screening OR that
each evaluation of a screening method (or of cancer-specific survival of cancers detected by
closely related methods collectively, when eval- screening is no better than that of cancers diag-
uated together) applies or applies most strongly. nosed in the absence of screening also provide
(iii) The cancer screening method is not evidence for lack of cancer prevention from the
classifiable as to its capacity to reduce the screening method.
incidence of cancer of the [target organ] OR A conclusion that the screening method may
to reduce mortality from cancer of the [target lack the capacity to reduce cancer incidence and/
organ] (Group C) or cancer mortality is limited to the screening
method or methods evaluated and the popu-
The available studies are of insufficient lations and life-stages, conditions and levels of
quality, consistency, or statistical precision to screening, and length of observation covered by
enable a conclusion to be drawn about the pres- the available studies. In addition, the possibility
ence or absence of a causal preventive association of a very small preventive effect at the levels of
between the screening method or methods and the intervention studied can never be excluded.
cancer incidence or mortality; OR there is some
evidence that the screening method or methods 7.2 Rationale
has a preventive effect, based on precancer or
clinically advanced cancer as outcomes, but not The reasoning that the Working Group uses
enough to qualify for Group B. The first of the to reach its evaluation is summarized so that the
above conditions includes: (a) there are relevant basis for the evaluation offered is transparent.
studies available, but all are of poor quality or This section includes concise statements of the
informativeness; and (b) there are relevant studies principal lines of argument that emerged in the
available of sufficient quality, but their results are deliberations of the Working Group, the conclu-
inconsistent or otherwise inconclusive. sions of the Working Group on the strength
of the evidence, an indication of the body of
(iv) The cancer screening method may lack evidence that was pivotal to these conclusions,
the capacity to reduce the incidence of cancer of and an explanation of the reasoning of the
the [target organ] OR to reduce mortality from Working Group in making the evaluations.
cancer of the [target organ] (Group D) Where relevant, it also includes reference to use
There are several high-quality studies that are of an intermediate outcome as an, or the, evalu-
mutually consistent in not showing a preventive ation outcome.
association between the screening method or In the rationale, the Working Group may
methods and the studied cancer at the observed draw attention to the fact that the evaluations
levels of use. The results from these studies alone should be interpreted in the light of specific
or combined should have narrow confidence circumstances that vary between countries,
intervals with upper limits above or close to the
null value (e.g. a relative risk of 1.0). Chance,
68
which influence the feasibility of implementa- Greenland S, O’Rourke K (2008). Meta-analysis. In:
tion of programmes based on the interventions Rothman KJ, Greenland S, editors. Modern epidemi-
ology. Philadelphia, PA, USA: Wolters Kluwer Health/
evaluated. Lippincott Williams & Wilkins; pp. 652–82.
Hernán MA, Hernández-Díaz S, Robins JM (2004). A
structural approach to selection bias. Epidemiology.
15(5):615–25. PMID:15308962 doi:10.1097/01.ede.0000
References 135174.63482.43
Hill AB (1965). The environment and disease: asso-
AMSTAR (2017). AMSTAR checklist. Assessing the ciation or causation? Proc R Soc Med. 58:295–300.
methodological quality of systematic reviews. PMID:14283879
Available from: https://amstar.ca/Amstar_Checklist. IARC (1997). Non-steroidal anti-inflammatory drugs.
php, accessed 6 March 2019. IARC Handb Cancer Prev. 1:1–202. Available from:
Axelson O, Steenland K (1988). Indirect methods of https://publications.iarc.fr/371.
assessing the effects of tobacco use in occupational IARC (2002a). Weight control and physical activity. IARC
studies. Am J Ind Med. 13(1):105–18. PMID:3344750 Handb Cancer Prev. 6:1–315. Available from: https://
doi:10.1002/ajim.4700130107 publications.iarc.fr/376.
Boyle P, Levin B, editors (2008). World cancer report IARC (2002b). Breast cancer screening. IARC Handb
2008. Lyon, France: International Agency for Research Cancer Prev. 7:1–229. Available from: https://
on Cancer. Available from: https://publications.iarc. publications.iarc.fr/377.
fr/395. IARC (2005). Use of screening for cervical cancer. In:
Caro JJ, Briggs AH, Siebert U, Kuntz KM; ISPOR-SMDM Cervix cancer screening. IARC Handb Cancer Prev.
Modeling Good Research Practices Task Force (2012). 10:117–62. Available from: https://publications.iarc.
Modeling good research practices – overview: a fr/380.
report of the ISPOR-SMDM Modeling Good Research IARC (2007). Tobacco control: reversal of risk after quit-
Practices Task Force-1. Med Decis Making. 32(5):667– ting smoking. IARC Handb Cancer Prev. 11:1–358.
77. PMID:22990082 doi:10.1177/0272989X12454577 Available from: https://publications.iarc.fr/381.
Cochrane AL (1972). Effectiveness and efficiency: random IARC (2008). Methods for evaluating tobacco control
reflections on health services. Nuffield Trust. Available policies. IARC Handb Cancer Prev. 12:1–300. Available
from: https://www.nuffieldtrust.org.uk/research/ from: https://publications.iarc.fr/382.
effectiveness-and-efficiency-random-reflections-on- IARC (2009). Evaluating the effectiveness of smoke-free
health-services. policies. IARC Handb Cancer Prev. 13:1–334. Available
Cogliano V, Baan R, Straif K, Grosse Y, Secretan B, from: https://publications.iarc.fr/383.
El Ghissassi F, et al. (2005). Transparency in IARC IARC (2011). Effectiveness of tax and price policies for
Monographs. Lancet Oncol. 6(10):747. doi:10.1016/ tobacco control. IARC Handb Cancer Prev. 14:1–366.
S1470-2045(05)70380-6 Available from: https://publications.iarc.fr/384.
Cogliano VJ, Baan RA, Straif K, Grosse Y, Secretan MB, IARC (2016a). Breast cancer screening. IARC Handb
El Ghissassi F, et al. (2004). The science and practice Cancer Prev. 15:1–469. Available from: https://
of carcinogen identification and evaluation. Environ publications.iarc.fr/545.
Health Perspect. 112(13):1269–74. PMID:15345338 IARC (2016b). Study designs to assess the effective-
doi:10.1289/ehp.6950 ness of screening. In: Breast cancer screening. IARC
Cole P, Morrison AS (1980). Basic issues in population Handb Cancer Prev. 15:287–94. Available from: https://
screening for cancer. J Natl Cancer Inst. 64(5):1263–72. publications.iarc.fr/545.
PMID:6767876 IARC (2018). Guidelines for Observers at IARC Hand-
Davies L, Petitti DB, Martin L, Woo M, Lin JS (2018). books Meetings. Available from: https://handbooks.
Defining, estimating, and communicating overdiag- iarc.fr/preambles/guidelines-for-observers/.
nosis in cancer screening. Ann Intern Med. 169(1):36– IARC (2019a). International Classification of Diseases for
43. PMID:29946705 doi:10.7326/M18-0694 Oncology, ICD-O-3 online. Lyon, France: International
Emery JD, Gray V, Walter FM, Cheetham S, Croager EJ, Agency for Research on Cancer. Available from: http://
Slevin T, et al. (2014). The Improving Rural Cancer codes.iarc.fr/usingicdo.php, accessed 8 January 2019.
Outcomes (IRCO) Trial: a factorial cluster-randomised IARC (2019b). WHO Classification of Tumours. Lyon,
controlled trial of a complex intervention to reduce time France: International Agency for Research on Cancer.
to diagnosis in rural patients with cancer in Western Available from: https://whobluebooks.iarc.fr/, accessed
Australia: a study protocol. BMJ Open. 4(9):e006156. 8 January 2019.
PMID:25231496 doi:10.1136/bmjopen-2014-006156
69
Irwig L, Houssami N, Armstrong B, Glasziou P (2006). Sanders GD, Neumann PJ, Basu A, Brock DW, Feeny D,
Evaluating new screening tests for breast cancer. Krahn M, et al. (2016). Recommendations for conduct,
BMJ. 332(7543):678–9. PMID:16565097 doi:10.1136/ methodological practices, and reporting of cost-effec-
bmj.332.7543.678 tiveness analyses: second panel on cost-effectiveness
Miles A, Cockburn J, Smith RA, Wardle J (2004). in health and medicine. JAMA. 316(10):1093–103.
A perspective from countries using organized doi:10.1001/jama.2016.12195 PMID:27623463
screening programs. Cancer. 101(5 Suppl):1201–13. Siemieniuk R, Guyatt G (2019). What is GRADE? BMJ
PMID:15316915 doi:10.1002/cncr.20505 Best Practice. Available from: https://bestpractice.bmj.
Moher D, Liberati A, Tetzlaff J, Altman DG, PRISMA com/info/toolkit/learn-ebm/what-is-grade/, accessed 6
Group (2009). Preferred reporting items for systematic March 2019.
reviews and meta-analyses: the PRISMA statement. Stewart BW, Kleihues P, editors (2003). World cancer
PLoS Med. 6(7):e1000097. PMID:19621072 doi:10.1371/ report 2003. Lyon, France: International Agency
journal.pmed.1000097 for Research on Cancer. Available from: https://
Petitti DB, Lin JS, Owens DK, Croswell JM, Feuer EJ publications.iarc.fr/397.
(2018). Collaborative modeling: experience of the Stewart BW, Wild CP, editors (2014). World cancer report
U.S. Preventive Services Task Force. Am J Prev 2014. Lyon, France: International Agency for Research
Med. 54(1S1):S53–62. PMID:29254526 doi:10.1016/j. on Cancer; Available from: https://publications.iarc.
amepre.2017.07.003 fr/396.
Porta M, editor (2008). A dictionary of epidemiology. Vandenbroucke JP, Broadbent A, Pearce N (2016).
5th ed. New York (NY), USA: Oxford University Causality and causal inference in epidemiology:
Press. Available from: http://www.oxfordreference. the need for a pluralistic approach. Int J Epidemiol.
com/view/10.1093/acref/9780195314496.001.0001/ 45(6):1776–86. PMID:26800751 doi:10.1093/ije/dyv341
acref-9780195314496. WHO (2017). Guide to cancer early diagnosis. Geneva,
Porta M, editor (2014). A dictionary of epidemiology. Switzerland: World Health Organization. Available
6th ed. New York (NY), USA: Oxford University from: https://www.who.int/cancer/publications/
Press. Available from: https://global.oup.com/ cancer_early_diagnosis/en/.
academic/product/a-dictionary-of-epidemiology- Wild CP, Cogliano VJ (2011). A plea on behalf of expert
9780199976737?cc=fr&lang=en&. evaluation and the experts involved. Int J Epidemiol.
Richardson DB, Laurier D, Schubauer-Berigan MK, 40(1):253–61. PMID:20223796 doi:10.1093/ije/dyq038
Tchetgen Tchetgen E, Cole SR (2014). Assessment and Young GP, Senore C, Mandel JS, Allison JE, Atkin WS,
indirect adjustment for confounding by smoking in Benamouzig R, et al. (2016). Recommendations for
cohort studies using relative hazards models. Am J a step-wise comparative approach to the evaluation
Epidemiol. 180(9):933–40. PMID:25245043 doi:10.1093/ of new screening tests for colorectal cancer. Cancer.
aje/kwu211 122(6):826–39. PMID:26828588 doi:10.1002/cncr.29865
Rothman KJ, Greenland S, Lash TL (2008). Causation
and causal inference. In: Modern epidemiology.
Philadelphia (PA), USA: Wolters Kluwer Health/
Lippincott Williams & Wilkins; pp. 5–31.
70
GENERAL REMARKS
In April 2016, an international group of sci- History of use of smokeless tobacco

entific and public health leaders from 21 coun- and areca nut
tries met in Kuala Lumpur, Malaysia, at the In-
ternational Conference on Betel Quid and Areca
Smokeless tobacco was originally used by
Nut, to identify research needs and discuss strat- the Indigenous populations of South America
egies to reduce the prevalence of use of betel for various purposes, including in spiritual cere-
quid and areca nut, and thereby the incidence of monies, as a poultice with potential therapeutic
oral cancers related to their use (Mehrtash et al., effects, and as an exchangeable good (Shafey
2017). Several members of the panel suggested et al., 2009). In the 15th century, Columbus
that IARC – in particular, the IARC Handbooks brought tobacco from his transatlantic expe-
of Cancer Prevention – perform a review and an ditions to Europe (Shafey et al., 2009). By the
evaluation of the current scientific literature on mid-16th century, adventurers and diplomats
several aspects of oral cancer prevention, with a such as Nicot began to popularize its use, with
specific focus on South-East Asia. the introduction of tobacco in France in 1556,
Volume 19 of the IARC Handbooks of Can- Portugal in 1558, Spain in 1559, and England in
cer Prevention series provides a first-time evalu- 1565 (CNN.com, 2000). In India, tobacco was
ation of primary and secondary prevention ap- introduced by the Portuguese in the 17th century
proaches for oral cancer. The expert knowledge (Prasad, 2007). Despite changes in consump-
summarized in this Handbook will play a major tion patterns with the global spread of tobacco
role as a resource for policy-makers involved in through the centuries (from smokeless tobacco
the regulation of smokeless tobacco and will help in the 18th century to cigars in the 19th century
fight a major public health problem. In addition, and cigarettes in the 20th century), tobacco is
this Handbook aligns with the World Health Or- still currently used in a smokeless form (Shafey
ganization (WHO) mission of tobacco control, et al., 2009).
including the WHO Framework Convention on The chewing of areca nut is deeply embedded
Tobacco Control (FCTC), and with IARC’s mis- in the sociocultural history of South-East Asia
sion of cancer research for cancer prevention and many parts of the Western Pacific. It is
with a particular focus on low- and middle-in- mentioned in the Mahāvamsa, a historical chron-
come countries. icle of Sri Lanka written in 504 BCE (Krenger,
1942), and in ancient writings of Hinduism
in about 600 BCE (Bhishagratna, 1907) and in
71
writings of mainland China during the Tang between outcomes by type of smokeless tobacco
dynasty (7th to 9th centuries CE). Cultivation and includes both oral snuff and chewing tobacco
of the Areca catechu palm tree has been wide- products.
spread across South-East Asia and South Asia for
millennia, initially in the Philippines and then
gradually spreading across the Western Pacific Distinguishing between products
Islands. that contain smokeless tobacco
or areca nut or both
Distinguishing between smokeless
Smokeless tobacco and areca nut may be used
tobacco products either on their own or in combination. Thus,
the resultant products may be categorized as
Oral smokeless tobacco products are tradi- containing only smokeless tobacco, only areca
tionally sold in various forms, but they can be nut, or both smokeless tobacco and areca nut.
broadly categorized as snuff (powdered or ground However, a lack of clarity in reporting (Theilmann
tobacco) or chewing tobacco (leaf, plug, or twist) et al., 2022) the product categories has been
(Stanfill et al., 2011). Smokeless tobacco products observed in many studies, possibly leading to
may contain different concentrations of tobac- inappropriate interpretation and evaluation
co-specific N′-nitrosamines, depending on their of the evidence. For example, in many studies,
preparation and processing (NCI and CDC, 2014). products that contain both smokeless tobacco
N′-nitrosonornicotine (NNN), one of the most and areca nut are reported as smokeless tobacco
abundant tobacco-specific N′-nitrosamines, is (or chewing tobacco). In other instances, the
formed by N-nitrosation of tobacco alkaloids, presence or absence of tobacco in the betel quid –
particularly during tobacco curing; NNN is clas- a preparation that always contains areca nut – is
sified by the IARC Monographs as carcinogenic not specified. In both cases, identification of the
to humans (Group 1) (IARC, 2007; Ammann specific product(s) relies heavily on knowledge
et al., 2016). The water content of smokeless of the practices in the region(s) where the study
tobacco products also varies; the water content was conducted (Gupta and Warnakulasuriya,
of chewing tobacco (which ranges between 7% 2002). Hence, it is essential that either the correct
and 21%) is lower than that of moist snuff but product category (as mentioned above) is speci-
higher than that of dry snuff (IARC, 2007). The fied or the specific products are clearly listed in
variation in water content under the same manu- the study details.
facturing and storage conditions has the poten-
tial to influence the level of NNN.
Despite the differences between smokeless Differences in cessation interven-
tobacco products and the potential differences
in toxicant exposure and carcinogenic poten-
tions among youth and adults
tial, most epidemiological studies do not distin-
guish between smokeless tobacco products, The impact of an intervention to quit use of
and this makes it difficult to evaluate cancer smokeless tobacco or areca nut on adults and
risks by product type. For the same reason, youth differs because of age, perception of health
the Working Group’s evaluation of smokeless risks associated with tobacco use, and the impact
tobacco (without areca nut) does not distinguish of tobacco advertising. Initiation of use is mainly
at ages 13–14 years, and most of the initiation
72
General remarks
happens before age 18 years. There is a paucity of Research gaps in oral cancer
data in the literature on the efficacy of interven- prevention
tions in preventing initiation of use. Also, young
people generally do not perceive that tobacco
Globally, research efforts for oral cancer
kills or causes serious diseases. Hence, interven-
trail behind those for most other common
tions based on communicating long-term health
cancer types. A comprehensive, well-funded
risks may not be salient to adolescents. However,
research strategy is needed to assess changes in
attitudes and behaviours regarding tobacco use
the patterns of oral potentially malignant disor-
among youth are influenced by advertising in
ders (OPMDs) and cancer, especially in parts
any form. Although performing such an evalu-
of the world where the burden of oral cancer
ation was outside the scope of this Handbook,
is increasing. The reasons for the geographical
mass media anti-tobacco advertisements in the
variations in the incidence of oral cancer must
form of audiovisual spots, radio spots, print
be better understood, with a focus on dispar-
media, and educational awareness campaigns
ities in the socioeconomic status of the global
can be effective in promoting cessation.
population and lifestyle habits related to use of
tobacco, use of areca nut, and alcohol consump-
tion. Research on the association between use of
Framework of evaluation of primary smokeless tobacco or areca nut and oral cancers
prevention interventions in lower-middle-income countries is lacking. In
addition, new prevention and cessation interven-
The impact of preventive interventions on tion models need to be developed and tested at
risk behaviours may take more than a decade to the population and individual levels.
produce any significant beneficial effect on the The Working Group also identified the
future incidence of cancer. For this reason, it is following in high-risk populations: a lack of
necessary to monitor intermediate outcomes to knowledge of signs, symptoms, and risk factors
assess the benefits of quitting risk habits. For for oral cancer; inconsistencies in the assess-
this Handbook, IARC used the new Preamble, ment of risk behaviours; and gaps in the tech-
developed in 2019 (IARC, 2019), with a two-step nical practice of clinical oral examination. With
evaluation: step 1 to assess whether a community respect to screening, research gaps include the
programme or an intervention directed at an following: identifying approaches in the popu-
individual leads to cessation of use of smokeless lation to encourage and sustain participation
tobacco and/or areca nut, and step 2 to evaluate among the hard-to-reach, high-risk population;
whether quitting an exposure leads to a reduction assessing the cost–effectiveness of standard clin-
in oral cancer incidence or mortality. However, ical oral examination as a screening approach for
as mentioned above, the lack of consistency in oral cancer compared with the addition of new
the terminology that has been used in the liter- point-of-care diagnostics; and improving the
ature to describe the products prevented the overall 5-year survival rate, which remains about
Working Group from making a full evaluation, 50%. Finally, an increase in research efforts to
from intervention to cancer outcome, for any control the use of smokeless tobacco products,
product category or specific product. and strict implementation of the WHO FCTC
recommendations, are desirable goals to reduce
the global burden of oral cancer.
73
References nut. Lancet Oncol. 18(12):e767–75. doi:10.1016/S1470-

2045(17)30460-6 PMID:29208442
NCI and CDC (2014). Smokeless tobacco and public
Ammann JR, Lovejoy KS, Walters MJ, Holman MR (2016). health: a global perspective (NIH Publication No.
A survey of N′-nitrosonornicotine (NNN) and total 14-7983). Bethesda (MD), USA: U.S. Department of
water content in select smokeless tobacco products Health and Human Services, Centers for Disease
purchased in the United States in 2015. J Agric Food Control and Prevention and National Institutes
Chem. 64(21):4400–6. doi:10.1021/acs.jafc.6b00922 of Health, National Cancer Institute. Available
PMID:27192054 from: https://cancercontrol.cancer.gov/brp/tcrb/
Bhishagratna KKL, editor (1907). An English translation smokeless-tobacco.
of the Sushruta Samhita, based on the original Sanskrit Prasad VM (2007). Case study of tobacco cultivation
text. Calcutta, India: SL Bhaduri. and alternate crops in India. Study conducted as
CNN.com (2000). A brief history of tobacco. Available a technical document for the first meeting of the
from: http://edition.cnn.com/US/9705/tobacco/ Ad Hoc Study Group on Alternative Crops estab-
history/. lished by the Conference of the Parties to the WHO
Gupta PC, Warnakulasuriya S (2002). Global epide- Framework Convention on Tobacco Control, February
miology of areca nut usage. Addict Biol. 7(1):77–83. 2007. Available from: https://www.who.int/docs/
doi:10.1080/13556210020091437 PMID:11900626 default-source/searo/india/tobacoo/india-case-study.
IARC (2007). Smokeless tobacco and some tobacco-spe- pdf?sfvrsn=143f1fae_2.
cific N-nitrosamines. IARC Monogr Eval Carcinog Risks Shafey O, Eriksen M, Ross H, Mackey J (2009). The
Hum. 89:1–626. Available from: https://publications. tobacco atlas. 3rd ed. Atlanta (GA), USA: American
iarc.fr/107 PMID:18335640 Cancer Society.
IARC (2019). IARC Handbooks of Cancer Prevention. Stanfill SB, Connolly GN, Zhang L, Jia LT, Henningfield
Preamble for primary prevention. Lyon, France: JE, Richter P, et al. (2011). Global surveillance of oral
International Agency for Research on Cancer. tobacco products: total nicotine, unionised nicotine
Available from: https://handbooks.iarc.fr/documents- and tobacco-specific N-nitrosamines. Tob Control.
handbooks/hb-preamble-primary-prevention.pdf. 20(3):e2. doi:10.1136/tc.2010.037465 PMID:21109685
Krenger W (1942). History and culture of betel chewing. Theilmann M, Lemp JM, Winkler V, Manne-Goehler J,
Synthesis and preparation of betel. [in German] Ciba Marcus ME, Probst C, et al. (2022). Patterns of tobacco
Z. 84:2922–41. use in low and middle income countries by tobacco
Mehrtash H, Duncan K, Parascandola M, David A, product and sociodemographic characteristics:
Gritz ER, Gupta PC, et al. (2017). Defining a global nationally representative survey data from 82 coun-
research and policy agenda for betel quid and areca tries. BMJ. 378:e067582. doi:10.1136/bmj-2021-067582
PMID:36041745
74
LIST OF ABBREVIATIONS
ADH alcohol dehydrogenase

AI artificial intelligence
AJCC American Joint Committee on Cancer
ALDH aldehyde dehydrogenase
ARCAGE Alcohol-Related Cancers and Genetic Susceptibility in Europe
ASIR age-standardized incidence rate
ASMR age-standardized mortality rate
AUC area under the curve
BMI body mass index
CI confidence interval
COE clinical oral examination
DOI depth of invasion
EGFR epidermal growth factor receptor
ENE extranodal extension
EPIC European Prospective Investigation into Cancer and Nutrition
EU European Union
FAD flavin adenine dinucleotide
FCTC Framework Convention on Tobacco Control
FVL fluorescence visualization loss
FVR fluorescence visualization retained
GATS Global Adult Tobacco Survey
GSPS Global School Personnel Survey
GYTS Global Youth Tobacco Survey
H&E haematoxylin and eosin
HDI Human Development Index
HNC head and neck cancer
HPV human papillomavirus
HR hazard ratio
HRME high-resolution microendoscope
IARC International Agency for Research on Cancer
ICARE Investigation of Occupational and Environmental Causes of Respiratory Cancers
ICD-O International Classification of Diseases for Oncology
IL-6 interleukin 6
INHANCE International Head and Neck Cancer Epidemiology
75
LOH loss of heterozygosity

miRNA microRNA
MMP-1 matrix metalloproteinase-1
mRNA messenger RNA
MSE mouth self-examination
NBI narrow-band imaging
NIH-AARP United States National Institutes of Health-American Association of Retired Persons
OCT optical coherence tomography
OED oral epithelial dysplasia
OLP oral lichen planus
OPMDs oral potentially malignant disorders
OR odds ratio
OSCC oral squamous cell carcinoma
OSF oral submucous fibrosis
PAF population attributable fraction
PAS periodic acid–Schiff
PVL proliferative verrucous leukoplakia
RCT randomized controlled trial
RERI relative excess risk due to interaction
ROC receiver operating characteristic
RR relative risk
SCC squamous cell carcinoma
SEER Surveillance, Epidemiology, and End Results
SLT smokeless tobacco
STEPS WHO STEPwise Approach to Surveillance
TNF-α tumour necrosis factor α
TNM tumour–node–metastasis
UICC Union for International Cancer Control
WCRF World Cancer Research Fund
WHO World Health Organization
76
1. ORAL CANCER AND ORAL POTENTIALLY
MALIGNANT DISORDERS
1.1 Anatomy of the oral cavity and (a) Lips

the oropharynx The lips surround the oral aperture, marking
the external boundary of the mouth, and are
1.1.1 Anatomy of the oral cavity used for speech, mastication, swallowing, and
controlling the size of the oral aperture.
The oral cavity is the entrance to the gastroin- The upper lip extends from the inferior
testinal tract. It is bounded anteriorly by the lips, border of the nose and laterally up to the naso-
posteriorly by the faucial arches anterior to the labial grooves. The lower lip is bordered by the
tonsils, laterally by the cheeks (buccal mucosae), labiomental groove. The upper and lower lips
superiorly by the palate, and inferiorly by the meet at the oral commissures.
muscular floor. The space between the labial The lips are composed of a muscle that
mucosae of the lips or the buccal mucosae of the is covered externally by skin and internally
cheeks and the teeth is defined as the oral vesti- by non-keratinized labial mucosa, with the
bule (labial or buccal vestibule) (Fig. 1.1). vermilion zone in between. The vermilion zone
The oral mucous membrane is covered by the has keratinized epithelium and is highly vascu-
stratified squamous epithelium, which comprises larized and densely innervated. Numerous
four different layers and protects the inside of the mucous glands are present on the labial mucosa.
oral cavity. The oral mucosa is subdivided into The lymphatic drainage of the lips is primarily to
masticatory mucosa (keratinized), lining mucosa the submandibular lymph nodes and to a lesser
(non-keratinized), and specialized mucosa. The extent to the submental, intraparotid, or internal
mucosa has the capacity to undergo constant jugular lymph nodes (Sinnatamby, 2011; Nanci,
regeneration (Sinnatamby, 2011; Nanci, 2017; 2017; Berkovitz et al., 2018; Standring, 2020).
Berkovitz et al., 2018; Standring, 2020).
The subsites of the oral cavity are described (b) Tongue
below according to the International Clas- The tongue is a muscular organ that occu-
sification of Diseases for Oncology (ICD-O) pies the floor of the mouth. It is attached to the
coding: lip (C00), tongue (C02), gingivae (C03), mandible (lower jawbone) and the hyoid bone by
floor of the mouth (C04), palate (C05), and buccal the root. The anterior two thirds of the structure
mucosa and oral commissures (C06). is free to move.
The dorsum is divided into an anterior oral
part and a posterior pharyngeal part, which
77
IARC HANDBOOKS OF CANCER PREVENTION - 19
Fig. 1.1 Anatomy of the oral cavity
Superior lip
Superior labial
frenulum
Palatine
raphe
Teeth
Hard palate
Soft palate Oropharynx
Uvula
Tongue
Lingual
frenulum
Gingivae
(gums)
Inferior labial
Inferior lip
frenulum
© AdobeStock.com/Ilusmedical
forms the base of the tongue. The anterior two mastication, swallowing, speech, oral cleansing,
thirds of the dorsum is covered by specialized and taste.
mucosa that contains numerous papillae, some of The tongue is highly vascularized; lingual
which bear taste buds. The posterior third slopes veins are visible on the inferior surface. The
down towards the epiglottis and has a nodular anterior two thirds of the tongue drains into
appearance because of the underlying lingual the submental and submandibular nodes, which
tonsils. Two other important anatomical areas of empty into the deep cervical lymph nodes. The
the tongue are the lateral borders and the ventral posterior third of the tongue drains directly into
surface (undersurface) of the tongue. the deep cervical lymph nodes. The lymphatic
The extrinsic muscles stabilize and move drainage is significant because some areas of the
the tongue, and the intrinsic muscles maintain tongue drain into bilateral cervical lymph nodes
its shape. The main functions of the tongue are (Sinnatamby, 2011; Nanci, 2017; Berkovitz et al.,
2018; Standring, 2020).
78
Oral cancer prevention
(c) Gingivae (gums) (e) Hard palate

The gingiva is the highly keratinized mucosa The palate forms the roof of the oral cavity and
that immediately surrounds the neck of an consists of two parts: the hard palate and the soft
erupted tooth and is firmly attached to the alve- palate. The hard palate is part of the oral cavity,
olar margins of the jaws. At the gingival crest, and the soft palate is part of the oropharynx.
the epithelium slopes towards the tooth to form The palatine processes of the maxillae (upper
a sulcus and is attached to the tooth surface. jawbone) and the horizontal plates of the palatine
The part of the gingiva facing the oral cavity is bones form the hard palate, which is bounded
masticatory mucosa, and the change from alve- anteriorly by the maxillary teeth (upper teeth)
olar mucosa to gingival mucosa is identifiable by and continues posteriorly to the soft palate. The
an abrupt colour change of the tissue. Healthy palatine raphe extends anteroposteriorly in the
gingiva has some stippling on its surface. The midline, and an irregular set of rugae radiates
lymphatic drainage of the gingivae is to the from it in the anterior part of the hard palate.
submandibular lymph nodes. In addition, the The incisive papilla lies at the anterior end of
lower anterior gingiva drains into the submental the hard palate, which contains the opening of
nodes (Sinnatamby, 2011; Nanci, 2017; Berkovitz the incisive canal. The hard palate has a thick
et al., 2018; Standring, 2020). keratinized mucosa, which is tightly bound to
the periosteum anteriorly and contains minor
(d) Floor of the mouth salivary glands in the posterior submucosa.
The floor of the mouth is a horseshoe-shaped The lymphatic drainage is primarily to the deep
region situated between the movable part of the cervical lymph nodes (Sinnatamby, 2011; Nanci,
tongue and the mylohyoid muscles. The lingual 2017; Berkovitz et al., 2018; Standring, 2020).
frenulum is a mucosal fold that arises near the
base of the tongue and extends onto the infe- (f) Buccal mucosa
rior surface of the tongue. The protuberance The buccal mucosa is the mucosal lining of
at the anterior floor of the mouth is called the the cheeks, extending from the line of contact
sublingual papilla or caruncle; this is where the of the opposing lips to the pterygomandibular
submandibular salivary ducts open into the oral raphe. It extends to the line of attachment of the
cavity. On either side laterally and backward are alveolar mucosa superiorly and inferiorly, which
the sublingual folds, which cover the subman- forms the anterolateral boundary of the oral
dibular ducts and the sublingual salivary glands. vestibule. The buccal mucosa has a non-kerati-
The covering epithelium is non-keratinized and nized epithelium and is firmly attached to the
is much thinner than for other subsites of the underlying muscle. The submucosa contains
oral cavity. The lymphatic drainage is mainly minor salivary glands. A white line coinciding
to the submandibular lymph nodes; the ante- with the occlusal plane, called the linea alba, may
rior part drains into the submental nodes be present. The parotid ducts of the parotid gland
(Sinnatamby, 2011; Nanci, 2017; Berkovitz et al., on either side pierce the buccal mucosa opposite
2018; Standring, 2020). the second maxillary molar tooth and present
as the parotid papillae. The lymphatic drainage
is primarily to the submandibular lymph nodes
(Sinnatamby, 2011; Nanci, 2017; Berkovitz et al.,
2018; Standring, 2020).
79
Fig. 1.2 Anatomy of the pharynx
Nasopharynx
Oropharynx
Pharynx
Hypopharynx Larynx
Oral cavity
Trachea
© Nicolas Primola/Shutterstock.com
1.1.2 Anatomy of the oropharynx and the constrictor muscles. The palatine tonsils project
soft palate into the lateral wall of the oropharynx from the
tonsillar fossa. The oropharynx is covered by a
The oropharynx is a tube-shaped fibromus- non-keratinized stratified squamous epithelium
cular structure behind the oral cavity, contin- (Sinnatamby, 2011; Nanci, 2017; Berkovitz et al.,
uous with the nasopharynx superiorly and the 2018; Standring, 2020).
hypopharynx inferiorly (Fig. 1.2). The oropharynx The soft palate is a mobile flap that extends
has functional roles in both the respiratory backward from the hard palate and fuses with
system and the digestive system. It extends from the lateral wall of the oropharynx. It is made
the lower surface of the soft palate to the upper up of five paired muscles and an aponeurosis.
border of the epiglottis and communicates The soft palate can be raised to make contact
with the oral cavity anteriorly. The posterior with the posterior wall of the oropharynx to
wall of the oropharynx is formed by the three close off the nasopharynx during swallowing.
80
The non-keratinized mucosa covers the oral (b) Geographical variations in incidence and
side and the posterior part of the nasal side, mortality
and the respiratory mucosa covers the anterior In 2020, the incidence rates of oral cancer
part of the nasal side. The submucosa of both (including lip cancer) were highest in Melanesia
surfaces contains mucous glands and taste buds. and South Asia (Ferlay et al., 2020; Miranda-
Lymphoid follicles are scattered on the oral Filho and Bray, 2020). The rates (ASIR, per
surface. The uvula hangs down at the midline 100 000, in both sexes) were highest in Papua
of the posterior end of the soft palate and helps New Guinea (21.2), followed by Pakistan (10.1),
in phonation (Sinnatamby, 2011; Nanci, 2017; India (9.8), Sri Lanka (9.7), and Bangladesh (9.5)
Berkovitz et al., 2018; Standring, 2020). (Ferlay et al., 2020). For oral cancer mortality
rates, the pattern was similar. The rates (ASMR,
1.2 Global burden of oral cancer, per 100 000, in both sexes) were highest in Papua
oropharyngeal cancer, and oral New Guinea (8.3), followed by Pakistan (6.4),
Bangladesh (5.6), India (5.4), and Sri Lanka (4.5)
potentially malignant disorders (Ferlay et al., 2020) (Fig. 1.4).
1.2.1 Oral cancer and oropharyngeal cancer Both ASIR and ASMR were consistently
higher in men than in women across the world
(a) Global incidence and mortality (Ferlay et al., 2020; Miranda-Filho and Bray,
Oral cancer, along with oropharyngeal 2020).
cancer, is among the most common cancer types In 2020, the incidence rates of oropharyn-
globally. geal cancer were highest in Europe. The rates
In 2020, there were an estimated 377 713 (ASIR, per 100 000, in both sexes) were highest in
new cases of oral cancer worldwide, with global Denmark (5.0), France (4.3), and Romania (4.3).
age-standardized incidence rates (ASIR) of 6.0 For oropharyngeal cancer mortality, the rates
per 100 000 men and 2.3 per 100 000 women. (ASMR, per 100 000, in both sexes) were highest
There were an estimated 177 757 deaths from oral in Slovakia (2.5), followed by the Republic of
cancer, with global age-standardized mortality Moldova (2.3) and Romania (2.3) (Ferlay et al.,
rates (ASMR) of 2.8 per 100 000 men and 1 per 2020).
100 000 women.
In 2020, there were an estimated 98 412 new (c) Socioeconomic status
cases of oropharyngeal cancer worldwide, with For oral cancer (including lip cancer), the
global ASIR of 1.8 per 100 000 men and 0.4 per ASIR and ASMR were highest in countries with
100 000 women. There were an estimated 48 143 medium levels of the Human Development
deaths from oropharyngeal cancer, with global Index (HDI); for oropharyngeal cancer, the
ASMR of 0.89 per 100 000 men and 0.17 per ASIR was highest in countries with very high
100 000 women (Ferlay et al., 2020). HDI, and the ASMR was highest in countries
For both oral cancer and oropharyngeal with medium HDI (Ferlay et al., 2020) (Fig. 1.5).
cancer, the incidence increases with age (Fig. 1.3). HDI was found to be negatively associated with
the annual percentage change in the ASIR and
ASMR for oral cancer (Ren et al., 2020).
A meta-analysis of 41 case–control studies
revealed that low socioeconomic status increased
the risk of oral cancer (pooled adjusted odds
81
Fig. 1.3 Age-specific incidence curves in the world population for oral cancer (A) and
oropharyngeal cancer (B), 2020
From Ferlay et al. (2020).
ratio [OR], 3.41; 95% confidence interval [CI], of age, sex, centre, cigarette smoking, and alcohol
2.14–5.44; n = 2), as did low occupational consumption (Conway et al., 2015).
social class (pooled adjusted OR, 1.41; 95% CI,
1.10–1.79; n = 4) and low educational attainment (d) Time trends in incidence
(pooled adjusted OR, 1.74; 95% CI, 1.33–2.27; During the past two decades, incidence rates
n = 17) (Conway et al., 2008). A large pooled of oral and oropharyngeal cancers combined
analysis by the International Head and Neck have decreased in several countries in North
Cancer Epidemiology (INHANCE) consortium America, South-East Asia, and Europe, espe-
found an association between low educational cially in males. However, in females, incidence
attainment and increased risk of oral cancer rates have increased mainly in the European
(OR, 1.33; 95% CI, 1.02–1.75) and oropharyngeal countries, and in males, incidence rates have
cancer (OR, 1.88; 95% CI, 1.23–2.88), independent increased in the United Kingdom, Japan, and
82
Fig. 1.4 Global distribution of estimated age-standardized (World) incidence rates (A) and
mortality rates (B) per 100 000 for oral cancer in both sexes, 2020
83
Fig. 1.5 Estimated age-standardized (World) incidence and mortality rates per 100 000 for oral
cancer (A) and oropharyngeal cancer (B), by Human Development Index (HDI) level, 2020
84
Fig. 1.6 Time trends in age-standardized (World) incidence rates per 100 000 for oral cancer and
oropharyngeal cancer combined, by country, in males (red) and females (green)
85
Fig. 1.6 (continued)
Czechia (Fig. 1.6; Bosetti et al., 2020; Ferlay et al., India (Fig. 1.7; Miranda-Filho and Bray, 2020).
2020; Lin, 2020). Incidence rates of oropharyn- Incidence rates of tongue cancer have increased
geal cancer specifically have increased in several in the USA and Thailand (Argirion et al., 2019;
countries in the Americas, Europe, and Asia, Kim and Kim, 2020).
especially in males (Bosetti et al., 2020; Menezes
et al., 2021). (e) Projections of incidence and mortality
In most countries, incidence rates of can- Table 1.1 shows estimates of the incidence
cer of the oral cavity (excluding the lip and and mortality for oral cancer and oropharyn-
the tongue) have decreased more in males geal cancer in 2020 and projected to 2040, by
than in females. In females, incidence rates HDI category and overall. Globally, the projected
have decreased in Thailand, Colombia, and increase from 2020 to 2040 in the estimated
86
Fig. 1.7 Estimated annual percentage change (EAPC) of the trends in age-standardized rates of
mouth cancer in selected registry populations by sex, in 1998–2012, sorted in descending order
according to EAPC in men
* National
** Regional
Reprinted from Miranda-Filho and Bray (2020), Copyright 2020, with permission from Elsevier.
number of new cases per year is 49.6% for oral of excessive exposure to ultraviolet radiation
cancer and 40.2% for oropharyngeal cancer. For from sunlight. The incidence rates are highest
both oral cancer and oropharyngeal cancer, the in Australia in both sexes, followed by Spain
highest increases by 2040 in the numbers of new and Poland in males and the Netherlands and
cases and deaths are expected to occur in coun- Norway in females (Fig. 1.8A). Although inci-
tries with low HDI (Table 1.1) (Ferlay et al., 2020). dence rates of lip cancer have decreased in most
countries, incidence rates have increased in
(f) Lip cancer females in Germany, the Netherlands, Norway,
Incidence rates of lip cancer are relatively China, Slovakia, and Japan and in males in
high in certain parts of the world as a result India (Fig. 1.8B; Miranda-Filho and Bray, 2020).
87
Table 1.1 Global burden of oral cancer and oropharyngeal cancer: estimated annual numbers of
incident cases and deaths, by HDI category and overall, in 2020 and projected to 2040
HDI categorya Population in 2020 Number of new cases Increase Number of deaths Increase
(millions) (%) (%)
2020 2040 2020 2040
Oral cancer
Very high HDI 1 564 118 036 147 172 24.7 37 048 48 590 31.2
High HDI 2 909 72 418 112 182 54.9 34 765 57 958 66.7
Medium HDI 2 327 177 018 285 228 61.1 99 662 161 437 62.0
Low HDI 990 10 126 20 163 99.1 6 251 12 554 100.8
World 7 791 377 598 564 745 49.6 177 726 280 539 57.8
Oropharyngeal cancer
Very high HDI 1 564 47 971 56 233 17.2 18 592 23 522 26.5
High HDI 1 564 20 614 30 097 46.0 11 248 17 532 55.9
Medium HDI 2 327 27 932 47 869 71.4 17 053 29 241 71.5
Low HDI 990 1 839 3 727 102.7 1 230 2 510 104.1
World 7 791 98 356 137 926 40.2 48 123 72 805 51.3
HDI, Human Development Index.
a The four tiers of HDI are: low (< 0.55), medium (≥ 0.55 to < 0.7), high (≥ 0.7 to < 0.8), and very high (≥ 0.8).
Created using data from Ferlay et al. (2020).
1.2.2 Oral potentially malignant disorders CI, 2.91–4.67%), and Europe (3.07%; 95% CI,
1.64–4.93%) (Mello et al., 2018).
An oral potentially malignant disorder Globally, the highest prevalence is observed
(OPMD) is defined as any oral mucosal for oral submucous fibrosis (4.96%; 95% CI,
abnormality that is associated with a statisti- 2.28–8.62%). Other common OPMDs include
cally increased risk of developing oral cancer leukoplakia (4.11%; 95% CI, 1.98–6.97%), actinic
(Warnakulasuriya et al., 2007). OPMDs share cheilitis (2.08%; 95% CI, 0.94–3.67%), erythro-
common risk factors with invasive carcinoma plakia (0.17%; 95% CI, 0.07–0.32%) (Mello et al.,
of the oral cavity. OPMDs include leukoplakia, 2018), and oral lichen planus (1.01%; 95% CI,
erythroplakia, oral submucous fibrosis, oral 0.74–1.32%) (González-Moles et al., 2021).
lichen planus, actinic keratosis (actinic cheilitis),
palatal lesion in reverse smokers (in reverse
smoking, the smoker places the lit end of the 1.3 Oral neoplasia
cigarette, rather than the unlit end, into their
mouth and inhales the smoke), oral lupus erythe- 1.3.1 Classification and natural history of
matosus, dyskeratosis congenita, oral lichenoid OPMDs and oral cancer
lesion, and oral graft-versus-host disease (Warna- Oral cancer includes cancers of the lip, other
kulasuriya and Greenspan, 2020). and unspecified parts of the tongue (excluding
The overall global prevalence of OPMDs is the lingual tonsils), gum, floor of the mouth,
4.47% (95% CI, 2.43–7.08%), with geographical palate, and other and unspecified parts of the
variations; the highest prevalence is observed in mouth (Conway et al., 2018).
Asia (10.54%; 95% CI, 4.60–18.55%), followed by The term OPMD was introduced in 2005,
South America and the Caribbean (3.93%; 95% replacing the terms “oral precancerous/premalig-
CI, 2.43–5.77%), the Middle East (3.72%; 95% nant lesions and conditions” (Warnakulasuriya
88
Fig. 1.8 (A) Bar chart of age-standardized incidence rates of lip cancer in selected countries, by
sex, all ages, in 2008–2012. (B) Estimated annual percentage change (EAPC) of the trends in age-
standardized rates of lip cancer in selected registry populations by sex, in 1998–2012, sorted in
descending order according to EAPC in men
A B
* National registry Age-standardized rate (world) per 100 000 * National registry
** Regional registries ** Regional registries
Reprinted from Miranda-Filho and Bray (2020), Copyright 2020, with permission from Elsevier.
et al., 2007). OPMDs comprise a wide range of Araújo et al., 2014; González-Moles et al., 2019;
disorders (Box 1.1) with varying rates of malig- Warnakulasuriya and Greenspan, 2020). The
nant transformation into oral cancer, of which diagnosis of an OPMD significantly increases
oral squamous cell carcinoma (OSCC) is the most the risk of developing oral cancer during a life-
common type (Holmstrup et al., 2006; Speight time (Warnakulasuriya et al., 2007; Reibel et al.,
et al., 2018; Farah et al., 2019). In 2020, the World 2017; Speight et al., 2018; Warnakulasuriya and
Health Organization (WHO) Collaborating Greenspan, 2020).
Centre for Oral Cancer recommended a list of
OPMDs, which include leukoplakia, prolifera- (a) Clinical presentation of OPMDs
tive verrucous leukoplakia (PVL), erythroplakia, Leukoplakia is a predominantly white plaque
oral submucous fibrosis, oral lichen planus of questionable risk having excluded other
(OLP), actinic keratosis (actinic cheilitis), nico- known diseases that carry no increased risk
tinic stomatitis in reverse smokers, oral lupus of cancer (Warnakulasuriya et al., 2007). This
erythematosus, and dyskeratosis congenita. Oral OPMD has a wide range of clinical appear-
lichenoid lesion and oral graft-versus-host dis- ances, ranging from homogeneous to non-ho-
ease were added to the list on the basis of the avail- mogeneous, including nodular leukoplakia,
able evidence on their malignant potential (de verrucous leukoplakia, and erythroleukoplakia
89
Box 1.1 Classified oral potentially malignant disorders (OPMDs)

Erythroplakia
Erythroleukoplakia
Leukoplakia
Proliferative verrucous leukoplakia
Oral submucous fibrosis
Palatal lesion associated with reverse smoking
Oral lichenoid lesiona
Oral lichen planus
Actinic keratosis (actinic cheilitis)
Smokeless tobacco keratosisb
Oral graft-versus-host disease
Oral lupus erythematosus
Familial cancer syndromes including Fanconi anaemia, dyskeratosis congenita, xeroderma pigmentosum, Li–Fraumeni
syndrome, Bloom syndrome, ataxia–telangiectasia, and Cowden syndrome
aOral lesion resembling lichen planus but lacking typical clinical or histopathological appearances.
bRisk varies with tobacco type.
Adapted from WHO Classification of Tumours Editorial Board (2023).
(Warnakulasuriya and Greenspan, 2020). Several migrans, desquamative gingivitis, erosive OLP,
other white lesions should be excluded to arrive oral lupus erythematosus, and vesiculobullous
at the clinical diagnosis of leukoplakia, such as disorders (Reichart and Philipsen, 2005).
white sponge naevus, acute pseudomembranous Most lesions of oral lichen planus present as
candidiasis, frictional keratosis, OLP, chronic white striae (reticular or annular) or plaques;
hyperplastic candidiasis, leukoedema, chem- some have papular, atrophic, erosive, bullous, or
ical injury, uremic stomatitis, nicotinic stoma- ulcerative features. The lesions are usually present
titis, skin grafts, and oral hairy leukoplakia bilaterally (Warnakulasuriya and Greenspan,
(Warnakulasuriya, 2018). 2020). Incipient PVL often mimics OLP lesions
Proliferative verrucous leukoplakia presents both clinically and histologically (Gilligan et al.,
as multiple white patches at different sites in the 2021); this leads to diagnostic challenges.
oral cavity (usually on the gingiva, palate, and Common signs and symptoms of oral
alveolar mucosa), with a preponderance in elderly submucous fibrosis are a burning sensation
women. PVLs start as flat lesions, and most of when eating spicy food, diffuse blanching of oral
them progress to a verrucous appearance. In the mucosa, and restricted mouth opening. In addi-
early stages, PVL may mimic OLP clinically and tion, restriction in tongue movement, palpable
histologically (McParland and Warnakulasuriya, fibrous bands, a leathery feeling of the mucosa,
2021; Thompson et al., 2021). depapillation of the tongue, shrunken uvula, and
Erythroplakia is a predominantly fiery red sunken cheeks are present to various degrees
patch that cannot be characterized clinically (Tilakaratne et al., 2006).
or pathologically as any other definable disease
(Warnakulasuriya et al., 2007). Other definable (b) Histopathological spectrum of OPMDs
red lesions should be excluded to arrive at the Histopathological features vary depending
clinical diagnosis of erythroplakia, such as on the type of OPMD. However, the presence of
erythematous candidiasis, inflammatory condi- variable levels of epithelial dysplasia is the most
tions, denture-induced stomatitis, erythema
90
important histopathological feature common to substantially depending on the study population,

all OPMDs, and this is a fairly reliable biological risk habits, and site in the oral cavity, and from
marker, which guides treatment stratification study to study (Reibel, 2003; Bouquot et al., 2006;
based on the risk of malignant transformation. Napier and Speight, 2008).
Epithelial dysplasia in the oral mucosa is graded PVL and erythroplakia have the highest
into three categories: mild, moderate, and severe. malignant transformation rates (30–50%), and
Grading of epithelial dysplasia using this scale is OLP has the lowest (~1–2%) (Warnakulasuriya
subjective and leads to significant intra-examiner and Greenspan, 2020). The risk of transforma-
and inter-examiner variability (Tilakaratne et al., tion of leukoplakia depends on the clinical type
2019). In 2017, the fourth edition of the WHO and grade of epithelial dysplasia (Mehanna et al.,
Classification of Head and Neck Tumours intro- 2009). Globally, the malignant transformation
duced a new binary grading system: low-risk and rate for leukoplakia was reported as 1.36% per
high-risk epithelial dysplasia (Reibel et al., 2017). year (95% CI, 0.69–2.03%) by Petti (2003) and as
The histopathological spectrum of leuko- 9.8% (95% CI, 7.9–11.7%) in the systematic review
plakias varies from cases of keratosis without and meta-analysis by Aguirre-Urizar et al.
dysplasia to mild, moderate, or severe dysplasia. (2021) based on 5-year data. The natural history
Erythroplakia is a high-risk lesion because most of leukoplakia is a dynamic rather than a static
cases at diagnosis are either severe epithelial process with respect to malignant transforma-
dysplasia or in situ OSCC. PVL is a lesion with tion. The malignant transformation rates of oral
minimal dysplasia, although about 50% of PVLs submucous fibrosis vary widely across studies,
transform into OSCC. Early cases of PVL have ranging from 7% to 13% (Tilakaratne et al., 2006;
histopathological features similar to those of Ekanayaka and Tilakaratne, 2016).
OLP, which may lead to misdiagnosis (Thompson Over time, OPMDs may persist unchanged,
et al., 2021). OLP has orthokeratinized or para- increase in size, regress in size, or even completely
keratinized surface epithelium with a band-like resolve (Fig. 1.9), which has been shown in many
lymphocytic infiltrate in the upper corium and follow-up studies (Mehta et al., 1972; Gupta et al.,
associated basal cell destruction and apoptosis. 1980; Silverman et al., 1984; Holmstrup et al.,
Epithelial dysplasia may be present in a minority 2006; Speight et al., 2018; Farah et al., 2019). Even
of OLP lesions, and this feature increases the in the absence of significant epithelial dysplasia,
risk of malignant transformation. Oral submu- some OPMDs can progress to OSCC with time;
cous fibrosis has characteristic histopathological therefore, lifetime clinical follow-up is highly
features, such as atrophy of the surface epithelium recommended (Villa et al., 2019).
and hyalinization and fibrosis of the submucosa,
which extend deep into the underlying connec- (d) Clinical features of oral cancer
tive tissue and muscle as the disease progresses. The clinical features of oral cancer vary
Atrophic surface epithelium may have features of depending on the site and the stage of clinical
epithelial dysplasia in some cases (Utsunomiya presentation (Bagan et al., 2010; Dissanayaka et
et al., 2005). al., 2012). The two most common sites for oral
cancer are the tongue and the buccal mucosa.
(c) Malignant transformation of OPMDs Other sites of involvement are the floor of
OPMDs are a heterogeneous group of lesions, the mouth, the gingivae, and the palate
and the rates of transformation to cancer vary (Warnakulasuriya, 2009). The lesions have a
from 1.4% to 49.5% (Iocca et al., 2020). Rates variable size, ranging in diameter from a few
of malignant transformation of OPMDs vary
91
Fig. 1.9 Natural history of oral potentially malignant disorders and oral cancer
92

Normal mucosa Keratosis without Mild epithelial dysplasia
dysplasia
Moderate epithelial
dysplasia
Oral potentially Molecularly/ Static

malignant disorders biologically benign
Severe epithelial
Molecularly/ dysplasia
biologically
premalignant
Oral squamous cell

carcinoma
Epithelial dysplasia
Some regress
No progression Regional metastasis
Progression
Distant metastasis
Created by the Working Group.

Fig. 1.10 Molecular events in the natural history of oral cancer
SIA
ERPL A
HYP
L
R MA EPIGEN
ETIC
EV
NO
(HYPER ENT
METH
YLA S
TIO
N) EP
IG
FOXM1 (H
IS EN
ET
D
MGMT MO TO
RARB2 CMTM3 DI NE I
TSPYL5 FI
C
TRPC4
CA
R
EV
MAP6
TI
O
EN
ZNF582
CE
S
GE G
SI
GIO IN
TS
NS
NE
EGFLAM
AN DUC
LYS4
)
CLDNII
NKX2-3 LYS9
AN
IN
RBP4 LYS18
EP
LYS2T
IGE
VEGF EDNRB
(M
PDPN
Y
PAX1
NE T
ICR
ORAL C
FGF p14ARK
E
HBP17 p16INK4a
ORN
TIV
IC EV
HIF-1 miR-31
miR-21
LICA
As)
L IT Y
miR-146
ENTS
miR-211
miR-204
ENABLING REP
IMMORTA
miR-155
LTERT miR-24
LTERC
GENETIC CON
DNA ANEUPLOIDY
LOH 3p14
S
ABNO A
9p21
4q
MMP-2 17p
E-CADHERIN MMP-9 8p
ACTI
R
SYNDECAN-1 -CATENIN 11q
M
LAMININ COX-2
13q
TENT
COLLAGEN XVII CD44V6 Cyclin D1
VAT
L
AND
EP-CAN c-Fos
PERLECAN c-Jun
ING
MMP-7 c-Myc LMLH1

MET
Bmi-1
INV SIS
BCL-2 TGF
SURVIVIN
A ST
SMAD
AS
HSP60 C-erbB1 (HER1) ABCG2

A
TY
IO
P
AB E
HSP27 C-erbB2 (HER2/neu) ALDH
M
N
C-erbB3 (HER3)
ILI
P-AKT CD133
IN ENO
HSP70 CDK2 LGR5
METALLOTHIONEIN CDK4 LGR4
G
ST
Cyclin E
CIP2A
GEMININ Y-H2AX
Ki-67
EV
GR
p53
W A BrdU
p63 ER
O
TH SIO NC
LS
AgNOR
p73
SU N O EL
CA M C
PCNA
MDM2
MCM2-7
PP F E
RE
SS ST
L
OR E
S
SUS AG
TAIN
ING PR DAM E
A GRO OLIFERATIVE DNA PONS
WTH
SIGNALLING RES
A
I S
LOH, loss of heterozygosity.

Created by the Working Group.
millimetres to several centimetres in advanced from an OPMD. The base of ulcerated tumours
cases. is indurated, and the surrounding mucosa has
In the initial stages, oral malignant lesions everted margins because of proliferation of the
present as well-demarcated erythroleukoplakic epithelium. Early cancers are asymptomatic, but
lesions consisting of red, white, or red and advanced tumours can be very painful. Tongue
white areas with a slight roughness along with cancer causes difficulty in swallowing and
reduced elasticity or induration of the soft tissue. speaking, and restricted movements. Cancer of
As the disease advances, there is ulceration and/ the buccal mucosa can lead to severe trismus
or nodularity and fixation to underlying tissues. when it has invaded into muscles. Enlarged and
The tumours can be either exophytic or endo- fixed cervical lymph nodes due to locoregional
phytic, and many of them may have residual spread is a late presentation of the disease (Bagan
red and white areas or a nodular and/or gran- et al., 2010).
ular appearance, indicating their possible origin
93
(e) Histopathology of oral cancer at an early stage, and the highest mortality rate is
The histopathological hallmark of OSCC is reported in patients with tongue cancer (Su et al.,
the invasion of malignant epithelial cells into the 2019). Positive regional lymph nodes, particu-
underlying connective tissue. When the tumour larly with extracapsular spread, have a direct
cells resemble the surface normal squamous negative effect on prognosis (Abdel-Halim et al.,
epithelium, with marked keratin formation, 2021). Although the 5-year survival rate of OSCC
the cancer is categorized as well-differentiated is reported to be about 50%, recent data show an
OSCC. At the other end of the spectrum, when improvement to 66% in some centres (Liu et al.,
the tumour cells do not bear any resemblance 2021).
to the squamous cells and there is no evidence
of keratin formation, the cancer is categorized 1.3.2 Stage at diagnosis and stage-related
as poorly differentiated OSCC. The tumours in survival
between these two extremes are categorized as Prognosis of cancers of the lip, oral cavity,
moderately differentiated OSCC. In addition to and oropharynx depends mainly on the stage of
the conventional types, some subtypes of OSCC the disease at diagnosis. Table 1.2 shows survival
have also been described; these include basaloid, rates for these cancer types by country or terri-
adenoid (acantholytic), adenosquamous, papil- tory in five continents in 2006–2014 (IARC,
lary, spindle cell, cuniculatum, and verrucous 2022). Heterogeneity across countries is high;
carcinoma. 5-year survival rates range from 0% to 64%
Histopathological parameters that must (median, 39%) for patients with cancer of the lip
be contained in a pathology report include the or oral cavity and from 0% to 67% (median, 32%)
level of differentiation, vascular and perineural for patients with oropharyngeal cancer.
invasion, pattern of invasion, depth of invasion, The extent of the disease can be classified
and immune response. In addition, the clearance as localized (tumours confined to the organ of
distance of excision margins and lymph node origin without invasion into the surrounding
status should be included in a report of surgical tissue or organs and without involvement of
excision of the primary tumour with neck dissec- any regional or distant lymph nodes or organs),
tion (i.e. removal of the lymph nodes in the neck). regional (tumours invading the surrounding
Numerous molecular events have been described tissue or organs, with or without the involvement
with respect to oral carcinogenesis (Dionne et al., of the regional lymph nodes, but not involving
2015; Nikitakis et al., 2018; Farah, 2021; Fig. 1.10). non-regional lymph nodes or organs), or with
(f) Prognosis of oral cancer distant metastasis (spreading to the non-regional
lymph nodes or distant organs; or unknown)
Prognosis of oral cancer depends on multiple (WHO Classification of Tumours Editorial
factors, including tumour-, host-, and treat- Board, 2023). Overall, cancer of the lip or oral
ment-related factors. The most significant prog- cavity is more frequently diagnosed with local-
nostic factors are the stage of disease, depth of ized stage, compared with oropharyngeal cancer,
invasion, pattern of invasion, lymphovascular which is more frequently diagnosed with regional
invasion, nodal status, and distant metastases disease. [Limitations of the study are significant,
(Dissanayaka et al., 2012; De Silva et al., 2018). including a high proportion of unclassified
The stage at diagnosis and the mortality rate vary cancers (~10% to 50%) and the variability in the
according to the primary site of the tumour; for number of patients analysed per country, which
example, cancer of the lower lip is often diagnosed
94
Table 1.2 Survival (at 1 year, 3 years, and 5 years) of oral cancer and oropharyngeal cancer, by
country or territory in 2006–2014, for both sexes combined
Country or territory Survival (%)

Oral cancer Oropharyngeal cancer
1 year 3 years 5 years 1 year 3 years 5 years
Algeria 94 31 0 50 0 0
Argentina 62 42 35 57 38 34
Bahrain 71 52 52 75 75 38
Brazil 77 55 49 58 39 31
Chile 61 43 34 52 36 29
China 72 53 47 54 35 30
Colombia 56 39 39 50 0 0
Costa Rica 73 57 52 67 49 42
Ecuador 65 50 45 80 46 31
India 71 47 40 61 32 24
Israel 82 66 58 85 63 55
Republic of Korea 80 62 57 81 65 58
Malaysia 58 36 31 68 53 47
Martinique, France 65 45 39 71 41 41
Peru 65 45 37 72 57 44
Puerto Rico, USA 65 27 16 60 27 15
Saudi Arabia 81 64 64 100 67 67
Seychelles 45 29 19 N/A N/A N/A
South Africa 49 29 18 44 31 31
Thailand 51 32 26 49 28 23
Turkey 83 68 59 74 51 40
Uruguay 60 37 31 56 38 32
N/A, not available.
Compiled from IARC (2022).
is very small in some cases (ranging from 13 to distant disease, respectively (American Cancer
3453).] Society, 2023).
The Surveillance, Epidemiology, and End The treatment of cancer of the lip, oral cavity,
Results (SEER) database tracks 5-year relative and oropharynx is driven mainly by staging of the
survival rates for oral cancer and oropharyngeal disease. Since its conception in 1959, the Union
cancer in the USA. In patients with oral cancer, for International Cancer Control/American Joint
based on different anatomical subsites (lip, Committee on Cancer (UICC/AJCC) tumour–
tongue, or floor of the mouth), the 5-year relative node–metastasis (TNM) staging system has
survival rates were 73–94%, 42–70%, and 23–41% become the main modality of tumour staging
for localized, regional, and distant disease, and is used to tailor the treatment of patients
respectively; survival was worse for patients with (Tirelli et al., 2018a). New editions of the AJCC
cancer of the floor of the mouth than for those TNM staging system are regularly published to
with tongue cancer. In patients with oropharyn- improve the ability to predict patient outcomes.
geal cancer, the 5-year relative survival rates were The eighth edition, which was published in 2017
59%, 62%, and 29% for localized, regional, and (AJCC, 2017), had two major changes in TNM
95
96

Table 1.3 Tumour–node–metastasis staging system for carcinomas of the oral cavity
Primary tumour (T) Clinical N (cN) Pathological N (pN) Distant metastasis (M)
T T criteriaa N N criteria N N criteria M category M criteria
category categoryb categoryb
TX Primary tumour NX Regional lymph nodes NX Regional lymph nodes cannot be assessed No distant
cannot be cannot be assessed M0 metastasis
assessed
T0 No evidence of N0 No regional lymph node N0 No regional lymph node metastasis Distant
M1
primary tumour metastasis metastasis
Tis Carcinoma in situ N1 Metastasis in a single N1 Metastasis in a single ipsilateral lymph node,
ipsilateral lymph node, ≤ 3 cm in greatest dimension and ENE(−)
≤ 3 cm in greatest
dimension and ENE(−)
T1 Tumour ≤ 2 cm N2 Metastasis in a single N2 Metastasis in a single ipsilateral lymph node
with DOI ≤ 5 mm ipsilateral lymph node ≤ 3 cm in greatest dimension and ENE(+); or
> 3 cm and ≤ 6 cm in Metastasis in a single ipsilateral lymph node
greatest dimension and > 3 cm and ≤ 6 cm in greatest dimension and
ENE(−); or ENE(−); or
Metastases in multiple Metastases in multiple ipsilateral lymph
ipsilateral lymph nodes, nodes, none > 6 cm in greatest dimension,
none > 6 cm in greatest and ENE(−); or
dimension, and ENE(−); or Metastases in bilateral or contralateral lymph
Metastases in bilateral or node(s), none > 6 cm in greatest dimension,
contralateral lymph nodes, and ENE(−)
none > 6 cm in greatest
dimension, and ENE(−)
T2 Tumour ≤ 2 cm, N2a Metastasis in a single N2a Metastasis in a single ipsilateral lymph node
with DOI > 5 mm ipsilateral lymph node ≤ 3 cm in greatest dimension and ENE(+); or
and ≤ 10 mm; or > 3 cm and ≤ 6 cm in Metastasis in a single ipsilateral lymph node
Tumour > 2 cm greatest dimension and > 3 cm and ≤ 6 cm in greatest dimension and
and ≤ 4 cm, with ENE(−) ENE(−)
DOI ≤ 10 mm
T3 Tumour > 2 cm N2b Metastases in multiple N2b Metastases in multiple ipsilateral lymph
and ≤ 4 cm with ipsilateral lymph nodes, nodes, none > 6 cm in greatest dimension,
DOI > 10 mm; or none > 6 cm in greatest and ENE(−)
Tumour > 4 cm dimension, and ENE(−)
with DOI
≤ 10 mm
Table 1.3 (continued)
T T criteria a N N criteria N N criteria M category M criteria
T4 Moderately N2c Metastases in bilateral or N2c Metastases in bilateral or contralateral lymph
advanced or very contralateral lymph nodes, node(s), none > 6 cm in greatest dimension,
advanced local none > 6 cm in greatest and ENE(−)
disease dimension, and ENE(−)
T4a Moderately N3 Metastasis in a lymph N3 Metastasis in a lymph node > 6 cm in greatest
advanced local node > 6 cm in greatest dimension and ENE(−); or
disease dimension and ENE(−); or Metastasis in a single ipsilateral lymph node
Tumour > 4 cm Metastasis in any lymph > 3 cm in greatest dimension and ENE(+); or
with DOI node(s) and clinically overt Metastasis in multiple ipsilateral,
> 10 mm; or ENE(+) contralateral, or bilateral lymph nodes, any
Tumour invades ENE(+); or
adjacent Metastasis in a single contralateral lymph
structures only node of any size and ENE(+)
(e.g. through
cortical bone of
the mandible
or maxilla, or
involves the
maxillary sinus or
skin of the face)
Note: Superficial
erosion of bone/
tooth socket
(alone) by a
gingival primary
is not sufficient to
classify a tumour
as T4.

T4b Very advanced N3a Metastasis in a lymph N3a Metastasis in a lymph node > 6 cm in greatest
local disease node > 6 cm in greatest dimension and ENE(−)
Tumour invades dimension, and ENE(−)
masticator space,
pterygoid plates,
or skull base
and/or encases the
internal carotid
artery
97
98

T T criteriaa N N criteria N N criteria M category M criteria
T4b N3b Metastasis in any lymph N3b Metastasis in a single ipsilateral lymph node
(cont.) node(s) and clinically overtc > 3 cm in greatest dimension and ENE(+); or
ENE(+) Metastasis in multiple ipsilateral,
contralateral, or bilateral lymph nodes, any
ENE(+); or
Metastasis in a single contralateral lymph
node of any size and ENE(+)
AJCC, American Joint Committee on Cancer; DOI, depth of invasion; ENE, extranodal extension; TNM, tumour–node–metastasis.
a DOI is depth of invasion and not tumour thickness.
b
A designation of “U” or “L” may be used for any N category to indicate metastasis above the lower border of the cricoid (U) or below the lower border of the cricoid (L). Similarly,
clinical and pathological ENE should be recorded as ENE(−) or ENE(+).
c
The presence of skin involvement or soft tissue invasion with deep fixation or tethering to underlying muscle or adjacent structures or clinical signs of nerve involvement is classified as
clinical ENE.
Adapted from AJCC (2017). The original source for this information is the AJCC cancer staging manual, 8th edition, published by Springer International Publishing. Corrected at 4th
printing, 2018.
Table 1.4 Tumour–node–metastasis staging system for carcinomas of the oral cavity: prognostic
stage groups
T category N category M category Stage group
Tis N0 M0 Stage 0
T1 N0 M0 Stage I
T2 N0 M0 Stage II
T3 N0 M0 Stage III
T1, T2, T3 N1 M0 Stage III
T4a N0, N1 M0 Stage IVA
T1, T2, T3, T4a N2 M0 Stage IVA
Any T N3 M0 Stage IVB
T4b Any N M0 Stage IVB
Any T Any N M1 Stage IVC
Adapted from AJCC (2017). The original source for this information is the AJCC cancer staging manual, 8th edition, published by Springer
International Publishing. Corrected at 4th printing, 2018.
categorization compared with previous editions Besides disease staging, many other factors
(Amin et al., 2017) (Table 1.3 and Table 1.4): may affect the prognosis of individual patients:
inclusion of the depth of invasion (DOI) of access to specialized care, associated comorbidi-
the tumour (≤ 5 mm, 5–10 mm, and > 10 mm) ties, and the quality of treatment planning, which
affects the T categorization, and inclusion of the is multidisciplinary in nature and is strongly
extranodal extension (ENE) affects the N cate- linked to the experience of the team (Hansen
gorization. The T1–3 but not the T4 classification et al., 2020). Finally, it is important to note that
is dependent on both the size of the tumour and one quarter to one third of deaths in patients
the DOI. Also, extrinsic muscle involvement has with head and neck squamous cell carcinoma
been excluded as a criterion for T4 staging of are attributable to a second primary malignancy
tongue cancer. Finally, the absence of ENE is a in the field of cancerization; this may affect the
prerequisite to classify N stage as N1, N2, or N3a upper aerodigestive tract again, the oesophagus,
disease, except if there is ENE of less than 3 cm or the lung, which are among the most frequent
in diameter in a single node (pN2a) (Zanoni and anatomical sites (Braakhuis et al., 2002; Baxi
Patel, 2020). et al., 2014).
Based on data from the United States
National Cancer Database and staging with 1.3.3 Treatment and management of OPMDs
the eighth edition of the TNM staging system, and oral cancer
the 5-year overall survival rate of patients with
oral cancer who received treatment was 78.8% (a) Treatment and management of OPMDs
(median survival not reached) for stage 0, 72.2% OPMDs are heterogeneous in their clinical
(median survival not reached) for stage I, 57.5% presentation. Some OPMDs remain stable for
(median survival, 5.70 years) for stage II, 55.1% many years or even regress; some eventually
(median survival, 5.59 years) for stage III, 39.7% transform into oral cancer (see Section 1.3.1).
(median survival, 3.08 years) for stage IVA, 27.1% Therefore, one of the main challenges of clinical
(median survival, 1.45 years) for stage IVB, and management is to identify such high-risk lesions
15.8% (median survival, 1.27 years) for stage IVC (Lingen et al., 2017).
(Cramer et al., 2018).
99
After the clinical diagnosis of an OPMD such as excision or ablation using various tech-
(Warnakulasuriya et al., 2021), a biopsy is recom- niques, including cold blade or electrocautery,
mended for histopathological diagnosis, which is laser, cryotherapy, and photodynamic therapy
the current reference standard for confirmation (Birur et al., 2022). Surgical excision was shown
of diagnosis, treatment guidance, and prognos- to decrease the rate of malignant transforma-
tication (Lingen et al., 2017). The histopatholog- tion of oral dysplasia but not totally eliminate it
ical diagnosis of oral epithelial dysplasia, which (Mehanna et al., 2009).
is routinely classified by grade (mild, moderate,
and severe), has both intra-rater and inter-rater (b) Treatment and management of oral cancer
variability, which is linked to pathologists’ The different modalities of treatment of oral
training and experience. Patients with OPMDs cancer are surgery, radiotherapy, chemotherapy,
that harbour high-grade dysplasia are at a and immunotherapy. Treatment planning is
greater risk for development of OSCC than are done at a multidisciplinary level; the patient is
patients with OPMDs with low-grade dysplasia. evaluated by a surgeon, a radiation oncologist,
Different in vivo optical imaging techniques may and a medical oncologist.
reduce diagnostic variability, but they have not Patients with early-stage and locally
been thoroughly evaluated (see Section 4.1.6). advanced oral cancer (stage I and stage II) are
Predictive biomarkers, such as loss of heterozy- typically offered surgical resection. Ipsilateral
gosity (LOH) at specific chromosomal sites and and sometimes bilateral neck dissection may
aneuploidy, have been suggested, but none has be recommended. Depending on the depth of
entered routine clinical use (William et al., 2009; invasion and the presence of lymphovascular or
Woo, 2019; Vermorken et al., 2021). perineural invasion, postoperative radiotherapy
There is no evidence-based international to the primary site and the neck (unilateral or
consensus on management algorithms for bilateral) may be recommended.
OPMDs. After diagnosis, the management of Management of locoregionally advanced
OPMDs may include one or more strategies, oral cancer (stage III and stage IVA–B) requires
depending on the grade of dysplasia and other multimodality treatment: surgical resection
clinical factors. These include preventive strate- of the primary tumour and neck dissection,
gies (e.g. lifestyle risk modification: cessation of followed by postoperative radiotherapy or che-
tobacco use and/or alcohol consumption and/or moradiotherapy (Pignon et al., 2000; Shaw et al.,
use of areca nut, improvement of diet), disease 2020). Patients who experience recurrent disease
monitoring or surveillance (i.e. a watchful despite these treatments may be candidates for
waiting approach), medical interventions (i.e. targeted anti-epidermal growth factor receptor
use of topical or systemic agents, chemopre- (anti-EGFR) monoclonal antibodies and,
vention), surgical management, and others more recently, immunotherapy (Bernier, 2016;
(Warnakulasuriya, 2020; Kerr and Lodi, 2021; Vermorken et al., 2021).
Birur et al., 2022). Consensus guidelines for Certain adjunct methods used in secondary
clinical management of patients with OPMDs, prevention of oral cancer may also have utility
focusing on leukoplakia or erythroplakia, oral for tertiary prevention. In some studies, auto-
submucous fibrosis, and OLP, have recently been fluorescence and narrow-band imaging (see
proposed (Birur et al., 2022). For low-risk lesions, Section 4.1.3) have demonstrated utility to guide
clinical management may be limited to lifelong surgical margin assessment for the excision of
close surveillance, as an alternative to potentially oral cancer (Farah et al., 2016; Poh et al., 2016;
morbid, repeated, multistep surgical treatments, Farah, 2018; Guillaud et al., 2018; Tirelli et al.,
100
2018b; Schorn et al., 2020). Among vital staining Birur PN, Patrick S, Warnakulasuriya S, Gurushanth K,
techniques (see Section 4.1.3), toluidine blue and Raghavan SA, Rath GK, et al. (2022). Consensus guide-
lines on management of oral potentially malignant
Lugol’s iodine have demonstrated utility in the disorders. Indian J Cancer. 59(3):442–53. doi:10.4103/
surveillance of patients with a history of oral ijc.IJC_128_21 PMID:36412324
cancer, when used by experts for tertiary preven- Bosetti C, Carioli G, Santucci C, Bertuccio P, Gallus S,
Garavello W, et al. (2020). Global trends in oral and
tion (Epstein et al., 2003; Simões et al., 2017). pharyngeal cancer incidence and mortality. Int J Cancer.
147(4):1040–9. doi:10.1002/ijc.32871 PMID:31953840
Bouquot JE, Speight PM, Farthing PM (2006). Epithelial
References dysplasia of the oral mucosa – diagnostic problems and
prognostic features. Curr Diagn Pathol. 12(1):11–21.
doi:10.1016/j.cdip.2005.10.008
Abdel-Halim CN, Rosenberg T, Larsen SR, Høilund- Braakhuis BJ, Tabor MP, Leemans CR, van der Waal I,
Carlsen PF, Sørensen JA, Rohde M, et al. (2021). Snow GB, Brakenhoff RH (2002). Second primary
Histopathological definitions of extranodal extension: tumors and field cancerization in oral and oropharyn-
a systematic review. Head Neck Pathol. 15(2):599–607. geal cancer: molecular techniques provide new insights
doi:10.1007/s12105-020-01221-4 PMID:32918710 and definitions. Head Neck. 24(2):198–206. doi:10.1002/
Aguirre-Urizar JM, Lafuente-Ibáñez de Mendoza I, hed.10042 PMID:11891950
Warnakulasuriya S (2021). Malignant transformation of Conway DI, Brenner DR, McMahon AD, Macpherson
oral leukoplakia: systematic review and meta-analysis LMD, Agudo A, Ahrens W, et al. (2015). Estimating
of the last 5 years. Oral Dis. 27(8):1881–95. doi:10.1111/ and explaining the effect of education and income on
odi.13810 PMID:33606345 head and neck cancer risk: INHANCE consortium
AJCC (2017). AJCC cancer staging manual. 8th ed. pooled analysis of 31 case-control studies from 27
American Joint Committee on Cancer, American countries. Int J Cancer. 136(5):1125–39. doi:10.1002/
College of Surgeons. Chicago (IL), USA; Cham, ijc.29063 PMID:24996155
Switzerland: Springer Cham. Conway DI, Petticrew M, Marlborough H, Berthiller J,
American Cancer Society (2023). Survival rates for oral Hashibe M, Macpherson LMD (2008). Socioeconomic
cavity and oropharyngeal cancer. Available from: inequalities and oral cancer risk: a systematic review
https://www.cancer.org/cancer/types/oral-cavity-and- and meta-analysis of case-control studies. Int J Cancer.
oropharyngeal-cancer/detection-diagnosis-staging/ 122(12):2811–9. doi:10.1002/ijc.23430 PMID:18351646
survival-rates.html. Conway DI, Purkayastha M, Chestnutt IG (2018). The
Amin MB, Greene FL, Edge SB, Compton CC, Gershenwald changing epidemiology of oral cancer: definitions,
JE, Brookland RK, et al. (2017). The eighth edition trends, and risk factors. Br Dent J. 225(9):867–73.
AJCC cancer staging manual: continuing to build a doi:10.1038/sj.bdj.2018.922 PMID:30412558
bridge from a population-based to a more “personal- Cramer JD, Reddy A, Ferris RL, Duvvuri U, Samant S
ized” approach to cancer staging. CA Cancer J Clin. (2018). Comparison of the seventh and eighth edition
67(2):93–9. doi:10.3322/caac.21388 PMID:28094848 American Joint Committee on Cancer oral cavity
Argirion I, Zarins KR, Defever K, Suwanrungruang staging systems. Laryngoscope. 128(10):2351–60.
K, Chang JT, Pongnikorn D, et al. (2019). Temporal doi:10.1002/lary.27205 PMID:29658104
changes in head and neck cancer incidence in Thailand de Araújo RL, Lyko KF, Funke VA, Torres-Pereira CC
suggest changing oropharyngeal epidemiology in (2014). Oral cancer after prolonged immunosuppression
the region. J Glob Oncol. 5(5):1–11. doi:10.1200/ for multiorgan chronic graft-versus-host disease. Rev
JGO.18.00219 PMID:30860955 Bras Hematol Hemoter. 36(1):65–8. doi:10.5581/1516-
Bagan J, Sarrion G, Jimenez Y (2010). Oral cancer: clin- 8484.20140016 PMID:24624039
ical features. Oral Oncol. 46(6):414–7. doi:10.1016/j. De Silva RK, Siriwardena BSMS, Samaranayaka A,
oraloncology.2010.03.009 PMID:20400366 Abeyasinghe WAMUL, Tilakaratne WM (2018). A
Baxi SS, Pinheiro LC, Patil SM, Pfister DG, Oeffinger KC, model to predict nodal metastasis in patients with oral
Elkin EB (2014). Causes of death in long-term survi- squamous cell carcinoma. PLoS One. 13(8):e0201755.
vors of head and neck cancer. Cancer. 120(10):1507–13. doi:10.1371/journal.pone.0201755 PMID:30091996
doi:10.1002/cncr.28588 PMID:24863390 Dionne KR, Warnakulasuriya S, Zain RB, Cheong SC
Berkovitz BKB, Holland GR, Moxham BJ (2018). Oral (2015). Potentially malignant disorders of the oral
anatomy, histology and embryology. Amsterdam, cavity: current practice and future directions in the
Netherlands: Elsevier. clinic and laboratory. Int J Cancer. 136(3):503–15.
Bernier J, editor (2016). Head and neck cancer: multimo- doi:10.1002/ijc.28754 PMID:24482244
dality management. 2nd ed. Switzerland: Springer.
101
Dissanayaka WL, Pitiyage G, Kumarasiri PVR, Liyanage a systematic review and meta-analysis. Oral Dis.
RL, Dias KD, Tilakaratne WM (2012). Clinical 27(4):813–28. doi:10.1111/odi.13323 PMID:32144836
and histopathologic parameters in survival of oral Guillaud M, MacAulay CE, Berean KW, Bullock M,
squamous cell carcinoma. Oral Surg Oral Med Oral Guggisberg K, Klieb H, et al. (2018). Using quantita-
Pathol Oral Radiol. 113(4):518–25. doi:10.1016/j. tive tissue phenotype to assess the margins of surgical
oooo.2011.11.001 PMID:22668430 samples from a pan-Canadian surgery study. Head Neck.
Ekanayaka RP, Tilakaratne WM (2016). Oral submu- 40(6):1263–70. doi:10.1002/hed.25106 PMID:29451953
cous fibrosis: review on mechanisms of malignant Gupta PC, Mehta FS, Daftary DK, Pindborg JJ, Bhonsle
transformation. Oral Surg Oral Med Oral Pathol Oral RB, Jalnawalla PN, et al. (1980). Incidence rates of
Radiol. 122(2):192–9. doi:10.1016/j.oooo.2015.12.018 oral cancer and natural history of oral precancerous
PMID:27289264 lesions in a 10-year follow-up study of Indian villagers.
Epstein JB, Feldman R, Dolor RJ, Porter SR (2003). The Community Dent Oral Epidemiol. 8(6):283–333.
utility of tolonium chloride rinse in the diagnosis of doi:10.1111/j.1600-0528.1980.tb01302.x PMID:6937277
recurrent or second primary cancers in patients with Hansen CC, Egleston B, Leachman BK, Churilla TM,
prior upper aerodigestive tract cancer. Head Neck. DeMora L, Ebersole B, et al. (2020). Patterns of multi-
25(11):911–21. doi:10.1002/hed.10309 PMID:14603451 disciplinary care of head and neck squamous cell
Farah CS (2018). Narrow band imaging-guided resection carcinoma in Medicare patients. JAMA Otolaryngol
of oral cavity cancer decreases local recurrence and Head Neck Surg. 146(12):1136–46. doi:10.1001/
increases survival. Oral Dis. 24(1–2):89–97. doi:10.1111/ jamaoto.2020.3496 PMID:33090191
odi.12745 PMID:29480612 Holmstrup P, Vedtofte P, Reibel J, Stoltze K (2006).
Farah CS (2021). Molecular landscape of head and neck Long-term treatment outcome of oral premalig-
cancer and implications for therapy. Ann Transl Med. nant lesions. Oral Oncol. 42(5):461–74. doi:10.1016/j.
9(10):915. doi:10.21037/atm-20-6264 PMID:34164549 oraloncology.2005.08.011 PMID:16316774
Farah CS, Dalley AJ, Nguyen P, Batstone M, Kordbacheh F, IARC (2022). SURVCAN: Cancer survival in countries
Perry-Keene J, et al. (2016). Improved surgical margin in transition. Lyon, France: International Agency for
definition by narrow band imaging for resection of Research on Cancer. Available from: https://survival.
oral squamous cell carcinoma: a prospective gene iarc.fr/survcan/.
expression profiling study. Head Neck. 38(6):832–9. Iocca O, Sollecito TP, Alawi F, Weinstein GS, Newman JG,
doi:10.1002/hed.23989 PMID:25545703 De Virgilio A, et al. (2020). Potentially malignant disor-
Farah CS, Shearston K, Nguyen AP, Kujan O (2019). ders of the oral cavity and oral dysplasia: a systematic
Oral carcinogenesis and malignant transformation. review and meta-analysis of malignant transformation
In: Brennan PA, Aldridge T, Dwivedi RC, edi- rate by subtype. Head Neck. 42(3):539–55. doi:10.1002/
tors. Premalignant conditions of the oral cavity. hed.26006 PMID:31803979
Singapore: Springer Singapore; pp. 27–66. Kerr AR, Lodi G (2021). Management of oral poten-
doi:10.1007/978-981-13-2931-9_3 tially malignant disorders. Oral Dis. 27(8):2008–25.
Ferlay J, Ervik M, Lam F, Colombet M, Mery L, Piñeros doi:10.1111/odi.13980 PMID:34324758
M, et al. (2020). Global Cancer Observatory: Cancer Kim Y-J, Kim JH (2020). Increasing incidence and
Today. Lyon, France: International Agency for Research improving survival of oral tongue squamous cell carci-
on Cancer. Available from: https://gco.iarc.fr/today. noma. Sci Rep. 10(1):7877. doi:10.1038/s41598-020-
Gilligan G, Garola F, Piemonte E, Leonardi N, Panico 64748-0 PMID:32398670
R, Warnakulasuriya S (2021). Lichenoid prolifera- Lin Y-H (2020). Trends in incidence of cancer of oral
tive leukoplakia, lichenoid lesions with evolution to cavity, oropharynx and hypopharynx in Taiwan:
proliferative leukoplakia or a continuum of the same data from the nationwide cancer registry, 1979–2016.
precancerous condition? A revised hypothesis. J Clin Otolaryngol. 45(2):302–3. doi:10.1111/coa.13488
Oral Pathol Med. 50(2):129–35. doi:10.1111/jop.13133 PMID:31789484
PMID:33225541 Lingen MW, Abt E, Agrawal N, Chaturvedi AK, Cohen E,
González-Moles MA, Ruiz-Ávila I, González-Ruiz L, Ayén D’Souza G, et al. (2017). Evidence-based clinical practice
Á, Gil-Montoya JA, Ramos-García P (2019). Malignant guideline for the evaluation of potentially malignant
transformation risk of oral lichen planus: a systematic disorders in the oral cavity: a report of the American
review and comprehensive meta-analysis. Oral Oncol. Dental Association. J Am Dent Assoc. 148(10):712–727.
96:121–30. doi:10.1016/j.oraloncology.2019.07.012 e10. doi:10.1016/j.adaj.2017.07.032 PMID:28958308
PMID:31422203 Liu TPJ, Fisher BM, Chua B, Clark JR, Low TH, Batstone
González-Moles MÁ, Warnakulasuriya S, González- MD (2021). Survival outcomes following modern
Ruiz I, González-Ruiz L, Ayén Á, Lenouvel D, et al. multidisciplinary management of oral squamous
(2021). Worldwide prevalence of oral lichen planus: cell carcinoma in Australia. Oral Surg Oral Med
102
Oral Pathol Oral Radiol. 131(1):92–8. doi:10.1016/j. visualization-guided surgery for early-stage oral
oooo.2020.09.012 PMID:33187950 cancer. JAMA Otolaryngol Head Neck Surg. 142(3):209–
McParland H, Warnakulasuriya S (2021). Lichenoid 16. doi:10.1001/jamaoto.2015.3211 PMID:26769431
morphology could be an early feature of oral prolif- Reibel J (2003). Prognosis of oral pre-malignant lesions:
erative verrucous leukoplakia. J Oral Pathol Med. significance of clinical, histopathological, and molec-
50(2):229–35. doi:10.1111/jop.13129 PMID:33185900 ular biological characteristics. Crit Rev Oral Biol
Mehanna HM, Rattay T, Smith J, McConkey CC (2009). Med. 14(1):47–62. doi:10.1177/154411130301400105
Treatment and follow-up of oral dysplasia – a systematic PMID:12764019
review and meta-analysis. Head Neck. 31(12):1600–9. Reibel J, Gale N, Hille J (2017). Oral potentially malignant
doi:10.1002/hed.21131 PMID:19455705 disorders and oral epithelial dysplasia. In: El-Naggar
Mehta FS, Shroff BC, Gupta PC, Daftary DK (1972). Oral AK, Chan JKC, Grandis JR, Takata T, Slootweg PJ,
leukoplakia in relation to tobacco habits. A ten-year editors. WHO classification of head and neck tumours.
follow-up study of Bombay policemen. Oral Surg Oral 4th ed. Lyon, France: International Agency for Research
Med Oral Pathol. 34(3):426–33. doi:10.1016/0030- on Cancer; pp. 112–5.
4220(72)90319-2 PMID:4505755 Reichart PA, Philipsen HP (2005). Oral erythroplakia –
Mello FW, Miguel AFP, Dutra KL, Porporatti AL, a review. Oral Oncol. 41(6):551–61. doi:10.1016/j.
Warnakulasuriya S, Guerra ENS, et al. (2018). oraloncology.2004.12.003 PMID:15975518
Prevalence of oral potentially malignant disorders: a Ren ZH, Hu CY, He HR, Li YJ, Lyu J (2020). Global
systematic review and meta-analysis. J Oral Pathol Med. and regional burdens of oral cancer from 1990 to
47(7):633–40. doi:10.1111/jop.12726 PMID:29738071 2017: results from the global burden of disease study.
Menezes FDS, Fernandes GA, Antunes JLF, Villa LL, Cancer Commun (Lond). 40(2–3):81–92. doi:10.1002/
Toporcov TN (2021). Global incidence trends in head cac2.12009 PMID:32067418
and neck cancer for HPV-related and -unrelated Schorn L, Rana M, Madry A, Ipaktchi R, Möllmann H,
subsites: a systematic review of population-based Gellrich NC, et al. (2020). Does autofluorescence help
studies. Oral Oncol. 115:105177. doi:10.1016/j. detect recurrent squamous cell carcinoma? A prospec-
oraloncology.2020.105177 PMID:33561611 tive clinical study. Oral Surg Oral Med Oral Pathol Oral
Miranda-Filho A, Bray F (2020). Global patterns and trends Radiol. 130(3):258–63. doi:10.1016/j.oooo.2020.04.809
in cancers of the lip, tongue and mouth. Oral Oncol. PMID:32593620
102:104551. doi:10.1016/j.oraloncology.2019.104551 Shaw RJ, O’Connell JE, Bajwa M (2020). Basic surgical
PMID:31986342 principles and techniques. In: Warnakulasuriya
Nanci A, editor (2017). Ten Cate’s oral histology: devel- S, Greenspan JS, editors. Textbook of oral
opment, structure and function. 9th ed. Maryland cancer: prevention, diagnosis and management.
Heights (MO), USA: Mosby, Elsevier. Cham, Switzerland: Springer Nature; pp. 253–82.
Napier SS, Speight PM (2008). Natural history of poten- doi:10.1007/978-3-030-32316-5_20
tially malignant oral lesions and conditions: an over- Silverman S Jr, Gorsky M, Lozada F (1984). Oral
view of the literature. J Oral Pathol Med. 37(1):1–10. leukoplakia and malignant transformation. A
doi:10.1111/j.1600-0714.2007.00579.x PMID:18154571 follow-up study of 257 patients. Cancer. 53(3):563–8.
Nikitakis NG, Pentenero M, Georgaki M, Poh CF, Peterson doi:10.1002/1097-0142(19840201)53:3<563::AID-CN-
DE, Edwards P, et al. (2018). Molecular markers associ- CR2820530332>3.0.CO;2-F PMID:6537892
ated with development and progression of potentially Simões CA, Durazzo MD, de Aquino Xavier FC,
premalignant oral epithelial lesions: current knowl- Gallottini MHC, Lourenço SV, Pinto Júnior DDS, et
edge and future implications. Oral Surg Oral Med al. (2017). Lugol chromoscopy in the follow-up of head
Oral Pathol Oral Radiol. 125(6):650–69. doi:10.1016/j. and neck carcinoma. Ann Maxillofac Surg. 7(2):188–93.
oooo.2018.03.012 PMID:29709496 doi:10.4103/ams.ams_95_17 PMID:29264284
Petti S (2003). Pooled estimate of world leukoplakia prev- Sinnatamby CS (2011). Last’s anatomy: regional and
alence: a systematic review. Oral Oncol. 39(8):770–80. applied. 11th ed. London, UK: Churchill Livingstone,
doi:10.1016/S1368-8375(03)00102-7 PMID:13679200 Elsevier.
Pignon JP, Bourhis J, Domenge C, Designé L (2000). Speight PM, Khurram SA, Kujan O (2018). Oral poten-
Chemotherapy added to locoregional treatment tially malignant disorders: risk of progression to
for head and neck squamous-cell carcinoma: three malignancy. Oral Surg Oral Med Oral Pathol Oral
meta-analyses of updated individual data. MACH-NC Radiol. 125(6):612–27. doi:10.1016/j.oooo.2017.12.011
Collaborative Group. Meta-Analysis of Chemotherapy PMID:29396319
on Head and Neck Cancer. Lancet. 355(9208):949–55. Standring S, editor (2020). Gray’s anatomy: the anatom-
doi:10.1016/S0140-6736(00)90011-4 PMID:10768432 ical basis of clinical practice. 42nd ed. Amsterdam,
Poh CF, Anderson DW, Durham JS, Chen J, Berean Netherlands: Elsevier.
KW, MacAulay CE, et al. (2016). Fluorescence
103
Su WW, Su CW, Chang DC, Chuang SL, Chen SL, Warnakulasuriya S (2009). Global epidemiology of
Hsu CY, et al. (2019). Impact of varying anatomic oral and oropharyngeal cancer. Oral Oncol. 45(4–
sites on advanced stage and survival of oral cancer: 5):309–16. doi:10.1016/j.oraloncology.2008.06.002
9-year prospective cohort of 27 717 cases. Head Neck. PMID:18804401
41(5):1475–83. doi:10.1002/hed.25579 PMID:30652378 Warnakulasuriya S (2018). Clinical features and pres-
Thompson LDR, Fitzpatrick SG, Müller S, Eisenberg E, entation of oral potentially malignant disorders. Oral
Upadhyaya JD, Lingen MW, et al. (2021). Proliferative Surg Oral Med Oral Pathol Oral Radiol. 125(6):582–90.
verrucous leukoplakia: an expert consensus guideline doi:10.1016/j.oooo.2018.03.011 PMID:29673799
for standardized assessment and reporting. Head Neck Warnakulasuriya S (2020). Potentially malignant dis-
Pathol. 15(2):572–87. doi:10.1007/s12105-020-01262-9 orders of the oral cavity. In: Warnakulasuriya
PMID:33415517 W, Greenspan JS, editors. Textbook of oral
Tilakaratne WM, Jayasooriya PR, Jayasuriya NS, De Silva cancer: prevention, diagnosis and management.
RK (2019). Oral epithelial dysplasia: causes, quan- Cham, Switzerland: Springer Nature; pp. 141–58.
tification, prognosis, and management challenges. doi:10.1007/978-3-030-32316-5_12
Periodontol 2000. 80(1):126–47. doi:10.1111/prd.12259 Warnakulasuriya S, Greenspan JS (2020). Epidemi-
PMID:31090138 ology of oral and oropharyngeal cancers. In:
Tilakaratne WM, Klinikowski MF, Saku T, Peters TJ, Warnakulasuriya S, Greenspan JS, editors. Textbook
Warnakulasuriya S (2006). Oral submucous fibrosis: of oral cancer: prevention, diagnosis and manage-
review on aetiology and pathogenesis. Oral Oncol. ment. Cham, Switzerland: Springer Nature; pp. 5–21.
42(6):561–8. doi:10.1016/j.oraloncology.2005.08.005 doi:10.1007/978-3-030-32316-5_2
PMID:16311067 Warnakulasuriya S, Johnson NW, van der Waal I (2007).
Tirelli G, Gatto A, Boscolo Nata F, Bussani R, Piccinato A, Nomenclature and classification of potentially malig-
Marcuzzo AV, et al. (2018a). Prognosis of oral cancer: nant disorders of the oral mucosa. J Oral Pathol Med.
a comparison of the staging systems given in the 7th 36(10):575–80. doi:10.1111/j.1600-0714.2007.00582.x
and 8th editions of the American Joint Committee PMID:17944749
on Cancer Staging Manual. Br J Oral Maxillofac Warnakulasuriya S, Kujan O, Aguirre-Urizar JM, Bagan
Surg. 56(1):8–13. doi:10.1016/j.bjoms.2017.11.009 JV, González-Moles MÁ, Kerr AR, et al. (2021). Oral
PMID:29198482 potentially malignant disorders: a consensus report
Tirelli G, Piovesana M, Marcuzzo AV, Gatto A, Biasotto from an international seminar on nomenclature and
M, Bussani R, et al. (2018b). Tailored resections in oral classification, convened by the WHO Collaborating
and oropharyngeal cancer using narrow band imaging. Centre for Oral Cancer. Oral Dis. 27(8):1862–80.
Am J Otolaryngol. 39(2):197–203. doi:10.1016/j. doi:10.1111/odi.13704 PMID:33128420
amjoto.2017.11.004 PMID:29150027 WHO Classification of Tumours Editorial Board (2023).
Utsunomiya H, Tilakaratne WM, Oshiro K, Maruyama S, Head and neck tumours. 5th ed. Lyon, France:
Suzuki M, Ida-Yonemochi H, et al. (2005). Extracellular International Agency for Research on Cancer (WHO
matrix remodeling in oral submucous fibrosis: its Classification of Tumours series, Vol. 9). (forthcoming)
stage-specific modes revealed by immunohistochem- Available from: https://publications.iarc.fr/.
istry and in situ hybridization. J Oral Pathol Med. William WN Jr, Heymach JV, Kim ES, Lippman SM
34(8):498–507. doi:10.1111/j.1600-0714.2005.00339.x (2009). Molecular targets for cancer chemoprevention.
PMID:16091118 Nat Rev Drug Discov. 8(3):213–25. doi:10.1038/nrd2663
Vermorken JB, Budach V, Leemans CR, Machiels J-P, PMID:19247304
Nicolai P, O’Sullivan B, editors (2021). Critical issues in Woo SB (2019). Oral epithelial dysplasia and premalig-
head and neck oncology: key concepts from the seventh nancy. Head Neck Pathol. 13(3):423–39. doi:10.1007/
THNO meeting. Cham, Switzerland: Springer Nature. s12105-019-01020-6 PMID:30887394
doi:10.1007/978-3-030-63234-2 Zanoni DK, Patel SG (2020). New AJCC: how does it impact
Villa A, Hanna GJ, Kacew A, Frustino J, Hammerman PS, oral cancers? Oral Oncol. 104:104607. doi:10.1016/j.
Woo SB (2019). Oral keratosis of unknown significance oraloncology.2020.104607 PMID:32208339
shares genomic overlap with oral dysplasia. Oral Dis.
25(7):1707–14. doi:10.1111/odi.13155 PMID:31295753
104
2. REDUCING INCIDENCE OF
CANCER OR PRECANCER
2.1 Established risk factors studies in Europe typically reported a 3-fold

increase in risk for current smokers or ever-
2.1.1 Tobacco smoking smokers compared with never-smokers (range of
The carcinogenicity of tobacco smoking RR, 2.11–3.53). Similarly increased risks of oral
was first established by the IARC Monographs cancer were reported for smoked tobacco prod-
programme in 1985, including evidence on “the ucts other than cigarettes (i.e. cigars, pipe, bidi)
occurrence of malignant tumours of the respira- (Balaram et al., 2002; Wyss et al., 2013).
tory tract” (IARC, 1986). Subsequent evalua- In non-alcohol users, the INHANCE consor-
tions have individually listed the oral cavity, tium reported a lower-magnitude pooled risk
oropharynx, and hypopharynx among the estimate (odds ratio [OR], 1.35; 95% confidence
multiple affected anatomical sites (IARC, 2004b, interval [CI], 0.9–2.01) associated with ever ciga-
2012b). In most countries, tobacco smoking is rette smoking (Hashibe et al., 2007), whereas
the leading cause of oral cancer and oral cancer a multicentre population-based case–control
death (Chang et al., 2015a; Inoue-Choi et al., study in France reported a higher-magnitude
2019). risk estimate (OR, 3.2; 95% CI, 1.9–5.3) (Radoï
et al., 2015).
(a) Risk of oral cancer The most prevalent tumour histology in the
oral cavity is squamous cell carcinoma, and
Observational studies that reported pooled
observational studies have reported strong
relative risk (RR), meta-RR, or single RR esti-
associations with tobacco smoking, whether
mates of oral cancer incidence or mortality,
including all histology subtypes diagnosed
whether associated with ever or current cigarette
(Hashibe et al., 2007; Wyss et al., 2013) or only
smoking, consistently showed statistically signif-
squamous cell carcinoma (Lee et al., 2009;
icantly elevated risk estimates (Supplementary
Maasland et al., 2014).
Table S2.1, web only; available from https://
Reported RRs of oral cancer death in current
publications.iarc.fr/617). A meta-analysis
cigarette smokers (hazard ratio [HR], 5.32; 95%
of studies published in the 1990s (Gandini
CI, 2.95–9.58) and in daily cigarette smokers
et al., 2008), a pooled analysis from the Inter-
(HR, 6.23; 95% CI, 3.42–11.33) were of large
national Head and Neck Cancer Epidemiology
magnitude (Inoue-Choi et al., 2019). Significantly
(INHANCE) consortium (Wyss et al., 2013), and
increased risks of oral cancer death, with esti-
more recent multi-country (Agudo et al., 2012)
mates varying between 4.0 and 7.9, were also
and single-country (Maasland et al., 2014) cohort
105
reported in primary cigar smokers, including in (Toporcov et al., 2015) (Supplementary Table S2.1,
people who reported no inhalation (Chang et al., web only; available from https://publications.
2015a). iarc.fr/617). A suggested trend of increasing risk
RR estimates for oropharyngeal cancer asso- of oral cancer with decreasing age at initiation
ciated with ever or current cigarette smoking of tobacco smoking appeared to be driven by
have shown larger variations than those for oral longer duration of smoking or higher cumula-
cancer, with RR of 3.01 (95% CI, 2.71–3.35) in tive pack-years of smoking (age at initiation and
the INHANCE consortium (Wyss et al., 2013), duration of use are highly correlated), because
5.95 (95% CI, 3.41–10.4) and 8.53 (95% CI, statistical adjustment for these factors elim-
3.38–21.55) in studies in Europe (Agudo et al., inated the originally observed trend (Chang
2012; Maasland et al., 2014), and 1.63 (95% CI, et al., 2019). Geographically, studies in North
1.08–2.45) in a study in the USA (Stingone et al., and South America (Szymańska et al., 2011) and
2013) (Supplementary Table S2.1, web only; avail- in Europe (Bosetti et al., 2008) have consistently
able from https://publications.iarc.fr/617). reported positive and significant associations of
cigarette smoking with risks of oral cancer and
(i) Smoking intensity, duration, and
oropharyngeal cancer.
pack-years
The risk of oral cancer increases with (b) Risk of OPMDs
increasing frequency (number of cigarettes Tobacco smoking is associated with the occur-
smoked per day), duration (in years), and cumu- rence of oral potentially malignant disorders
lative pack-years of smoking, showing significant (OPMDs), specifically leukoplakia and eryth-
dose–response trends (IARC, 2012b; Toporcov roplakia, and their malignant transformation,
et al., 2015) (Supplementary Table S2.2, web including epithelial dysplasia (Warnakulasuriya
only; available from https://publications.iarc. et al., 2010; Li et al., 2011; van der Waal, 2014;
fr/617). Elevated risks of oral cancer associated Mello et al., 2018a). Increased risk of oral submu-
with current smoking are also evident even at a cous fibrosis (OSF) was also reported (Lee et al.,
low daily dose (2 cigarettes) (Polesel et al., 2008). 2003) (Supplementary Table S2.1, web only; avail-
Also, a more pronounced effect for the duration able from https://publications.iarc.fr/617).
of smoking than for frequency was observed
for oral and pharyngeal cancers combined (Di (c) Population attributable fraction
Credico et al., 2019).
Among studies that reported population
The risk of oropharyngeal cancer also
attributable fractions (PAFs), there were vari-
increases with increasing frequency, duration,
ations in the anatomical site of the cancer,
and cumulative pack-years of smoking, showing
the definitions of tobacco products, and the
significant dose–response trends (IARC, 2012b;
geographical span of the populations comprised.
Toporcov et al., 2015) (Supplementary Table S2.2,
Studies reported estimated PAFs of cigarette
web only; available from https://publications.
smoking for oral cancer of 33% (95% CI, 23–48%;
iarc.fr/617).
Agudo et al., 2012), 21.6% (95% CI, 15.9–25.8%;
(ii) Demographic characteristics Anantharaman et al., 2011), and 24.8% (95%
Effect estimates from large studies show that CI, 19.6–31.1%; Hashibe et al., 2009), and for
the association of smoking with oral cancer is oropharyngeal cancer of 49% (95% CI, 36–69%;
retained when the population is stratified by Agudo et al., 2012) and 29.7% (95% CI, 24.6–33.1%;
sex (Agudo et al., 2012) and age at diagnosis Anantharaman et al., 2011). Estimates from
those studies had at a minimum overlapping
106
95% CIs; this points to the sizeable proportion dose–risk response was less apparent for oral
of oral and oropharyngeal cancers that are due cancer and for duration of drinking (Hashibe
to tobacco smoking, mainly cigarette smoking. et al., 2007) (Supplementary Table S2.3, web only;
For OPMDs, in particular leukoplakia, the PAF available from https://publications.iarc.fr/617).
can be even higher (e.g. 56.4% in Taiwan, China; Three systematic reviews and meta-analyses
Lee et al., 2003). investigated risks of increasing alcohol intake
associated with oral and pharyngeal cancers
(d) Interaction between tobacco smoking and combined (Tramacere et al., 2010; Turati et al.,
alcohol consumption 2013; Bagnardi et al., 2015). When measured
Studies assessing the joint effect of tobacco in drinks per day, the pooled RR was 1.21
smoking and other established risk factors on the (95% CI, 1.10–1.33) for ≤ 1 drink per day and
risk of oral cancer are discussed in Section 2.1.7. increased to 5.24 (95% CI, 4.36–6.30) for heavy
alcohol consumption (≥ 4 drinks per day); when
2.1.2 Alcohol consumption measured in grams of ethanol per day, the pooled
RR ranged from 1.29 (95% CI, 1.25–1.32)
(a) Risk of cancer for 10 g ethanol per day to 13.02 (95% CI,
Consumption of alcoholic beverages has been 9.87–17.18) for 125 g ethanol per day. Bagnardi
classified as carcinogenic to humans (Group 1) et al. (2015) reported pooled risks associated with
by the IARC Monographs programme, causing oral and pharyngeal cancer with increasing
cancers of the oral cavity and pharynx, among alcohol consumption, with RRs of 1.13
multiple other sites (IARC, 2010, 2012b). The risks (95% CI, 1.00–1.26) for light drinking, 1.83
of oral and oropharyngeal cancer associated with (95% CI, 1.62–2.07) for moderate drinking,
alcohol consumption become more apparent in and 5.13 (95% CI, 4.31–6.10) for heavy drinking.
relation to dose–response and in combination Risks were broadly similar in men and in women,
with smoking (Supplementary Table S2.3, web for heavy drinking versus non-drinking or occa-
only; available from https://publications.iarc. sional drinking.
fr/617). Smoking-adjusted estimates for oral and (ii) Total exposure and frequency of exposure
pharyngeal cancer range from a 4-fold to a 9-fold
Lubin et al. (2009) assessed the risk of oral
increased risk; in non-smokers, “the majority of
cancer by total exposure and by frequency of
the studies found a strong association with alco-
use. For equal drink-years (a function of the
holic beverage consumption among non-smokers
frequency of alcohol use per day and the dura-
with a dose–response relationship” (IARC, 2010).
tion of drinking in years), higher alcohol intake
Similar risk estimates were reported across
for a shorter duration conferred a greater risk
types of alcoholic beverages (Purdue et al., 2009;
compared with lower alcohol intake for a longer
IARC 2012b; Turati et al., 2013) (Supplementary
duration [these data are not shown in the table].
publications.iarc.fr/617). (iii) Gene polymorphisms and ethnic
differences
(i) Drinking intensity and duration
Gene polymorphisms of alcohol dehydro-
In non-tobacco users, there was a clear dose–
genase (ADH) and aldehyde dehydrogenase
risk response with increased frequency of alcohol
(ALDH), two important enzymes in alcohol
consumption (drinks per day) for oropharyngeal
metabolism, have been well described; individ-
and hypopharyngeal cancers combined (OR for
uals with some of these gene polymorphisms are
≥ 5 drinks per day, 5.50; 95% CI, 2.26–13.4); the
107
at increased risk of oral cancer associated with 1.2–1.7) for ever versus never alcohol consump-
alcohol consumption (IARC, 2012b). Individuals tion. A stronger alcohol–risk association was
with homozygous ADH1B*1/*1 and ADH1C*1/*1 observed for erythroplakia (OR, 3.0; 95% CI,
genotypes are at increased risk of oral cancers 1.6–5.7) (Hashibe et al., 2000b). The two case–
(Hashibe et al., 2006; Marichalar-Mendia et al., control studies in Taiwan (China) with data on
2010). ALDH2*1/*2 heterozygotes are also at alcohol consumption and OSF had quite different
increased risk of head and neck cancer (HNC) findings: an OR of 0.68 (95% CI, 0.28–1.64) in
(Boccia et al., 2009). The ALDH2*2 variant allele men (Yang et al., 2010) and an OR of 1.8 (95% CI,
is prevalent in up to 30% of East Asian popula- 1.1–3.1) (Lee et al., 2003). No systematic reviews
tions (IARC, 2012b). A significantly increased risk or meta-analyses were identified that assessed
of oral cancer in individuals with ALDH2*1/*2 the risks of alcohol consumption associated with
genotype was shown in the Japanese population OPMDs.
(Nomura et al., 2000).
In their systematic review, Turati et al. (2013) (c) Interaction of alcohol consumption with
reported minimal differences with respect to other risk factors
geographical area both for drinking overall and Studies assessing the joint effect of alcohol
for heavy drinking (≥ 4 drinks per day); the RR consumption and other established risk factors
was lowest for Asia (4.75; 95% CI, 3.14–7.17) and on the risk of oral cancer are discussed in
highest for Europe (5.63; 95% CI, 4.09–7.77). Section 2.1.7.
Voltzke et al. (2018) investigated ethnic
differences in the relationship between alcohol 2.1.3 Smokeless tobacco use
consumption and risk of oral and oropharyngeal
cancer in the USA. They reported consistently In this Handbook, the term “smokeless
stronger risk estimates for Blacks than for Whites tobacco” refers to products containing tobacco
(Supplementary Table S2.4, web only; available but not including areca nut or other non-tobacco
from https://publications.iarc.fr/617). components of betel quid. The composition and
use of these products are presented in Section 3.1
(b) Risk of OPMDs and in Table 3.1.
A total of 11 case–control studies investigated (a) Risk of oral cancer
the association between alcohol consumption
and risk of OPMDs (Supplementary Table S2.5, Use of smokeless tobacco has been classi-
web only; available from https://publications. fied as carcinogenic to humans (Group 1) by the
iarc.fr/617). Estimates of risk of any OPMDs for IARC Monographs programme (IARC, 2007a,
alcohol consumption ranged from 0.63 (95% CI, 2012b). Meta-analyses have reported RRs for oral
0.33–1.21) (Li et al., 2011) to 1.4 (95% CI, 0.7–2.7) and pharyngeal cancers combined ranging from
(Thomas et al., 2003) to 2.7 (95% CI, 1.2–6.3) 1.3 to 1.8 (Weitkunat et al., 2007; Boffetta et al.,
(Amarasinghe et al., 2010b). Estimates of risk 2008; Lee and Hamling, 2009; IARC, 2012b).
of leukoplakia for alcohol consumption ranged Since then, one meta-analysis (Asthana et al.,
from an OR of 0.22 (95% CI, 0.12–0.37) (Petti 2019), one pooled analysis (Wyss et al., 2016), and
and Scully, 2006) to 1.8 (95% CI, 1.1–2.8) (Lee three hospital-based case–control studies that
et al., 2003) and to 3.00 (95% CI, 10.27–33.50) for were not included in either the meta-analysis or
frequent alcohol drinkers (Shiu et al., 2000). In the pooled analysis (Nasher et al., 2014; Quadri
the largest case–control study in India, Hashibe et al., 2015; Gupta et al., 2017) have confirmed
et al. (2000a) reported an OR of 1.4 (95% CI, the increased risk (Supplementary Table S2.6,
108
web only; available from https://publications. months), and retention time of the product
iarc.fr/617). in the mouth (see Supplementary Table S2.7,
Risk estimates by type of smokeless tobacco web only; available from https://publications.
products vary greatly. Asthana et al. (2019) iarc.fr/617). These results were consistent when
reported smoking-adjusted ORs ranging from smoking was accounted for or when restricted to
0.86 (95% CI, 0.58–1.29) for snus/moist snuff to never-smokers.
1.20 (95% CI, 0.80–1.81) for nasal snuff/dipping
and 4.18 (95% CI, 2.37–7.38) for oral snuff. Risk (c) Population attributable fractions
estimates for other smokeless tobacco products Based on the GLOBOCAN 2002 incidence
were also elevated, such as for naswar (OR, data, the proportion of cases attributable to
11.8; 95% CI, 8.4–16.4; Khan et al., 2019) and smokeless tobacco use was estimated to be
for shammah (OR, 20.14; 95% CI, 8.23–49.25; 68.2% in men and 13.6% in women in the Sudan,
Quadri et al., 2015; and 39; 95% CI, 14–105; 52.5% in men and 51.6% in women in India,
Nasher et al., 2014). 50.6% in men and women in other countries in
Smoking-adjusted summary risk estimates Asia (including Bangladesh, Bhutan, Indonesia,
are generally higher in women than in men Myanmar, Nepal, Pakistan, and Sri Lanka), 6.6%
(Weitkunat et al., 2007; Asthana et al., 2019). in men in the USA, and 1.6% in men in Canada
Clear and significant positive dose–response (Boffetta et al., 2008). These estimates are similar
relationships were reported between duration of to those of a more recent report (NCI and CDC,
use (in years), frequency of chewing (times per 2014).
day), smokeless tobacco retention time in the
mouth (in minutes), and risk of oral cancer (see 2.1.4 Chewing areca nut products (including
Supplementary Table S2.7, web only; available betel quid) with added tobacco
from https://publications.iarc.fr/617).
There was no clear association of smokeless Areca nut products (including betel quid)
tobacco use with oropharyngeal cancer, with RRs with added tobacco include a variety of products
close to 1 in ever-smokers and in never-smokers with compositions and names that may differ
(Wyss et al., 2016; Supplementary Table S2.6, web depending on the geographical area where they
only; available from https://publications.iarc. are used. For more detailed information on the
fr/617). products, see Section 3.1.
(b) Risk of OPMDs (a) Risk of oral cancer

Numerous studies have consistently shown Chewing areca nut products (including betel
an increased risk of OPMDs, particularly leuko- quid) with added tobacco is an established risk
plakia, in current users or ever-users of snuff factor for oral cancer and pharyngeal cancer.
or chewing tobacco compared with never-users With the terminology of “betel quid with added
(Supplementary Table S2.6, web only; avail- tobacco”, these products have been classified as
able from https://publications.iarc.fr/617). The carcinogenic to humans (Group 1) by the IARC
direction of the risk association was similar Monographs programme (IARC, 2004a, 2012b).
by country and type of product chewed (snuff, The RRs for ever-chewers versus never-chewers
naswar, shammah, chewing tobacco, and other ranged from 2.1 (95% CI, 2.1–3.4) to 45.9 (95% CI,
products), and a clear dose–response relation- 25.0–84.1), and the highest RR was reported in
ship was demonstrated in terms of frequency women (IARC, 2012b). Since then, one meta-anal-
of chewing (times per day), duration of use (in ysis, a large number of case–control studies, and
109
a few cross-sectional studies, conducted mainly (c) Population attributable fractions

in the Indian subcontinent, have confirmed the In high-prevalence geographical areas, the
clear relationship between areca nut products PAF of chewing betel quid with added tobacco
with added tobacco and increased risk of oral for oral cancer and OPMDs may be very high.
cancer (see Supplementary Table S2.8; web only; In India, the PAF for oral cancer was estimated
available from https://publications.iarc.fr/617). to be 49.5% for both sexes, and higher in women
The risk is higher in women (14.6; 95% CI, (63.2%) than in men (44.7%) (Guha et al., 2014).
7.6–27.8) than in men (5.4; 95% CI, 3.9–7.4) (Guha For OPMDs, the PAF was estimated to be 84% in
et al., 2014). A clear and significant dose–response Sri Lanka (Amarasinghe et al., 2010a).
relationship was reported between the quantity
and duration of chewing areca nut with added
2.1.5 Chewing areca nut products (including
tobacco and the risk of oral cancer (Madathil
et al., 2016; Supplementary Table S2.9, web only; betel quid) without tobacco
available from https://publications.iarc.fr/617). Areca nut products (including betel quid)
without tobacco include a variety of products
(b) Risk of OPMDs with specific compositions and names that may
Evidence has accumulated on the association differ depending on the geographical area where
between chewing areca nut products (including they are used. For more detailed information on
betel quid) with added tobacco and the risk of the products, see Section 3.1.
OPMDs (Supplementary Table S2.8, web only;
available from https://publications.iarc.fr/617). (a) Risk of oral cancer
Risk estimates for chewers versus never-chewers Chewing areca nut products (including betel
for combinations of OPMDs ranged from 1.4 quid) without tobacco is an established risk
(95% CI, 0.5–3.7) to 50.5 (95% CI, 21.5–119.5). factor for oral cancer. The IARC Monographs
When OPMDs were considered separately, risk programme classified separately “betel quid
estimates adjusted for tobacco smoking and without added tobacco” (IARC, 2004a, 2012b)
alcohol consumption ranged from 6.1 (95% CI, and areca nut (IARC, 2012b) as carcinogenic to
1.8–21.3) to 55.6 (95% CI, 27.4–112.7) for OSF humans (Group 1). Since then, one meta-analysis,
and from 2.5 (95% CI, 1.1–5.6) to 10.0 (95% a very large number of case–control studies, and
CI, 8.3–12.0) for leukoplakia. However, when a few cohort studies, mainly in Taiwan (China)
the corresponding estimates were restricted to and some in India, have confirmed the clear rela-
non-smokers and non-drinkers, the ORs for the tionship between chewing areca nut products
different types of OPMDs showed less discrep- without tobacco and increased risk of oral cancer
ancy (Jacob et al., 2004). (Supplementary Table S2.10 (web only; available
Significant dose–response relationships were from https://publications.iarc.fr/617).
reported between chewing areca nut with added Guha et al. (2014) reported meta-RRs for
tobacco and the risk of OPMDs, in terms of oral cancer of 11.0 (95% CI, 4.9–24.8) for Taiwan
frequency of chewing (times per day), duration of (China) and 2.4 (95% CI, 1.8–3.2) for the Indian
use (in years), and age at the start of the chewing subcontinent. Meta-RRs were also calculated for
habit (Supplementary Table S2.9, web only; avail- cancer at specific subsites of the oral cavity for the
able from https://publications.iarc.fr/617). Indian subcontinent; the highest estimates were
The combined effects of betel quid chewing reported for the cancer of the palate: 5.1 (95%
with other established risk factors are discussed CI, 1.1–24.9). Guha et al. (2014) also reported a
in Section 2.1.7.
110
meta-RR of 2.6 (95% CI, 1.7–3.9) for oropharyn- Taiwan (China), the PAF for oral cancer was
geal cancer. estimated to be 57.3% for both sexes (Guha et al.,
Significant dose–response relationships were 2014). For OPMDs, the PAFs were estimated to
reported between chewing areca nut products be 85.4% for OSF and 73.2% for leukoplakia, in
without tobacco and the risk of oral cancer (Yang the southern part of the main island (Lee et al.,
et al., 2014; Hu et al., 2020) or oral cancer death 2003).
(Wen et al., 2010) in terms of quantity, frequency
of use, and duration of use (Supplementary 2.1.6 HPV16 infection
publications.iarc.fr/617). (a) Risk of cancer
The IARC Monographs programme (IARC,
(b) Risk of OPMDs 2012a) determined that there is sufficient evidence
Evidence has accumulated on the association in humans for the carcinogenicity of human
between chewing areca nut products (including papillomavirus type 16 (HPV16); the virus causes
betel quid) without tobacco and the risk of oral cancer and oropharyngeal cancer (IARC,
OPMDs (Supplementary Table S2.10, web only; 2012a). The association of HPV16 infection with
available from https://publications.iarc.fr/617). risk of cancer is heterogeneous in terms of the
Risk estimates for chewers versus non-chewers anatomical site (oral cavity vs oropharynx) as
for a combination of OPMDs grouped together well as the method of assessment of HPV expo-
ranged from 8.8 (95% CI, 3.2–24.5) to 25.3 (95% sure (oral HPV16 DNA, systemic HPV16 L1
CI, 20.8–30.7). When OPMDs were considered antibodies, and systemic HPV16 E6 antibodies).
separately, risk estimates adjusted for tobacco HPV16 infection is associated with a moderately
smoking and alcohol consumption ranged elevated risk of oral cancers; ORs are generally < 5
from 4.5 to 65.9 for OSF and from 3.7 to 22.3 for oral HPV16 DNA prevalence and HPV16 L1
for leukoplakia. In a study where estimates were or E6 seropositivity (Supplementary Table S2.12,
restricted to non-smokers and non-drinkers, web only; available from https://publications.
the ORs for men and women combined were iarc.fr/617).
22.2 (95% CI, 11.3–43.7) for leukoplakia, 29.0 HPV16 infection is strongly associated with
(95% CI, 5.6–149.5) for erythroplakia, and 56.2 risk of oropharyngeal cancers; the risk esti-
(95% CI, 21.8–144.8) for OSF (Jacob et al., 2004; mates from case–control studies range from
Supplementary Table S2.10, web only; available 14 to > 100 for oral HPV16 DNA prevalence,
from https://publications.iarc.fr/617). from 1.1 to > 100 for HPV16 L1 seropositivity, and
Significant dose–response relationships were from 10 to > 200 for HPV16 E6 seropositivity.
reported between chewing areca nut without Reported risk estimates from prospective
tobacco and the risk of OPMDs, in terms of cohort studies were > 20 for oral HPV16 DNA
frequency of chewing, duration of use, and prevalence, 2–14 for HPV16 L1 seroposi-
age at the start of chewing (see Supplementary tivity, and 98–274 for HPV16 E6 seroposi-
Table S2.11, web only; available from https:// tivity (Supplementary Table S2.12, web only;
publications.iarc.fr/617). available from https://publications.iarc.fr/617).
Importantly, HPV16 E6 seropositivity precedes
(c) Population attributable fractions diagnosis of oropharyngeal cancer by several
In high-prevalence geographical areas, the decades, underscoring the temporality of HPV16
PAF of chewing betel quid without tobacco for exposure and cancer incidence (Kreimer et al.,
oral cancer and OPMDs may be very high. In 2013, 2017, 2019).
111
(b) Risk of OPMDs (greater-than-multiplicative) of the individual

A recent systematic review and meta-analysis risk estimates. A summary of statistical interac-
reported an HPV16 prevalence of 10.8% in OPMDs, tions across these established risk factors is given
primarily leukoplakia, with a similar prevalence in Supplementary Table S2.13 (web only; avail-
in dysplastic and non-dysplastic lesions (de la Cour able from https://publications.iarc.fr/617).
et al., 2021). [The reporting studies have generally (a) Interactions between tobacco smoking and
used only HPV16 DNA detection, which does alcohol consumption
not indicate either an established or active HPV
infection, or HPV causality in cancers.] Several studies have reported a great-
er-than-multiplicative interaction between
(c) Population attributable fractions tobacco smoking and alcohol consumption for
Globally, the PAF of HPV is ~2% for oral the risk of oral cancers and pharyngeal cancers
cancers and ~31% for oropharyngeal cancers, (which included cancers of the oropharynx,
and most of the cancers are caused by HPV16 hypopharynx, and other pharynx) (Blot
infection (de Martel et al., 2017). There is a wide et al., 1988; Barón et al., 1993; Hayes et al., 1999;
geographical heterogeneity in HPV etiological Schlecht et al., 1999; Anantharaman et al., 2011).
fractions for oropharyngeal cancers, ranging In a meta-analysis of seven observational studies
from estimates of 40% to > 50% in North America, in India and seven studies in Taiwan (China),
Europe, Australia and New Zealand, Japan, and Petti et al. (2013) found a 6.3-fold increased risk
the Republic of Korea to estimates of < 15% in in oral cancer for tobacco smoking and alcohol
most other parts of the world (Ndiaye et al., 2014; consumption combined, showing an at least
de Martel et al., 2017). This heterogeneity may additive effect.
reflect differences in sexual behaviours that are A pooled analysis of 17 case–control studies
relevant for acquisition of oral HPV infection in Europe and the USA from the INHANCE
(e.g. lifetime and recent oral sex behaviours) as consortium (Hashibe et al., 2009) reported a
well as the relative contributions of HPV infec- greater-than-multiplicative interaction between
tion compared with tobacco use and alcohol tobacco use (smoking and chewing) and alcohol
consumption across countries and geographical consumption for the risk of oral cancer (multi-
regions (Heck et al., 2010). plicative interaction parameter, 3.09; 95% CI,
1.82–5.23) and the risk of pharyngeal cancers
(multiplicative interaction parameter, 1.90;
2.1.7 Combined effects of established risk
95% CI, 1.41–2.56). The interaction was also
factors greater-than-multiplicative with high expo-
Tobacco smoking, alcohol consumption, sure to both smoking (> 20 cigarettes per day)
smokeless tobacco use, chewing areca nut prod- and alcohol consumption (> 3 drinks per day);
ucts with or without tobacco, and HPV16 infec- the ORs for joint exposure were 15.49 (95% CI,
tion are independent risk factors for OPMDs, 7.24–33.14) for oral cancers and 14.29 (95% CI,
oral cancers, and oropharyngeal cancers. 7.26–28.15) for pharyngeal cancers. Tobacco use
Combined exposure to more than one of these and alcohol consumption collectively accounted
carcinogens can confer a risk that is at least the for PAFs of 67.1% for oral cancers (23.5% from
sum of the individual risks for each of these the tobacco–alcohol interaction effect) and 74.3%
carcinogens (risk additivity) or can confer a for pharyngeal cancers (24.6% from the tobacco–
risk that exceeds the sum (greater-than-addi- alcohol interaction effect).
tive) or that exceeds the multiplication product
112
(b) Interactions with smokeless tobacco use across betel quid chewing, smoking, and alcohol
Few studies reported formal statistical evalu- consumption. A statistically significant great-
ations of interaction effects of smokeless tobacco er-than-additive interaction was observed
use with tobacco smoking or with alcohol between betel quid chewing and tobacco
consumption on the risk of OPMDs, oral cancers, smoking (RERI, 5.48; 95% CI, 1.06–8.20),
or oropharyngeal cancers. The few available and a non-significant additive interaction was
studies reported the absence of statistical inter- observed between betel quid chewing and
action (i.e. consistency with risk additivity) with alcohol consumption (RERI, 1.34; 95% CI, −1.29
tobacco smoking or with alcohol consumption to 4.50). Importantly, a statistically significant
on the risk of oral cancers (Winn et al., 1981). greater-than-additive three-way interaction was
observed across betel quid chewing, smoking,
(c) Interactions with chewing betel quid with and alcohol consumption (RERI, 28.36; 95%
or without tobacco CI, 22.92–33.74). Furthermore, the extent of the
three-way greater-than-additive interaction was
Reports of effect modification of the risk
similar in studies in India (RERI, 38.11; 95%
conferred by chewing betel quid with or without
CI, 30.05–41.62) and studies in Taiwan, China
tobacco by tobacco smoking and/or alcohol
(RERI, 36.42; 95% CI, 24.87–53.68). Betel
consumption have been inconsistent (IARC,
quid chewing, tobacco smoking, and alcohol
2012b). Some studies have reported the absence
consumption collectively accounted for 74.9% of
of statistical interaction (i.e. consistency with
oral cancers (68.4% from joint effects of all three
risk additivity) between ever chewing betel quid
exposures).
and ever smoking or ever alcohol consump-
tion for the risk of oral cancers (Subapriya (d) Interactions with HPV16 infection
et al., 2007; Muwonge et al., 2008). Some studies
have reported a greater-than-additive interac- Reports are sparse for interactions of HPV16
tion between ever chewing betel quid and ever infection with other risk factors for the risk of
smoking in non-drinkers on the risk of oral OPMDs or oral cancers. Most previous evalua-
cancers (Sankaranarayanan et al., 1989). Some tions of the interaction of HPV16 infection (as
studies have reported a greater-than-multiplica- determined by oral HPV16 DNA or systemic
tive interaction between ever chewing betel quid HPV16 L1 or HPV16 E6 antibodies) with smoke-
without tobacco and ever smoking on the risk of less tobacco, chewing betel quid with or without
oral and pharyngeal cancers (Znaor et al., 2003). A tobacco, smoking, and alcohol consumption
few studies have also reported a greater-than-ad- have included oropharyngeal cancers and have
ditive interaction between ever chewing betel been conducted in Europe and North and South
quid without tobacco and ever smoking on the America. Perhaps because of the geographical
risk of OPMDs, particularly leukoplakia (Lee clustering of these studies, most of the studies
et al., 2003). have primarily addressed the interaction of
Petti et al. (2013) conducted a meta-analysis HPV16 infection with tobacco smoking and
that included 14 studies – 7 in India (without alcohol consumption. Results for the interac-
separation of chewing betel quid with or without tion of HPV16 infection with other risk factors
tobacco) and 7 in Taiwan, China (chewing betel have been very inconsistent in the literature:
quid without tobacco) – to evaluate two-way and studies have reported a lack of statistical inter-
three-way additive interactions, as measured by action between HPV16 infection and smoking
relative excess risk due to interaction (RERI) or alcohol consumption on an additive scale
(D’Souza et al., 2007; Anantharaman et al.,
113
2016) or a multiplicative scale (Herrero et al., 15 years conferred a higher risk of oral cancer
2003; Farsi et al., 2017), the presence of a great- (OR, 2.07; 95% CI, 1.54–2.79) compared with
er-than-additive interaction between HPV16 L1 non-exposed people (Mariano et al., 2022).
antibodies and smoking (Schwartz et al., 1998),
greater-than-additive interactions between oral (b) Indoor air pollution
HPV16 DNA and alcohol consumption (Smith The IARC Monographs programme classified
et al., 2004), and less-than-multiplicative inter- indoor emissions from household combustion of
actions between HPV16 E6 antibodies and coal as carcinogenic to humans (Group 1), with
smoking (Ribeiro et al., 2011) and between HPV16 sufficient evidence for lung cancer (IARC, 2012b).
L1 antibodies and smoking and HPV16 L1 anti- More recently, a meta-analysis of 4 studies found
bodies and alcohol consumption (Applebaum a significant risk from household air pollution
et al., 2007). [Despite this inconsistency, smoking for the development of oral cancer (OR, 2.44;
and heavy alcohol consumption are associated 95% CI, 1.87–3.19) (Josyula et al., 2015). Notably,
with increased risk of both HPV16-positive and a high incidence of oral cancer was reported in
HPV16-negative oropharyngeal cancers and, at chefs engaged in regular cooking (Foppa and
the very least, should be considered to be inde- Minder, 1992). Indoor air pollution could be a
pendent risk factors for oropharyngeal cancers.] risk factor that increases risk in women more
than in men.
2.2 Additional potential risk factors (c) Heavy metals in soil
for oral cancer Most of the studies on heavy metals in soil
A proportion of oral cancers cannot be and risk of oral cancer are from Taiwan (China),
attributed to the major established risk factors particularly from Changhua County, which has
(Sections 2.1.1–2.1.6), particularly oral cancers a higher environmental heavy metal concentra-
that occur in women and young people. There tion than the other counties. Studies pointed
is a substantial amount of literature on several to arsenic and nickel in farm soils as new
other putative risk factors, for some of which risk factors for oral cancer (Su et al., 2010).
there is only little evidence. Significant associations between oral cancer
and blood levels of nickel and/or chromium
2.2.1 Environmental factors have been reported after controlling for poten-
tial confounders (Chiang et al., 2011; Yuan et al.,
(a) Second-hand smoke 2011). Also, Tsai et al. (2017) reported that 68.8%
The most recent evaluation by the of leukoplakia with subsequent malignant trans-
IARC Monographs programme (IARC, 2012b) formation occurred in people exposed to high
confirmed that second-hand tobacco smoke levels of nickel in soil.
(also called environmental tobacco smoke, passive
(d) Occupational exposures
smoking, or involuntary smoking) is carcinogenic
to humans (Group 1), although evidence for oral Increased risks due to occupational exposure
cancer was sparse. A recent meta-analysis of five to heavy metals were reported, for oral cancer
case–control studies reported a positive associ- due to exposure to metal dust containing chro-
ation between exposure to second-hand smoke mium and nickel (OR, 3.4; 95% CI, 1.7–7.0) (Tisch
and risk of oral cancer (overall OR, 1.51; 95% et al., 1996) and for risk of tongue cancer due to
CI, 1.20–1.91). A duration of exposure of > 10 or exposure to chromium(VI) compounds (Tisch
114
and Maier, 1996). A recent systematic review properly adjust for confounding and interactions.
analysed risk of HNC and occupational expo- One case–control study, in the USA, reported
sure to formaldehyde, wood dust, metal, coal an increased risk of HNC in regular marijuana
particles, and asbestos, but it included only few users (Zhang et al., 1999), whereas an analysis
studies on oral cancer (Awan et al., 2018). from the INHANCE consortium (Marks et al.,
2014) found no such risk.
2.2.2 Lifestyle factors
(d) Opium consumption
(a) Maté drinking The IARC Monographs programme recently
Maté is a beverage prepared from the leaves evaluated the carcinogenicity of opium consump-
of the Ilex paraguariensis plant and is usually tion, smoked or ingested (IARC, 2021). One
drunk very hot with a metal straw in Argentina, ecological study, one case–control study, and
southern Brazil, Chile, Paraguay, and Uruguay. one large case series (Fahmy et al., 1983; Razmpa
The IARC Monographs programme concluded et al., 2014; Rashidian et al., 2016) reported that
that drinking very hot beverages – at temper- opium use was associated with increased risk
atures above 65 °C – is probably carcinogenic of oral cancer; however, these studies had some
to humans (Group 2A) (IARC, 2018). Two limitations, and the evidence was considered to
meta-analyses reported a significant associ- be inadequate (Warnakulasuriya et al., 2020).
ation between maté drinking and oral cancer
(OR, 2.11; 95% CI, 1.39–3.19) (Dasanayake et al., (e) Mouthwash use
2010) and oral and oesophageal cancers (OR, Several case–control studies have examined
1.49; 95% CI, 1.08–2.05) (Mello et al., 2018b). the risk of mouthwash use for the causation of
The 2018 World Cancer Research Fund (WCRF) oral cancer. Several reviews and meta-analyses
reported that the evidence suggesting that greater were performed, which reported conflicting
consumption of maté increases the risk of oral evidence (Lewis and Murray, 2006; McCullough
cancer is limited (WCRF, 2018). and Farah, 2008; La Vecchia, 2009; Gandini et al.,
2012; Currie and Farah, 2014). A risk quantitative
(b) Khat chewing meta-analysis (Gandini et al., 2012) and an inde-
Khat (Catha edulis Forsk), also known as qat, pendent expert group assembled by the United
is consumed in Yemen and in East Africa, partic- States Food and Drug Administration (FDA,
ularly in Somalia and Ethiopia, as well as in the 2003) found no excess risk of oral cancer from
global diaspora from this region. Although khat use of mouthwash containing or not containing
chewing has detrimental effects on teeth and the alcohol. However, daily use of mouthwash over
periodontium, a systematic review (El-Zaemey a prolonged period (> 35 years) was suggested to
et al., 2015) and a narrative review (Al-Maweri cause oral cancer by an international consortium
et al., 2018) did not demonstrate any significant (Boffetta et al., 2016). [It is likely that people with
association between khat use and oral cancer. oral cancer may use mouthwashes to mask their
halitosis or to control symptoms of the disease.
(c) Cannabis smoking In many of the case–control studies, reverse
Evidence is lacking on the association causation was not considered.]
between smoking of cannabis (also called
marijuana) and oral cancer. Cannabis smoking
is often combined with heavy tobacco use and
alcohol consumption, which makes it difficult to
115
2.2.3 Demographic factors (mostly in the USA) also found that ill-fitting
dentures substantially increased the risk of oral
Studies conducted in the United Kingdom cancer (OR, 3.90; 95% CI, 2.48–6.13) (Manoharan
and in several countries in Europe indicate that et al., 2014).
most patients with oral cancer have lower socio-
economic status, live in low-resource settings, or (b) Oral hygiene
have jobs with low occupational social prestige
Several studies have provided evidence that
(Woolley et al., 2006; Conway et al., 2008, 2021).
advanced periodontal disease due to poor oral
Also, patients with oral cancer living in deprived
hygiene may be an independent risk factor for
areas had an increased risk of death from oral
oral cancer and HNC (Guha et al., 2007; Meyer
cancer (RR, 1.28; 95% CI, 1.11–1.47) compared
et al., 2008). Bleeding gums (OR, 3.94; 95% CI,
with people living in affluent areas (Edwards and
2.49–6.25) and dental check-ups only at the
Jones, 1999).
time of pain (OR, 3.84; 95% CI, 2.38–6.20) were
In contrast, a study in Brazil reported no
both associated with significantly increased
significant risk of oral cancer in people with lower
risk after adjustment for potential confounders
education levels (OR, 1.71; 95% CI, 0.74–3.96)
(Gupta et al., 2017). The INHANCE consortium
(Andrade et al., 2015). A study in Scotland
reported a strong association of poor oral health
was also inconclusive regarding the individual
with oral cancer (OR for worst oral health vs best
components of socioeconomic status and the risk
oral health, 3.12; 95% CI, 2.08–4.68) (Hashim
of HNC (Conway et al., 2010).
et al., 2016). Three meta-analyses reported that
periodontal disease (OR, 3.08; 95% CI, 1.60–3.93)
2.2.4 Orodental factors (Zeng et al., 2013a), tooth loss (OR, 1.72; 95% CI,
(a) Chronic mechanical irritation 1.26–2.36) (Zeng et al., 2013b), and infrequent
tooth brushing (OR, 1.73; 95% CI, 1.36–2.20)
Chronic mechanical irritation to the oral
(Zeng et al., 2015) were associated with increased
mucosa may, over a period of time, lead to
risk of oral cancer or head and neck squamous
OPMDs and oral cancer (Piemonte et al., 2010,
cell carcinoma.
2018). Because of loss of the protective barrier
of the mucosa, chronic mechanical irritation (c) Oral infections
arising from dental factors could facilitate the
Several reviews have examined the published
entry of carcinogens or infections into deeper
evidence on the relationship between the oral
layers of the squamous epithelium (Gilligan
microbiome and oral squamous cell carci-
et al., 2017).
noma (OSCC) (Whitmore and Lamont,
Poor dentition (faulty restorations, malpo-
2014; Gholizadeh et al., 2016; Perera et al.,
sitioned teeth, or sharp or broken teeth due to
2016; Chen et al., 2017). In multiple studies,
decay or fractures) and ill-fitting prosthesis have
significantly higher levels of Porphyromonas spp.
been associated with risk of oral cancer in several
and Fusobacterium spp. were found in OSCC
case–control studies (Lockhart et al., 1998; Velly
tissues than in healthy mucosa (Nagy et al.,
et al., 1998; Rosenquist, 2005; Vaccarezza et al.,
1998; Katz et al., 2011; Pushalkar et al., 2012). The
2010; Bektas-Kayhan et al., 2014; Huang et al.,
presence of specific species of bacteria in tumour
2015; Li et al., 2015; Chen et al., 2018; Piemonte
tissue (Zhang et al., 2020) adds strength to the
and Lazos, 2018) (Supplementary Table S2.14,
specificity of these studies.
High lipopolysaccharide levels in cancerous
iarc.fr/617). A meta-analysis based on 9 studies
conditions were indicative of Gram-negative
116
bacteria found in the subgingival microflora, 2.2.5 Systemic factors

which have lipopolysaccharide in their cell
wall, thus causing lipopolysaccharide-induced (a) Immunosuppression
inflammation (Kavarthapu and Gurumoorthy, Immunosuppression has also been shown to
2021). A systematic review of 14 in vitro studies be a mechanism that can lead to cancer (Baan
and 3 studies in animal models proposed a role et al., 2019). A few case series of secondary oral
of Porphyromonas gingivalis in the development cancer after allogeneic haematopoietic cell trans-
of OSCC through epithelial–mesenchymal tran- plantation or after renal transplantation have
sition of malignant cells, neoplastic prolifera- been published (King et al., 1995; van Leeuwen
tion, and tumour invasion (Lafuente Ibáñez de et al., 2009; Santarone et al., 2021). The studies of
Mendoza et al., 2020). Laprise et al. (2019) and van Leeuwen et al. (2009)
A nested case–control study conducted in confirmed that immunosuppressive agents
prospective studies in two populations in the (azathioprine and cyclosporine) used after organ
USA found that abundance of Corynebacterium transplantation may increase susceptibility to lip
and Kingella was associated with a decreased and oral cancer.
risk of head and neck squamous cell carcinoma, Patients with inflammatory bowel disorders
whereas Parvimonas micra and Neisseria sicca (e.g. Crohn disease) who may take long-term
were associated with a decreased risk of oral immunosuppressive agents (e.g. azathioprine)
cancer. However, an unnamed Actinomyces was may be at increased risk of tongue or oral cancer
associated with an increased risk of oral cancer (Li et al., 2003; Katsanos et al., 2016).
(Hayes et al., 2018). In a study conducted during the pandemic of
Several studies and meta-analyses have inves- HIV infection before the era of combined antire-
tigated the presence of Epstein–Barr virus in oral troviral therapy (cART), patients diagnosed with
carcinoma, with a reported prevalence ranging HIV disease did not have an increased risk of oral
from 0% to 100% (Acharya et al., 2015; She et al., cancer (Hille and Johnson, 2017). However, the
2017; de Lima et al., 2019). A meta-analysis of rate of HPV-associated HNC is higher in people
8 case–control studies reported a significant living with HIV (Beachler and D’Souza, 2013).
positive association between Epstein–Barr virus
infection and oral lichen planus (OLP) (Ashraf (b) Obesity, underweight, and body mass
et al., 2020). index
Candida is frequently present in oral biopsy Obesity is an established risk factor for
samples of moderate and severe dysplasia, and many cancer types (Arnold et al., 2016).
significant dysplastic changes have been noted The 2018 WCRF report, which analysed 25
in the epithelium of candidal leukoplakia studies, reported that obesity marked by BMI,
harbouring Candida species (McCullough et al., waist circumference, and waist-to-hip ratio
2002; Shukla et al., 2019). A recent systematic probably increased the risk of oral and pharyn-
review on candidal leukoplakia (Shukla et al., geal cancers (WCRF, 2018). In contrast, in a
2019) identified three studies, which reported pooled data analysis from 15 case–control
malignant transformation ratios of 2.5%, 6.5%, studies, ORs were increased in underweight
and 28.7%. (BMI < 18.5 kg/m2) compared with normal weight
(BMI, 18.5–24.9 kg/m2) and decreased in over-
weight and obese categories (BMI ≥ 25 kg/m2)
for oral cancer and other HNC; ORs were similar
in men and women (Lubin et al., 2011). A more
117
recent study from the INHANCE consortium increased the risk of oral cancer (OR, 1.53; 95%
also found that low BMI (i.e. < 18.5 kg/m2) was CI, 1.11–2.11) (Negri et al., 2009).
associated with higher risk of HNC (Gaudet Of the many familial cancer syndromes,
et al., 2015). patients with Fanconi anaemia, xeroderma
A study in Sri Lanka found that low BMI pigmentosum, Li–Fraumeni syndrome, Bloom
(< 18.5 kg/m2) was a significant independent syndrome, ataxia–telangiectasia, and Cowden
risk factor for the development of OPMDs syndrome have shown an increased suscepti-
(Amarasinghe et al., 2013). bility to oral cancer due to genetic instability, and
those with Fanconi anaemia have the strongest
(c) Metabolic syndrome predisposition (Furquim et al., 2018; Amenábar
In two studies of people with metabolic et al., 2019). Dyskeratosis congenita (also called
syndrome (Chang et al., 2015b; Siewchaisakul Zinsser–Cole–Engman syndrome) is a rare
et al., 2020) the condition was found to be signif- hereditary condition with predisposition to
icantly associated with OPMDs. Three compo- leukoplakia of the tongue that could transform
nents of metabolic syndrome were reported to into cancer in early life (Handley and Ogden,
be significantly associated with OPMDs: central 2006).
obesity, hypertriglyceridaemia, and hypergly- A genome-wide association study of oral and
caemia (Siewchaisakul et al., 2020). pharyngeal cancers with 6034 cases and 6585
controls in Europe, North America, and South
(d) Haematinic and micronutrient deficiency America detected 8 loci (regions) contributing
Haematinic deficiency (e.g. deficiency of iron, to susceptibility to oral and pharyngeal cancers.
folate, or vitamin B12) can cause histopatholog- Oral cancer was associated with two new regions
ical changes in the oral mucosa and/or clinically (2p23.3 and 9q34.12) and with known cancer
detectable OPMDs, presumably by interfering loci (9p21 and 5p15.33). Oral and pharyngeal
in epithelial proliferation and/or maturation cancers combined were associated with loci at
(Ranasinghe et al., 1983). A recent study reported 6p21.32, 10q26.13, and 11p15.4 (Lesseur et al.,
significantly higher frequencies of haematinic 2016).
deficiencies and hyperhomocysteinaemia in The TP53 codon 72 polymorphism has
patients with OPMDs than in healthy controls been suggested to play a role in cancer suscep-
(Wu et al., 2019). tibility, and more specifically susceptibility to
HPV-associated cancers. An association between
2.2.6 Familial or genetic predisposition p53 gene variants and oral cancer susceptibility
was reported in India (Patel et al., 2013). A study
Sporadic case reports proposed that oral in Argentina reported that the frequency of TP53
cancer could be familial (Ankathil et al., 1996). codon 72 Pro72variant was higher in patients
A case–control study in Italy and Switzerland with OSCC and OPMDs than in controls (Zarate
reported that a family history of oral cancer, et al., 2017), and a study in China (Hou et al.,
pharyngeal cancer, or laryngeal cancer is a 2015) reported that p53 Arg72Pro polymor-
strong determinant of risk of oral and phar- phism together with HPV infection may jointly
yngeal cancer, independent of tobacco use and alter an individual’s susceptibility to oral cancer.
alcohol consumption (Garavello et al., 2008). A meta-analysis of 11 studies suggested that in
The INHANCE consortium reported that a the absence of HPV infection the TP53 codon
family history of cancer in first-degree relatives 72 polymorphism (Arg vs Pro) is not associated
with the risk of OSCC (Zeng et al., 2014).
118
2.3 Impact upon quitting 2.3.1 Tobacco smoking

For the evaluation of studies in humans on (a) Risk of oral cancer and oropharyngeal
the potential reduction in cancer risk due to cancer
reduction or cessation of exposure to a risk factor Volume 11 of the IARC Handbooks of Cancer
for oral cancer, intervention studies, cohort Prevention evaluated the scientific evidence
studies, case–control studies, and cross-sec- available until the first trimester of 2006 on the
tional studies were eligible for inclusion. The effects of smoking cessation on the risk of cancer
selection was limited to studies of established (IARC, 2007b). The Working Group concluded
risk factors, i.e. tobacco smoking, consumption that for oral and pharyngeal cancer, the risk
of alcoholic beverages, use of smokeless tobacco, “is lower in former smokers than in otherwise
and chewing of areca nut (including betel quid) similar current smokers”, the relative reduction
with added tobacco or without tobacco [hereafter in risk increases with duration of quitting, and
described as the exposure]. Only studies that the RR after ≥ 2 decades of smoking cessation
evaluated separately the effect on cancer of the returns to that in never-smokers (IARC, 2007b).
oral cavity, or of the oral cavity and the pharynx
(i) Overview of studies
combined (oropharynx and/or hypopharynx)
were included. Studies of cancer incidence and The Working Group assessed all the avail-
cancer mortality were eligible for inclusion. able studies published since 2006. Studies that
In addition, studies on OPMDs, such as oral reported risk estimates in former smokers by
leucoplakia or erythroplakia, were included as time since quitting smoking were considered
supporting evidence. to be more informative and included individual
Only those studies that compared former cohort studies (Freedman et al., 2007; Maasland
exposure and current exposure with never expo- et al., 2014), a pooled analysis of 17 case–control
sure, and former exposure with current exposure, studies (Marron et al., 2010), a pooled analysis of
were included. Studies that compared former 2 case–control studies (Bosetti et al., 2008), and 4
exposure versus never exposure but not current individual case–control studies (De Stefani et al.,
exposure versus never exposure were excluded. 2007; Lee et al., 2009; Varela-Lema et al., 2010;
No studies reported on reduction of exposure Radoï et al., 2013a). Two mortality cohort studies
and risk of cancer or OPMDs. that included former smokers but did not report
For the evaluation of cessation of chewing risk estimates by duration of smoking cessation
areca nut with added tobacco and chewing areca were identified (Ide et al., 2008; Christensen
nut without tobacco, in addition to the analyses et al., 2018). Most of the studies included male
in published studies, Working Group performed and female participants; two studies included
primary analyses of unpublished data on the only male participants (De Stefani et al., 2007;
associations with risk of oral cancer or risk of Varela-Lema et al., 2010).
OPMDs. Table 2.15 shows the number of analyses The studies varied with respect to the defini-
for each exposure, by study design; some studies tions of study population, cancer outcome, and
contributed evidence to more than one group. former smoker, the categorization of time since
quitting smoking, the reference group used to
estimate RRs, and the extent of adjustment for
potential confounders. The definition of former
smoker, when available, varied from having quit
smoking ≥ 6 months before enrolment to having
119
Table 2.15 Number of studies that assess quitting exposure to the risk factor and reduction in
risk of oral cancer or OPMDs
Risk factor Type of studies Number of studies
Oral cavity or OPMDs
oral cavity and
pharynx
Tobacco smoking Cohort 4 1
Case–control 4 6
Cross-sectional 0 1
Pooled analysis (of case–control studies) 2 0
Meta-analysis 0 0
Alcoholic beverage Cohort 3 0
consumption Case–control 6 7
Pooled analysis (of case–control studies) 1 0
Meta-analysis 0 0
Smokeless tobacco use Cohort 2 4
Case–control 4 2
Cross-sectional 0 2
Pooled analysis 0 0
Meta-analysis by the Working Group (of cohort studies 1 1
and case–control studies)
Chewing areca nut products Cohort (published/primary analysisa) 2/1 1
(including betel quid) with Case–control (published/primary analysisa) 3/1 2
added tobacco Pooled analysis 0 0
Meta-analysis (of cohort studies and case–control studies) 1 0
Chewing areca nut products Cohort (published/primary analysisa) 0/3 0/3
(including betel quid) Case–control (published/primary analysisa) 4/1 3/1
without tobacco Cross-sectional 0 2
Meta-analysis (of case–control studies) 1 0
OPMDs, oral potentially malignant disorders.
a Primary analyses of unpublished data performed by the Working Group.
quit > 2 years before enrolment. The duration risk of oropharyngeal cancer alone or risk of
of smoking cessation was reported in at least oropharyngeal cancer death. In most studies, the
two categories, usually using a cut-off point smoked tobacco product was cigarettes.
of 10 years; few studies used more categories (ii) Cohort studies
of duration of smoking cessation. Few studies
controlled for cumulative smoking or presented See Table 2.16.
estimates by time since quitting smoking Freedman et al. (2007) reported on the associ-
stratifying by quantity smoked or cumulative ation of smoking status and HNC in men and in
smoking. Only the pooled analysis and three women in the prospective United States National
case–control studies used current smokers as Institutes of Health-American Association of
the reference group to assess reductions in RR Retired Persons (NIH-AARP) Diet and Health
associated with quitting smoking. Outcomes of Study, which enrolled 476 211 participants from
oral cancer, oropharyngeal and hypopharyngeal October 1995 until the end of 2000. Former
cancer, pharyngeal cancer, and oral and pharyn- smokers were defined as people who had quit
geal cancer were used to report RRs associated smoking > 1 year before the date of completing
with smoking cessation. No studies reported
120
Table 2.16 Cessation of tobacco smoking and risk of oral cancer and/or pharyngeal cancer – cohort studies
Reference Study population, number Cancer end-point Smoking Number of RR or HR Adjustments/comments

Location of participants, follow-up and smoking cases/number in (95% CI)
period cessation cohort
metrics
Cancer incidence
Freedman Prospective NIH-AARP SCC of the oral Cigarette Oral cancer: HR: Current smokers included
et al. (2007) Diet and Health Study, cavity (lips, tongue, smoking: regular smokers and people
USA following up 283 691 gums, palate, floor Men: who stopped smoking within
men and 192 520 women; of the mouth, and Never-smokers 54 1.0 (ref) the year before enrolment
aged 50–71 yr, in 6 states other parts of the Estimates adjusted for age
Current smokers 71 2.99 (2.05–4.38)
of the USA, from 1995 mouth) and oro- at entry into cohort, BMI,
until the end of 2000; 759 hypopharynx Former smokers 104 1.00 (0.72–1.40) education level, alcohol
head and neck cancers, (oropharynx, tonsils, Duration of cessation (yr): consumption, vigorous
310 oral cancers, and hypopharynx, 1–4 18 2.49 (1.45–4.28) physical activity, usual activity
139 oropharyngeal/ pyriform sinus, 5–9 17 1.29 (0.74–2.25) throughout the day, fruit
hypopharyngeal cancers and pharynx not ≥ 10 69 0.83 (0.58–1.19) intake, vegetable intake, and
were diagnosed otherwise specified) total energy
Ptrend < 0.001
Women:
Never-smokers 14 1.0 (ref)
Current smokers 42 7.57 (4.02–14.28)
Former smokers 25 2.10 (1.08–4.06)
Duration of cessation (yr):
1–4 8 6.18 (2.57–14.86)
5–9 4 1.88 (0.62–5.75)
≥ 10 13 1.53 (0.72–3.27)
Ptrend < 0.001
Cigarette Oro/hypopha-
smoking: ryngeal cancer:
Men:

Former smokers 49 1.52 (0.86–2.70)
121
122

metrics
Freedman Duration of cessation (yr):
et al. (2007) 1–4 8 3.42 (1.45–8.07)
(cont.) 5–9 13 3.05 (1.45–6.40)
≥ 10 28 1.10 (0.59–2.05)
Ptrend < 0.001
Women:
Current smokers 16 11.39 (3.21–40.40)
Former smokers 14 5.29 (1.50–18.61)
1–4 4 12.57 (2.78–56.86)
5–9 3 6.11 (1.22–30.60)
≥ 10 7 3.81 (0.98–14.89)
Ptrend < 0.001
Maasland The Netherlands Cohort Microscopically Smoking: Oral cancer: Former smoker status
et al. (2014) Study, initiated in 1986, confirmed SCC Never-smokers 29 1.0 (ref) not defined, but from
The enrolled 120 852 men and of the head and Current smokers 57 2.03 (1.16–3.56) categorization of the variable
Netherlands women aged 55–69 yr neck, including the “years since quitting” recorded
Former smokers 24 –
from 204 municipal oral cavity and the at baseline, it is evident that
population registers in the oropharynx and Duration of cessation (yr): people who quit within the
Netherlands. In 17.3 yr of hypopharynx > 0 – < 10 11 0.84 (0.39–1.83) year of enrolment or earlier
follow-up, 110 oral cancers 10 – < 20 8 0.78 (0.32–1.86) were considered former
and 83 oropharyngeal/ ≥ 20 5 0.63 (0.22–1.81) smokers. Estimates adjusted
hypopharyngeal cancers Ptrend < 0.004 for age (years), sex, and alcohol
were diagnosed consumption (grams of ethanol
Smoking: Oro/hypopha- per day; continuous). Analysis
ryngeal cancer: by duration of cessation also
Never-smokers 6 1.0 (ref) adjusted by pack-years of
Current smokers 55 8.10 (3.14–20.87) cigarette smoking (continuous)
Former smokers 22 –
metrics
Maasland Duration of cessation (yr)
et al. (2014) > 0 – < 10 8 2.48 (0.77–7.93)
(cont.) 10 – < 20 8 3.29 (1.04–10.39)
≥ 20 6 3.35 (0.97–11.55)
Ptrend < 0.001
Cancer mortality
Ide et al. The Japan Collaborative Annual Smoking status: Oral and Current or former smokers not
(2008) Cohort Study for Evaluation ascertainment of pharyngeal defined
Japan of Cancer Risk covered 45 oral and pharyngeal cancer deaths: RR of death adjusted for
geographical areas in the cancer deaths, Men: age, alcohol consumption,
country, enrolling 46 465 identified by ICD-10 Non-smokers 5 1.0 (ref) consumption of green tea,
men and 64 327 women aged codes C01–C14, preference for salty foods, and
40–79 yr in 1988–1990, with excluding C07–C08 consumption of green and
12.5 yr of follow-up and (salivary gland Former smokers 7 0.9 (0.3–3.0) yellow vegetables
identification of 52 oral and cancer) and C11
pharyngeal cancer deaths (nasopharyngeal
(41 in men) cancer)
Christensen The National Longitudinal Lip, oral, and Exclusive Oral and Former smokers were defined
et al. (2018) Mortality Study pharyngeal cancer cigarette pharyngeal as people who had ever smoked
USA included a representative deaths (ICD-10 codes smoking: cancer deaths: ≥ 100 cigarettes but were
sample of civilian, non- C00–C14) Never-smokers 31 1.0 (ref) non-smokers at the time of the
institutionalized men and Current smokers 79 9.02 (5.78–14.09) baseline survey
women aged 35–80 yr Risk of death (HR) adjusted
Former smokers 50 2.70 (1.66–4.39)
(n = 357 420) who completed for age, sex, race/ethnicity,
the Tobacco Use Supplement education level, and year of
of the national Current survey
Population Survey starting Estimates not adjusted for

in 1985, with death alcohol consumption, and
ascertainment until the end therefore probably confounded
of 2011
BMI, body mass index; CI, confidence interval; HR, hazard ratio; ICD, International Classification of Diseases; NIH-AARP, United States National Institutes of Health-American
Association of Retired Persons; ref, reference; RR, relative risk; SCC, squamous cell carcinoma; yr, year or years.
123
the baseline questionnaire, which also recorded Maasland et al. (2014) reported on the
time since quitting smoking. Netherlands Cohort Study, which was initiated
The RRs of oral cancer in former smokers in 1986 and enrolled 120 852 men and women
decreased progressively with increasing time aged 55–69 years from 204 Dutch municipal
since quitting smoking in men (from HR for population registers. Follow-up for cancer inci-
1–4 years since quitting, 2.49; 95% CI, 1.45–4.28 dence, extended until 2003, was done through
to HR for > 10 years since quitting, 0.83; 95% CI, annual record linkage to the Netherlands
0.58–1.19) and in women (from HR for 1–4 years Cancer Registry and the nationwide network of
since quitting, 6.18; 95% CI, 2.57–14.86 to HR pathology registries. Former smoker status was
for > 10 years since quitting, 1.53; 95% CI, not defined. The RR estimates for oral cancer in
0.72–3.27); these estimates were lower than the former smokers by time since quitting smoking
RRs in current male smokers (HR, 2.99; 95% were < 1, and the CIs included 1. A tendency of
CI, 2.05–4.38) and current female smokers (HR, decreasing RR with increasing duration of quit-
7.57; 95% CI, 4.02–14.28). RRs of oral cancer were ting was observed, from RR for > 0 to < 10 years
steadily higher in women than in men, whether in since quitting, 0.84 (95% CI, 0.39–1.83) to RR
former smokers or in current smokers compared for ≥ 20 years since quitting, 0.63 (95% CI,
with never-smokers. [A larger proportion of oral 0.22–1.81); for current smokers, RR was 2.03
cancers in men (23%) than in women (17%) were (95% CI, 1.16–3.56; Ptrend < 0.004). A similar
diagnosed in never-smokers, which may suggest tendency of decreasing RR with increasing dura-
that there are factors increasing the background tion of quitting was observed for oropharyngeal
risk in men more than in women, and this differ- and hypopharyngeal cancer; the magnitude
ential appears to lower the magnitude of the RRs of the RR at any duration of quitting was still
compared with never-smokers reported in men elevated in former smokers with respect to never-
with respect to the RRs reported in women.] smokers but was substantially lower than the RR
The elevated RRs of oropharyngeal and in current smokers.
hypopharyngeal cancer in former smokers Two cohort studies reported risk of death
compared with never-smokers decreased with in former smokers and current smokers using
increasing time since quitting smoking in men non-smokers as the reference group (Ide
(from HR for 1–4 years since quitting, 3.42; 95% et al., 2008; Christensen et al., 2018). The Japan
CI, 1.45–8.07 to HR for > 10 years since quitting, Collaborative Cohort Study for Evaluation of
1.10; 95% CI, 0.59–2.05) and in women (from Cancer Risk, conducted in 45 geographical areas
HR for 1–4 years since quitting, 12.6; 95% CI, in the country, enrolled 46 465 men and 64 327
2.78–56.86 to HR for > 10 years since quitting, women who were followed up for an average of
3.81; 95% CI, 0.98–14.89); although these esti- 12.5 years (Ide et al., 2008). In men, the RR of oral
mates remained elevated, they were of lower and pharyngeal cancer death in former smokers
magnitude than the RRs in current male smokers compared with non-smokers was 0.9 (95% CI,
(HR, 5.29; 95% CI, 2.88–9.73) and current female 0.3–3.0), and the risk of death in current smokers
smokers (HR, 11.39; 95% CI, 3.21–40.40). was more than twice that in non-smokers (RR,
[The Working Group noted that this is one of 2.6; 95% CI, 1.0–6.7). In women, the risk of oral
the very few studies that investigated the associ- and pharyngeal cancer death in current smokers
ation with quitting smoking separately in men compared with non-smokers was substantially
and in women, and cautioned about interpreting higher (RR, 8.2; 95% CI, 2.1–32.1). [The Working
differences in RR by sex.] Group noted the lack of a definition of former
smoker and the absence of deaths in female
124
former smokers, which precluded the generation in risk became more pronounced the longer
of a mortality risk estimate. No estimates by time the cessation interval (Ptrend < 0.01). In recent
since quitting were available.] quitters (from 13 months to 4 years since quit-
The National Longitudinal Mortality Study ting), the OR was 0.65 (95% CI, 0.52–0.80). With
includes a representative sample of the civilian, ≥ 20 years since quitting, the RR decreased to
non-institutionalized population of the USA, 0.19 (95% CI, 0.15–0.24), a RR similar in magni-
including men and women. For the analysis tude and precision to the RR reported for never-
reported by Christensen et al. (2018), cohort smokers (OR, 0.19; 95% CI, 0.14–0.27). Similarly,
members who completed the tobacco use for oropharyngeal and hypopharyngeal cancers
questionnaire included 357 420 participants combined, the magnitude of the reduction in
(excluding exclusive smokeless tobacco users risk increased progressively with longer time
and users of multiple types of tobacco). Former since quitting (Ptrend < 0.01). The reduction
smokers were defined as people who had ever in risk was already evident in recent quitters
smoked ≥ 100 cigarettes but were non-smokers (OR for > 1–4 years since quitting, 0.72; 95%
at the time of the survey. The definition of former CI, 0.52–1.00) and became more pronounced
smoker did not specify the duration of cessation. the longer the cessation interval, until the RR
The RR of death from oral and pharyngeal reached that in never-smokers after ≥ 20 years of
cancer in former smokers was almost 3 times that cessation. [The Working Group recognized the
in never-smokers (RR, 2.70; 95% CI, 1.66–4.39) large sample size of this pooled study based on
and was much lower than the RR of death in harmonized data collected in countries encom-
current smokers (RR, 9.02; 95% CI, 5.78–14.09). passing a wide geographical distribution. Risk
[The Working Group noted that, given that study estimates were adjusted for alcohol consump-
participants were classified as former smokers or tion and cumulative smoking for oral cancer
current smokers at baseline and cancer mortality and oro-hypopharyngeal cancer by time since
was ascertained years later, changes in smoking quitting smoking. Current smokers were used as
status during follow-up could have introduced the reference group, and reduction in risk was
misclassification of exposure in the cohort, reported in a dose-dependent manner, including
which could lead to underestimation or overesti- cessation intervals of ≥ 20 years.]
mation of the reported risks. Risk estimates may In addition to the data included in the pooled
be confounded by lack of adjustment for alcohol analysis (Marron et al., 2010), Bosetti et al. (2008)
consumption.] reported RR estimates of oral and pharyngeal
cancers combined in former smokers by age at
(iii) Case–control studies
quitting using current smokers as the reference
See Table 2.17. group and using data from two hospital-based
Marron et al. (2010) reported on a large case–control studies in Italy. The risk of oral
individual-level data pooled analysis of 17 and pharyngeal cancer decreased with quitting
case–control studies exploring the association smoking irrespective of the age at quitting, and
of smoking cessation and HNC within the the magnitude of the reduction in risk decreased
INHANCE consortium, reporting ORs for oral progressively with lowering of the age at quitting
cancer and oro-hypopharyngeal cancer by time smoking, from OR for quitting at age 55–64 years
since quitting smoking using current smokers of 0.48 (95% CI, 0.34–0.66) to OR for quitting at
as the reference group. The risk of oral cancer age < 35 years of 0.14 (95% CI, 0.08–0.26). [The
decreased with quitting smoking compared Working Group noted that this is the only study
with continuing smoking, and the reduction
125
126

Table 2.17 Cessation of tobacco smoking and risk of oral cancer and/or pharyngeal cancer – case–control studies
Reference Study population, number of Cancer end- Smoking and smoking Number of RR or OR Adjustments/comments
Location participants, study period point cessation metrics cases/controls (95% CI)
Marron et al. Pooled analysis of the Invasive tumour Oral cancer: Former smokers include
(2010) INHANCE consortium of the oral cavity, Smoking: meta-OR: people who had quit smoking
Central of 17 hospital-based and oropharynx, Current smokers 2256/5183 1.0 (ref) cigarettes, cigars, or pipe for
Europe, population-based case– hypopharynx, > 1 year as of date of diagnosis
Former smokers 583/5009 0.30 (0.26–0.34)
France, Italy, control studies (including or oral cavity or date of interview
Switzerland, men and women) accruing or pharynx Duration of cessation Risk estimates adjusted for
Latin a total of 3302 oral cancer not otherwise (yr): age, sex, race/ethnicity, study
America, cases, 3989 oropharyngeal or specified > 1–4 156/620 0.65 (0.52–0.80) centre, education level, pack-
Puerto Rico, hypopharyngeal cancer cases, 5–9 129/836 0.43 (0.32–0.58) years of tobacco smoking,
USA and 16 377 controls. Most cases 10–19 144/1582 0.25 (0.21–0.31) and frequency of alcohol
were diagnosed with SCC ≥ 20 154/1971 0.19 (0.15–0.24) consumption
Never-smokers 463/6186 0.19 (0.14–0.27)
Ptrend < 0.01
Oro/hypopha-
ryngeal cancer:
Current smokers 2565/5183 1.0 (ref)
Former smokers 957/5009 0.41 (0.32–0.53)
Duration of cessation
(yr):
> 1–4 260/620 0.72 (0.52–1.00)
5–9 198/836 0.51 (0.38–0.67)
10–19 272/1582 0.36 (0.27–0.49)
≥ 20 281/1971 0.29 (0.19–0.43)
Never-smokers 467/6186 0.25 (0.15–0.42)
Ptrend < 0.01
Radoï et al. Multicentre population-based Incident and Oral cancer: Former smokers were people
(2013a) case–control study of UADT histology- or Any smoking: who had stopped smoking
France and lung cancer (ICARE) cytology- Never-smokers 62/1262 1.0 (ref) for ≥ 2 yr before the study
conducted in 10 departments in confirmed SCC interview. Current smokers
Current smokers 537/820 9.8 (7.0–16.6)
France with cancer registration of the oral cavity included people who had
(2002–2007), including men including the Former smokers 171/1464 – stopped recently (within < 2 yr
and women. Of 968 oral cancer floor of the Duration of cessation of the date of the interview)
cases contacted, 792 (82%) mouth, mobile (yr): Estimates adjusted for age,
completed the questionnaire tongue, base of 2–9 90/318 3.9 (2.7–5.9) sex, area of residence, pack-
and 772 cases aged ≤ 75 yr the tongue, soft 10–19 42/384 2.1 (1.3–3.3) years of smoking, (continuous
were included. Controls were palate, gums, 20–29 22/413 1.3 (0.7–2.2) variable), and alcohol
randomly selected from the hard palate, and consumption (categories of
≥ 30 15/346 1.6 (0.9–3.0)
population by random-digit other parts of the grams per day)
dialling; 3555 (80.6%) were mouth (ICD-10
included codes C01–C06)
De Stefani Hospital-based case–control Microscopically Smoking: Oral cancer: Current smokers include
et al. (2007) study enrolling study confirmed SCC Current smokers 261/639 1.0 (ref) people who smoked at the
Montevideo participants (men only) in of the mouth and Former smokers, time of the interview or had
(Uruguay) 4 hospitals (1988–2000), pharynx duration of cessation quit smoking ≤ 1 yr before the
including 335 oral cancer (yr): date of the interview. Smokers
and 441 pharyngeal cancer who had quit > 1 yr before
≤9 47/182 0.65 (0.44–0.94)
cases and 1501 controls with the interview were considered
non-neoplastic conditions 10–19 10/146 0.16 (0.08–0.32) former smokers
not related to tobacco use or ≥ 20 9/160 0.15 (0.07–0.31) Estimates adjusted for age,
alcohol consumption Never-smokers 8/374 0.08 (0.04–0.16) residence, urban/rural status,
Ptrend < 0.0001 hospital, year at diagnosis,
Pharyngeal education level, family history
cancer: of cancer among first-degree
relatives, occupation, total
Current smokers 340/639 1.0 (ref)
consumption of vegetables and

Former smokers, fruits, maté intake, and alcohol
duration of cessation consumption. No adjustment
(yr): for intensity or duration of
≤9 63/182 0.64 (0.45–0.91) smoking
10–19 18/146 0.22 (0.13–0.39)
≥ 20 15/160 0.22 (0.12–0.40)
Never-smokers 5/374 0.04 (0.01–0.10)
Ptrend < 0.0001
127
128

Lee et al. Multicentre hospital-based Histology- or Oral and Former smokers included
(2009) case–control study (ARCAGE) cytology- pharyngeal people who had stopped
Czech of aerodigestive tract cancer, confirmed SCC cancer: smoking ≥ 12 months before
Republic, including men and women, of the oral cavity Any smoking: enrolment
Croatia, enrolling 993 cases of oral or or the pharynx Never-smokers 109/712 1.0 (ref) Estimates adjusted for age,
France, oropharyngeal cancer and (excluding the sex, education level, centre,
Current smokers 660/715 5.83 (4.50–7.54)
Germany, 2221 controls (1987–1992; nasopharynx) and alcohol consumption
Greece, 2002–2005) with conditions Former smokers 205/741 1.80 (1.37–2.37) frequency (continuous) and
Ireland, Italy, not related to tobacco use or Duration of cessation duration (continuous)
Norway, alcohol consumption. In this (yr):
Spain, and analysis, 974 cases and 2168 < 20 yr, > 0–20 pack- 40/132 2.13 (1.40–3.25)
the United controls were included years
Kingdom < 20 yr, > 20 pack-years 106/247 3.05 (2.19–4.25)
≥ 20 yr, > 0–20 pack- 40/263 1.09 (0.73–1.64)
years
≥ 20 yr, > 20 pack-years 19/95 1.49 (0.84–2.63)
Current smokers:
> 0–20 pack-years 102/219 3.42 (2.45–4.78)
21–40 pack-years 257/258 6.65 (4.95–8.93)
> 40 pack-years 298/244 8.46 (6.22–11.5)
Ptrend < 0.001
Varela-Lema Hospital-based case–control Incident Oral and Former smokers defined
et al. (2010) study enrolling men aged histopathology- pharyngeal as people who had quit
Galicia > 20 yr with newly diagnosed confirmed cancer: > 6 months before the date of
(Spain) oral or pharyngeal cancer and primary oral Smoking: the study interview
controls from consecutive or pharyngeal Current smokers 73/67 1.0 (ref) Estimates adjusted for age,
patients to undergo surgery cancer (ICD-10 lifetime tobacco consumption,
not related to tobacco use or codes C00–C14), Former smokers, alcohol consumption in grams
alcohol consumption at the excluding the lip duration of cessation per week, high-risk occupation,
same hospital (1996–2000), (yr): and education level
including 92 cancer cases and 1–10 10/31 0.6 (0.2–1.5)
230 controls > 10 7/50 0.3 (0.1–0.9)
Bosetti et al. Data from 2 multicentre Incident Oral and Former smokers defined
(2008) hospital-based case–control histologically pharyngeal as people who had stopped
Milan, studies on UADT cancer confirmed oral cancer: smoking ≥ 12 months before
Pordenone, conducted in 1984–1997 in and pharyngeal Smoking: enrolment and at age < 65 yr
Rome (Italy) northern and central Italy. cancer Current smokers 712/1176 1.0 (ref) Reference group included
Analysis shown restricted to current smokers and former
Former smokers, age at
enrolled men aged < 75 yr. 961 smokers who had quit at age
quitting:
cases of oral and pharyngeal ≥ 65 yr
cancer and 2824 controls 55–64 yr 75/203 0.48 (0.34–0.66) Estimates adjusted for age,
included. This study population 45–54 yr 90/301 0.36 (0.27–0.48) centre, education level, and
is included in the INHANCE 35–44 yr 45/279 0.20 (0.14–0.29) alcohol consumption
consortium data set, but < 35 yr 13/162 0.14 (0.08–0.26)
analysis by age at quitting is not
reported in Marron et al. (2010)
ARCAGE, Alcohol-Related Cancers and Genetic Susceptibility in Europe; CI, confidence interval; ICARE, Investigation of Occupational and Environmental Causes of Respiratory
Cancers; ICD, International Classification of Diseases; INHANCE, International Head and Neck Cancer Epidemiology; OR, odds ratio; ref, reference; RR, relative risk; SCC, squamous
cell carcinoma; UADT, upper aerodigestive tract; yr, year or years.

129
identified that documents the impact of age at CI, 0.07–0.31). Similarly, the risk of pharyngeal
quitting on the RR reduction.] cancer in former smokers was lower than that
Radoï et al. (2013a) reported on a multi- in current smokers within 9 years of quitting
centre population-based case–control study smoking (OR, 0.64; 95% CI, 0.45–0.91) and
(the Investigation of Occupational and continued to decrease with longer time since
Environmental Causes of Respiratory Cancers quitting (OR for ≥ 20 years of cessation, 0.22;
[ICARE] study) of upper aerodigestive tract 95% CI, 0.12–0.40). [The Working Group noted
cancer, including oral cancer, conducted in the high participation rates of eligible cases and
2002–2007 in 10 departments in France with controls, the generation of fully adjusted risk
cancer registration, including male and female estimates, including by alcohol consumption,
participants. The ICARE study documented time and the use of a definition of former smoker that
since quitting smoking in former smokers and classified smokers quitting within 1 year of the
used never-smokers as the reference group. The date of cancer diagnosis or interview as current
RR of oral cancer in former smokers decreased smokers rather than former smokers, which
in magnitude with increasing time since quit- reduced the possible distortion of risk estimates
ting smoking but remained significantly elevated by exposure misclassification. The Working
with up to 19 years since quitting (OR for Group also observed that the ORs for former
2–9 years of quitting, 3.9; 95% CI, 2.7–5.9; OR for smokers by time since quitting smoking were
10–19 years of quitting, 2.1; 95% CI, 1.3–3.3; OR not adjusted for intensity, duration, or cumula-
for 20–29 years of quitting, 1.3; 95% CI, 0.7–2.2; tive past smoking.]
OR for ≥ 30 years of quitting, 1.6; 95% CI, 0.9–3.0) Lee et al. (2009) reported on a multicentre
[no trend reported]. The estimates were markedly international hospital-based case–control study
lower than the RR in current smokers (OR, 9.8; of aerodigestive tract cancer (the Alcohol-
95% CI, 7.0–16.6). [The Working Group acknowl- Related Cancers and Genetic Susceptibility
edged the moderate sample size of this study, in Europe [ARCAGE] study), which collected
which used robust definitions of former smoker information on smoking and smoking cessation
and current smoker and enrolled participants interval in former smokers. The study enrolled
from a wide geographical distribution in France, male and female cases and controls in 10 coun-
and also generated risk estimates adjusted for tries (the Czech Republic, Croatia, France,
alcohol consumption and cumulative smoking Germany, Greece, Ireland, Italy, Norway, Spain,
by time since quitting smoking but used never- and the United Kingdom) in 2002–2005 (with
smokers as the reference group.] the exception of cases and controls in France,
De Stefani et al. (2007) conducted a male- recruited earlier). The RR of oral and pharyngeal
only hospital-based case–control study assessing cancer in former smokers decreased with longer
the effects of tobacco smoking and alcohol duration of cessation in people with equivalent
consumption on the occurrence of oral and cumulative pack-years of smoking, and the RR
pharyngeal cancers in Montevideo, Uruguay, in current smokers with similar smoking history
in 1988–2000. The risk of oral cancer in former was markedly higher (P < 0.001). For instance,
smokers was lower than that in current smokers for former smokers with > 0 to 20 pack-years of
within 9 years of quitting smoking (OR, 0.65; smoking, the ORs were 2.13 (95% CI, 1.40–3.25)
95% CI, 0.44–0.94) and decreased markedly with for < 20 years of quitting and 1.09 (95% CI,
longer time since quitting (OR for 10–19 years of 0.73–1.64) for ≥ 20 years of quitting, compared
smoking cessation, 0.16; 95% CI, 0.08–0.32; OR with an OR of 3.42 (95% CI, 2.45–4.78) in current
for ≥ 20 years of smoking cessation, 0.15, 95% smokers. The reduction in risk with increasing
130
time since quitting was observed for both cate- quit smoking for only ≥ 6 months by the time of
gories of cumulative smoking (0–20 pack-years enrolment, and risk estimates were combined for
and > 20 pack-years), but the magnitude of the oral and pharyngeal cancers.]
risk estimates was higher in former smokers
with higher cumulative smoking. [The Working (b) Risk of OPMDs
Group acknowledged the large size of this multi- See Table 2.18.
centre study based on European populations,
(i) Overview of studies
and the calculation of risk estimates by time
since quitting, using two categories of cumu- A group of studies, limited in sample size,
lative smoking; however, the study reported addressing cessation of tobacco smoking and
RR estimates for oral and pharyngeal cancers incidence of OPMDs was available to the Working
combined, precluding the identification of risk Group. These included one cohort study (Gupta
of oral cancer alone.] et al., 1995), six case–control studies (Macigo
Varela-Lema et al. (2010) reported on a et al., 1996; Hashibe et al., 2000a, b; Shiu et al.,
hospital-based case–control study in Santiago 2000; Fisher et al., 2005; Amarasinghe et al.,
de Compostela, Galicia, Spain, in 1996–2000 2010a; Li et al., 2011), and one cross-sectional
investigating the association between tobacco study (Pivovar et al., 2017). Most of these studies
smoking and oral cancer and/or pharyngeal included male and female participants; two
cancer in men. A total of 92 cases and 230 studies were based only on men (Gupta et al.,
controls were included in the analysis, which 1995; Pivovar et al., 2017). Most of these studies
combined cases of oral and pharyngeal cancer reported the RR of OPMDs or a specific OPMD
and considered two categories for time since quit- (i.e. leukoplakia or erythroplakia) in former
ting smoking: 1–10 years and > 10 years. Using smokers using never-smokers as the reference
current smokers as the reference group, the risk of group, and one study described effect estimates
oral and pharyngeal cancer decreased in former by time since quitting smoking (Macigo et al.,
smokers with > 10 years of quitting (OR, 0.3; 95% 1996).
CI, 0.1–0.9). This study also provided ORs using (ii) Intervention study
never-smokers as the reference group, gener- Gupta et al. (1995) reported on a very large
ating very high ORs in former smokers (OR, 4.8; cohort study in Ernakulam District in Kerala,
95% CI, 2.9–73.5) and in current smokers (OR, India, with a 10-year follow-up (Table 2.18).
34.5; 95% CI, 7.5–157.8), which included light Men accrued 77 681 person-years of observa-
and heavy consumers of alcohol; heavy alcohol tion, and women accrued 32 544 person-years
consumers were over-represented in current of observation. The prevailing risk factors in the
smokers. [Adjustment by alcohol consumption study population were bidi smoking and betel
may not have entirely controlled for the risk-po- quid chewing, along with commercial cigarette
tentiating effect of dual exposure to these two risk smoking. The study calculated age-adjusted inci-
factors, particularly in current consumers. The dence rates separately for each type of OPMD, and
Working Group acknowledged the reporting the ratio of leukoplakia incidence was estimated
of cancer risk estimates by time since quitting between former smokers and current smokers.
smoking in this small study. However, this study In men, who reported smoking more frequently
did not include any description of matching of than women, the age-adjusted incidence of
controls to cases, and participation rates in cases leukoplakia was 24 per 100 000 (1 incident case)
and controls were not mentioned. The defini- in former bidi smokers and 155 per 100 000 (80
tion of former smoker included people who had
131
132

Table 2.18 Cessation of tobacco smoking and risk of OPMDs
Reference Study design and population End-point Exposure category Number Risk estimate/ Adjustments/comments
Location of study prevalence or
participants/ incidence ratio
cases/controls/ (95% CI)
age-adjusted
incidence
Gupta et al. Cohort of 12 212 male and Leukoplakia Bidi smoking: Men (cases/ Incidence ratio: Stopping smoking defined
(1995) female tobacco users aged ≥ 15 yr age-adjusted as quitting bidi or cigarette
Kerala, identified in a baseline house-to- incidence): smoking for > 6 months at the
Trivandrum house survey (1977–1978) and Stopped 1/24 per 0.15 (N/A) time of the 10-year survey.
(India) recontacted annually for tobacco 100 000 Duration of cessation not
Intervention control education. Incidence Continued 80/155 per – reported
study of OPMDs at the 10-yr follow- 100 000 Incidence rates age-adjusted
up visit is reported by tobacco Large sample size of men
cessation Men accrued 77 681 and women at high risk of
person-years, and women accrued developing OPMDs. The
32 544 person-years proportion of person-years
accrued of tobacco cessation
was higher in women (14.4%,
mainly chewing) than in men
(6.5%, mainly bidi smoking).
Risk estimates reported
without a measure of precision
age-adjusted
incidence
Macigo et al. Community-based case–control Clinically Industrial Cases/controls: Leukoplakia: Definition of former smoker
(1995, 1996) study of cases of leukoplakia, diagnosed cases cigarette smoking: not provided
Meru District including men and women aged of leukoplakia Never-smokers 18/78 1.0 (ref) RRs not adjusted for potential
(Kenya) 21–75 yr residing for ≥ 5 yr Former smokers 5/31 0.7 (0.2–2.3) confounders (i.e. alcohol
in the Githongo sublocation consumption)
Current smokers 62/32 8.4 (4.1–17.4)
of Meru District (n = 85), and Well-defined clinical
age-, sex-, and sampling cluster- Kiraiku hand- diagnostic criteria and
matched controls (n = 141), rolled cigarette histological confirmation
including administration of smoking:
structured questionnaire and oral Never-smokers 42/120 1.0 (ref)
examination Former smokers 29/17 4.9 (2.3–10.4)
Current smokers 14/4 10.0 (2.9–43.4)
Time smoking
before quitting
(yr):
≤ 10 24/15 4.6 (2.1–10.2)
> 10 5/2 7.1 (1.1–76.6)
Duration of
cessation (yr):
≤4 6/2 8.6 (1.4–88.7)
5–9 12/7 4.9 (1.7–14.9)
≥ 10 11/8 3.9 (1.4–11.6)

133
134

age-adjusted
incidence
Hashibe et al. Community-based case–control Leukoplakia or Tobacco smoking: Cases/controls: Leukoplakia: Former smoker not defined,
(2000a, b) study nested in an intervention erythroplakia Never-smokers 428/35 591 1.0 (ref) and duration of smoking
Kerala (India) trial screening male and female diagnosed by a Former smokers 46/1815 1.7 (1.0–2.7) cessation not reported
residents aged ≥ 35 yr and dentist Former smoking-associated
Occasional 19/764 2.0 (1.4–2.8)
identifying 49 174 eligible study leukoplakia and erythroplakia
smokers
participants examined at home. effect estimates adjusted for
3585 people with suspicious Current smokers 434/9602 3.4 (2.8–4.1) age, sex, education level, BMI,
OPMDs or cancer lesions referred Erythroplakia: years of chewing, and years of
to the dentist or the oncologist. Never-smokers 428/35 591 1.0 (ref) alcohol consumption
The study included 927 cases Former smokers NR 1.6 (0.8–2.9) Large sample size, leukoplakia
of leukoplakia, 100 cases of Current smokers, NR 1.2 (0.6–2.4) lesions confirmed by a
erythroplakia, and 47 773 controls 1–20×/day dentist, and effect estimates
fully adjusted for important
Current smokers, NR 2.3 (1.1–5.1)
confounders
21–40×/day
Shiu et al. Hospital-based case–control study Cohort of Cigarette smoking: Cases/controls: Leukoplakia: Former smoker not defined,
(2000) of 100 randomly selected cases leukoplakia Never-smokers NR 1.0 (ref) and duration of smoking
Taiwan of leukoplakia out of a cohort of cases clinically Former smokers NR 1.04 cessation not reported
(China) 580 patients with leukoplakia diagnosed (0.24–4.59) ORs adjusted for alcohol
diagnosed at a single institution according to consumption and areca nut
Current smokers NR 3.22 (1.06–9.78)
in 1988–1998, and 100 age-, sex-, WHO definition chewing
and date of diagnosis-matched Effect estimates are adjusted
controls randomly selected for important confounders.
from patients diagnosed with Incomplete reporting, and
periodontal disease at the same estimates with low precision
hospital
age-adjusted
incidence
Fisher et al. Community-based case–control Leukoplakia Tobacco smoking: Cases/controls: Leukoplakia:* Former smoker not defined,
(2005) study of cases (n = 90) identified histologically Never-smokers 38/25 1.0 (ref) and duration of smoking
West Virginia at a leukoplakia tissue registry confirmed as Former smokers 30/29 0.71 (0.27–1.86) cessation not reported.
(USA) and controls (n = 78) at the ICD-9 code Leukoplakia ORs adjusted
Current smokers 22/24 0.48 (0.17–1.33)
surgical biopsy service supporting 528.6 with for age, sex, smokeless
the tissue registry but with a hyperkeratosis tobacco use, daily alcohol
diagnosis of periapical cyst (ICD- with or without consumption, and dental
9 code 522.8) and no diagnosis of epithelial atypia prostheses use. [*Results
leukoplakia or dysplasia shown correspond to model
assessing smokeless tobacco
use]
Cases with histological
confirmation, and effect
estimates fully adjusted, but
small sample size
Amarasinghe Community-based case–control Suspected cases Tobacco smoking: Cases/controls: Leukoplakia: Former smokers included
et al. (2010a) study built on a randomly selected of leukoplakia Never-smokers 43/NR 1.0 (ref) ever-smokers who had quit
Sabaragamuwa multistage cross-sectional sample identified during Former smokers 6/NR 0.5 (0.2–1.6) > 1 calendar year before the
Province (Sri (n = 1029) of people aged > 30 yr screening referred date of diagnosis or interview
Occasional 6/NR 0.8 (0.3–2.5)
Lanka) drawn to assess the prevalence of to the hospital for Smoking-related effect
smokers
OPMDs in a rural setting. People histopathological estimates adjusted for sex,
with suspected OPMDs on oral confirmation Current smokers 15/NR 0.7 (0.3–1.6) age, education level, BMI,
examination were considered occupation, β-carotene-
cases (n = 102), and screenees free containing total fruit and
of oral mucosa abnormalities were vegetable portions, betel
considered controls quid chewing, and alcohol

consumption
Cases with histological
confirmation, and risk
estimates fully adjusted, but
incomplete exposure reporting
and small number of exposed
cases. Study in a population
where chewing is common
135
136

age-adjusted
incidence
Li et al. (2011) Case–control study identifying Histopathological Tobacco smoking: Cases/controls OPMD, OR: Former smoker defined as
Puerto Rico men and women aged ≥ 30 yr diagnosis of oral with benign a person who was an ever-
(USA) with an oral cavity examination hyperkeratosis, lesions: smoker and quit smoking
histopathology report generated epithelial Never-smokers 38/99 1.0 (ref) for > 1 calendar year before
in 2003–2007 at pathology hyperplasia, Former smokers 17/30 1.47 (0.67–3.21) the year of diagnosis. No
laboratories in Puerto Rico. and epithelial information on duration of
Current smokers 31/26 4.32 (1.99–9.38)
People with benign oral lesions dysplasia in cessation was reported
(n = 155) were considered people with no Estimates adjusted for age,
controls, and those with OPMDs prior history of sex, education level, fruit and
(n = 86) were considered cases oral lesions vegetable intake, and alcohol
consumption (4 levels)
Cases histologically
confirmed, and interviewer
blinded on case–control
status of responders. Original
specific OPMDs in cases not
reported
Pivovar et al. Cross-sectional study to screen OPMDs and Tobacco smoking: Screened Prevalence Former smokers included
(2017) for oral cancer in high-risk men oral cancer first OPMD- ratio: ever-smokers with a smoking
Curitiba, (former or current smokers) aged diagnosed by a positive/ history of ≥ 20 yr and who
Paraná (Brazil) 50–65 yr registered in a primary dentist on clinical negative: had quit < 5 yr before the
health-care programme; 233 were grounds and Former smokers 13/76 1.0 (ref) interview. Model generating
ever-smokers, and 202 completed suspected lesions Current smokers 44/69 2.66 (NR) leukoplakia prevalence
the oral examination at the dentist with histological ratios in current smokers to
Leukoplakia:
analysis. former smokers adjusted for
Former smokers 6/83 1.0 (ref) family income and history
Current smokers 34/79 4.31 of compliance with clinical
(1.76–10.57) examinations
Histological confirmation.
No adjustment for alcohol
consumption
BMI, body mass index; CI, confidence interval; ICD, International Classification of Diseases; N/A, not available; NR, not reported; OPMDs, oral potentially malignant disorders; OR,
odds ratio; ref, reference; RR, relative risk; WHO, World Health Organization; yr, year or years.
incident cases) in current bidi smokers, gener- Hashibe et al. (2000a) reported on a large
ating an incidence ratio of 0.15. [The Working community-based case–control study embedded
Group noted that although the incidence ratio in a randomized intervention trial in Kerala,
was reported without an estimate of precision India, screening for oral cancer in male and
and without taking alcohol consumption into female residents. The study included 927 cases
account, such a large decrease in the incidence of leukoplakia confirmed by a dentist and 47 773
of leukoplakia after quitting bidi smoking, in screened people free of oral diseases (controls).
a population known to have low or no alcohol The RR of leukoplakia in former smokers
consumption, is probably not due to chance or compared with never-smokers (OR, 1.7; 95% CI,
confounding.] 1.0–2.7) was lower than that in current smokers
(OR, 3.4; 95% CI, 2.8–4.2); the effect estimates
(iii) Case–control studies
were controlled for important confounders. In
One hospital-based case–control study (Shiu a related publication from the same population
et al., 2000) and five community-based case– (Hashibe et al., 2000b), the association between
control studies (Macigo et al., 1996; Hashibe cigarette smoking and erythroplakia was inves-
et al., 2000a, b; Fisher et al., 2005; Amarasinghe tigated (100 cases). The RR of erythroplakia in
et al., 2010a; Li et al., 2011) were identified, former cigarette smokers compared with never-
including participants from India, Kenya, Puerto smokers (OR, 1.6; 95% CI, 0.8–2.9) was lower than
Rico, Sri Lanka, Taiwan (China), and the USA the RR in current smokers who reported smoking
(Table 2.18). 21–40 times per day (OR, 2.3; 95% CI, 1.1–5.1) but
In a community-based case–control study not lower than the RR in current smokers who
in Meru District in north-eastern Kenya, 85 reported smoking 1–20 times per day (OR, 1.2;
leukoplakia cases and 141 controls were iden- 95% CI, 0.6–2.4). [The Working Group noted that
tified in a house-to-house survey of eligible the studies did not provide definitions of former
residents (Macigo et al., 1995, 1996). The RR of smoker or current smoker and did not present
leukoplakia in former smokers of commercial leukoplakia or erythroplakia effect estimates by
cigarettes compared with never-smokers was number of years since quitting smoking.]
< 1 (OR, 0.7; 95% CI, 0.2–2.3); this estimate is Shiu et al. (2000) randomly selected 100 cases
substantially lower than that in current smokers of leukoplakia in a cohort of 435 cases diagnosed
(OR, 8.4; 95% CI, 4.1–17.4). In contrast, the RR in 1988–1998 at a medical institution in Taiwan
of leukoplakia in former smokers of kiraiku (China) and 100 matched controls. Leukoplakia
hand-rolled cigarettes compared with never- risk estimates were calculated using never-
smokers was markedly elevated (OR, 4.9; 95% CI, smokers as the reference group. Multivariate
2.3–10.4) but was lower than the RR in current analysis adjusting for alcohol intake and betel
smokers of kiraiku cigarettes (OR, 10.0; 95% quid chewing generated RR estimates in former
CI, 2.9–43.4). The risk of leukoplakia remained smokers (OR, 1.04; 95% CI, 0.24–4.59) of lower
elevated in former smokers with > 10 years of magnitude than the RR estimates in current
kiraiku smoking cessation (OR, 3.9; 95% CI, smokers (OR, 3.22; 95% CI, 1.06–9.78). [The
1.4–11.6). [The Working Group noted the omis- Working Group noted that the study did not
sion of definitions of former smoker and current provide definitions of former smoker and current
smoker. Furthermore, effect estimates associated smoker and did not present effect estimates by
with smoking were not adjusted for important number of years since quitting smoking.]
confounders, including alcohol consumption, a Fisher et al. (2005) reported on a case–control
behaviour that is socially accepted in Kenya.] study in West Virginia (USA) including cases
137
of leukoplakia (n = 90; response rate of eligible rather than by clinical entity or diagnosis, so that
people, 55%) and controls with a periapical cyst the dysplasia observed microscopically may have
(n = 78; response rate of eligible people, 50%) emerged from leukoplakia or from erythroplakia
identified at the same tissue registry in 2001– originally detected in the mouth.]
2002. The fully adjusted RRs of leukoplakia in (iv) Cross-sectional studies
former smokers (OR, 0.71; 95% CI, 0.27–1.86)
and in current smokers (OR, 0.48; 95% CI, Pivovar et al. (2017) reported on a cross-sec-
0.17–1.33) were < 1. [The Working Group noted tional study within the framework of oral cancer
the very modest response rate in cases and in screening in primary health care in the city of
controls, which raises concerns of selection bias. Curitiba in the state of Paraná in southern Brazil.
Furthermore, the Working Group acknowledged The prevalence of OPMDs and leukoplakia
the omission of a definition of former smoker in former smokers and current smokers was
and the use of a control group with a pathology adjusted for family income and history of compli-
condition that was not described in any detail; ance with clinical examinations. The prevalence
this control group was probably not appropriate. of leukoplakia was markedly higher in current
Also, cases and controls differed by socioeco- smokers than in former smokers (prevalence
nomic status or by education level, factors that ratio, 4.31; 95% CI, 1.76–10.57) (Table 2.18). [The
were not taken into account and that may influ- Working Group noted that this study compared
ence the level of smoking. Finally, the restriction former smokers with current smokers but calcu-
of controls to people with a periapical cyst may lated the leukoplakia prevalence ratio using
have indirectly selected for controls with preva- former smokers rather than current smokers as
lent smoking.] the reference group.]
The very small study by Amarasinghe et al.
(2010a) included few cases, and all ORs were < 1; 2.3.2 Alcohol consumption
it was considered uninformative. This section summarizes the findings from
The community-based case–control study of observational case–control studies, cohort
Li et al. (2011) identified men and women aged studies, and a pooled analysis that investigated
≥ 30 years with an oral cavity examination histo- the effect of cessation of alcohol consumption
pathology report generated in 2003–2007 at and duration of alcohol cessation on the risks
pathology laboratories in Puerto Rico and lacking of oral cancer and OPMDs. These included the
a previous history of oral diseases. People with pooled analysis from the INHANCE consortium
benign oral conditions (n = 155) were consid- with data from 13 case–control studies (Marron
ered controls, and those with OPMDs (n = 86), et al., 2010), three cohort studies (Ide et al., 2008;
defined as oral epithelial dysplasia, oral hyper- Cancela et al., 2009; Im et al., 2021), and two
keratosis, or epithelial hyperplasia without individual case–control studies, one published
epithelial dysplasia, were considered cases. The before the INHANCE analysis (Takezaki et al.,
effect estimate for OPMDs in former smokers 1996) and one published since the INHANCE
compared with never-smokers (OR, 1.47; 95% analysis (Andrade et al., 2015).
CI, 0.67–3.21) was lower than that for current
smokers compared with never-smokers (OR, (a) Risk of oral cancer
4.32; 95% CI, 1.99–9.38). [The Working Group See Table 2.19.
noted that this case–control study, which clearly The INHANCE Consortium investigated the
defined former smoker, was the only study that effects of quitting alcohol consumption on the
defined OPMDs by histopathology features,
138
Table 2.19 Cessation of alcoholic beverage consumption and risk of oral cancer and/or pharyngeal cancer
Reference Study population, Oral cancer or Exposure Number of OR Adjustments/comments

Location number of precancer end- category cases/controls (95% CI)
participants, study point
period
Pooled analysis of case–control studies
Marron et al. INHANCE Invasive tumour Oral cancer: Former drinkers were defined
(2010) consortium pooled of the oral cavity, Current drinkers 1131/5715 1.0 (ref) as people who had quit
International analysis of case– oropharynx, Duration of cessation (yr): drinking the following alcoholic
(multiple control studies, hypopharynx, beverages: wine, beer, liquor,
> 1–4 132/504 0.81 (0.61–1.07)
studies in including men and or oral cavity and aperitifs. People who had
(France, Italy women; 2615 oral or pharynx not 5–9 149/576 0.77 (0.52–1.15) stopped drinking for > 1 yr were
Switzerland, cancer cases, 3989 otherwise specified 10–19 174/801 0.66 (0.47–0.92) classified as former drinkers.
Latin/Central oropharyngeal or ≥ 20 155/763 0.45 (0.26–0.78) The number of years that former
America, hypopharyngeal Never-drinkers 737/3674 0.65 (0.36–1.16) drinkers had quit drinking
USA) cancer cases, and Ptrend = 0.05 was determined from age at
12 359 and 12 593 reference date (interview or
controls, respectively < 1 drink/day: diagnosis date) and age at which
(~1990s to early Current drinkers 256/2250 1.0 (ref) they had stopped drinking
2000s) Duration of cessation (yr): Analysis adjusted for age, sex,
> 1–4 30/144 1.51 (0.80–2.87) race/ethnicity, study centre,
education level, and pack-years
5–9 22/204 1.06 (0.39–2.88)
of tobacco smoking
10–19 40/307 0.80 (0.37–1.75)
≥ 20 57/338 0.98 (0.54–1.77)
Never-drinkers 727/3238 0.86 (0.39–1.89)
1–2 drinks/day:
Current drinkers 234/1539 1.0 (ref)
> 1–4 24/149 0.67 (0.33–1.35)
5–9 36/154 1.22 (0.43–3.43)
10–19 30/205 0.34 (0.15–0.80)

≥ 20 29/186 0.59 (0.22–1.57)
Never-drinkers 717/3144 0.58 (0.26–1.28)
139
140

period
Marron et al. ≥ 3 drinks/day:
(2010) Current drinkers 589/1554 1.0 (ref)
(cont.) Duration of cessation (yr):
> 1–4 77/206 0.79 (0.54–1.14)
5–9 90/207 0.85 (0.51–1.41)
10–19 102/279 0.82 (0.50–1.34)
≥ 20 69/232 0.43 (0.28–0.67)
Never-drinkers 727/3580 0.19 (0.09–0.39)
Ptrend = 0.06
Oro/hypopha-
ryngeal cancer:
Alcohol cessation:
> 1–4 213/505 1.04 (0.73–1.48)
5–9 240/576 0.95 (0.61–1.49)
10–19 340/802 1.15 (0.92–1.43)
≥ 20 221/763 0.74 (0.50–1.09)
Never-drinkers 406/3693 0.65 (0.42–1.02)
Ptrend = 0.18
< 1 drink/day:
> 1–4 29/144 2.02 (1.07–3.80)
5–9 28/205 1.44 (0.65–3.16)
10–19 67/309 1.49 (0.96–2.34)
≥ 20 60/338 1.16 (0.65–2.05)
Never-drinkers 406/3693 0.97 (0.59–1.58)
period
Marron et al. 1–2 drinks/day:
(2010) Current drinkers 335/1808 1.0 (ref)
(cont.) Duration of cessation (yr):
> 1–4 38/152 1.09 (0.65–1.82)
5–9 33/156 1.09 (0.55–2.16)
10–19 55/205 1.06 (0.67–1.68)
≥ 20 45/186 0.80 (0.47–1.37)
Never-drinkers 400/3599 0.49 (0.30–0.81)
≥ 3 drinks/day:
> 1–4 141/206 1.05 (0.69–1.59)
5–9 174/207 1.12 (0.60–2.08)
10–19 213/279 1.15 (0.73–1.81)
≥ 20 115/232 0.77 (0.45–1.30)
Never-drinkers 397/3580 0.19 (0.10–0.37)
Ptrend < 0.01

141
142

period
Case–control studies
Huang et al. Hospital-based ICD-classified Oral cancer: Age, sex, education, cigarette
(2017) case–control study, primary Non-drinkers 195/517 1.0 (ref) smoking (pack-year categories),
Taiwan including men and pathologically (never + occasional) and betel quid chewing (pack-
(China) women; 509 oral confirmed Former drinkers 61/109 0.77 (0.51–1.17) year categories)
cancer cases, 118 squamous cell Selection of hospital-based
Current drinkers 253/314 1.29 (0.97–1.73)
oropharynx cases, carcinoma of the controls with conditions
and 89 hypopharynx oral cavity Oropha- thought to be unrelated to
cases (2010–2016) ryngeal cancer: smoking or alcohol use
Non-drinkers 29/517 1.0 (ref) No adjustment for past amount
(never + occasional) of alcohol consumed or
Former drinkers 20/109 2.83 (1.39–5.76) duration of smoking cessation
Hypopharyn-
geal cancer:
Non-drinkers 4/517 1.0 (ref)
(never + occasional)
Former drinkers 19/109 14.02 (4.38–44.85)
Andrade et al. Hospital-based case– Histopathologically Oral cancer: Moderate-sized case–control
(2015) control study, with confirmed oral Non-drinkers 27/113 1.0 (ref) study
Brazil data abstracted from squamous cell Former drinkers 56/57 2.73 (1.73–4.31) Alcohol consumption categories
medical records, carcinoma not defined
including men and Crude ORs, no adjustment
women; 127 oral Duration of cessation (yr):
cancer cases and 381 ≥ 10 20/41 1.0 (ref)
controls (2002–2012) < 10 36/16 4.61 (2.08–10.22)
period
De Stefani Hospital-based Microscopically Oral cancer: Adjusted for age, residence,
et al. (2007) case–control study, confirmed Non-drinkers 34/527 1.0 (ref) urban/rural status, hospital,
Uruguay including men only; squamous cell Former drinkers 91/317 3.0 (1.9-4.7) diagnosis year, education,
335 oral cancer cases carcinoma of the first-degree family history of
Current drinkers 210/657 3.4 (2.3-5.2)
and 441 pharyngeal mouth or pharynx cancer, total vegetable and fruit,
cancer cases Pharyngeal and maté intake, occupation,
(1998–2000) cancer: smoking status, years since
Non-drinkers 33/527 1.0 (ref) quitting smoking and current
Former drinkers 116/317 3.9 (2.5-6.1) cigarettes/day
Current drinkers 292/657 4.5 (3.0-6.8) Selection of hospital-based
controls with conditions
thought to be unrelated to
smoking or alcohol use
No adjustment for past amount
of alcohol consumed
Zheng et al. Hospital-based Histologically Tongue cancer: Adjusted for tobacco, years of
(1997) case–control study, confirmed tongue Non-drinkers 64/72 1.0 (Ref.) education and matching factors
China including men and cancer Selection of hospital-based
women; 111 tongue Former drinkers 7/6 1.20 (0.58–2.50) controls with conditions
cancer and 111 sex- Current drinkers 40/33 0.94 (0.28–3.22) thought to be unrelated to
and age-matched smoking or alcohol use
controls (1988–1989) No adjustment for past amount
of alcohol consumed or
duration of smoking cessation
Takezaki et al. Hospital-based Histologically Oral, oropha- Alcohol consumption defined
(1996) case–control study, confirmed, ICD- ryngeal, and and standardized
Japan including men classified primary hypopha- Crude ORs, no adjustment
and women; 203 oral cancer, ryngeal cancer: Not clear what reference group

oral cancer cases, oropharyngeal is – possibly never-drinkers
35 oropharyngeal cancer, and
cancer cases, and hypopharyngeal 0 (never quit) 138/13 811 1.2 (0.9–1.6)
28 hypopharyngeal cancer 0–4 9/320 2.4 (1.1–5.1)
cancer cases 5–14 4/180 1.7 (0.6–4.8)
≥ 15 4/62 3.4 (1.2–9.9)
143
144

period
Ko et al. Hospital-based Histologically Oral cancer: Adjusted for education,
(1995) case–control study, confirmed, ICD- Non-drinkers 25/89 1.0 (ref) occupation, cigarette smoking
Taiwan including men and classified oral Former drinkers 14/37 1.0 (0.3–3.3) and betel chewing status
(China) women; 107 oral cancer, No details on selection of
cancer cases and 200 hospital-based controls
controls (1992–1993) provided
No adjustment for past amount
of alcohol consumed or
duration of smoking cessation
Cohort studies
Im et al. Cohort study. 209 Cancer of mouth or Mouth or Analysis adjusted for age,
(2021) 237 men, aged throat by ICD-10 throat cancer study area, education, income,
China 30–79 years, with codes (C00-C14, incidence: smoking, physical activity, fruit
no previous history C32) Abstention 23/42 479 1.00 (0.65–1.53) intake, BMI, and family history
of cancer; follow- Ex-regular drinkers 12/18 061 1.06 (0.60–1.87) of cancer
up time from 2004 Floating standard errors were
Occasional drinkers 39/78 963 1.33 (0.96–1.86)
until January 2017 used to estimate the confidence
(median 10 years); Current regular drinkers 66/69 734 1.89 (1.46–2.45) intervals
incident cancer Abstention is the reference
cases ascertained by category
linkage with cancer No adjustment for past amount
registries and the of alcohol consumed or
National Health duration of smoking cessation
insurance databases No data on alcohol and oral
cancer risk among women
provided
period
Cancela et al. Cohort study. Oral cancer was Oral cancer: Small numbers of cases and
(2009) Trivandrum Oral defined by ICD-10 Incidence HR: deaths, and consequently wide
India Cancer Screening codes C02 (other (person-years): CIs
Study RCT, with and unspecified Never-drinkers 61/178 932 1.00 (ref) Individuals who had never
cancer registry parts of the tongue), consumed alcohol during their
Current drinkers 52/85 022 1.49 (1.01–2.21)
follow-up of C03 (gum), C04 lifetime were categorized as
incidence and (floor of the mouth), Former drinkers 21/19 127 1.90 (1.13–3.18) never, those who were currently
mortality, including C05 (palate), and Mortality consuming alcohol or those who
men and women C06 (other and (person-years): had stopped drinking alcohol
aged 35–100 yr; unspecified parts of Never-drinkers 43/179 134 1.00 (ref) for < 6 months were categorized
32 347 participants the mouth) Current drinkers 34/85 158 1.76 (1.08–2.86) as current, and those who had
recruited in 1996. Former drinkers 14/19 212 2.04 (1.08–3.86) quit drinking ≥ 6 months before
In 10 yr of follow-up the time of the interview were
(1996–2006), 134 categorized as former
oral cancer cases Analyses adjusted for age,
were diagnosed, and education level, religion,
91 oral cancer deaths occupation, standard of living,
were registered betel quid chewing habits,
smoking habits, intake of
vegetables, and intake of fruits
Ide et al. Cohort study. Oral and Men: Oral and Non-drinker and former
(2008) 110 792 participants, pharyngeal cancer pharyngeal drinker were not defined
Japan including men deaths were cancer Small numbers and wide CIs
(46 465) and women identified by ICD- mortality: Adjusted for age (continuous),
(64 327) aged 10 codes C01–C14, Non-drinkers 5/77 513 1.0 (ref) smoking status (never, former,
40–79 yr, recruited excluding C07–C08 Former drinkers 2/23 423 1.2 (0.2–6.0) current), consumption of green
in 1988–1990. In (salivary gland tea (≥ 1 cups per day, < 1 cup
Current drinkers 34/319 502 2.0 (0.8–5.1)
12.5 yr of follow-up, cancer) and C11 per day, unknown), preference

52 deaths: 25 from (nasopharyngeal for salty foods (like, normal
oral cancer and 27 cancer) or dislike, unknown), and
from pharyngeal consumption of green and
cancer yellow vegetables (daily or not)
BMI, body mass index; CI, confidence interval; HR, hazard ratio; ICD, International Classification of Diseases; INHANCE, International Head and Neck Cancer Epidemiology; OR,
odds ratio; RCT, randomized controlled trial; ref, reference; yr, year or years.
145
risks of oral cancer (based on 12 studies) and cavity cancer than current drinkers relative to
oropharyngeal and hypopharyngeal cancers never-drinkers (Im et al., 2021). [This relatively
(based on 13 studies) by performing a robust small cohort study did not adjust for past alcohol
pooled analysis with comprehensive adjustment consumption or smoking in the analysis.]
for confounding factors (Marron et al., 2010). The individual case–control study published
Cessation of alcohol consumption was asso- before the INHANCE analysis (Takezaki et al.,
ciated with a reduced risk of oral cancer (OR, 1996) showed an increase in risk associated with
0.60; 95% CI, 0.43–0.84). The reduction in risk long duration of quitting (OR for > 15 years of
after alcohol cessation increases with duration quitting, 3.4; 95% CI, 1.2–9.9). [The numbers of
of cessation, with the risk decreasing by > 50% participants in each category were very small,
by 20 years of quitting for oral cancer (OR, 0.45; and the estimates were not adjusted for poten-
95% CI, 0.26–0.78) and by about 25% by 20 years tial confounders, including smoking.] More
of quitting for oropharyngeal and hypopharyn- recently, a small hospital-based case–control
geal cancers combined (OR, 0.74; 95% CI, study in Brazil (Andrade et al., 2015) reported
0.50–1.09). Further subgroup analyses showed that cessation of alcohol consumption for
that the effects of quitting on the risk of oral < 10 years compared with cessation for ≥ 10 years
cancer were more pronounced in former heavy conferred a large increased risk (OR, 4.61; 95% CI,
drinkers (≥ 3 drinks per day) and the RR reduc- 2.08–10.22). [The categories of alcohol consump-
tion became significant after ≥ 20 years of quit- tion were not defined, and the crude estimates
ting (OR, 0.43; 95% CI, 0.28–0.67); there was no were not adjusted for any potential confounding
relationship with duration of consumption. For factors.]
oropharyngeal and hypopharyngeal cancers, the In addition, four other hospital-based case–
relationship with previous frequency of alcohol control studies (Ko et al., 1995; Zheng et al.,
consumption and duration of quitting was less 1997; De Stefani et al., 2007; Huang et al., 2017)
clear. reported only risks associated with former
Three cohort studies analysed the risk asso- drinking relative to never drinking. In all four
ciated with former alcohol consumption [none studies, the risk associations for former drinking
of them reported duration of alcohol cessa- relative to never drinking were lower than those
tion]. In a cohort in India, former drinkers for current drinking relative to never drinking;
had a higher risk of oral cancer incidence and ORs ranged from 0.77 to 3.0 for former drinking
death than current drinkers relative to never- and from 1.2 to 3.4 for current drinking (relative
drinkers (Cancela et al., 2009). [This study had to never drinking).
small numbers of cases and deaths, well-defined
categories of alcohol consumption, and robust (b) Risk of OPMDs
analyses.] In a cohort in Japan, the RR of oral See Table 2.20.
and pharyngeal cancer death associated with No studies were identified that showed the
former drinking relative to non-drinking in men effect of duration of alcohol cessation on the risk
was lower than the RR in current drinkers (Ide of OPMDs. Seven case–control studies reported
et al., 2008). [This study had small numbers of risk estimates for former drinkers relative to
deaths. No categories of alcohol consumption never-drinkers alongside estimates for current
were defined, but the analyses were adjusted for drinkers relative to never-drinkers. [The studies
potential confounders.] In a recent cohort study generally had small sample sizes and were of
of men in China, former drinkers relative to varying quality.] Two studies reported OPMD
never-drinkers had a lower RR for lip and oral outcomes combined, one reported multiple
146
Table 2.20 Cessation of alcoholic beverage consumption and risk of OPMDs
Reference Study population, number of Oral cancer or precancer Exposure Number of OR Adjustments/
Location participants, study period end-point category cases/controls (95% CI) comments
Li et al. (2011) Case–control study, including Histopathological diagnosis OPMDs: Never-drinker and
Puerto Rico (USA) men and women aged of oral hyperkeratosis, Never-drinkers 41/73 1.0 (ref) former drinker
≥ 30 yr. People with benign epithelial hyperplasia, Ever-drinkers 45/82 0.63 (0.33–1.21) not defined. Small
oral lesions (n = 155) were and epithelial dysplasia in numbers and wide CIs
Former drinkers 14/22 0.63 (0.25–1.57)
considered controls, and people with no prior history Adjusted for age (4
those with OPMDs (n = 86) of oral lesions Current drinkers 31/60 0.63 (0.32–1.26) levels), sex, education
were considered cases level (3 levels), fruit
(2003–2007) and vegetable intake
(4 levels), and current
smoking
Amarasinghe et al. Community-based case– Suspected cases of OPMDs: Former drinkers
(2010a) control study. Randomly leukoplakia identified Non-drinkers 39/551 1.0 (ref) also include current
Sri Lanka selected multistage cross- during screening Monthly, weekly, 27/114 2.7 (1.2–6.3) occasional drinkers
sectional sample (n = 1029) referred to the hospital and daily Adjusted for sex, age,
including men and women for histopathological drinkers socioeconomic status,
aged > 30 yr. People with confirmation β-carotene-containing
Former, 35/63 1.1 (0.5–2.6)
suspected OPMDs on fruits and vegetables
occasional
oral examination were portion, BMI,
drinkers
considered cases (n = 102), smoking, betel quid
and screenees free of oral chewing, and alcohol
mucosa abnormalities were consumption
considered controls
Lee et al. (2003) Community-based case– Leukoplakia or OSF. Leukoplakia: Adjusted for
Taiwan (China) control study, including men Histologically confirmed Never-drinkers 72/349 1.0 (ref) education level and
and women aged ≥ 15 yr. and diagnosed according to Former drinkers 9/40 1.1 (0.5–2.4) occupation
Cases of leukoplakia or OSF WHO definitions
(1994 and 1995). 219 OPMD
cases and 876 age- and OSF:
Never-drinkers 55/266 1.0 (ref)

sex-matched controls were
enrolled Former drinkers 7/27 1.4 (0.6–3.4)
147
148

Thomas et al. (2003) Community-based case– Multiple OPMDs diagnosed Multiple Large sample size.
India control study nested in an by a dentist OPMDs: Confirmed diagnosis
intervention trial screening Non-drinkers 91/40 801 1.0 (ref) by dentist
men and women aged ≥ 35 yr Occasional 4/2743 1.1 (0.4–3.2) Adjusted for age, sex,
and identifying 49 174 eligible drinkers education level, BMI,
study participants examined smoking (continuous,
at home. 3585 people with pack-years), tobacco
suspicious OPMDs or Former drinkers 7/1475 1.8 (0.7–4.5) chewing (continuous,
cancer lesions referred to duration in years),
the dentist or the oncologist. fruit intake (low or
The study included 927 cases high), and vegetable
of leukoplakia, 100 cases of intake (low or high)
erythroplakia, 115 people
with multiple OPMDs, and
47 773 controls
Hashibe et al. Community-based case– Leukoplakia diagnosed by Leukoplakia: Large sample size.
(2000a) control study described above a dentist Non-drinkers 619/40 801 1.0 (ref) Confirmed diagnosis
India in Thomas et al. (2003) Occasional 65/2743 1.2 (0.9–1.6) by dentist
drinkers Adjusted for age, sex,
education level, BMI,
smoking, and tobacco
Former drinkers 78/1475 1.4 (1.1–1.9) chewing
Hashibe et al. Community-based case– Erythroplakia diagnosed by Erythroplakia: Large sample size.
(2000b) control study described above a dentist Non-drinkers 62/40 801 1.0 (ref) Confirmed diagnosis
India in Thomas et al. (2003) Occasional 3/2743 0.9 (0.3–3.1) by dentist
drinkers Adjusted for age, sex,
education level, BMI,
smoking (continuous,
Former drinkers 14/1475 4.8 (2.4–9.7) pack-years), and
chewing tobacco
(continuous, duration
in years)
Macigo et al. (1996) Community-based case– Clinically diagnosed cases Leukoplakia: Alcohol consumption
Kenya control study of cases of of leukoplakia (147 lesions Never-drinkers 26/62 1.0 (ref) not defined
leukoplakia, including men in 85 cases). Only 5 cases Current drinkers 39/47 2.0 (1.0–3.9) RRs not adjusted for
and women aged 21–75 yr had non-homogeneous potential confounders
Former drinkers 36/49 1.5 (0.7–3.3)
residing for ≥ 5 yr in the lesions. Biopsies (i.e. alcohol
Githongo sublocation of obtained in 49 cases, and consumption)
Meru District (n = 85), and histopathology examination Crude estimates, not
age-, sex-, and sampling revealed no cancer and 11 adjusted for potential
cluster-matched controls cases of moderate to severe confounders
(n = 141), including dysplasia
administration of structured
questionnaire and oral
examination
BMI, body mass index; CI, confidence interval; OPMDs, oral potentially malignant disorders; OR, odds ratio; OSF, oral submucous fibrosis; ref, reference; RR, relative risk; WHO, World
Health Organization; yr, year or years.

149
OPMDs, three reported leukoplakia, one due to smoking. Neither study found an associ-
reported erythroplakia, and one reported OSF. ation between snuff use (former or current) and
Relative to never-drinkers, the RR estimates for risk of oral cancer; they reported risk estimates
OPMDs were generally (in 4 of 7 studies) lower for former users of 0.7 (95% CI, 0.1–5.0) (Luo
in former drinkers than in current drinkers, for et al., 2007) and 1.0 (95% CI, 0.3–3.5) (Boffetta
former drinkers ranging from 1.1 to 1.8 and for et al., 2005). [The Working Group noted the
current drinkers ranging from 1.3 to 2.7. In the small number of incident oral cancers in former
three studies that reported leukoplakia outcomes, snuff users in both studies: 1 event in the study
the risk estimates in former drinkers compared in Sweden (Luo et al., 2007) and 3 events in
with never-drinkers ranged from 1.1 to 1.5 and the study in Norway (Boffetta et al., 2005). The
those in current drinkers compared with never- Working Group noted the absence of repeat
drinkers ranged from 1.6 to 2.0; however, the CIs assessment of status of snuff use as an important
were wide and overlapping. limitation in these studies, particularly given the
long follow-up period. In addition, neither of the
2.3.3 Smokeless tobacco use studies adjusted for alcohol consumption.]
Four case–control studies examined the
(a) Risk of oral cancer risk of oral cancer in former users of smoke-
Six informative observational studies that less tobacco. Of these, three were conducted in
reported on the association between former use Sweden (Lewin et al., 1998; Schildt et al., 1998;
of smokeless tobacco and risk of oral cancer, Rosenquist, 2005) and examined oral snuff use.
including two cohort studies and four case– Exposure data were collected using question-
control studies, were identified by the Working naires, and cancer outcome data were obtained
Group. In most of these studies, the former use through linkage to hospital or cancer registries.
category was defined at study entry, and often no Controls from population-based registries were
information was provided with respect to dura- matched to cases. All three studies accounted
tion of cessation. Studies were well powered with for potential confounding due to smoking either
sufficient sample size to estimate overall effects by statistical adjustment or by providing strat-
but tended to have small numbers of former ified estimates in never-smokers. Two studies
users. There were no studies that provided found 1.5–1.8-fold non-statistically significant
risk estimates for former users compared with increased risk of oral cancer in former oral snuff
current users, and none that provided risk esti- users compared with never-users, whereas no
mates by time since quitting smokeless tobacco association was observed in current users (OR,
use. Detailed information on the six identified 0.7 and 1.0). In the third study (Rosenquist, 2005),
observational studies is presented in Table 2.21 using never-users as the reference group, former
and Table 2.22. users had a lower risk of oral cancer (OR, 0.3;
Two cohort studies, one in Sweden (Luo et al., 95% CI, 0.1–0.9) compared with current users
2007) and one in Norway (Boffetta et al., 2005), (OR, 1.1; 95% CI, 0.5–2.5). [All three studies were
examined the association between oral snuff conducted in Sweden, where reported associa-
use and risk of oral cancer. Data on snuff use tions between current snuff use and risk of oral
were collected using questionnaires, and cancer cancer are weak. A role for reverse causation in
outcome data were obtained through linkage to the observed elevated estimates cannot be ruled
cancer registries. The follow-up period between out.]
exposure and outcome was 12–35 years. Both The fourth case–control study, conducted in
analyses accounted for potential confounding Yemen, found a significantly elevated risk of oral
150
Table 2.21 Cessation of smokeless tobacco use and risk of oral cancer – cohort studies
Reference Study population, number of Outcome Exposure Number RR Comments

Location participants, study period, follow-up assessed categories of cases (95% CI)
period (number of cases)
Luo et al. (2007) Cohort study of 279 897 male Oral cancer Snus use: Association between snus use
Sweden construction workers in the Swedish (ICD-7 codes Never-users of any 50 1.0 (ref) and oral cancer was adjusted
building industry in 1978–1992 140, 141, 143, tobacco for age and BMI
Detailed information on smoking and and 144 not Former users 1 0.7 (0.1–5.0) Former snus user was defined
snus use collected through personal including on entry into study; changes
Current users 9 0.9 (0.4–1.8)
interview cancers of the in habit were not accounted
Oral cancer incidence data collected salivary glands, for
thorough complete linkage to pharynx, or Very small number of
population and health registries larynx) exposed cases
12-yr follow-up (until 2004)
Boffetta et al. (2005) Cohort study in Norwegian general Oral and Snus use: Adjusted for age and smoking
Norway population that included a probability pharyngeal Never-users 25 1.00 (ref) Former users were defined
sample of the general adult population cancer (ICD-7 Former users 3 1.04 (0.31–3.50) at entry into study, with no
from the 1960 census who were alive codes 141–148) repeat assessment
Current users 6 1.13 (0.45–2.83)
on 1 January 1966 No clear definition of former
Questionnaires were mailed for users
collection of data on smokeless
tobacco use
35-yr follow-up completed through
cancer registry linkage until 2001
Study included only men (n = 10 136)
BMI, body mass index; CI, confidence interval; ICD, International Classification of Diseases; N/A, not available; ref, reference; RR, relative risk; yr, year or years.

151
152

Table 2.22 Cessation of smokeless tobacco use and risk of oral cancer – case–control studies
Reference Study population, number of Outcome Exposure Number OR (95% CI) Comments
Location participants, study period, assessed category of cases/
follow-up period controls
Lewin et al. (1998) Registry-based case–control study Oral cancer Oral moist snuff: Estimates adjusted for age,
Stockholm Included men registered in Never-users 103/550 1.0 (ref) region, smoking, and alcohol
(Sweden) hospital-based or population- Former users 15/41 1.8 (0.9–3.7) consumption
based registries in 2 geographical
Current users 10/50 1.0 (0.5–2.2)
regions, aged 40–79 yr in 1988–
1990 Ever-users 25/91 1.4 (0.8–2.4)
128 oral cancer cases, 756
randomly selected controls
matched on age, sex, region, and
vital status
Exposure data collected through
personal interview
Schildt et al. Population-based case–control Oral cancer Oral moist snuff: Estimates adjusted for age,
(1998) study (ICD-7 codes Never-users 287/282 1.0 (ref) sex, and county of residence.
Sweden Oral cancer cases confirmed by 140, 141, Former users 28/18 1.5 (0.8–2.9) Smoking was not adjusted for,
histopathology and registered in 143–145). but stratified estimates were
Current users 23/54 0.7 (0.4–1.1)
4 northern regions of Sweden in provided
1980–1989 In never-smokers: A former smoker or former
Controls from population Never-users 124/144 1.0 (ref) snuff user was defined as a
registries matched on age, sex, Former users 9/4 1.8 (0.9–3.5) person who had quit smoking
county, and vital status and year of Current users 19/23 0.7 (0.4–1.2) or snuff use ≥ 1 yr before the
death where applicable diagnosis
Questionnaires mailed to collect
information on tobacco use
(smoking and moist snuff)
Rosenquist (2005) Hospital-based case–control study Oral cancer Oral snuff: ORs adjusted for smoking and
Sweden in Sweden in 2000–2004 Never-users 112/255 1.0 (ref) total alcohol consumption.
132 oral cancer cases (91 men) Former users 7/34 0.3 (0.1–0.9) Further adjustment for HPV
identified from 2 hospitals status had minor effects
Current users 13/31 1.1 (0.5–2.5)
reflecting 80% participation rate A former snuff user was
of cases in the region. 320 controls defined as a person who had
(215 men) matched on age, sex, quit the habit ≥ 6 months
and county from the population before the interview
registry. Data were collected by
interview; oral examination and
HPV testing were completed
Reference Study population, number of Outcome Exposure Number OR (95% CI) Comments
Location participants, study period, assessed category of cases/
follow-up period controls
Nasher et al. Hospital-based case–control Oral cancer Shammah: Estimates were adjusted for
(2014) study. Cases were confirmed by in users of Never-users 11/98 1.0 (ref) age, sex, EBV status, and
Yemen histopathology shammah Former users 7/8 12.6 (3.3–48.2) tobacco smoking
Oral cancer cases and age- and dipping
Current users 42/14 39 (14–105)
sex-matched controls
CI, confidence interval; EBV, Epstein–Barr virus; HPV, human papillomavirus; ICD, International Classification of Diseases; ref, reference; RR, relative risk; yr, year or years.

153
cancer in former shammah users compared with 1997; Shulman et al., 2004). Sinusas et al. (1992)
non-users (OR, 12.6; 95% CI, 3.3–48.2), which found a prevalence of leukoplakia in former users
was significantly lower than that in current equivalent to that in never-users (6%). Current
users (OR, 39; 95% CI, 14–105) (Nasher et al., users had a much higher prevalence of lesions
2014). [The Working Group noted that the esti- (37%), corresponding to a > 9-fold increase
mates were based on a small number of former compared with former smokeless tobacco
chewers; no definition was provided with respect users and non-users (Sinusas et al., 1992). [In
to duration of cessation, and the estimates were two studies (Ernster et al., 1990; Sinusas et al.,
not adjusted for alcohol consumption.] 1992), chewing tobacco use and snuff use were
combined to generate risk estimates; it is likely
(b) Risk of OPMDs that snuff and chewing tobacco may reflect differ-
Four cohort studies, two case–control studies, ential risks towards oral cancer. In the other two
and two cross-sectional studies have examined studies (Tomar et al., 1997; Shulman et al.,
the association between former use of smoke- 2004), multiple OPMDs were grouped together;
less tobacco and risk of OPMDs. [Most of these because some of these may not be etiologically
studies were well powered with sufficient sample related to smokeless tobacco use, these results
size to estimate overall effects, but they tended should be interpreted with caution.]
to have small numbers of former users.] Many of Two case–control studies were identified, one
these studies reported risk estimates using never- in the USA and one in Uzbekistan. In the study
users as the reference group, and some studies in Uzbekistan, risk estimates for former users
reported only the prevalence of lesions across of naswar (nass) were slightly lower than those
exposed groups (Table 2.23). for current users when compared with never-
The four cohort studies were all conducted users (Evstifeeva and Zaridze, 1992). The study
in the USA: two were in baseball players, and in the USA (Fisher et al., 2005) reported risks
two were large population-based cohorts. Three of leukoplakia for smokeless tobacco use and
of the four studies diagnosed leukoplakia as the snuff use separately. For both products, higher
outcome of interest at baseline entry into the risk estimates were found for current users than
study, whereas Shulman et al. (2004) diagnosed for former users compared with never-users.
oral mucosal lesions. Histopathological confir- [Both studies accounted for smoking and other
mation was indicated in only one study (Ernster potential confounding factors, but neither of the
et al., 1990). All four studies examined use of studies defined former use with respect to the
oral snuff and chewing tobacco; Sinusas et al. duration of cessation. In addition, the criteria for
(1992) also examined use of moist snuff. In these identification of leukoplakia and pre-leukoplakia
studies, former users were categorized at study were not defined in the study in Uzbekistan.]
entry as past users, with no further definition One cross-sectional study, conducted in
with regard to duration of cessation, except in Uzbekistan, reported percentages of leukoplakia
the study of Ernster et al. (1990), in which former and pre-leukoplakia that were similar for former
users were defined as past users who had used and current naswar use: 11.5% for former use
smokeless tobacco more than once per month in and 12% for current use, compared with 2.2%
the past and who had quit use ≥ 1 month ago. in never-users (Zaridze et al., 1986). [No defini-
Three studies found no increased risk in former tion was provided for former users.] The other
users of smokeless tobacco compared with never- cross-sectional study, conducted in Yemen,
users and found increased risk estimates for included 346 people diagnosed with leukopla-
current users (Ernster et al., 1990; Tomar et al., kia-like lesions based on the Axell criteria. Khat
154
Table 2.23 Cessation of smokeless tobacco use and risk of OPMDs
Reference Study population, number of Outcome Exposure Number of RR Comments

Location participants, study period, assessed category participants/ (95% CI)
follow-up period cases/controls (%
with OPMDs)
Cohort studies
Ernster Cohort of 1109 baseball players Leukoplakia Smokeless Analysis adjusted for age, race,
et al. (1990) who underwent training in (as per the tobacco: smoking, alcohol consumption, and
USA 1988 (median age, 18 yr), of Greer and Never-users 493 (1.4%) 1.0 (ref) dental hygiene
whom 75% used snuff and 21% Poulson Former users 138 (1.4%) 1.0 (0.2–5.0) Smokeless tobacco use defined at
chewed tobacco criteria) entry
Current 88(17.2%) 14.5 (5.7–36.7)
Leukoplakia was identified Snuff users had a significantly
chewing
by dentists on entry and was increased prevalence of leukoplakia
tobacco users
biopsy-confirmed compared with chewing tobacco
Current snuff 304 (55.6%) 86.9 (39.9–189.5) users, OR: 4.4 (2.4–9.3)
users Former users were those who had
used smokeless tobacco more than
once a month in the past but had not
used it within the previous month
Sinusas Cohort of 206 professionals in Leukoplakia [Crude estimates based No definition was given for former
et al. (1992) baseball organization, of whom (Greer and on reported numbers] users
Florida 42.7% were current users and Poulson Moist snuff No adjustment was made for
(USA) 16.5% were former users of and Axell and chewing smoking, but only 7 of the 206
smokeless tobacco (moist snuff criteria) tobacco: participants were smokers (3.4%); 4
and chewing tobacco) Non-users 79 (6.0%) 1.0 (ref) also used smokeless tobacco
Former users 32 (5.9%) [0.99 (0.18–5.35)]
Current 24 (7.6%) [1.32 (0.24–7.22)]
seasonal users
Current year- 39 (37.1%) [9.32 (3.29–26.37)]
round users
Tomar Cohort of 17 206 children aged Oral mucosal Chewing RRs adjusted for age, smoking, and
et al. (1997) 12–17 yr who participated lesions tobacco: alcohol consumption

USA in the 1986–1987 National classified Never-users 5195 (3.0%) 1.0 (ref) Former users of snuff or chewing
Survey of Oral Health in using Greer Former users 527 (6.0%) 1.3 (0.7–2.2) tobacco were defined as those who
schoolchildren in the USA, of and Poulson reported that they had ever used
Current users 273 (19.6%) 2.5 (1.3–5.0)
whom 3.1% used any smokeless and Axell these products but were not using
tobacco, 2.0% used any snuff, criteria as Snuff: them at the time of the survey
and 1.5% used any chewing “white or Never-users 5359 (1.9%) 1.0 (ref)
tobacco whitish oral Former users 329 (5.6%) 2.4 (1.0–6.1)
soft-tissue Current users 307 (34.9%) 18.4 (8.5–39.8)
lesions”
155
156

with OPMDs)
Shulman A sample of 17 235 people 48 different Smokeless Analyses adjusted for age, sex,
et al. aged ≥ 17 yr from the Third oral mucosal tobacco: denture status, race, and smoking
(2004) National Health and Nutrition lesion types Never-users 8143 (23.8%) 1.00 (ref) Specific type of smokeless tobacco
USA Examination Survey, 1988– classified Former users 183 (12.8%) 0.53 (0.25–1.13) used was not indicated
1994 (NHANES III) who based on Definition of former users is unclear
Current users 371 (60.3%) 3.90 (2.75–5.55)
underwent oral examination the WHO Oral lesions considered included
by dentists for identification of Guide to denture-related (8.4%) and tobacco-
oral lesions epidemiology related lesions (smokeless tobacco-
Lifestyle data were collected by and diagnosis related and nicotine stomatitis)
interview of oral (4.7%)
mucosal
diseases and
conditions
Cross-sectional studies
Zaridze Cross-sectional study in Leukoplakia Nass: Estimates are provided for never-
et al. (1986) Uzbekistan and pre- Never-users 625 (2.2%) 1.0 (ref) smokers
Uzbekistan 1569 people from a population- leukoplakia Former users 26 (11.5%) NR It is unclear whether these estimates
based cohort of men invited for were adjusted for potential
Current users 525 (12%) 5.6 (3.4–9.5)
medical examination by local confounding factors
authority; 42% used nass Definition of former users is unclear
Oral lesions clinically
diagnosed, and exposure
assessed by interview
Al-Tayar Cross-sectional study in 2014 Leukoplakia- Shammah: Analyses were adjusted for age,
et al. (2015) in Dawan Valley, Yemen, like lesions Never-users 248 (NR) 1.00 (ref) education level, and frequency of
Yemen involving 346 male residents (Axell Former users 30 (NR) 3.65 (1.40–9.50) shammah use
aged ≥ 18 yr. An interview- criteria) Former shammah users were those
Current users 68 (NR) 12.99 (6.34–26.59)
based questionnaire was used 80 individuals who had previously
to collect demographic, oral leukoplakias consumed shammah but stopped
hygiene, and shammah use were their consumption for ≥ 1 yr
information diagnosed
Smokers and khat users were at grade 1–4
excluded and 266 at
grade 0
with OPMDs)
Evstifeeva Case–control study in a Leukoplakia Nass: Analyses adjusted for age, smoking,
and region of Uzbekistan with and pre- Never-users 66/282 1.00 (ref) and alcohol consumption
Zaridze high incidence of oral and leukoplakia Former users 7/13 3.00 (1.08–8.32)
(1992) oesophageal cancer
Current users 118/171 3.86 (2.60–5.72)
Uzbekistan 191 men with leukoplakia and
466 controls
Data on use of nass quid,
cigarette smoking, and alcohol
consumption were collected by
interview from 1569 men
Fisher et al. Hospital-based case–control Leukoplakia Smokeless ORs adjusted for age, sex, smoking,
(2005) study in the USA (ICD-9 tobacco: alcohol consumption, and denture
West 90 cases (54 men) aged classification) Never-users 55/64 1.00 (ref) status
Virginia ≥ 18 yr with leukoplakia Former users 19/9 2.73 (0.69–10.84)
(USA) with histopathological
Current users 16/5 9.21 (1.49–57.00)
confirmation of hyperkeratosis
were compared with 78 (37 Snuff:
men) controls with periapical Never-users 64/71 1.00 (ref)
cysts from the same surgical Former users 8/5 0.98 (0.17–5.61)
pathology unit Current users 15/2 30.08 (2.67–338.48)
Smokeless tobacco and related
data collected by postal
questionnaires
CI, confidence interval; ICD, International Classification of Diseases; NHANES, National Health and Nutrition Examination Survey; NR, not reported; OPMDs, oral potentially
malignant disorders; OR, odds ratio; ref, reference; RR, relative risk; WHO, World Health Organization.

157
users and smokers were excluded. Past history of Group undertook primary data analyses from
shammah use was reported to increase the risk of unpublished data from one cohort study and
oral cancer by > 3-fold (3.65; 95% CI, 1.40–9.50). one case–control study, both conducted in India
Current shammah use further increased the risk and providing information on incidence of oral
of oral cancer in this population (Al-Tayar et al., cancer in relation to time since chewing cessa-
2015). [The study appears to be of limited power tion (Table 2.26).
because of a small number of former users. Also, One intervention study and three follow-up
duration of cessation was not defined.] studies focusing on assessing the relationship
Although the body of evidence appeared between cessation of chewing areca nut with
inconsistent with regard to the direction and added tobacco and the incidence of leukoplakia
the magnitude of risk of OPMDs, the RR esti- and OSF at the 5-year and 10-year follow-ups
mates for former users of smokeless tobacco (Gupta et al., 1986; Murti et al., 1990; Gupta
were generally lower than those for current users et al., 1992, 1995) were available to the Working
when compared with never-users as the reference Group (Table 2.27). Two additional case–control
group within each study. To clarify whether the studies focused on the incidence of OPMDs as the
distribution of covariance within individual outcome (Amarasinghe et al., 2010a; Worakhajit
studies could explain or potentially reveal under- et al., 2021) (Table 2.28).
lying risk trends, the Working Group undertook
additional analysis (Table 2.24). First, the RR in (a) Studies of oral cancer
former users compared with current users was (i) Evidence from the published literature
estimated for each study based on the Dirichlet– See Table 2.25.
multinomial distribution method (Gelman et al., The two reports of Jayalekshmi et al. (2009,
1995). Next, the recalculated risk estimates and 2011) were based on a large cohort established as
95% CIs were used to derive the variance and a part of the cancer registry in Karunagappally
covariance matrices of case and control popu- in Kerala, India. The cohort included 66 277
lations based on the tri-gamma distribution of men and 78 140 women; by 2005, 160 cases of
the corresponding variables, which were then oral cancer in men and 92 in women were iden-
approximated. [The meta-estimate reflected tified from the cancer registry. The association
nearly 70% reduction in RR for former users between chewing areca nut with added tobacco
compared with current users of smokeless and risk of oral cancer was examined overall for
tobacco (OR, 0.30; 95% CI, 0.14–0.46).] both men and women, as well as in men who were
never and current bidi smokers. In men, the risk
2.3.4 Chewing areca nut products (including of oral cancer in former chewers (OR, 2.1; 95%
betel quid) with added tobacco CI, 1.3–3.6) was comparable to that in current
chewers (OR, 2.4; 95% CI, 1.7–3.3). Among never
Two prospective cohort studies (Jayalekshmi
bidi smokers, the RR estimate in former chewers
et al., 2009, 2011), one nested case–control study
(OR, 3.2; 95% CI, 1.1–9.6) was lower than that
(Muwonge et al., 2008), two case–control studies
in current chewers (OR, 5.4; 95% CI, 3.0–9.0); in
(Balaram et al., 2002; Znaor et al., 2003), and a
current bidi smokers, the risk estimate for former
recent meta-analysis (Gupta et al., 2022) assessed
chewers was not significantly elevated compared
the effect of cessation of chewing areca nut with
with never-chewers (OR, 1.3; 95% CI, 0.6–2.9).
added tobacco on the incidence of oral cancer
In women, a 9-fold increased risk of oral cancer
(Table 2.25). To complement the evidence avail-
was reported in former chewers (OR, 9.2; 95% CI,
able from the published literature, the Working
158
Table 2.24 Cessation of smokeless tobacco use and risk of OPMDs – recalculation of the relative
risk for former chewers versus current chewers, and meta-analysis of results
Reference Study design Effect size for chewing habit Effect size for chewing
(versus never-chewers) habit with consideration of
covariance
Former chewers Current chewers Former chewers versus
Estimate (95% CI) Estimate (95% CI) current chewers
Estimate (95% CI)
Ernster et al. (1990) Cohort 1.0 (0.2–5.0) 14.5 (5.7–36.7) 0.07 (0.01–0.44)
Sinusas et al. (1992) Cohort 0.99 (0.18–5.35) 9.32 (3.29–26.37) 0.11 (0.02–0.48)
Tomar et al. (1997) Cohort 1.3 (0.7–2.2) 2.5 (1.3–5.0) 0.52 (0.31–0.87)
Shulman et al. (2004) Cohort 0.5 (0.3–1.1) 3.9 (2.8–5.6) 0.14 (0.06–0.30)
Al-Tayar et al. (2015) Cross-sectional 3.7 (1.4–9.5) 13.0 (6.3–26.6) 0.28 (0.11–0.73)
Evstifeeva and Zaridze Case–control 3.0 (1.1–8.3) 3.9 (2.6–5.5) 0.77 (0.28–2.14)
(1992)
Fisher et al. (2005) Case–control 2.7 (0.7–10.8) 9.2 (1.5–57.0) 0.30 (0.05–1.69)
Results of meta-analysis
Random-effect model 0.30 (0.14–0.46)
Fixed-effect model 0.34 (0.22–0.45)
CI, confidence interval; OPMDs, oral potentially malignant disorders.
4.6–18.1), whereas a nearly 5-fold increased risk a case–control study nested in the cohort of a
was reported in current chewers (OR, 5.5; 95% randomized controlled study in Trivandrum,
CI, 3.3–9.0) compared with never-chewers. This India (Sankaranarayanan et al., 2000). In this
study also examined risk of oral cancer by time study, the RR of chewing areca nut with added
since quitting chewing areca nut with added tobacco for the incidence of oral cancer was 4.3
tobacco. In men, ≥ 10 years of quitting appeared (95% CI, 3.1–6.1) in current chewers and 11.9
to reduce risks to levels comparable to those in (95% CI, 7.0–20.4) in former chewers compared
never-chewers, with differences in estimates that with never-chewers. [The Working Group noted
were not statistically significant. No such reduc- that the higher risk reported for former chewers
tion was noted in women (Jayalekshmi et al., could result from reverse causation.]
2009, 2011). [The higher risk in former chewers A matched case–control study enrolled 591
compared with current chewers in women is cases of oral cancer and 582 controls who were
difficult to understand and cannot be attri- frequency-matched (on age, sex, and centre)
buted to reverse causation, because the risk of in three centres in Bangalore, Madras, and
oral cancer in those with ≥ 10 years of quitting Trivandrum in southern India (Balaram et al.,
was still higher than that in current chewers. 2002). In men, the risk of oral cancer in former
Estimates were not adjusted for tobacco smoking chewers decreased progressively with increasing
and alcohol consumption, although these behav- time since chewing cessation compared with
iours were reported to be rare in women in this current chewers, reaching a reduction of 25%
population.] (RR, 0.75; 95% CI, 0.23–2.52) ≥ 10 years after
Muwonge et al. (2008) enrolled 282 incident cessation. In women, on contrast, the risk of oral
oral cancer cases and 1410 matched controls in cancer was higher for ≥ 10 years of cessation than
159
160

Table 2.25 Cessation of chewing of areca nut products (including betel quid) with added tobacco and risk of oral cancer –
observational studies
Reference Study population, sample Study design, number of Exposure category RR (95% CI) Comments
Location selection, response rate participants, study period, Number of exposed
follow-up time cases/controls
Cohort studies
Jayalekshmi Women aged 30–84 yr in Prospective cohort study Women: Poisson regression model was used
et al. (2009) Karunagappally, Kerala, designed to link 78 140 enrolled Never-chewers: 25 1.0 (ref) to calculate relevant estimates
India were enrolled with house- women participating in the Former chewers: 14 9.2 (4.6–18.1) Adjusted for age and family income
to-house surveys to have baseline survey with the cancer Estimates not adjusted for tobacco
Current chewers: 53 5.5 (3.3–9.0)
baseline information registry. Baseline information smoking and alcohol consumption;
The response rate was 93% was collected on lifestyle, Duration of cessation however, according to the authors
including tobacco chewing, (yr): these habits are rare in women in
and sociodemographic factors Current chewers: 53 1.0 (ref) this population
in 1990–1997. By the end of < 10: 7 1.7 (0.8–3.7)
2005, 92 oral cancer cases were ≥ 10: 4 2.6 (0.9–7.2)
identified Never-chewers: 25 0.2 (0.1–0.3)
Jayalekshmi Men aged 30–84 yr in The same prospective cohort Men (cases/person-yr): Poisson regression model was used
et al. (2011) Karunagappally, Kerala, study was designed as above, Overall: to calculate relevant estimates
India were enrolled with house- but the target participants were Never-chewers: 64 1.0 (ref) Adjusted for age and family income.
to-house surveys to have 66 277 men. By the end of 2005, Former chewers: 19 2.1 (1.3–3.6) Estimates not adjusted for alcohol
baseline information 160 oral cancer cases were consumption
The response rate was 93% identified
In never bidi smokers:
Never-chewers: 18 1.0 (ref)
Former chewers: 4 3.2 (1.1–9.6)
In current bidi smokers:
Never-chewers: 38 1.0 (ref)
Former chewers: 7 1.3 (0.6–2.9)
(yr):
Current chewers: 75 1.0 (ref)
< 10: 12 1.1 (0.6–2.0)
≥ 10: 2 0.3 (0.1–1.2)
Never-chewers: 64 0.4 (0.3–0.6)
Balaram et al. Patients with incident oral Matched case–control study Men: The sex-related differences in the
(2002) cancer and their hospital- conducted in 1996–1999 Duration of cessation results may be attributed to selection
India based matched controls in Case group: 591 incident cases (yr): bias for women, who may be less
Bangalore, Madras, and of oral cancer Current chewers: 120/37 1.0 (ref) likely to go to hospitals, because
Trivandrum centres Control group: 582 hospital < 10: 45/14 1.02 (0.45–2.29) the proportion of ever-chewers
controls, frequency-matched in women in such hospital-based
≥ 10: 14/6 0.75 (0.23–2.52)
to cases on age and sex and on controls was lower than that in
centre (relatives and friends of Women: women in the general population
patients admitted to hospitals Duration of cessation
because of diseases other than (yr):
oral cancer in Bangalore and Current chewers: 203/29 1.0 (ref)
Madras, and outpatients in < 10: 31/6 0.72 (0.23–2.21)
Trivandrum) ≥ 10: 17/3 0.97 (0.23–4.11)
Confounding factors adjusted
for in the logistic regression
model were age, location,
education level, and only for
men: tobacco smoking (never/
ever) and alcohol consumption
(never/ever)
Znaor et al. Male patients with Case–control study conducted Duration of cessation
(2003) oral cancer as cases in in 1993–1999 (yr):
India Chennai (Tamil Nadu) and Case group: 1563 oral cancer Current chewers: 1.0 (ref)
Trivandrum (Kerala) cases 640/460
Control group: 1711 male 2–4: 93/41 1.15 (0.75–1.77)
patient controls from both
5–9: 59/20 1.60 (0.92–2.81)
centres and 1927 male healthy
hospital visitors in Chennai 10–14: 30/19 0.71 (0.37–1.35)

Confounding factors ≥ 15: 30/19 0.67 (0.36–1.26)
adjusted for in the logistic
regression model were age,
location, education level,
tobacco smoking, and alcohol
consumption (never/ever)
161
162

Muwonge People aged ≥ 35 yr in Nested case–control design Overall: The high risk in former chewers
et al. (2008) Trivandrum District based on data from a Never-chewers: 80/915 1.0 (ref) compared with current chewers
India randomized control trial Current chewers: 4.3 (3.1–6.1) was observed in both sexes in this
for oral cancer screening 160/445 study. It may be the result of reverse
conducted in 1996–2004 in causation (i.e. the more severe cases
Former chewers: 42/50 11.9 (7.0–20.4)
Trivandrum (Kerala) are more likely to quit chewing)
Case group: 282 incident oral Men:
cancer cases Never-chewers: 64/561 1.0 (ref)
Control group: 1410 controls Current chewers: 78/222 2.7 (1.8–4.2)
matched on sex, age (± 1 yr), Former chewers: 21/32 5.9 (3.0–11.7)
area of residence, and screening Women:
participation
Never-chewers: 16/354 1.0 (ref)
Current chewers: 9.5 (5.0–18.0)
82/223
Former chewers: 21/18 39.0 (15.0–101.8)
Meta-analysis
Gupta et al. 2 cohort and 4 case– Never-chewers 1.0 (ref) 6 of the 7 studies were restricted to
(2022) control studies and one Former chewers 6.87 (4.10–11.52) men or provided sex-specific results;
case–control study nested Current chewers 6.29 (3.83–10.33) 4 studies did not provide a clear
in a randomized trial definition of former users; 2 case–
control studies provided relative
(yr):
risks of oral cancer by duration of
< 10 1.21 (0.90–1.63) cessation
> 10 0.72 (0.48–1.07)
CI, confidence interval; ref, reference; RR, relative risk; yr, year or years.
for < 10 years of cessation. [A selection bias may (ii) Evidence from primary data analyses
explain the results in women – who may be less See Table 2.26.
likely to go to hospitals – because the proportion Data collected at two sites in India were used
of ever-chewers in women in such hospital-based for primary analysis by the Working Group to
controls was lower than that in women in the assess the impact of quitting chewing betel quid
general population.] with added tobacco on the risk of oral cancer.
Another case–control study, conducted in The first primary analysis used data derived
1993–1999 at the cancer institute in Chennai from the cluster-randomized controlled trial in
(Tamil Nadu) and the Regional Cancer Centre Trivandrum, India (Sankaranarayanan et al.,
in Trivandrum (Kerala), India, enrolled 1563 2000). The data were from a cohort of 191 870
male oral cancer cases and 3638 male hospital participants aged ≥ 35 years enrolled in 1996–
controls (Znaor et al., 2003). The risk estimate 2006. Incident oral cancer was ascertained
of oral cancer compared with current chewers until 31 December 2009; the average follow-up
decreased by 29% (RR, 0.71; 95% CI, 0.37–1.35) period was 7 years. The main exposure of interest
for 10–14 years of cessation and by 33% (RR, 0.67; included the chewing status (current, former,
95% CI, 0.36–1.26) for ≥ 15 years of cessation. and never) and duration of cessation. The major
[The selection of the control group was different: confounders were adjusted for in the Cox propor-
the hospital control from both centres and an tional hazards regression model. Per year of quit-
additional healthy control from only one of the ting chewing betel quid with added tobacco, the
two centres. In addition, compared with cases of risk of oral cancer decreased significantly (HR,
oral cancer, the control group was younger and 0.97; 95%, 0.96–0.99). However, for participants
educated. Although these demographic char- with > 15 years of cessation, the risk of oral
acteristics were considered in the multivariate cancer remained high (HR, 2.5; 95% CI, 1.6–3.7)
analysis, residual confounding may still exist.] compared with current chewers. Compared with
In the last days of the Working Group people with < 2 years of cessation, those with
meeting, a meta-analysis was made available > 10 years of cessation had a lower risk of oral
to the Working Group that combined seven cancer (HR for 10–15 years, 0.8; 95% CI, 0.3–2.0;
reports to assess the potential benefit of long- HR for > 15 years, 0.7; 95% CI, 0.4–1.4), although
term cessation of chewing areca nut with added this was not statistically significant. [Duration
tobacco (Gupta et al., 2022). [The meta-analysis of cessation was imputed using current age and
includes all the cohort and case–control studies duration of chewing, which may explain the wide
reported above (Balaram et al., 2002; Znaor 95% CIs. There are issues with identifiability and
et al., 2003; Muwonge et al., 2008; Jayalekshmi collinearity of time since quitting with duration
et al., 2009, 2011).] The meta-RR of oral cancer and age. The median age was different between
for former chewers with < 10 years of cessation current and former chewers: 52 years for current
compared with current chewers was increased chewers and 62–65 years for the several catego-
(1.21; 95% CI, 0.90–1.63) and for former chewers ries of former chewers.]
with > 10 years of cessation was decreased (0.72; The second primary analysis used data
95% CI, 0.48–1.07). [The increased risk after derived from cancer hospitals in India. A
< 10 years of cessation could be due to reverse case–control study design was applied. Cases
causation. The sample size was still insufficient were patients with oral cavity cancer diag-
to reach statistical significance in the reversal of nosed in the cancer hospital from three cities:
risk of oral cancer after long-term cessation.] Mumbai, Varanasi, and Guwahati. Controls
163
164

Table 2.26 Cessation of chewing of areca nut products (including betel quid) with added tobacco and risk of oral cancer –
primary data analyses performed by the Working Group
Study Study Study design, study period, number Cancer end- Exposure OR (95% CI) Comments
Location population of participants, information on betel point category
quid chewing, and confounders Number of
considered cases/controls
Kerala oral cancer Kerala, India Cluster-randomized trial of 191 870 Oral cancer 202 cases in Difference between categories
screening trial participants aged ≥ 35 yr who were incidence ever-chewers, 65 of duration of cessation was
(several previous recruited in 1996–2006 and followed during 7 yr of cases in former not statistically significant
publications) up until 31 December 2009. Data on follow-up chewers Duration of cessation was
exposures collected at baseline was Compared with imputed using current age
used for analysis current chewers: and duration of chewing.
The analysis was Cox proportional Current There are issues with
hazards regression chewers: identifiability and collinearity
The key exposure was duration of of duration of cessation with
202/37 419 1.0 (ref)
cessation of chewing of betel quid duration of use and age
(primarily with added tobacco). This Duration of There is age confounding
metric was derived using simple, cessation (yr): between current and former
single-value imputation of age at < 2: 13/567 3.7 (2.1–6.5) chewers (the median age is
initiation of chewing (10-year birth 2–5: 6/195 5.1 (2.2–11.8) 52 yr for current chewers, and
cohort and sex-specific, estimated 5–10: 12/390 5.1 (2.8–9.4) 62, 62, 62, 59, and 65 yr for
from GATS India 2009–2010), 10–15: 7/435 3.1 (1.4–6.5) former chewers with < 2, 2–5,
duration of chewing, and age at 5–10, 10–15, and > 15 yr of
> 15: 27/1854 2.5 (1.6–3.7)
study participation. Individuals with cessation, respectively)
negative duration of cessation were Compared with
excluded from analyses quitting < 2 yr:
Analyses were adjusted for age, sex, Duration of
education level, chewing duration cessation (yr):
and intensity, smoking duration and <2 1.0 (ref)
intensity, and alcohol consumption 2–5 1.4 (0.5–3.8)
duration and intensity (days per week
5–10 1.4 (0.6–3.1)
of alcohol consumption)
Two sets of analyses were conducted: 10–15 0.8 (0.3–2.0)
(1) analyses restricted to ever- > 15 yr 0.7 (0.4–1.4)
chewers (n = 40 860), and (2) analyses Per year of 0.97 (0.96–0.99)
restricted to former chewers (n = 3441) cessation
Study Study Study design, study period, number Cancer end- Exposure OR (95% CI) Comments
Location population of participants, information on betel point category
quid chewing, and confounders Number of
considered cases/controls
Unpublished, Tata Mumbai, Hospital-based case–control study of Buccal 391 cancers in Inverse relationship between
Memorial Centre, Varanasi, patients with buccal mucosa cancer mucosa current chewers, the categories of duration
Mumbai. Study is and and controls, matched on 5-year age, cancers 99 cancers in of cessation and the risk of
part of a GWAS of Guwahati, sex, and site former chewers, buccal mucosa cancer
buccal cancers India The main exposure was duration of 1367 controls
cessation of chewing Compared with
Logistic regression analyses were current chewers:
adjusted for age, sex, study site, alcohol Current
consumption intensity, smoking chewers:
duration and intensity, and chewing
391/969 1.0 (ref)
duration and intensity
Duration of
cessation (yr):
< 1: 15/146 3.1 (1.4–7.1)
2–5: 25/136 1.5 (0.8–2.8)
5–10: 14/41 1.1 (0.4–2.5)
≥ 10: 45/75 0.7 (0.3–1.5)
Per year of 0.98 (0.95–1.02)
cessation
CI, confidence interval; GATS, Global Adult Tobacco Survey; GWAS, genome-wide association study; OR, odds ratio; ref, reference; yr, year or years.

165
matched on age (5-year band), sex, and site were tobacco was also higher in the intervention
selected from the hospital. The main exposures cohort than in the control cohort (28% vs 9%).
of interest were the status of chewing betel quid The education programme showed significant
with added tobacco and duration of cessa- effectiveness in reducing the risk of leukoplakia:
tion. The confounding factors were adjusted for reported rate ratios were 0.51 [95% CI, 0.28–0.93]
in the logistic regression analysis. There were in men and 0.19 [95% CI, 0.11–0.30] in women for
391 cancers in current chewers, 99 cancers in chewers, and 0.20 [95% CI, 0.13–0.30] in men and
former chewers, and 1367 matched controls. A 0.19 [95% CI, 0.02–2.12] in women for chewers
2% reduction in risk of oral cavity cancer was who also smoked. [Because this study was not
calculated per year of cessation (OR, 0.98; 95% randomized, the effectiveness may be affected by
CI, 0.95–1.02). The risk of oral cavity cancer was unadjusted confounding factors, such as demo-
lower in former chewers with > 10 years of cessa- graphic characteristics. Age was not adjusted
tion (OR, 0.7; 95% CI, 0.3–1.5) compared with for, and only stratification by sex was provided.
current chewers. [Neither estimate was statisti- A second unadjusted confounding factor was
cally significant.] baseline socioeconomic status, which may have
differed between the intervention cohort and
(b) Studies of OPMDs the control cohort (recruited 10 years earlier
(i) Intervention study than the intervention cohort). Also, cases in the
See Table 2.27. intervention cohort included individuals who
The intervention study (Gupta et al., 1986) had reduced the intensity of chewing, who had
enrolled tobacco chewers and smokers older stopped chewing, and those who had continued
than 15 years in three districts in India in 1966: chewing.]
Ernakulam District in Kerala, Srikakulam Murti et al. (1990) reported on the cohorts in
District in Andhra Pradesh, and Bhavnagar Ernakulam District, focusing on the incidence
District in Gujarat. This is currently the only of OSF, with a follow-up period of 10 years. The
study worldwide that was designed to evaluate education programme resulted in a RR reduc-
the effectiveness of an education programme for tion of OSF incidence of 62% (RR, 0.38; [95% CI,
tobacco users in reducing incidence of OPMDs. 0.06–2.24]) in men and 37% (RR, 0.63; [95% CI,
The intervention arm received primary preven- 0.25–1.65]) in women for chewers. [The major
tion in the form of an education programme limitation of this study is the small number of
with professional advice provided by dentists OSF events in chewers.]
and trained social scientists, as well as radio Gupta et al. (1992) also reported on a 10-year
broadcasts and newspaper articles. Ernakulam follow-up of the cohorts, focusing on leukoplakia.
District was the only one of the three districts The incidence of leukoplakia was reduced signifi-
in which chewing betel quid with added tobacco cantly, by 37% (RR, 0.63; [95% CI, 0.37–1.06]) in
was the main habit in the population; therefore, men and by 55% (RR, 0.45; [95% CI, [0.32–0.63])
only results from that district were relevant here. in women for chewers, and by 63% (RR, 0.37;
After the 5-year follow-up, the proportion of [95% CI, 0.25–0.54]) in men for chewers who also
individuals who had stopped chewing betel quid smoked. In a later report, Gupta et al. (1995) also
with added tobacco was higher in the interven- reported on a 10-year follow-up by comparing the
tion cohort than in the control cohort (9% vs 3%), incidence of leukoplakia and of OLP between the
and the proportion of individuals who reduced “stopped” category (former chewers who stopped
the intensity of chewing betel quid with added chewing ≥ 6 months ago) and “all others” (other
categories combined) using the intervention
166
Table 2.27 Cessation of chewing of areca nut products (including betel quid) with added tobacco and risk of OPMDs –
intervention study
Reference Study population, sample Study design, number of OPMD end- Exposure RR Comments
Location selection, response rate participants, intervention, point category 95% CI)
study period, follow-up Number of cases,
time intervention/
control
Gupta et al. Tobacco users (chewers Prospective study with Incidence of Chewing only: 10 yr of follow-up of the
(1986, 1992, and smokers) aged ≥ 15 yr intervention cohort and leukoplakia Men: [32/25] 0.63 [(0.37–1.06)] main study focusing
1995); Murti in 3 districts in India control cohort Women: [60/72] 0.45 [(0.32–0.63)] on Ernakulam District
et al. (1990) Two distinct cohorts Intervention cohort in Kerala conducted by
Mixed chewing
Kerala, India were selected in each (n = 12 212) and control Gupta et al. (1986)
and smoking:
district through house-to- cohort (n = 6075) in Chewing betel quid
house surveys to have an Ernakulam District Men: [44/68] 0.37 [(0.25–0.54)] with added tobacco
interview and a clinical Recruitment in 1976–1985 Women: 0.52 [(0.01–29.85)] was the main habit
mouth examination at for intervention cohort, and in the population in
baseline and regular in 1966–1977 for control Ernakulam District in
follow-up cohort Kerala
≥ 97% follow-up rate for 10-yr follow-up Results are crude
the intervention cohort, Intervention was an incidence stratified by
and 84–95% follow-up rate education programme sex, not adjusted for age.
for the control cohort through professional advice Only 5-year age-adjusted
(dentist and social scientist) incidence was reported
and social media for total tobacco use
Higher stoppage of chewing (rather than different
(15.1% vs 2.3% for men; categories)
18.4% vs 7.8% for women) Baseline socioeconomic
and of mixed chewing and status may have differed
smoking (3.8% vs 2.0% between the intervention
for men;13.2% vs 3.8% for cohort and the control
women) cohort (10 yr earlier than
the intervention cohort)
81–93% follow-up rate for Intervention cohort Incidence of OSF Men: 2/3 0.38 (0.06–2.24) Murti et al. (1990).

the intervention cohort, (n = 6341 chewers) and Women: 9/8 0.63 (0.25–1.65) Analysis restricted to
and 71–75% follow-up rate control cohort (n = 3809 chewers only. Events
for the control cohort chewers) of OSF are too rare to
have sufficient statistical
power
167
168

Reference Study population, sample Study design, number of OPMD end- Exposure RR Comments
Location selection, response rate participants, intervention, point category 95% CI)
study period, follow-up Number of cases,
time intervention/
control
Gupta et al. Incidence of OLP “stopped”/“all others”: Gupta et al. (1995).
(1986, 1992, Men: 1/30 0.02 [(0.00–0.13)] Part of the main study
1995); Murti Women: 18/90 1.29 [(0.78–2.14)] focusing on Ernakulam
et al. (1990) District in Kerala
Incidence of “stopped”/“all others”:
(cont.) conducted by Gupta
leukoplakia Men: 4/33 0.81 [(0.29–2.28]) et al. (1986) with 10 yr of
Women: 5/52 0.30 [(0.12–0.75)] follow-up
Only chewers in the
intervention cohort
were considered. The
“stopped” category
included former chewers
who stopped chewing
≥ 6 months ago
CI, confidence interval; OLP, oral lichen planus; OPMDs, oral potentially malignant disorders; OSF, oral submucous fibrosis; RR, relative risk; vs, versus.
cohort only. Cessation of chewing betel quid by duration of quitting chewing was available to
with added tobacco significantly reduced the judge the strength of the results.]
incidence of leukoplakia, by 19% (RR, 0.81; 95%
CI, 0.29–2.28) in men and 70% (RR, 0.30; 95% CI, 2.3.5 Chewing areca nut products (including
25–88%) in women for former chewers, whereas betel quid) without tobacco
there was no effect of chewing cessation in
reducing the incidence of OLP. [There was a lack Published evidence on the impact of quitting
of statistical power for OLP incidence because of chewing areca nut products without tobacco on
too few OLP events.] the risk of oral cancer consisted of four case–
control studies with data in former chewers and
(ii) Observational studies current chewers compared with never-chewers
See Table 2.28. (Table 2.29). In addition, a recent meta-analysis
In a case–control study in Sri Lanka, chewers of 14 case–control studies provided estimates of
of betel quid were categorized as daily, occa- oral cancer incidence after cessation of chewing
sional, and former chewers (Amarasinghe et al., areca nut without tobacco (Gupta et al., 2022).
2010a). Two thirds of the chewers used betel quid To complement the evidence available from the
with added tobacco: 82% among the cases and published literature, the Working Group under-
32% among the controls. The incidence of leuko- took primary analyses from unpublished data
plakia, OSF, and OLP were used as outcomes. from three large cohort studies and one case–
For daily chewers, the risk of OPMDs increased control study providing information on inci-
10-fold (OR, 10.6; 95% CI, 3.6–31.0) compared dence of oral cancer (Table 2.30) in relation to
with never-chewers. For former chewers, the time since chewing cessation and age at quitting.
incidence of OPMDs increased 2-fold (OR, Published evidence on the impact of quitting
2.4; 95% CI, 0.4–14.5), similarly to occasional chewing areca nut products without tobacco
chewers (OR, 2.0; 95% CI, 0.4–9.4). The Working on the risk of OPMDs consisted of three case–
Group calculated a lower, non-significant RR control studies and two cross-sectional studies
of OPMDs for former chewers compared with (Table 2.31). Similarly, the Working Group
current chewers [OR, 0.23; 95% CI, 0.03–1.79]. undertook primary analyses from unpublished
[The Working Group noted two major limitations data from three large cohort studies and one
of this study: (i) the results were for a mixture of case–control study providing information on
chewers of betel quid with and without tobacco, incidence of OPMDs in relation to time since
and (ii) no information on the time since quitting chewing cessation and age at quitting (Table 2.32).
was available.]
A case–control study in northern Thailand (a) Studies on oral cancer
(Worakhajit et al., 2021) was conducted in 2019– (i) Evidence from the published literature
2021 to investigate the relationship between betel See Table 2.29.
quid chewing and risk of OPMDs. This study Ko et al. (1995) reported on a hospital-based
enrolled 562 cases (people with identified OPMD) matched case–control study that assessed the
and 886 controls (people without OPMD). Using independent effects of use of betel quid without
those with < 5 years of quitting as the reference tobacco, cigarette smoking, and alcohol consump-
group, those with ≥5 years of quitting had a tion on oral cancer, as well as the synergistic effect
slightly lower, but not statistically significantly of these behaviours. [Information on time since
so, RR of OPMDs (OR, 0.94; 95% CI, 0.22–3.92). chewing cessation was lacking.] Current chewers
[Not enough information on the number of cases were defined as those chewing ≥ 1 quid daily for
169
170

Table 2.28 Cessation of chewing of areca nut products (including betel quid) with added tobacco and risk of OPMDs –
Reference Study Study design, number of OPMDs end- Exposure OR Interpretation/comments

Location population, participants, study period, follow- point category (95% Cl)
sample up time Cases/controls
selection,
response rate
Amarasinghe People aged Two-phase designed study Leukoplakia, Compared Study based on a screening
et al. (2010a) ≥ 30 yr in Phase 1: Cross-sectional community OSF, and OLP with non- programme for OPMDs. Results
Sri Lanka Sabaragamuwa survey with a house-to-house combined chewers: were for a mixture of chewers
Province method to screen for OPMDs for Non-chewers: 1.0 (ref) with and without combined use
1029 people randomly selected by 4/277 of tobacco
a multistage, stratified, clustered Former chewers: 2.4 (0.4–14.5)
sampling technique 2/36
Phase 2: Nested case–control study
Occasional 2.0 (0.4–9.4)
with a case group (n = 101) who were
chewers: 3/83
identified as having OPMDs (i.e.
leukoplakia, erythroplakia, OSF, Daily chewers: 10.6 (3.6–31.0)
OLP) and a control group (n = 728) 92/332
without OPMDs from Phase 1
Adjusted for sex, age, education
level, occupation, BMI, tobacco
smoking, and alcohol consumption
Worakhajit People aged Case–control study design, OPMDs Duration of
et al. (2021) ≥ 40 yr in north- conducted in 2019–2021 cessation (yr):
Thailand eastern Thailand Community-based screening at <5 1.00 (ref)
the village level for 392 396 people ≥5 0.94
with an oral cancer risk screening (0.22–3.92)
questionnaire administered by
health-care volunteers
1448 people aged ≥ 40 yr were
enrolled, including 562 with
identified OPMD as the case group
and 886 without OPMD as the
control group
BMI, body mass index; CI, confidence interval; OLP, oral lichen planus; OPMDs, oral potentially malignant disorders; OSF, oral submucous fibrosis; ref, reference; yr, year or years.
Table 2.29 Cessation of chewing of areca nut products (including betel quid) without tobacco and risk of oral cancer –
case–control studies
Reference Study Study design, study period, number of Cancer Exposure category OR Comments
Location population participants, information on betel quid end-point Number of cases/ (95% CI)
chewing, and confounders considered controls
Ko et al. Patients at a Hospital-based matched case–control study in Oral cancer Non-chewers: 31/153 1.0 (ref) Information
(1995) medical centre 1992–1993 Current chewers: 6.9 (3.1–15.2) on duration of
Taiwan in Kaohsiung, Case group: 107 patients with oral cancer with 71/42 cessation was
(China) southern Taiwan diagnosis confirmed by histopathology Former chewers: 5/5 4.7 (0.9–22.7) lacking
(China) Control group: 200 age- and sex-matched Insufficient
controls consisting of non-carcinoma patients statistical power
treated during the same period because of too few
Chewers chewing ≥ 1 quid daily for ≥ 1 yr were former chewers
defined as current chewers
Confounding factors adjusted for in the
multivariate analysis included education level,
occupation, alcohol consumption, cigarette
smoking, residence, marriage status, religion,
ethnicity, and dietary habits
Thomas Cases were Case–control study in 1985–1987 Oral cancer Non-chewers: 2/9 1.0 (ref) This study had an
et al. (2007) patients with Case group: 143 patients with first diagnosis Current daily 1.29 (0.25–6.51) extremely high
Papua New oral cancer of clinically apparent oral squamous cell chewers: 124/375 prevalence of ever
Guinea hospitalized in carcinoma Current occasional 0.98 (0.17–5.74) betel quid chewing
6 hospitals, and Control group: 477 controls were those chewers: 8/37
controls were admitted or related to someone admitted to the
Former chewers: 0.57 (0.10–3.28)
those related same hospital
9/56
to someone Frequency-matching was performed on age,
admitted to the sex, and geographical location Ever-chewers: 1.10 (0.22–5.51)
same hospitals Confounding factors in the multivariate 141/468
analysis included age, sex, province, residence,
income, education level, and frequency of
smoking

171
172

Lee et al. Patients with Multicentre case–control study in 2001–2007 Oral Oral cancer: Information
(2012) carcinoma Case group: Of the enrolled patients with cancer and Non-chewers: 1.0 (ref) on duration of
Taiwan on the upper cancer, 810 with oral cancer and 231 with pharyngeal 136/2002 cessation was
(China) aerodigestive pharyngeal cancer cancer Current chewers: 16.7 (12.1–23.0) lacking
tract to Control group: 2250 age- and sex-matched 450/160 The possibility of
gastrointestinal controls selected from the same hospital reverse causation is
tract at 2 during the same period. Confounding factors a concern
224/88
medical centres in the multivariate analysis included sex, The pharyngeal
in Kaohsiung, age, ethnicity, education level, drink-years of Pharyngeal cancer: cancer group
southern Taiwan alcohol consumption, pack-years of cigarette Non-chewers: 1.0 (ref) included
(China) smoking, and consumption of vegetables and 55/2002 oropharyngeal and
fruits Current chewers: 9.3 (6.1–14.2) hypopharyngeal
147/160 cancers
29/88
Wu et al. Male patients Hospital-based case–control study in 2010– Oral Oral cancer: Control group
(2016) at one medical 2014 cancer and Non-chewers: 67/446 1.0 (ref) selected from the
Taiwan centre in Tainan Case group: 487 male patients aged ≥ 20 yr pharyngeal Current chewers: 8.05 (5.10–12.71) otolaryngology
(China) City with a new diagnosis of head and neck cancer. cancer 113/66 and stomatology
Of them, 313 had oral cancer and 119 had oro- departments
hypopharyngeal cancer may not be
133/105
Control group: 617 male controls matched to representative
the cases on age and from the same department Duration of of the general
as the cases but undergoing surgery for cessation (yr): population in their
non-cancerous disease not related to alcohol Current chewers: 1.0 (ref) risk of oro- and
consumption, betel quid use, or smoking, and 113/66 hypopharyngeal
without history of cancer diagnosis 0.0–9.9: 67/59 0.72 (0.44–1.17) cancer
Confounding factors in the multivariate 10.0–19.9: 48/23 1.42 (0.77–2.61)
analysis included age, education level, cigarette ≥ 20: 15/25 0.34 (0.16–0.73)
smoking (pack-year categories), and alcohol
Per year of cessation 0.976 (0.952–1.001)
consumption (frequency)
Oro- and hypo-
pharyngeal cancer:
Non-chewers: 31/446 1.0 (ref)
Current chewers: 4.80 (2.57–8.99)
45/66
43/105
Duration of
cessation (yr):
Current chewers: 1.0 (ref)
45/66
0.0–9.9: 28/56 0.74 (0.39–1.42)

10.0–19.9: 9/23 0.63 (0.24–1.61)
≥ 20: 6/25 0.26 (0.09–0.78)
Per year of cessation 0.967 (0.933–1.001)
173
174

Meta-analysis
Gupta et al. 4 case–control Tobacco smoking was adjusted for in all the Oral cancer Non-chewers 1.0 (ref) Most studies
(2022) studies studies, and alcohol consumption was adjusted included only or
for in all studies except one Former chewers 5.61 (2.24–14.04) predominantly
Current chewers 7.89 (3.90–15.98) male participants.
The duration
of cessation for
defining former
chewers was > 1 yr
in one study and
> 6 months in one
study; two studies
did not mention
this aspect
CI, confidence interval; OR, odds ratio; ref, reference; yr, year or years.
≥ 1 year. Compared with never-chewers, the OR 2.0–6.1) for former chewers and 9.3 (95% CI,
for the risk of oral cancer in former chewers was 6.1–14.2) for current chewers, compared with
lower (4.7; 95% CI, 0.9–22.7) than that in current never-chewers. [The Working Group calculated
chewers (6.9; 95% CI, 3.1–15.2). [The Working that the OR for former chewers versus current
Group calculated that the OR for oral cancer in chewers was 0.92 (95% CI, 0.61–1.39) for oral
former chewers versus current chewers was 0.68 cancer and 0.38 (95% CI, 0.20–0.70) for pharyn-
(95% CI, 0.12–3.79). The Working Group noted geal cancer. Because this is a hospital-based case–
three limitations: (i) selecting controls from the control study with study participants recruited
ophthalmology and physical check-up depart- from patients, the possibility that patients quit
ments may have a tendency to enrol few chewers, chewing after knowing the diagnosis of oral
and this selection bias may lead to overestima- cancer cannot be ruled out.]
tion of the risk of oral cancer for current chewers In another hospital-based case–control study
and former chewers; (ii) many confounders may to investigate the association between betel quid
have required adjustment; and (iii) few former chewing and the risk of HNC at different sites
chewers led to insufficient statistical power.] (Wu et al., 2016), 487 male cancer patients and
In a hospital-based case–control study in 617 age- and sex-matched controls were enrolled
Papua New Guinea (Thomas et al., 2007), daily in 2010–2014. Information obtained by ques-
chewing of betel quid resulted in the highest tionnaire included data for the three categories
RR of oral cancer (OR, 1.29; 95% CI, 0.25–6.51) of betel quid chewers – current, former (stopped
compared with occasional chewing (OR, 0.98; > 6 months ago), and never. Time since cessation
95% CI, 0.17–5.74) and with former chewing for the former chewers was expressed as a contin-
(OR, 0.57; 95% CI, 0.10–3.28). [The Working uous variable in years or an ordinal variable in
Group calculated an OR for oral cancer in former 10-year categories (0–9.9 years, 10–19.9 years,
chewers compared with current chewers of 0.44 and ≥ 20 years). For oral cancer, the OR for
(95% CI, 0.04–4.73). There were very few never- former chewers (6.43; 95% CI, 4.25–9.73) was
chewers (the reference group): 1.4% (2 of 143) in lower than that for current chewers (8.05; 95%
the case group and 1.9% (9 of 477) in the control CI, 5.10–12.71) compared with never-chewers.
group. In addition, because controls were selected [This resulted in a 20% reduction in RR of oral
from patients who had a diagnosis unrelated to cancer (OR, 0.80; 95% CI, 0.51–1.24), calculated
oral cancer but potentially related to other betel by the Working Group.] A significant trend
quid-related diseases, this may lead to underesti- with duration of cessation was noted, with a RR
mation of the risk of oral cancer.] reduction that was significant for ≥ 20 years of
A multicentre case–control study was betel quid cessation for oral cancer (OR, 0.34;
conducted in Taiwan (China) to assess the effect 95% CI, 0.16–0.73) and for pharyngeal cancer
of consumption of betel quid without tobacco on (including oropharyngeal and hypopharyngeal
the risk of aerodigestive tract cancers at different cancers) (OR, 0.26; 95% CI, 0.09–0.78), but the
anatomical sites, with adjustment for age, risk was still greater than that in never-chewers.
ethnicity and education level (Lee et al., 2012). Each year of cessation of betel quid chewing was
Compared with never-chewers, the OR for the associated with a 2.4% RR reduction (OR, 0.976;
risk of oral cancer in former chewers was 15.3 95% CI, 0.952–1.001) for oral cancer and a 3.3%
(95% CI, 10.6–22.0) and in current chewers was RR reduction (OR, 0.967; 95% CI, 0.933–1.001)
16.7 (95% CI, 12.1–23.0). For pharyngeal cancer for pharyngeal cancer. [The strength of this study
(including oropharyngeal and hypopharyngeal is to address a non-linear dose–response rela-
cancers), the estimated ORs were 3.5 (95% CI, tionship between the amount and the duration
175
of chewing and duration of cessation associated cessation was collected at entry. The study design
with HNC including oral cancer and pharyngeal and implementation were very similar across
cancer by using a spline regression method. The studies. The three cohorts were followed up over
study has three main limitations. First, because time to ascertain OPMDs and oral cancers. The
the control group was selected from the otolaryn- case–control study was derived from one of the
gology and stomatology departments, the source studies in Taiwan (China) on OPMDs and oral
population for the control group may be different cancer in collaboration with the United States
from that for the case group; thus, selection bias National Cancer Institute.
cannot be ruled out. Second, recall bias in the Results from three cohort studies showed
retrieval of information on chewing behaviour statistically significant trends of reduced risk
cannot be avoided. Third, the findings may have of oral cancer with an increase in time since
been affected by other unadjusted confounding quitting (Ptrend < 0.01). The most significant
factors, such as occupation, although age, educa- reduction was noted for ≥ 20 years of quitting
tion level, alcohol consumption, and smoking in Keelung and Tainan and for ≥ 10 years of
had been controlled for.] quitting in Changhua. The RR reductions per
The recently published meta-analysis (Gupta year of cessation were all statistically significant:
et al., 2022) combined data on chewing areca 6.7% (95% CI, 1.9–11.2%) in Keelung, 2.6% (95%
nut without tobacco from the four case–control CI, 0.8–4.4%) in Changhua, and 2.3% (95% CI,
studies described above. The risk estimate for oral 0.1–4.5%) in Tainan. With respect to age at quit-
cancer in former chewers (meta-RR, 5.61; 95% ting, the younger the age at quitting, the lower
CI, 2.24–14.04) was lower than that in current the risk of oral cancer, as shown by the signifi-
chewers (7.89; 95% CI, 3.90–15.98) compared with cant increasing trends per year of advancing age
never-chewers. [The analysis could not report on at quitting, 13% in Keelung and 3% in Changhua,
duration of cessation, because information on and a non-significant 1% in Tainan. Notably, the
duration of cessation is lacking for most of the results from the two cohort studies in the areas
published studies.] where the prevalence of areca nut chewing is high
– Tainan (in the southern part) and Changhua
(ii) Evidence from primary data analyses
(in the central part) – showed that quitting areca
See Table 2.30. nut chewing before age 40 years led to a signif-
Data on duration of cessation and age at icant reduction in the risk of oral cancer. [For
quitting from three prospective cohort studies each cohort, a time-dependent Cox regression
and one case–control study were available model was used to consider dynamic change of
for primary analysis by the Working Group. duration of quitting during follow-up. Relevant
The three cohort studies were derived from confounding factors have been well controlled to
three community-based integrated screening avoid recall bias.]
programmes for common cancer types (inclu- For the case–control study, analyses restricted
ding oral cancer) in three cities in Taiwan to ever-chewers resulted in a statistically signif-
(China): Keelung, Changhua, and Tainan, repre- icant relative reduction in risk per year of cessa-
senting the northern, central, and southern parts tion, estimated as 7% (95% CI, 5–9%).
of the country, where areca nut is consumed [The Working Group also performed a meta-
unripe and without tobacco. Information on analysis that combined the information on the
demographic characteristics, education level, three user categories from the observational
duration and frequency of smoking, alcohol studies presented in Table 2.29 and Table 2.30.
consumption, age at quitting, and duration of Former chewers had a statistically significantly
176
Table 2.30 Cessation of chewing of areca nut products (including betel quid) without tobacco and risk of oral cancer –
primary data analyses performed by the Working Group
Study Study population Study design, study period, number End- Exposure OR (95% CI) Comments
Location of participants, information on betel point category
quid chewing, and confounders Number of cases/
considered controls
Unpublished; Community- Prospective cohort study Oral Never-chewers: 1.00 (ref) This is a large-scale
community- based integrated People attending the KCIS programme cancer 245/110 555 community-based
based integrated screening study in 1999–2018. This cohort was Former chewers: 2.40 (1.68–3.42) screening programme with
screening, for residents aged followed up to ascertain incident oral 57/4757 long-term follow-up for the
Keelung, Taiwan 30 yr in Keelung, cancer by linking with the national Current chewers: 3.02 (2.16–4.22) outcome of incident oral
(China) northern Taiwan cancer registry in Taiwan (China) 64/4034 cancer and information
(China) (KCIS until 31 December 2018 on betel quid chewing
Per year of 0.933 (0.888–0.981)
programme) A time-dependent Cox regression in Keelung, where the
cessation of betel
121 714 people were model was used to consider dynamic prevalence of betel quid
quid chewing
enrolled, and 372 change of duration of cessation during chewing is lower than
oral cancers were follow-up Duration of in other parts of Taiwan
ascertained during Confounding factors adjusted for were cessation (yr): (China)
follow-up age, sex, education level, smoking Current chewers 1.00 (ref)
(never, < 10, 10–19.9, 20–29.9, and ≥ 30 < 10: 39/2410 1.78 (1.09–2.90)
pack-years), and alcohol consumption 10–20: 5/973 0.75 (0.43–1.31)
(never, ever, current) ≥ 20: 1/465 0.16 (0.04–0.67)
Never-chewers 0.32 (0.23–0.45)
Ptrend < 0.0001
Per year of age at 1.13 (1.05–1.22)
quitting
Age at quitting
(yr):
Current chewers 1.00 (ref)
< 40: 18/2364 0.72 (0.42–1.22)
40–49: 14/986 0.82 (0.43–1.75)

≥ 50: 13/497 1.48 (0.80–2.76)
177
178

considered controls
community- based integrated People enrolled in 2005–2014 were cancer 109/83 537 community-based
based integrated screening study used to assess the impact of cessation Former chewers: 3.86 (2.73–5.46) screening programme,
screening, for residents aged of betel quid chewing on risk of oral 119/5149 in an area with a high
Changhua, 30 yr in Changhua, cancer. This cohort was followed Current chewers: 4.77 (3.31–6.89) prevalence of betel quid
Taiwan (China) central Taiwan up to ascertain incident oral cancer 82/2921 chewing
(China) (CHCIS by linking with the national cancer
Per year of 0.974 (0.956–0.992)
programme) registry until 31 December 2018
cessation of betel
92 246 people were Exposures include current chewers,
quid chewing
enrolled in the former chewers, and never-chewers;
CHCIS cohort, and time in years since cessation measured Duration of
311 oral cancers in continuous years cessation (yr):
were ascertained Confounding factors adjusted for in Current chewers 1.00 (ref)
during follow-up the Cox regression model included < 10: 61/1992 1.09 (0.75–1.59)
age, sex, education level, smoking 10–20: 28/1617 0.65 (0.44–0.97)
(never, < 10, 10–19.9, 20–29.9, and ≥ 30 ≥ 20: 17/1119 0.59 (0.37–0.93)
pack-years), and alcohol consumption
(never, seldom, 1–2 per wk, 3–5 per
wk, and daily drinkers) Ptrend = 0.0142
A time-dependent Cox regression Per year of age at 1.03 (1.00–1.05)
model was used to consider dynamic quitting
change of duration of cessation during Age at quitting
follow-up (yr):
Current chewers 1.00 (ref)
< 40: 25/1793 0.64 (0.40–1.00)
40–49: 37/1590 0.87 (0.58–1.29)
≥ 50: 53/1673 0.84 (0.58–1.21)
Ptrend = 0.3255
considered controls
community- based integrated People enrolled in 2004–2009 were cancer 232/116 869 community-based
based integrated screening study for used to assess the impact of cessation Former chewers: 3.20 (2.40–4.29) screening programme
screening, residents aged 40 yr of betel quid chewing on risk of oral 85/4838 in an area with a higher
Tainan, Taiwan in southern Taiwan cancer. This cohort was followed Current chewers: 4.34 (3.27–5.77) prevalence of betel quid
(China) (China) (CIS up to ascertain incident oral cancer 99/3806 chewing
programme) by linking with the national cancer
Per year of 0.977 (0.955–0.999)
125 977 people were registry until 31 December 2018
cessation of betel
enrolled in the CIS Exposures include current chewers,
quid chewing
cohort, and 417 former chewers, and never-chewers;
oral cancers were time in years since cessation measured Duration of
ascertained during in continuous years cessation (yr):
follow-up Confounding factors adjusted for in Current chewers: 1.00 (ref)
the Cox regression model included < 10: 48/2263 0.88 (0.58–1.36)
age, sex, education level, smoking 10–20: 23/1316 0.83 (0.57–1.12)
(never, < 10, 10–19.9, 20–29.9, and ≥ 30
≥ 20: 6/797 0.40 (0.22–0.75)
pack-years), and alcohol consumption
(never, seldom, 1–2 per wk, 3–5 per Never-chewers 0.22 (0.17–0.30)
wk, and daily drinkers) Ptrend = 0.0068
Per year of age at 1.01 (0.99–1.04)
quitting
Age at quitting
(yr):
< 40: 17/1598 0.51 (0.30–0.86)
40–59: 55/2741 0.80 (0.58–1.12)
≥ 60: 12/474 0.91 (0.49–1.69)

Ptrend = 0.2362
179
180

considered controls
Unpublished; Study conducted Hospital-based case–control study, Oral Per year of 0.93 (0.91–0.95)
US NCI Taiwan in 4 hospitals in with recruitment of controls, patients cancer cessation of betel
(China) OPMD Taiwan (China): with oral cancer or OPMDs (primarily quid chewing
and oral cancer NTUH Taipei, leukoplakia and some OSF) Ptrend < 0.001
study CMUH Taichung, Controls were frequency-matched to Duration of
CGMH-Linkou, case group (OPMDs and cancer) on cessation (yr):
and CGMH- age (5-year groups), sex, study site,
Kaohsiung ever-smoking, and ever-chewing
241/158
Participants recruited in 2013–2021;
recruitment of controls and OPMD < 2: 43/12 2.08 (1.06–4.09)
cases is continuing 2–5: 63/32 1.09 (0.67–1.76)
This analysis (conducted in November 5–10: 86/64 0.67 (0.44–1.01)
2021) included 388 controls and 549 10–15: 45/35 0.49 (0.28–0.84)
cancer cases. Analyses were restricted ≥ 15: 72/87 0.21 (0.12–0.37)
to ever-chewers. Multinomial logistic
Ptrend < 0.001
regression models (cancer vs control)
were adjusted for age, sex, education
level (≤ vs > high school), smoking
duration and intensity, alcohol
consumption (drinks per week),
and chewing duration and intensity.
Primary analyses based on duration
of cessation (as a continuous, linear
variable and a categorical variable:
quit ≤ 2 yr, 2–5 yr, 5–10 yr, 10–15 yr,
and ≥ 15 yr)
CGMH, Chang Gung Memorial Hospital; CHCIS, Changhua Community-Based Integrated Screening; CI, confidence interval; CIS, Community-Based Integrated Screening; CMUH,
China Medical University Hospital; KCIS, Keelung Community-Based Integrated Screening; NTUH, National Taiwan University Hospital; OPMDs, oral potentially malignant
disorders; OSF, oral submucous fibrosis; ref, reference; US NCI, United States National Cancer Institute; vs, versus; wk, week; yr, years or years.
lower risk of oral cancer (OR, 0.79; 95% CI, chewing, smoking, and alcohol consumption was
0.68–0.94) compared with current chewers.] collected via a structured questionnaire through
in-person interview. A total of 219 cases (leuko-
(b) Studies on OPMDs plakia or OSF) and 876 controls were included. The
(i) Evidence from the published literature OR for leukoplakia in former chewers compared
See Table 2.31. with never-chewers (7.1; 95% CI, 2.3–21.5) was
Shiu et al. (2000) established a leukoplakia significantly lower than that in current chewers
cohort, which consisted of 435 patients diagnosed (22.3; 95% CI, 11.3–43.8). Similar findings were
at one medical centre in Taiwan (China) in 1988– reported for OSF. [The Working Group calcu-
1998. To assess the role of betel quid chewing, lated the ORs for former chewers compared with
tobacco smoking, and alcohol consumption on current chewers as 0.32 (95% CI, 0.09–1.10) for
the risk of leukoplakia, the case group consisted of leukoplakia and 0.30 (95% CI, 0.06–1.58) for OSF.
100 patients with leukoplakia randomly selected The fact that oral examination was not performed
from the cohort, and the control group consisted in the control group may have introduced bias. It
of 100 patients with periodontal disease diagnosed is not clear whether the estimates were adjusted
in the same period and at the same medical centre, for tobacco smoking and alcohol consumption.]
matched on age, sex, and date of diagnosis. After A case–control study in Papua New Guinea
adjustment for smoking and alcohol consump- (Thomas et al., 2008) reported an OR for former
tion, with never-chewers as the reference group, chewers that was lower than that for occasional
the OR for leukoplakia in former chewers (2.38; chewers and daily chewers compared with
95% CI, 0.34–16.75) was much lower than that in never-chewers. [The Working Group noted that
current chewers (17.43; 95% CI, 1.94–156.27). [The a limitation of this study was the extremely
Working Group noted the extremely wide CIs. high prevalence of ever betel quid chewing; the
The Working Group estimated the OR for former proportion of never-chewers was only 0.5% (1 of
chewers as 0.14 (95% CI, 0.007–2.73) compared 197) in the case group and 6.9% (89 of 1282) in
with current chewers. This study enrolled the the control group.]
control group from the same medical centre in A cross-sectional community screening study
the same period as the case group to ensure that for oral cancer conducted in four Indigenous
both groups were from the same catchment area. communities in Taiwan (China) in people aged
Information was collected via telephone survey ≥ 35 years in 2005 reported on the association
for both groups, instead of using medical chart between betel quid chewing and leukoplakia
review; this can avoid differential misclassifica- and OSF (Yang et al., 2010). The ORs for former
tion bias because in the medical charts, infor- chewers were lower than those for current
mation on betel quid chewing, tobacco smoking, chewers for leukoplakia in women (OR, 7.8; 95%
and alcohol consumption was more likely to be CI, 3.8–16.0 vs 15.6; 95% CI, 8.3–29.4), for OSF in
queried at diagnosis of leukoplakia than at diag- men (OR, 13.5; 95% CI, 3.8–48.7 vs 22.9; 95% CI,
nosis of periodontal disease. However, the use of 7.3–71.7), and for OSF in women (OR, 9.3; 95% CI,
a control group derived from patients diagnosed 3.3–26.0 vs 13.0; 95% CI, 5.2–32.6). In contrast, for
with periodontal disease may be a concern.] leukoplakia in men, ORs for former chewers were
Lee et al. (2003) reported on a hospital-based similar to those for current chewers (OR, 6.7; 95%
case–control study on OPMDs, including CI, 3.2–13.9 vs 6.6; 95% CI, 3.5–12.3). [The ORs
leukoplakia and OSF, conducted in 1994–1995 calculated for former chewers compared with
in Taiwan (China). Information on betel quid current chewers were 0.50 (95% CI, 0.20–1.22) for
leukoplakia in women, 0.59 (95% CI, 0.12–2.96)
181
182

Table 2.31 Cessation of chewing of areca nut products (including betel quid) without tobacco and risk of OPMDs –
Reference Study population Study design, study period, number OPMDs Exposure category OR (95% CI) Comments
Location of participants, information on betel end-point Number of cases/
quid chewing, and confounders controls
considered
Shiu et al. Patients with Case–control study Leukoplakia Leukoplakia: All cases and controls
(2000) leukoplakia in a Case group: 100 cases randomly Never-chewers 1.0 (ref) were interviewed via
Taiwan medical centre in selected from a cohort of 435 patients Current chewers 17.43 (1.94–156.27) telephone survey, to
(China) Taipei and their with leukoplakia diagnosed in avoid information bias
Former chewers 2.38 (0.34–16.75)
matched controls 1988–1998 between the 2 groups
from patients Control group: 100 controls with
with periodontal periodontal disease diagnosed in
disease the same period and medical centre,
matched to cases on age at diagnosis
(± 3 yr), sex, and date of diagnosis
Confounding factors in the
multivariate analysis included
cigarette smoking and alcohol
consumption
Lee et al. Patients at a Matched case–control study Leukoplakia Leukoplakia: People in the control
(2003) medical centre in conducted in 1994–1995 and OSF Never-chewers: 28/390 1.0 (ref) group did not receive
Taiwan Kaohsiung and Case group: 219 patients with Current chewers: 22.3 (11.3–43.8) an oral inspection.
(China) their sex- and age- leukoplakia (n = 125) or OSF (n = 94) 91/88 This might result in a
matched controls newly diagnosed and histologically biased estimate
Former chewers: 6/22 7.1 (2.3–21.5)
from residents confirmed Data on duration of
in the Greater Control group: 876 sex- and age- OSF: cessation for former
Kaohsiung area matched controls from 1864 Never-chewers: 11/302 1.0 (ref) chewers were not
household units Current chewers: 40.7 (16.0–103.7) available
Confounding factors in the 78/62 Not clear whether
multivariate analysis included Former chewers: 5/12 12.1 (2.8–51.9) adjusted for tobacco
education level and occupation smoking and alcohol
consumption
considered
Thomas et al. People aged A case–control study nested in a Leukoplakia Never-chewers: 1/89 1.0 (ref) Extremely high
(2008) ≥ 18 yr from 2 cross-sectional study in 1992 Former chewers: 7/149 1.4 (0.2–13.0) prevalence of ever
Papua New census divisions Case group: 197 patients with Occasional chewers: 6.1 (0.8–48.7) betel quid chewing.
Guinea (East Coast Kara identified leukoplakia 26/256 The proportion of
Nalik and South Control group: 1282 controls never-chewers was
Daily chewers: 5.0 (0.6–39.1)
Lavongai) of New ascertained in the cross-sectional 0.5% (1 of 197) in the
163/788
Ireland Province study with no evidence of oral case group and 6.9%
squamous cell carcinoma, (89 of 1282) in the
leukoplakia, leukoedema, control group
erythroplakia, or commissural
ulceration
multivariate analysis included age,
sex, census division, and smoking
Cross-sectional studies
Yang et al. Community oral Cross-sectional study in 2005 Leukoplakia Leukoplakia: 224 Information on
(2010) cancer screening Participants aged ≥ 35 yr, including and OSF Men: duration of cessation
Taiwan programme in 494 Indigenous men and 892 Non-chewers 1.0 (ref) was lacking
(China) 4 Indigenous Indigenous women
Current chewers 6.57 (3.51–12.28)
communities and The proportion of ever-chewers
1 remote island in was 11.0%, and the proportion of Former chewers 6.70 (3.21–13.99)
Pingtung County current chewers was 24.4%. The Women:
corresponding proportions were Non-chewers 1.0 (ref)
13.4% and 29.4% for men and 14.6% Current chewers 15.63 (8.31–29.39)
and 35.2% for women Former chewers 7.78 (3.77–16.04)
OSF: 89
multivariate analysis included sex,
age, tobacco smoking, and alcohol Men:

consumption Non-chewers 1.0 (ref)
Former chewers 13.53 (3.76–48.65)
Women:
Non-chewers 1.0 (ref)
Former chewers 9.32 (3.34–26.00)
183
184

considered
Yen et al. Community- Cross-sectional study in 2003–2008 OPMDs OPMD cases Estimates provided in
(2011) based integrated 79 940 participants aged ≥ 20 yr; 502 (% lesion): the publication were
Taiwan screening OPMDs Non-chewers: 256 1.0 (ref) crude ORs
(China) programme in Confounding factors in the (3.4%)
Keelung City multivariate analysis included Current chewers: 180 25.25 (20.77–30.69)
metabolic syndrome, age, sex, (80%)
education level, tobacco smoking, and
Former chewers: 64 7.43 (5.64–9.80)
alcohol consumption
(25%)
Adjusted
Non-chewers 1.0 (ref)
Current chewers [9.2 (7.2–11.8)]
Former chewers [2.8 (2.0–3.8)]
Former vs current [0.30 (0.22–0.43)]
CHCIS, Changhua Community-Based Integrated Screening; CI, confidence interval; OPMDs, oral potentially malignant disorders; OR, odds ratio; OSF, oral submucous fibrosis; ref,
reference; vs, versus; yr, year or years.
for OSF in men, and 0.72 (95% CI, 0.20–2.58) for Tainan) and the case–control study in Taiwan
OSF in women. This cross-sectional study did not (China) are described above for oral cancer (see
provide information on duration of cessation, Section 2.3.5(a)(ii)).
and there is a possibility of reverse causation, The three cohort studies reported statistically
which may explain the results obtained in men.] significant trends of reduced RR of OPMDs with
The cross-sectional study of Yen et al. increasing time since quitting (Ptrend < 0.001). The
(2011) reported data on the risk of OPMDs in most significant reduction was noted for ≥ 5 years
the Keelung Community-Based Integrated of abstinence in Keelung and Changhua and for
Screening (KCIS) programme in Taiwan (China) ≥ 2 years of abstinence in Tainan. All the risk
in 2003–2008 in former and current chewers of reductions per year of cessation were statistically
betel quid aged ≥ 20 years. [The Working Group significant: 3.5% (95% CI, 2.3–4.6%) in Keelung,
recalculated adjusted ORs: the estimate for former 3.2% (95% CI, 2.2–4.2%) in Changhua, and 0.8%
chewers versus never-chewers (2.8; 95% CI, (95% CI, 0.5–1.1%) in Tainan. With respect to age
2.0–3.8) was lower than that for current chewers at quitting, the younger the age at quitting the
versus never-chewers (9.2; 95% CI, 7.2–11.8), lower the risk of OPMDs, with significant RR
giving an OR for former chewers versus current reductions per year of younger age at quitting
chewers of 0.30 (95% CI, 0.22–0.43). When former of 2% in Keelung, 1.4% in Changhua, and 2% in
chewers were stratified by duration of quitting, Tainan. When comparing former versus current
an inverse dose–response relationship was noted chewers, cessation of chewing areca nut products
between time since quitting and the risk of without tobacco led to a significant reduction in
OPMDs, with ORs of 0.39 (95% CI, 0.27–0.56) for the risk of OPMDs in all three cohorts.
< 10 years of quitting, 0.22 (95% CI, 0.10–0.44) In the case–control study in southern Taiwan
for 10–19 years of quitting, and 0.19 (95% CI, (China), analyses restricted to ever-chewers
0.06–0.60) for ≥ 20 years of quitting. This large- resulted in a statistically significant 5% reduction
scale community-based screening programme in RR per year of cessation (OR, 0.95; 95% CI,
provided stable estimates. This was an integrated 0.93–0.98).
screening programme that targeted multiple The Working Group performed a meta-anal-
neoplasms and chronic diseases, for which infor- ysis combining information on the three user
mation on general health was queried, instead of categories (current chewers, former chewers, and
focusing on oral health; therefore, participants never-chewers) from the observational studies
were less likely to avoid answering questions presented in Table 2.31 and Table 2.32. Former
about smoking and betel quid chewing. In addi- chewers had a statistically significantly lower
tion, all disease status data were measured or risk of OPMDs (OR, 0.55; 95% CI, 0.39–0.72)
collected upon screening activity. Information compared with current chewers.]
bias on both independent covariates and disease
outcomes could be ruled out.] 2.3.6 HPV16 infection
(ii) Evidence from primary data analyses Three types of HPV vaccines are currently
See Table 2.32. available: a bivalent vaccine, a quadrivalent
Data on duration of cessation and age at vaccine, and a nonavalent vaccine (Schiller
quitting from three prospective cohort studies and Lowy, 2012; Arbyn and Xu, 2018). All
and one case–control study were available for three target HPV16, the type that causes most
primary analysis by the Working Group. The HPV-associated oral and orophar yngeal
same three cohorts (in Keelung, Changhua, and cancers. HPV vaccines are prophylactic (i.e.
185
186

Table 2.32 Cessation of chewing of areca nut products (including betel quid) without tobacco and risk of OPMDs – primary
data analyses performed by the Working Group
Study Study population Study design, study period, End-point Exposure category OR (95% CI) Comments
Location number of participants, Number of cases/
information on betel quid controls
chewing, and confounders
considered
Unpublished; Community- Prospective cohort study OPMD Never-chewers: 1.00 (ref) This is a large-
community- based integrated People attending the KCIS (leukoplakia, 2124/111 486 scale community-
based screening study programme in 1999–2020 were erythroleukoplakia, Former chewers: 2.22 (2.00–2.46) based screening
integrated for residents aged used to assess the impact of erythroplakia, OSF, 611/4273 programme, in
screening, 30 yr in Keelung, quitting betel quid chewing on oral verrucous Current chewers: 3.43 (3.11–3.78) an area where the
Keelung, northern Taiwan risk of OPMDs. This cohort hyperplasia) 844/3229 prevalence of betel
Taiwan (China) (China) (KCIS) was followed up to ascertain quid chewing is lower
Per year of cessation 0.965
124 353 people incident OPMDs by linking with than in other parts of
of betel quid chewing (0.954–0.977)
were enrolled in the the national cancer screening Taiwan (China)
CHCIS cohort, and registry until 31 December 2020 Current chewers 1.00 (ref) Because of the
3630 OPMDs were Patients with oral cancer and Duration of cessation repeated attendance
ascertained during people with a diagnosis of (yr): to screening, both
follow-up OPMD before the prevalent < 2: 116/503 0.83 (0.57–1.19) prevalent and
screen in KCIS were excluded 2–5: 132/797 0.83 (0.65–1.07) incident OPMDs
Exposures included current 5–10: 97/800 0.75 (0.61–0.91) were included in the
chewers, former chewers, and analysis
10–15: 70/656 0.66 (0.55–0.81)
never-chewers; age at cessation
and time in years since cessation ≥ 15: 78/718 0.50 (0.42–0.60)
measured in continuous years Never-chewers 0.29 (0.26–0.32)
Further details on this study are Ptrend < 0.0001
given in Table 2.30. Confounding Per year of age at 1.02 (1.01–1.04)
factors adjusted for in the Cox quitting
regression model included age, Current chewers 1.00 (ref)
sex, education level, smoking
Age at quitting (yr):
(never, < 10, 10–19.9, 20–29.9,
and ≥ 30 pack-years), and alcohol < 40: 275/2174 0.58 (0.50–0.67)
consumption (never, ever, 40–49: 149/850 0.77 (0.64–0.92)
current) ≥ 50: 69/444 0.77 (0.60–0.99)
A time-dependent Cox Never-chewers 0.30 (0.27–0.33)
regression model was used to Ptrend = 0.0073
consider dynamic change of
duration of cessation during
follow-up
considered
Unpublished; Community- Prospective cohort study OPMD Never-chewers: 1.00 (ref) Large-scale
community- based integrated People enrolled in 2005–2014 (leukoplakia, 646/28 997 community-based
based screening study were used to assess the impact erythroleukoplakia, Former chewers: 1.55 (1.35–1.78) screening programme
integrated for residents aged of cessation of betel quid erythroplakia, OSF, 440/4429 in an area where
screening, 30 yr in Changhua, chewing on risk of oral cancer oral verrucous Current chewers: 2.57 (2.24–2.95) the prevalence of
Changhua, central Taiwan and OPMDs. This cohort was hyperplasia) 460/2315 betel quid chewing
Taiwan (China) (China) (CHCIS followed up to ascertain incident is higher than in
programme) oral cancer by linking with other parts of Taiwan
37 327 people were the national cancer until 31 (China)
enrolled in the December 2018 Current chewers 1.00 (ref)
CHCIS cohort, and Further details on this study are Duration of cessation
1548 OPMDs were given in Table 2.30 (yr):
ascertained during < 2: 55/314 0.84 (0.43–1.64)
follow-up 2–5: 79/613 0.90 (0.65–1.25)
5–10: 85/739 0.64 (0.50–0.80)
10–15: 82/960 0.65 (0.52–0.80)
≥ 15: 111/1434 0.53 (0.44–0.64)
Ptrend < 0.0001
Per year of age at 1.014
quitting (1.002–1.026)
< 40: 134/1564 0.55 (0.45–0.67)
40–49: 154/1348 0.67 (0.55–0.80)

≥ 50: 144/1439 0.61 (0.50–0.74)
Ptrend = 0.4313
187
188

considered
Unpublished; Community- Prospective cohort study OPMD Never-chewers: 1.00 (ref) Large-scale
community- based integrated People attending the CIS (leukoplakia, 745/95 516 community-based
based screening study for programme in 2004–2009 erythroleukoplakia, Former chewers: 1.94 (1.69–2.23) screening programme
integrated residents aged 40 yr were used to assess the impact erythroplakia, OSF, 363/4761 in an area where the
screening, in Tainan, southern of cessation of betel quid oral verrucous Current chewers: 2.95 (2.39–3.37) prevalence of betel
Tainan, Taiwan Taiwan (China) chewing on risk of oral cancer hyperplasia) 471/3767 quid chewing is the
(China) (CIS programme) and OPMDs. This cohort was highest in Taiwan
125 977 people followed up for incident oral (China)
were enrolled in the cancer by the using national
Tainan cohort, and cancer registry until 31 Current chewers 1.00 (ref)
1584 OPMDs were December 2018 Duration of cessation
ascertained during Patients with oral cancer were (yr):
follow-up excluded < 2: 63/464 1.06 (0.81–1.37)
Exposures included current 2–5: 73/801 0.76 (0.59–0.97)
chewers, former chewers, and 5–10: 66/912 0.62 (0.48–0.79)
never-chewer; time in years
10–15: 57/941 0.57 (0.43–0.75)
since cessation measured in
continuous years ≥ 15: 60/1117 0.61 (0.47–0.80)
Adjustments for confounding Never-chewers 0.34 (0.30–0.38)
factors in this study are given in Ptrend < 0.0001
Table 2.29 Per year of age at 1.02 (1.00–1.03)
quitting
< 40: 105/1566 0.54 (0.44–0.67)
40–49: 156/1693 0.76 (0.64–0.92)
≥ 50: 83/1392 0.75 (0.59–0.95)
Ptrend = 0.0002
considered
Unpublished; Study conducted Hospital-based case–control OPMD Per year of cessation 0.95 (0.93–0.97)
US NCI Taiwan in 4 hospitals in study, with recruitment of of betel quid chewing
(China) OPMD Taiwan (China): controls, patients with oral Ptrend < 0.001
and oral cancer NTUH Taipei, cancer or OPMDs (primarily Current chewers: 1.00 (ref)
study CMUH Taichung, leukoplakia and some OSF) 743/158
CGMH-Linkou, This analysis (conducted in
and CGMH- November 2021) included 388
(yr):
Kaohsiung controls and 1468 OPMDs
Further details on this study < 2: 106/12 1.78 (0.95–3.34)
are given in Table 2.29 on oral 2–5: 142/32 0.87 (0.57–1.34)
cancer 5–10: 167/64 0.48 (0.34–0.69)
10–15: 11/35 0.49 (0.31–0.79)
≥ 15: 199/87 0.30 (0.19–0.47)
Ptrend < 0.001
CGMH, Chang Gung Memorial Hospital; CHCIS, Changhua Community-Based Integrated Screening; CI, confidence interval; KCIS, Keelung Community-Based Integrated Screening;
NTUH, National Taiwan University Hospital; OPMDs, oral potentially malignant disorders; OR, odds ratio; OSF, oral submucous fibrosis; ref, reference; US NCI, United States National
Cancer Institute; yr, year or years.

189
vaccination prevents future acquisition of infec- the prevalence of oral or oropharyngeal infec-
tion) and not therapeutic (i.e. vaccination does tion with vaccine-targeted HPV types (including
not enable clearance of prevalent infection) HPV16) is 83–93% lower in vaccinated individ-
(Schiller and Lowy, 2012; Arbyn and Xu, 2018). uals than in unvaccinated individuals (Herrero
The key effector mechanism of vaccine efficacy et al., 2013; Chaturvedi et al., 2018; Schlecht
is through the generation of systemic immuno- et al., 2019). Second, emerging evidence indi-
globulin G (IgG) antibody responses against the cates herd protection from HPV vaccination
HPV L1 protein (Schiller and Lowy, 2012; Arbyn in the population with reduced prevalence of
and Xu, 2018). oral or oropharyngeal HPV infection in unvac-
The HPV vaccines have been shown to be cinated individuals (Chaturvedi et al., 2019;
safe, highly efficacious, and highly effective Mehanna et al., 2019). Third, there is a strong
in preventing infection with vaccine-targeted analogy from other anatomical sites with respect
HPV types (at the cervix, vagina, vulva, anus, to vaccine efficacy and effectiveness; analogous
penis, and oral cavity), anogenital warts, and decreases in HPV infections, HPV-associated
HPV-associated precancer end-points (at the precancers, and cancers at other anatomical sites
cervix, vagina, vulva, anus, and penis), and to (cervix, vagina, vulva, anus, and penis) have been
result in population-level reductions in the inci- consistently reported in vaccinated individuals
dence of cervical cancer (Drolet et al., 2019; Lei and populations.
et al., 2020; Kjaer et al., 2021). Future reductions in the incidence of
There is currently no empirical evidence HPV-associated oral cancer and oropharyngeal
that prophylactic HPV vaccination results in a cancer will depend on the extent of female and
reduction in the incidence of oral or oropharyn- male vaccination coverage in men and women,
geal cancer or in the incidence of OPMDs. This as well as achieved levels of herd immunity in
lack of evidence arises from the recency of a country or region. In regions with high levels
the introduction of HPV vaccines (in 2006 for of female and/or gender-neutral vaccination
women and 2011 for men in most countries) as coverage, it would be expected that over the
well as the current recommendations to vacci- next 10–15 years HPV vaccination will result in
nate young people (the routine recommendation population-level reductions in the incidence of
is for vaccination before sexual debut until age HPV-associated oral cancer and oropharyngeal
12–14 years and for catch-up vaccination until cancer.
age mid-20s in some countries) (WHO, 2019).
Because the latency interval between the acqui-
sition of oral or oropharyngeal HPV16 infection
and the development of HPV-associated oral or
oropharyngeal cancer spans several decades,
many more years of observation would be needed
for prophylactic HPV vaccination of both sexes
to result in a reduction in incidence of oral cancer
or oropharyngeal cancer (Gillison et al., 2015).
However, there is a compelling scien-
tific rationale that HPV vaccination would
reduce the incidence of HPV-associated oral
or oropharyngeal cancer in the future. First,
several observational studies have shown that
190
2.4 Preventive dietary agents (b) Tea

The evidence for the association between tea
This section presents the available evidence
drinking and cancers of the mouth, pharynx,
on dietary agents that may have a protective effect
and larynx was considered limited by the WCRF
on the development of oral cancer and OPMDs.
reports, and no conclusion could be reached as
to a protective or harmful effect (WCRF, 2016;
2.4.1 Preventive dietary agents for the WCRF/AICR, 2018).
development of oral cancer Current evidence comes from a pooled
(a) Coffee analysis of cases and controls from 9 studies
in the INHANCE consortium (Galeone et al.,
The 2018 WCRF report (WCRF/AICR,
2010), a meta-analysis of 14 case–control studies
2018) concluded that there is “limited sugges-
(Zhou et al., 2018), a meta-analysis of one cohort
tive evidence” that consumption of coffee may
study and four case–control studies (Filippini
decrease the risk of oral cancer.
et al., 2020), and one individual cohort study
Studies on the association between coffee
(Ren et al., 2010) (Supplementary Table S2.33,
drinking and the incidence of oral cancer has
been reviewed in two meta-analyses (Miranda
iarc.fr/617). These studies reported risk estimates
et al., 2017; He et al., 2020) and one pooled
for oral cancer associated with self-reported tea
analysis (Galeone et al., 2010) (Supplementary
consumption taking into account major risk
factors, including tobacco smoking and alcohol
publications.iarc.fr/617). Miranda et al. (2017)
consumption.
calculated a meta-OR for the association between
The pooled analysis, which included study
oral cancer and coffee drinking of 0.82 (95% CI,
participants from France, Italy, Puerto Rico,
0.58–1.16) using data from one cohort study (Ren
Switzerland, and the USA, generated a non-sta-
et al., 2010) and five case–control studies (Franco
tistically significant adjusted pooled estimate of
et al., 1989; Franceschi et al., 1992; Pintos et al.,
risk of oral cancer associated with tea drinking
1994; Bundgaard et al., 1995; Radoï et al., 2013b).
of 1.06 (95% CI, 0.88–1.27); the estimate was
He et al. (2020) included all the studies that were
slightly reduced when based on people drinking
part of the meta-analysis by Miranda et al. (2017),
> 1 cup of tea per day (OR, 0.94; 95% CI, 0.68–1.29)
alongside with one additional case–control study
(Galeone et al., 2010). In a meta-analysis of
and one cohort study. They calculated a meta-OR
studies conducted in Brazil, China, Denmark,
for oral cancer (OR, 0.79; 95% CI, 0.40–1.58) for
Egypt, France, India, and Italy that reported
coffee drinkers using data from four case–control
adjusted risk estimates for oral cancer, Zhou et al.
studies (Franco et al., 1989; Franceschi et al.,
(2018) generated an overall meta-estimate of risk
1992; Bundgaard et al., 1995; Radoï et al., 2013b).
of oral cancer associated with tea consumption
Galeone et al. (2010) provided a pooled analysis
(OR, 0.70; 95% CI, 0.61–0.81). In a dose–response
of nine case–control studies of the INHANCE
analysis including 8 of the 14 case–control
cohort. They found a significant 54% reduction
studies, the risk of oral cancer decreased by 6.2%
in RR for drinking > 4 cups per day versus none
per 1 cup increase per day (OR, 0.938; 95% CI,
(OR, 0.46; 95% CI, 0.30–0.71).
0.922–0.955). [This study presented additional
pooled risk estimates according to type of tea,
geographical region, sex, and age group.]
191
In their more recent systematic review of showed significant reduction in the overall risk
green tea drinking and cancer, Filippini et al. of oral cancer (Pavia et al., 2006).
(2020) reported a significant inverse association, The two cohort studies examined total
with a meta-estimate of risk of oral cancer asso- consumption of fruits and vegetables. The cohort
ciated with consumption of green tea comparing study in the USA, conducted in the late 1990s
the highest versus the lowest intake (meta-RR, (Freedman et al., 2008), reported reduced risk of
0.71; 95% CI, 0.62–0.82). oral cancer with increasing total consumption of
One cohort study in the USA (Ren et al., 2010) fruits and vegetables (HR per serving per 1000
reported non-statistically significant inverse calories, 0.93; 95% CI, 0.86–1.00). The cohort
associations, after adjustment for important study in the Netherlands (Maasland et al., 2015),
confounders, in the category of the largest in which participants were enrolled in 1986 and
number of cups of tea consumed (HR for ≥ 1 cup followed up for 20 years, reported a reduction in
of hot tea per day, 0.75; 95% CI, 0.53–1.06; HR RR with increasing frequency of total consump-
for ≥ 1 cup of iced tea per day, 0.89; 95% CI, tion of fruits and vegetables (RR per 2.5 g per day,
0.67–1.19; and HR for ≥ 5 cups of green tea per 0.95; 95% CI, 0.92–0.99; Ptrend = 0.005).
day, 0.44; 95% CI, 0.19–1.04). A significant reduction in RR associated
with increasing consumption of specific fruits or
(c) Fruits and vegetables vegetables was observed for raw green vegetables,
The preventive role of consumption of fruits citrus fruits, apples and pears, fresh tomatoes,
and vegetables on risk of oral cancer has been and carotene-rich foods in one or several of three
investigated in a large pooled analysis of 22 case–control studies conducted in Brazil (Franco
case–control studies (Chuang et al., 2012), a et al., 1989; Galvão De Podestá et al., 2019) and
meta-analysis of 15 case–control studies and India (Rajkumar et al., 2003). For non-starchy
one cohort study (Pavia et al., 2006), two cohort vegetables, the reduction in RR was modest
studies (Freedman et al., 2008; Maasland et al., (RR per 25 g per day, 0.95; 95% CI, 0.89–1.02 to
2015), and three additional case–control studies RR per serving per 1000 calories, 0.84; 95% CI,
(Supplementary Table S2.33, web only; available 0.73–0.95) (WCRF, 2018).
from https://publications.iarc.fr/617).
The 2018 WCRF systematic review (WCRF/ (d) Dietary fibre
AICR, 2018) reported a limited–suggestive Evidence on the association between consump-
decrease in risk of oral cancer associated with tion of dietary fibre and oral cancer is available
“healthy dietary patterns” and with “greater from one large pooled analysis of case–control
intake of non-starchy vegetables”. studies and two individual cohort studies
In the pooled analysis, in which intake of (Supplementary Table S2.33, web only; available
fruits and of vegetables were standardized into from https://publications.iarc.fr/617).
frequency quartiles, the highest relative to the The pooled analysis of 10 case–control studies
lowest consumption level conferred reduced in the INHANCE consortium (Kawakita et al.,
risks of oral cancer for fruits (OR, 0.46; 95% CI, 2017), with 559 cases and 12 248 controls enrolled
0.38–0.56) and for vegetables (OR, 0.69; 95% CI, in Asia, Europe, and North America, reported
0.61–0.79) (Chuang et al., 2012). Similarly, the reduced RR with consumption of dietary fibre;
meta-analysis found that each portion consumed the pooled OR for the highest versus the lowest
per day of fruit (OR, 0.49; 95% CI, 0.39, 0.63) quintile category was 0.39 (95% CI, 0.29–0.52)
and of vegetables (OR, 0.43; 95% CI, 0.31, 0.59) for oral cancer and 0.54 (95% CI, 0.45–0.64) for
oropharyngeal and hypopharyngeal cancers
192
combined. A cohort study from the NIH-AARP tobacco users, and most of the cases and controls
Diet and Health Study with 494 991 participants had lower socioeconomic status. In Gujarat and
found a borderline association between dietary Kerala, most of the cases were clinically diag-
fibre intake and risk of oral cancer in women nosed with leukoplakia or OSF, and in Andhra
(Ptrend = 0.055) but not in men (Ptrend = 0.576) (Lam Pradesh most were diagnosed with palatal lesions
et al., 2011). A more recent cohort study from the due to reverse smoking. The study in Andhra
Prostate, Lung, Colorectal, and Ovarian (PLCO) Pradesh reported an OR for fibre intake (grams
cancer screening trial in the USA, with 101 700 per day) of 0.96 (95% CI, 0.94–0.99; P = 0.007)
participants enrolled in 1992–2001, reported (Hebert et al., 2002).
a significant risk reduction for oral cavity and A case–control study in Sri Lanka (Amara-
pharyngeal cancer with a dose–response rela- singhe et al., 2013), with cases of leukoplakia
tionship for total fibre intake, insoluble fibre mainly, found a protective effect of consumption
intake, and soluble fibre intake (Kawakita et al., of > 2 portions per day of β-carotene-containing
2019). vegetables and fruits on development of OPMDs
(Supplementary Table S2.34, web only; avail-
(e) Mediterranean diet able from https://publications.iarc.fr/617). [The
People who adhere to the Mediterranean authors pointed to prevailing undernutrition in
diet, which is based on consumption of olive oil OPMD cases in this rural population with very
in addition to frequent intake of fish and seafood, low daily consumption of fruits and vegetables
vegetables, fruits, and cereals, have been shown (< 2 portions per day).]
to have a strong inverse association between In a hospital-based case–control study in
adherence to such a diet and risk of oral cancer Rome, Italy (Cianfriglia et al., 1998), partici-
(Trichopoulou and Lagiou, 1997; Petridou et al., pants were interviewed about dietary habits, and
2002; Filomeno et al., 2014). the survey included questions on foods that are
major sources of vitamin A and carotenoids.
2.4.2 Preventive dietary agents for the The consumption of foods rich in vitamin A –
development or progression of OPMDs butter, eggs, liver, spinach, and carrots – in the
control group was > 40% higher than that in
(a) Observational studies the cases (P < 0.001). Specifically, the estimated
In the mid-1990s, the Tata Institute of Funda- mean retinol intake in the control group was
mental Research (in Bombay, India) conducted significantly higher than that in the leukoplakia
several population-based case–control studies group (Supplementary Table S2.34, web only;
in three regions of India – Gujarat, Kerala, and available from https://publications.iarc.fr/617).
Andhra Pradesh – to examine the role of food Consumption of foods and nutrients rich in
and nutrition on the progression of OPMDs vitamins A, C, E, and B12, β-carotene, lycopene,
(Gupta et al., 1998, 1999; Hebert et al., 2002; folate, retinol, α-tocopherol, and antioxidant
Supplementary Table S2.34, web only; available mineral zinc have been found to be protective
from https://publications.iarc.fr/617). A food against the development of OPMDs.
frequency questionnaire was used that was
(b) Biochemical studies
specific to this population and was developed
and validated for collecting dietary information Several biochemical investigations have
needed to estimate exposure to 92 food items; studied the role of nutrients in blood (serum or
the data included the frequency and quantity plasma) in the development of OPMDs. All but
of consumption. All people interviewed were one (cross-sectional) studies were of case–control
193
design; five were in India, two in Japan, one in References

Finland, and one in the Islamic Republic of Iran
(Supplementary Table S2.35, web only; available Acharya S, Ekalaksananan T, Vatanasapt P, Loyha K,
from https://publications.iarc.fr/617). Phusingha P, Promthet S, et al. (2015). Association of
In the studies in India, serum levels of vita- Epstein-Barr virus infection with oral squamous cell
carcinoma in a case-control study. J Oral Pathol Med.
mins A, C, E, and B12, β-carotene, folate, retinol, 44(4):252–7. doi:10.1111/jop.12231 PMID:25169715
α-tocopherol, and antioxidant mineral zinc were Agudo A, Bonet C, Travier N, González CA, Vineis P,
lower in leukoplakia or OSF cases than in controls Bueno-de-Mesquita HB, et al. (2012). Impact of cigarette
smoking on cancer risk in the European Prospective
(Ramaswamy et al., 1996; Gupta et al., 2004; Bose Investigation into Cancer and Nutrition study. J Clin
et al., 2012; Basu and Guhan, 2015; Param et al., Oncol. 30(36):4550–7. doi:10.1200/JCO.2011.41.0183
2018). In men in Japan, serum levels of lycopene PMID:23169508
and β-carotene were significantly lower in leuko- Al-Maweri SA, Warnakulasuriya S, Samran A (2018). Khat
(Catha edulis) and its oral health effects: an updated
plakia cases than in healthy controls (Nagao et al., review. J Investig Clin Dent. 9(1):e12288. doi:10.1111/
2000). In the study in Finland, the prevalence jicd.12288 PMID:28834423
of leukoplakia cases was significantly higher in Al-Tayar BA, Tin-Oo MM, Sinor MZ, Alakhali MS
a group with low plasma levels of ascorbic acid (2015). Association between shammah use and oral
leukoplakia-like lesions among adult males in Dawan
(≤ 25 µmol/L) (Tuovinen et al., 1992). Valley, Yemen. Asian Pac J Cancer Prev. 16(18):8365–70.
Two case–control studies reported on serum doi:10.7314/APJCP.2015.16.18.8365 PMID:26745086
retinol and carotenoid levels in OLP cases (Nagao Amarasinghe HK, Johnson NW, Lalloo R, Kumaraarachchi
M, Warnakulasuriya S (2010b). Derivation and valida-
et al., 2001; Rezazadeh and Haghighat, 2021; tion of a risk-factor model for detection of oral poten-
Supplementary Table S2.35, web only; available tially malignant disorders in populations with high
from https://publications.iarc.fr/617). In the prevalence. Br J Cancer. 103(3):303–9. doi:10.1038/
study in the Islamic Republic of Iran (Rezazadeh sj.bjc.6605778 PMID:20628386
Amarasinghe HK, Usgodaarachchi U, Kumaraarachchi
and Haghighat, 2021), neither parameter was M, Johnson NW, Warnakulasuriya S (2013). Diet and
found to be a risk factor for the development risk of oral potentially malignant disorders in rural Sri
of OPMDs. In the study in Japan (Nagao et al., Lanka. J Oral Pathol Med. 42(9):656–62. doi:10.1111/
2001), serum retinol levels were elevated in OLP jop.12067 PMID:23601045
Amarasinghe HK, Usgodaarachchi US, Johnson NW,
cases. [The authors remarked that this could be Lalloo R, Warnakulasuriya S (2010a). Betel-quid
due to changes in dietary habits by cases after the chewing with or without tobacco is a major risk
development of oral symptoms. In a subgroup factor for oral potentially malignant disorders in Sri
Lanka: a case-control study. Oral Oncol. 46(4):297–301.
analysis, serum lycopene levels were low in 4 doi:10.1016/j.oraloncology.2010.01.017 PMID:20189448
cases with erosive lesions.] Amenábar JM, Torres-Pereira CC, Tang KD, Punyadeera C
Serum analysis of leukoplakia cases in several (2019). Two enemies, one fight: an update of oral cancer
of the included studies showed that significantly in patients with Fanconi anemia. Cancer. 125(22):3936–
46. doi:10.1002/cncr.32435 PMID:31390058
low antioxidant vitamin status and low serum Anantharaman D, Marron M, Lagiou P, Samoli E, Ahrens
zinc levels could promote the development of W, Pohlabeln H, et al. (2011). Population attributable
OPMDs. risk of tobacco and alcohol for upper aerodigestive
[The Working Group noted that 7-day food tract cancer. Oral Oncol. 47(8):725–31. doi:10.1016/j.
oraloncology.2011.05.004 PMID:21684805
dairies recorded after the detection of an OPMD Anantharaman D, Muller DC, Lagiou P, Ahrens W,
may be biased by the avoidance of certain foods Holcátová I, Merletti F, et al. (2016). Combined effects
because of new oral symptoms, especially in of smoking and HPV16 in oropharyngeal cancer. Int
J Epidemiol. 45(3):752–61. doi:10.1093/ije/dyw069
patients with OSF.] PMID:27197530
Andrade JOM, Santos CAST, Oliveira MC (2015).
Factors associated with oral cancer: a case-control
194
study in a population in the northeast of Brazil. Rev Basu S, Guhan VN (2015). Antioxidant changes in precan-
Bras Epidemiol. 18(4):894–905. doi:10.1590/1980- cerous and cancerous lesions of oral cavity. Int J Life Sci
5497201500040017 PMID:26982303 Res. 3(1):23–8.
Ankathil R, Mathew A, Joseph F, Nair MK (1996). Is oral Beachler DC, D’Souza G (2013). Oral HPV infection
cancer susceptibility inherited? Report of five oral and head and neck cancers in HIV-infected individ-
cancer families. Eur J Cancer B Oral Oncol. 32B(1):63–7. uals. Curr Opin Oncol. 25(5):503–10. doi:10.1097/
doi:10.1016/0964-1955(95)00055-0 PMID:8729621 CCO.0b013e32836242b4 PMID:23852381
Applebaum KM, Furniss CS, Zeka A, Posner MR, Smith Bektas-Kayhan K, Karagoz G, Kesimli MC, Karadeniz
JF, Bryan J, et al. (2007). Lack of association of alcohol AN, Meral R, Altun M, et al. (2014). Carcinoma of the
and tobacco with HPV16-associated head and neck tongue: a case-control study on etiologic factors and
cancer. J Natl Cancer Inst. 99(23):1801–10. doi:10.1093/ dental trauma. Asian Pac J Cancer Prev. 15(5):2225–9.
jnci/djm233 PMID:18042931 doi:10.7314/APJCP.2014.15.5.2225 PMID:24716961
Arbyn M, Xu L (2018). Efficacy and safety of prophylactic Blot WJ, McLaughlin JK, Winn DM, Austin DF, Greenberg
HPV vaccines. A Cochrane review of randomized RS, Preston-Martin S, et al. (1988). Smoking and
trials. Expert Rev Vaccines. 17(12):1085–91. doi:10.108 drinking in relation to oral and pharyngeal cancer.
0/14760584.2018.1548282 PMID:30495978 Cancer Res. 48(11):3282–7. PMID:3365707
Arnold M, Leitzmann M, Freisling H, Bray F, Romieu Boccia S, Hashibe M, Gallì P, De Feo E, Asakage T,
I, Renehan A, et al. (2016). Obesity and cancer: an Hashimoto T, et al. (2009). Aldehyde dehydroge-
update of the global impact. Cancer Epidemiol. 41:8–15. nase 2 and head and neck cancer: a meta-analysis
doi:10.1016/j.canep.2016.01.003 PMID:26775081 implementing a Mendelian randomization approach.
Ashraf S, Al-Maweri SA, Alaizari N, Umair A, Ariffin Cancer Epidemiol Biomarkers Prev. 18(1):248–54.
Z, Alhajj MN, et al. (2020). The association between doi:10.1158/1055-9965.EPI-08-0462 PMID:19124505
Epstein-Barr virus and oral lichen planus: a system- Boffetta P, Aagnes B, Weiderpass E, Andersen A (2005).
atic review and meta-analysis. J Oral Pathol Med. Smokeless tobacco use and risk of cancer of the
49(10):969–76. doi:10.1111/jop.13093 PMID:32746493 pancreas and other organs. Int J Cancer. 114(6):992–5.
Asthana S, Labani S, Kailash U, Sinha DN, Mehrotra R doi:10.1002/ijc.20811 PMID:15645430
(2019). Association of smokeless tobacco use and oral Boffetta P, Hayes RB, Sartori S, Lee YC, Muscat J, Olshan
cancer: a systematic global review and meta-analysis. A, et al. (2016). Mouthwash use and cancer of the head
Nicotine Tob Res. 21(9):1162–71. doi:10.1093/ntr/nty074 and neck: a pooled analysis from the International
PMID:29790998 Head and Neck Cancer Epidemiology Consortium.
Awan KH, Hegde R, Cheever VJ, Carroll W, Khan S, Eur J Cancer Prev. 25(4):344–8. doi:10.1097/
Patil S, et al. (2018). Oral and pharyngeal cancer risk CEJ.0000000000000179 PMID:26275006
associated with occupational carcinogenic substances: Boffetta P, Hecht S, Gray N, Gupta P, Straif K (2008).
systematic review. Head Neck. 40(12):2724–32. Smokeless tobacco and cancer. Lancet Oncol. 9(7):667–
doi:10.1002/hed.25486 PMID:30387891 75. doi:10.1016/S1470-2045(08)70173-6 PMID:18598931
Baan RA, Stewart BW, Straif K, editors (2019). Tumour Bose SC, Singh M, Vyas P, Singh M (2012). Plasma zinc
site concordance and mechanisms of carcinogenesis antioxidant vitamins, glutathione levels and total
(IARC Scientific Publications No. 165). Lyon, France: antioxidant activity in oral leukoplakia. Dent Res J
International Agency for Research on Cancer. Available (Isfahan). 9(2):158–61. doi:10.4103/1735-3327.95229
from: https://publications.iarc.fr/578. PMID:22623931
Bagnardi V, Rota M, Botteri E, Tramacere I, Islami F, Bosetti C, Gallus S, Peto R, Negri E, Talamini R, Tavani
Fedirko V, et al. (2015). Alcohol consumption and A, et al. (2008). Tobacco smoking, smoking cessa-
site-specific cancer risk: a comprehensive dose-re- tion, and cumulative risk of upper aerodigestive tract
sponse meta-analysis. Br J Cancer. 112(3):580–93. cancers. Am J Epidemiol. 167(4):468–73. doi:10.1093/
doi:10.1038/bjc.2014.579 PMID:25422909 aje/kwm318 PMID:18056925
Balaram P, Sridhar H, Rajkumar T, Vaccarella S, Herrero Bundgaard T, Wildt J, Frydenberg M, Elbrønd O, Nielsen
R, Nandakumar A, et al. (2002). Oral cancer in southern JE (1995). Case-control study of squamous cell cancer
India: the influence of smoking, drinking, paan- of the oral cavity in Denmark. Cancer Causes Control.
chewing and oral hygiene. Int J Cancer. 98(3):440–5. 6(1):57–67. doi:10.1007/BF00051681 PMID:7718736
doi:10.1002/ijc.10200 PMID:11920597 Cancela MC, Ramadas K, Fayette JM, Thomas G, Muwonge
Barón AE, Franceschi S, Barra S, Talamini R, La Vecchia C R, Chapuis F, et al. (2009). Alcohol intake and oral
(1993). A comparison of the joint effects of alcohol and cavity cancer risk among men in a prospective study
smoking on the risk of cancer across sites in the upper in Kerala, India. Community Dent Oral Epidemiol.
aerodigestive tract. Cancer Epidemiol Biomarkers Prev. 37(4):342–9. doi:10.1111/j.1600-0528.2009.00475.x
2(6):519–23. PMID:8268767 PMID:19486349
195
Chang C-C, Lin M-S, Chen Y-T, Tu LT, Jane S-W, Chen development. J Exp Clin Cancer Res. 17(3):331–6.
M-Y (2015b). Metabolic syndrome and health-re- PMID:9894771
lated behaviours associated with pre-oral cancerous Conway DI, Hovanec J, Ahrens W, Ross A, Holcatova
lesions among adults aged 20-80 years in Yunlin I, Lagiou P, et al. (2021). Occupational socioeco-
County, Taiwan: a cross-sectional study. BMJ Open. nomic risk associations for head and neck cancer in
5(12):e008788. doi:10.1136/bmjopen-2015-008788 Europe and South America: individual participant
PMID:26685025 data analysis of pooled case-control studies within
Chang CM, Corey CG, Rostron BL, Apelberg BJ (2015a). the INHANCE Consortium. J Epidemiol Community
Systematic review of cigar smoking and all cause Health. 75(8):779–87. doi:10.1136/jech-2020-214913
and smoking related mortality. BMC Public Health. PMID:33622804
15(1):39 0. d o i : 1 0 . 1 1 8 6 / s 1 2 8 8 9 - 0 1 5 -1617- 5 Conway DI, McMahon AD, Smith K, Black R, Robertson
PMID:25907101 G, Devine J, et al. (2010). Components of socioec-
Chang CP, Chang SC, Chuang SC, Berthiller J, Ferro G, onomic risk associated with head and neck cancer:
Matsuo K, et al. (2019). Age at start of using tobacco a population-based case-control study in Scotland.
on the risk of head and neck cancer: pooled analysis in Br J Oral Maxillofac Surg. 48(1):11–7. doi:10.1016/j.
the International Head and Neck Cancer Epidemiology bjoms.2009.03.020 PMID:19481316
Consortium (INHANCE). Cancer Epidemiol. 63:101615. Conway DI, Petticrew M, Marlborough H, Berthiller J,
doi:10.1016/j.canep.2019.101615 PMID:31586822 Hashibe M, Macpherson LMD (2008). Socioeconomic
Chaturvedi AK, Graubard BI, Broutian T, Pickard RKL, inequalities and oral cancer risk: a systematic review
Tong ZY, Xiao W, et al. (2018). Effect of prophylactic and meta-analysis of case-control studies. Int J Cancer.
human papillomavirus (HPV) vaccination on oral HPV 122(12):2811–9. doi:10.1002/ijc.23430 PMID:18351646
infections among young adults in the United States. J Currie S, Farah C (2014). Alcohol-containing mouthwash
Clin Oncol. 36(3):262–7. doi:10.1200/JCO.2017.75.0141 and oral cancer risk: a review of current evidence. OA
PMID:29182497 Alcohol. 2(1):4.1–4.9. Available from: https://espace.
Chaturvedi AK, Graubard BI, Broutian T, Xiao W, library.uq.edu.au/view/UQ:347891.
Pickard RKL, Kahle L, et al. (2019). Prevalence of oral D’Souza G, Kreimer AR, Viscidi R, Pawlita M, Fakhry C,
HPV infection in unvaccinated men and women in Koch WM, et al. (2007). Case-control study of human
the United States, 2009–2016. JAMA. 322(10):977–9. papillomavirus and oropharyngeal cancer. N Engl J
doi:10.1001/jama.2019.10508 PMID:31503300 Med. 356(19):1944–56. doi:10.1056/NEJMoa065497
Chen F, Lin L, Yan L, Liu F, Qiu Y, Wang J, et al. (2018). PMID:17494927
Nomograms and risk scores for predicting the risk of Dasanayake AP, Silverman AJ, Warnakulasuriya S (2010).
oral cancer in different sexes: a large-scale case-control Maté drinking and oral and oro-pharyngeal cancer:
study. J Cancer. 9(14):2543–8. doi:10.7150/jca.24431 a systematic review and meta-analysis. Oral Oncol.
PMID:30026853 46(2):82–6. doi:10.1016/j.oraloncology.2009.07.006
Chen J, Domingue JC, Sears CL (2017). Microbiota dysbi- PMID:20036605
osis in select human cancers: evidence of association de la Cour CD, Sperling CD, Belmonte F, Syrjänen S, Kjaer
and causality. Semin Immunol. 32:25–34. doi:10.1016/j. SK (2021). Human papillomavirus prevalence in oral
smim.2017.08.001 PMID:28822617 potentially malignant disorders: systematic review
Chiang CT, Chang TK, Hwang YH, Su CC, Tsai KY, and meta-analysis. Oral Dis. 27(3):431–8. doi:10.1111/
Yuan TH, et al. (2011). A critical exploration of blood odi.13322 PMID:32144837
and environmental chromium concentration among de Lima MAP, Teodoro IPP, Galiza LE, Filho PHBM,
oral cancer patients in an oral cancer prevalent area Marques FM, Pinheiro Junior RFF, et al. (2019).
of Taiwan. Environ Geochem Health. 33(5):469–76. Association between Epstein-Barr virus and oral
doi:10.1007/s10653-010-9346-2 PMID:20978824 carcinoma: a systematic review with meta-anal-
Christensen CH, Rostron B, Cosgrove C, Altekruse SF, ysis. Crit Rev Oncog. 24(4):349–68. doi:10.1615/
Hartman AM, Gibson JT, et al. (2018). Association CritRevOncog.2019031897 PMID:32421990
of cigarette, cigar, and pipe use with mortality risk in de Martel C, Plummer M, Vignat J, Franceschi S (2017).
the US population. JAMA Intern Med. 178(4):469–76. Worldwide burden of cancer attributable to HPV by
doi:10.1001/jamainternmed.2017.8625 PMID:29459935 site, country and HPV type. Int J Cancer. 141(4):664–
Chuang SC, Jenab M, Heck JE, Bosetti C, Talamini 70. doi:10.1002/ijc.30716 PMID:28369882
R, Matsuo K, et al. (2012). Diet and the risk of head De Stefani E, Boffetta P, Deneo-Pellegrini H, Ronco
and neck cancer: a pooled analysis in the INHANCE AL, Acosta G, Ferro G, et al. (2007). The effect of
consortium. Cancer Causes Control. 23(1):69–88. smoking and drinking in oral and pharyngeal
doi:10.1007/s10552-011-9857-x PMID:22037906 cancers: a case-control study in Uruguay. Cancer
Cianfriglia F, Manieri A, Di Gregorio DA, Di Iorio AM Lett. 246(1–2):282–9. doi:10.1016/j.canlet.2006.03.008
(1998). Retinol dietary intake and oral leukoplakia PMID:16624486
196
Di Credico G, Edefonti V, Polesel J, Pauli F, Torelli N, Filomeno M, Bosetti C, Garavello W, Levi F, Galeone C,
Serraino D, et al. (2019). Joint effects of intensity and Negri E, et al. (2014). The role of a Mediterranean diet
duration of cigarette smoking on the risk of head and on the risk of oral and pharyngeal cancer. Br J Cancer.
neck cancer: a bivariate spline model approach. Oral 111(5):981–6. doi:10.1038/bjc.2014.329 PMID:24937666
Oncol. 94:47–57. doi:10.1016/j.oraloncology.2019.05.006 Fisher MA, Bouquot JE, Shelton BJ (2005). Assessment
PMID:31178212 of risk factors for oral leukoplakia in West Virginia.
Drolet M, Bénard É, Pérez N, Brisson M, Ali H, Boily M-C, Community Dent Oral Epidemiol. 33(1):45–52.
et al.; HPV Vaccination Impact Study Group (2019). doi:10.1111/j.1600-0528.2004.00195.x PMID:15642046
Population-level impact and herd effects following the Foppa I, Minder CE (1992). Oral, pharyngeal and laryngeal
introduction of human papillomavirus vaccination cancer as a cause of death among Swiss cooks. Scand
programmes: updated systematic review and meta- J Work Environ Health. 18(5):287–92. doi:10.5271/
analysis. Lancet. 394(10197):497–509. doi:10.1016/ sjweh.1575 PMID:1439654
S0140-6736(19)30298-3 PMID:31255301 Franceschi S, Barra S, La Vecchia C, Bidoli E, Negri
Edwards DM, Jones J (1999). Incidence of and survival E, Talamini R (1992). Risk factors for cancer of
from upper aerodigestive tract cancers in the U.K.: the the tongue and the mouth. A case-control study
influence of deprivation. Eur J Cancer. 35(6):968–72. from northern Italy. Cancer. 70(9):2227–33.
doi:10.1016/S0959-8049(99)00095-7 PMID:10533480 doi:10.1002/1097-0142(19921101)70:9<2227::AID-CN-
El-Zaemey S, Schüz J, Leon ME (2015). Qat chewing CR2820700902>3.0.CO;2-Z PMID:1394055
and risk of potentially malignant and malignant oral Franco EL, Kowalski LP, Oliveira BV, Curado MP, Pereira
disorders: a systematic review. Int J Occup Environ RN, Silva ME, et al. (1989). Risk factors for oral cancer
Med. 6(3):129– 4 3. d oi :10 .15171/ij o e m . 2 015. 537 in Brazil: a case-control study. Int J Cancer. 43(6):992–
PM I D : 2 6174 9 9 0 1000. doi:10.1002/ijc.2910430607 PMID:2732011
Ernster VL, Grady DG, Greene JC, Walsh M, Robertson Freedman ND, Abnet CC, Leitzmann MF, Hollenbeck
P, Daniels TE, et al. (1990). Smokeless tobacco use AR, Schatzkin A (2007). Prospective investigation of
and health effects among baseball players. JAMA. the cigarette smoking–head and neck cancer asso-
264(2):218–24. doi:10.1001/jama.1990.03450020070029 ciation by sex. Cancer. 110(7):1593–601. doi:10.1002/
PMID:2355443 cncr.22957 PMID:17724671
Evstifeeva TV, Zaridze DG (1992). Nass use, cigarette Freedman ND, Park Y, Subar AF, Hollenbeck AR,
smoking, alcohol consumption and risk of oral and Leitzmann MF, Schatzkin A, et al. (2008). Fruit and
oesophageal precancer. Eur J Cancer B Oral Oncol. vegetable intake and head and neck cancer risk in a large
28B(1):29–35. doi:10.1016/0964-1955(92)90008-O United States prospective cohort study. Int J Cancer.
PMID:1422467 122(10):2330–6. doi:10.1002/ijc.23319 PMID:18092323
Fahmy MS, Sadeghi A, Behmard S (1983). Epidemiologic Furquim CP, Pivovar A, Amenábar JM, Bonfim C, Torres-
study of oral cancer in Fars Province, Iran. Pereira CC (2018). Oral cancer in Fanconi anemia:
Community Dent Oral Epidemiol. 11(1):50–8. review of 121 cases. Crit Rev Oncol Hematol. 125:35–40.
doi:10.1111/j.1600-0528.1983.tb01354.x PMID:6572124 doi:10.1016/j.critrevonc.2018.02.013 PMID:29650274
Farsi NJ, Rousseau MC, Schlecht N, Castonguay G, Galeone C, Tavani A, Pelucchi C, Turati F, Winn DM, Levi
Allison P, Nguyen-Tan PF, et al. (2017). Aetiological F, et al. (2010). Coffee and tea intake and risk of head
heterogeneity of head and neck squamous cell carci- and neck cancer: pooled analysis in the International
nomas: the role of human papillomavirus infections, Head and Neck Cancer Epidemiology consortium.
smoking and alcohol. Carcinogenesis. 38(12):1188–95. Cancer Epidemiol Biomarkers Prev. 19(7):1723–36.
doi:10.1093/carcin/bgx106 PMID:29029021 doi:10.1158/1055-9965.EPI-10-0191 PMID:20570908
FDA (2003). 21 CFR part 356. Oral health care drug prod- Galvão De Podestá OP, Peres SV, Salaroli LB, Cattafesta
ucts for over-the-counter human use; antigingivitis/ M, De Podestá JRV, von Zeidler SLV, et al. (2019).
antiplaque drug products; establishment of a mono- Consumption of minimally processed foods as protec-
graph; proposed rules. Fed Regist. 68(103):32241–3. tive factors in the genesis of squamous cell carcinoma of
Silver Spring (MD), USA: U.S. Food and Drug the head and neck in Brazil. PLoS One. 14(7):e0220067.
Administration, Department of Health and Human doi:10.1371/journal.pone.0220067 PMID:31344089
Services. Available from: https://www.govinfo.gov/ Gandini S, Botteri E, Iodice S, Boniol M, Lowenfels AB,
content/pkg/FR-2003-05-29/pdf/03-12783.pdf. Maisonneuve P, et al. (2008). Tobacco smoking and
Filippini T, Malavolti M, Borrelli F, Izzo AA, cancer: a meta-analysis. Int J Cancer. 122(1):155–64.
Fairweather-Tait SJ, Horneber M, et al. (2020). Green doi:10.1002/ijc.23033 PMID:17893872
tea (Camellia sinensis) for the prevention of cancer. Gandini S, Negri E, Boffetta P, La Vecchia C, Boyle P
Cochrane Database Syst Rev. 3(3):CD005004. (2012). Mouthwash and oral cancer risk quantitative
doi:10.1002/14651858.CD005004.pub3 PMID:32118296 meta-analysis of epidemiologic studies. Ann Agric
Environ Med. 19(2):173–80. PMID:22742785
197
Garavello W, Foschi R, Talamini R, La Vecchia C, Rossi M, 36 000 Indian tobacco users. Lancet. 1(8492):1235–9.
Dal Maso L, et al. (2008). Family history and the risk of doi:10.1016/S0140-6736(86)91386-3 PMID:2872391
oral and pharyngeal cancer. Int J Cancer. 122(8):1827– Gupta PC, Mehta FS, Pindborg JJ, Bhonsle RB, Murti PR,
31. doi:10.1002/ijc.23199 PMID:18076043 Daftary DK, et al. (1992). Primary prevention trial of
Gaudet MM, Kitahara CM, Newton CC, Bernstein L, oral cancer in India: a 10-year follow-up study. J Oral
Reynolds P, Weiderpass E, et al. (2015). Anthropometry Pathol Med. 21(10):433–9. doi:10.1111/j.1600-0714.1992.
and head and neck cancer: a pooled analysis of cohort tb00970.x PMID:1460581
data. Int J Epidemiol. 44(2):673–81. doi:10.1093/ije/ Gupta PC, Murti PR, Bhonsle RB, Mehta FS, Pindborg
dyv059 PMID:26050257 JJ (1995). Effect of cessation of tobacco use on
Gelman A, Carlin JB, Stern HS, Rubin DB (1995). Bayesian the incidence of oral mucosal lesions in a 10-yr
data analysis. New York (NY), USA: Chapman follow-up study of 12,212 users. Oral Dis. 1(1):54–8.
and Hall/CRC. doi:10.1201/9780429258411 doi:10.1111/j.1601-0825.1995.tb00158.x PMID:7553382
Gholizadeh P, Eslami H, Yousefi M, Asgharzadeh M, Gupta R, Mariano LC, Nethan ST, Kedar A, Sinha DN,
Aghazadeh M, Kafil HS (2016). Role of oral microbiome Warnakulasuriya S, et al. (2022). Risk reversal of oral,
on oral cancers, a review. Biomed Pharmacother. pharyngeal and oesophageal cancers after cessation of
8 4 : 552– 8. d oi :10 .1016 /j . bi o p h a . 2 016.09.0 82 betel quid users: a systematic review and meta-anal-
PMID:27693964 ysis. Ann Glob Health. 88(1):5. doi:10.5334/aogh.3643
Gilligan GM, Galindez Costa MFF, Moine L, Panico RL, PMID:35087705
Piemonte ED (2017). Could chronic mechanical irrita- Gupta S, Reddy MV, Harinath BC (2004). Role of oxida-
tion facilitate entry of human papillomavirus (HPV) tive stress and antioxidants in aetiopathogenesis and
facilitating oral HPV infection? Transl Res Oral Oncol. management of oral submucous fibrosis. Indian J
2:1–6. doi:10.1177/2057178X17746799 Clin Biochem. 19(1):138–41. doi:10.1007/BF02872409
Gillison ML, Chaturvedi AK, Anderson WF, Fakhry C PMID:23105446
(2015). Epidemiology of human papillomavirus-pos- Handley TPB, Ogden GR (2006). Dyskeratosis congenita:
itive head and neck squamous cell carcinoma. J Clin oral hyperkeratosis in association with lichenoid
Oncol. 33(29):3235–42. doi:10.1200/JCO.2015.61.6995 reaction. J Oral Pathol Med. 35(8):508–12. doi:10.1111/
PMID:26351338 j.1600-0714.2006.00434.x PMID:16918603
Guha N, Boffetta P, Wünsch Filho V, Eluf Neto J, Shangina Hashibe M, Boffetta P, Zaridze D, Shangina O, Szeszenia-
O, Zaridze D, et al. (2007). Oral health and risk of Dabrowska N, Mates D, et al. (2006). Evidence for an
squamous cell carcinoma of the head and neck and important role of alcohol- and aldehyde-metabolizing
esophagus: results of two multicentric case-control genes in cancers of the upper aerodigestive tract.
studies. Am J Epidemiol. 166(10):1159–73. doi:10.1093/ Cancer Epidemiol Biomarkers Prev. 15(4):696–703.
aje/kwm193 PMID:17761691 doi:10.1158/1055-9965.EPI-05-0710 PMID:16614111
Guha N, Warnakulasuriya S, Vlaanderen J, Straif K Hashibe M, Brennan P, Benhamou S, Castellsague X,
(2014). Betel quid chewing and the risk of oral and Chen C, Curado MP, et al. (2007). Alcohol drinking
oropharyngeal cancers: a meta-analysis with implica- in never users of tobacco, cigarette smoking in never
tions for cancer control. Int J Cancer. 135(6):1433–43. drinkers, and the risk of head and neck cancer:
doi:10.1002/ijc.28643 PMID:24302487 pooled analysis in the International Head and Neck
Gupta B, Bray F, Kumar N, Johnson NW (2017). Cancer Epidemiology Consortium. J Natl Cancer Inst.
Associations between oral hygiene habits, diet, tobacco 99(10):777–89. doi:10.1093/jnci/djk179 PMID:17505073
and alcohol and risk of oral cancer: a case-control study Hashibe M, Brennan P, Chuang SC, Boccia S, Castellsague
from India. Cancer Epidemiol. 51:7–14. doi:10.1016/j. X, Chen C, et al. (2009). Interaction between tobacco
canep.2017.09.003 PMID:28968558 and alcohol use and the risk of head and neck cancer:
Gupta PC, Hebert JR, Bhonsle RB, Murti PR, Mehta H, pooled analysis in the International Head and Neck
Mehta FS (1999). Influence of dietary factors on oral Cancer Epidemiology Consortium. Cancer Epidemiol
precancerous lesions in a population-based case-con- Biomarkers Prev. 18(2):541–50. doi:10.1158/1055-9965.
trol study in Kerala, India. Cancer. 85(9):1885–93. EPI-08-0347 PMID:19190158
PMID:10223226 Hashibe M, Mathew B, Kuruvilla B, Thomas G,
Gupta PC, Hebert JR, Bhonsle RB, Sinor PN, Mehta H, Sankaranarayanan R, Parkin DM, et al. (2000b).
Mehta FS (1998). Dietary factors in oral leukoplakia Chewing tobacco, alcohol, and the risk of erythro-
and submucous fibrosis in a population-based case plakia. Cancer Epidemiol Biomarkers Prev. 9(7):639–
control study in Gujarat, India. Oral Dis. 4(3):200–6. 45. PMID:10919731
doi:10.1111/j.1601-0825.1998.tb00279.x PMID:9972171 Hashibe M, Sankaranarayanan R, Thomas G,
Gupta PC, Mehta FS, Pindborg JJ, Aghi MB, Bhonsle Kuruvilla B, Mathew B, Somanathan T, et al.
RB, Daftary DK, et al. (1986). Intervention study (2000a). Alcohol drinking, body mass index
for primary prevention of oral cancer among and the risk of oral leukoplakia in an Indian
198
population. Int J Cancer. 88(1):129–34. doi:10.1002/1097- Hu Y, Zhong R, Li H, Zou Y (2020). Effects of betel quid,
0215(20001001)88:1<129::AID-IJC20>3.0.CO;2-U smoking and alcohol on oral cancer risk: a case-con-
PMID:10962450 trol study in Hunan Province, China. Subst Use
Hashim D, Sartori S, Brennan P, Curado MP, Wünsch- Misuse. 55(9):1501–8. doi:10.1080/10826084.2020.1750
Filho V, Divaris K, et al. (2016). The role of oral hygiene 031 PMID:32569534
in head and neck cancer: results from International Huang CC, Hsiao JR, Lee WT, Lee YC, Ou CY, Chang
Head and Neck Cancer Epidemiology (INHANCE) CC, et al. (2017). Investigating the association between
consortium. Ann Oncol. 27(8):1619–25. doi:10.1093/ alcohol and risk of head and neck cancer in Taiwan.
annonc/mdw224 PMID:27234641 Sci Rep. 7(1):9701. doi:10.1038/s41598-017-08802-4
Hayes RB, Ahn J, Fan X, Peters BA, Ma Y, Yang L, et al. PMID:28851901
(2018). Association of oral microbiome with risk for Huang J, He B, Chen F, Liu F, Yan L, Hu Z, et al. (2015).
incident head and neck squamous cell cancer. JAMA Association between oral hygiene, chronic diseases, and
Oncol. 4(3):358–65. doi:10.1001/jamaoncol.2017.4777 oral squamous cell carcinoma. [in Chinese] Zhonghua
PMID:29327043 Yu Fang Yi Xue Za Zhi. 49(8):688–92. PMID:26733026
Hayes RB, Bravo-Otero E, Kleinman DV, Brown LM, IARC (1986). Tobacco smoking. IARC Monogr Eval
Fraumeni JF Jr, Harty LC, et al. (1999). Tobacco Carcinog Risk Chem Hum. 38:1–421. Available from:
and alcohol use and oral cancer in Puerto Rico. https://publications.iarc.fr/56
C a n c e r C a u s e s C o n t r o l . 1 0 ( 1) : 2 7 – 3 3 . IARC (2004a). Betel-quid and areca-nut chewing and
doi:10.1023/A:1008 8 7 6 1 1 5797 PMID:10334639 some areca-nut derived nitrosamines. IARC Monogr
He T, Guo X, Li X, Liao C, Yin W (2020). Association Eval Carcinog Risks Hum. 85:1–334. Available from:
between coffee intake and the risk of oral cavity cancer: https://publications.iarc.fr/103 PMID:15635762
a meta-analysis of observational studies. Eur J Cancer IARC (2004b). Tobacco smoke and involuntary smoking.
Prev. 29(1):80–8. doi:10.1097/CEJ.0000000000000515 IARC Monogr Eval Carcinog Risks Hum. 83:1–1452.
PMID:31021885 Available from: https://publications.iarc.fr/101
Hebert JR, Gupta PC, Bhonsle RB, Mehta H, Zheng W, PMID:15285078
Sanderson M, et al. (2002). Dietary exposures and oral IARC (2007a). Smokeless tobacco and some tobacco-spe-
precancerous lesions in Srikakulam District, Andhra cific N-nitrosamines. IARC Monogr Eval Carcinog Risks
Pradesh, India. Public Health Nutr. 5(2):303–12. Hum. 89:1–626. Available from: https://publications.
doi:10.1079/PHN2002249 PMID:12020382 iarc.fr/107 PMID:18335640
Heck JE, Berthiller J, Vaccarella S, Winn DM, Smith EM, IARC (2007b). Tobacco control: reversal of risk after
Shan’gina O, et al. (2010). Sexual behaviours and the quitting smoking. IARC Handb Cancer Prev. 11:1–358.
risk of head and neck cancers: a pooled analysis in the Available from: https://publications.iarc.fr/381
International Head and Neck Cancer Epidemiology IARC (2010). Alcohol consumption and ethyl carba-
(INHANCE) consortium. Int J Epidemiol. 39(1):166– mate. IARC Monogr Eval Carcinog Risks Hum. 96:1–
81. doi:10.1093/ije/dyp350 PMID:20022926 1424. Available from: https://publications.iarc.fr/114
Herrero R, Castellsagué X, Pawlita M, Lissowska J, Kee F, PMID:21735939
Balaram P, et al.; IARC Multicenter Oral Cancer Study IARC (2012a). Biological agents. IARC Monogr Eval
Group (2003). Human papillomavirus and oral cancer: Carcinog Risks Hum. 100B:1–475. Available from:
the International Agency for Research on Cancer https://publications.iarc.fr/119 PMID:23189750
multicenter study. J Natl Cancer Inst. 95(23):1772–83. IARC (2012b). Personal habits and indoor combustions.
doi:10.1093/jnci/djg107 PMID:14652239 IARC Monogr Eval Carcinog Risks Hum. 100E:1–
Herrero R, Quint W, Hildesheim A, Gonzalez P, Struijk L, 575. Available from: https://publications.iarc.fr/122
Katki HA, et al.; CVT Vaccine Group (2013). Reduced PMID:23193840
prevalence of oral human papillomavirus (HPV) 4 IARC (2018). Drinking coffee, mate, and very hot bever-
years after bivalent HPV vaccination in a randomized ages. IARC Monogr Eval Carcinog Risks Hum. 116:1–
clinical trial in Costa Rica. PLoS One. 8(7):e68329. 499. Available from: https://publications.iarc.fr/566 </
doi:10.1371/journal.pone.0068329 PMID:23873171 eref> PMID:31310458
Hille J, Johnson NW (2017). The burden of oral cancer IARC (2021). Opium consumption. IARC Monogr Identif
in sub-Saharan Africa: an estimate as presented to the Carcinog Hazard Hum. 126:1–253. Available from:
Global Oral Cancer Forum, March 2016. Transl Res https://publications.iarc.fr/600
Oral Oncol. 2:1–13. Ide R, Mizoue T, Fujino Y, Hoshiyama Y, Sakata K,
Hou J, Gu Y, Hou W, Wu S, Lou Y, Yang W, et al. (2015). Tamakoshi A, et al.; JACC Study Group (2008).
P53 codon 72 polymorphism, human papillomavirus Cigarette smoking, alcohol drinking, and oral and
infection, and their interaction to oral carcinoma pharyngeal cancer mortality in Japan. Oral Dis.
susceptibility. BMC Genet. 16(1):72. doi:10.1186/ 14(4):314–9. doi:10.1111/j.1601-0825.2007.01378.x
s12863-015-0235-7 PMID:26123760 PMID:18449960
199
Im PK, Millwood IY, Kartsonaki C, Guo Y, Chen Y, consortium. Int J Cancer. 141(9):1811–21. doi:10.1002/
Turnbull I, et al.; China Kadoorie Biobank (CKB) ijc.30886 PMID:28710831
collaborative group (2021). Alcohol drinking and risks Khan Z, Suliankatchi RA, Heise TL, Dreger S (2019).
of liver cancer and non-neoplastic chronic liver diseases Naswar (smokeless tobacco) use and the risk of oral
in China: a 10-year prospective study of 0.5 million cancer in Pakistan: a systematic review with meta-
adults. BMC Med. 19(1):216. doi:10.1186/s12916-021- analysis. Nicotine Tob Res. 21(1):32–40. doi:10.1093/ntr/
02079-1 PMID:34530818 ntx281 PMID:29294113
Inoue-Choi M, Shiels MS, McNeel TS, Graubard BI, King GN, Healy CM, Glover MT, Kwan JTC, Williams
Hatsukami D, Freedman ND (2019). Contemporary DM, Leigh IM, et al. (1995). Increased prevalence of
associations of exclusive cigarette, cigar, pipe, and dysplastic and malignant lip lesions in renal-transplant
smokeless tobacco use with overall and cause-specific recipients. N Engl J Med. 332(16):1052–7. doi:10.1056/
mortality in the United States. JNCI Cancer Spectr. NEJM199504203321602 PMID:7898522
3(3):pkz036. doi:10.1093/jncics/pkz036 PMID:31321380 Kjaer SK, Dehlendorff C, Belmonte F, Baandrup L
Jacob BJ, Straif K, Thomas G, Ramadas K, Mathew B, (2021). Real-world effectiveness of human papillo-
Zhang ZF, et al. (2004). Betel quid without tobacco mavirus vaccination against cervical cancer. J Natl
as a risk factor for oral precancers. Oral Oncol. Cancer Inst. 113(10):1329–35. doi:10.1093/jnci/djab080
40(7):697–704. doi:10.1016/j.oraloncology.2004.01.005 PMID:33876216
PMID:15172639 Ko YC, Huang YL, Lee CH, Chen MJ, Lin LM, Tsai
Jayalekshmi PA, Gangadharan P, Akiba S, Koriyama C, CC (1995). Betel quid chewing, cigarette smoking
Nair RR (2011). Oral cavity cancer risk in relation to and alcohol consumption related to oral cancer
tobacco chewing and bidi smoking among men in in Taiwan. J Oral Pathol Med. 24(10):450–3.
Karunagappally, Kerala, India: Karunagappally cohort doi:10.1111/j.1600-0714.1995.tb01132.x PMID:8600280
study. Cancer Sci. 102(2):460–7. doi:10.1111/j.1349- Kreimer AR, Ferreiro-Iglesias A, Nygard M, Bender
7006.2010.01785.x PMID:21129124 N, Schroeder L, Hildesheim A, et al. (2019). Timing
Jayalekshmi PA, Gangadharan P, Akiba S, Nair RR, Tsuji of HPV16-E6 antibody seroconversion before
M, Rajan B (2009). Tobacco chewing and female oral OPSCC: findings from the HPVC3 consortium. Ann
cavity cancer risk in Karunagappally cohort, India. Oncol. 30(8):1335–43. doi:10.1093/annonc/mdz138
Br J Cancer. 100(5):848–52. doi:10.1038/sj.bjc.6604907 PMID:31185496
PMID:19259098 Kreimer AR, Johansson M, Waterboer T, Kaaks R, Chang-
Josyula S, Lin J, Xue X, Rothman N, Lan Q, Rohan TE, et Claude J, Drogen D, et al. (2013). Evaluation of human
al. (2015). Household air pollution and cancers other papillomavirus antibodies and risk of subsequent
than lung: a meta-analysis. Environ Health. 14(1):24. head and neck cancer. J Clin Oncol. 31(21):2708–15.
doi:10.1186/s12940-015-0001-3 PMID:25890249 doi:10.1200/JCO.2012.47.2738 PMID:23775966
Katsanos KH, Roda G, McBride RB, Cohen B, Colombel Kreimer AR, Johansson M, Yanik EL, Katki HA, Check
JF (2016). Increased risk of oral cancer in patients DP, Lang Kuhs KA, et al. (2017). Kinetics of the human
with inflammatory bowel diseases. Clin Gastroenterol papillomavirus type 16 E6 antibody response prior
Hepatol. 14(3):413–20. doi:10.1016/j.cgh.2015.09.041 to oropharyngeal cancer. J Natl Cancer Inst. 109(8):
PMID:26499929 doi:10.1093/jnci/djx005 PMID:28376197
Katz J, Onate MD, Pauley KM, Bhattacharyya I, Cha S La Vecchia C (2009). Mouthwash and oral cancer risk:
(2011). Presence of Porphyromonas gingivalis in gingival an update. Oral Oncol. 45(3):198–200. doi:10.1016/j.
squamous cell carcinoma. Int J Oral Sci. 3(4):209–15. oraloncology.2008.08.012 PMID:18952488
doi:10.4248/IJOS11075 PMID:22010579 Lafuente Ibáñez de Mendoza I, Maritxalar Mendia X,
Kavarthapu A, Gurumoorthy K (2021). Linking chronic García de la Fuente AM, Quindós Andrés G, Aguirre
periodontitis and oral cancer: a review. Oral Oncol. Urizar JM (2020). Role of Porphyromonas gingivalis in
121:105375. doi:10.1016/j.oraloncology.2021.105375 oral squamous cell carcinoma development: a system-
PMID:34140233 atic review. J Periodontal Res. 55(1):13–22. doi:10.1111/
Kawakita D, Lee YA, Gren LH, Buys SS, La Vecchia C, jre.12691 PMID:31529626
Hashibe M (2019). Fiber intake and the risk of head Lam TK, Cross AJ, Freedman N, Park Y, Hollenbeck
and neck cancer in the Prostate, Lung, Colorectal and AR, Schatzkin A, et al. (2011). Dietary fiber and grain
Ovarian (PLCO) cohort. Int J Cancer. 145(9):2342–8. consumption in relation to head and neck cancer in
doi:10.1002/ijc.32162 PMID:30693489 the NIH-AARP Diet and Health Study. Cancer Causes
Kawakita D, Lee YA, Turati F, Parpinel M, Decarli A, Control. 22(10):1405–14. doi:10.1007/s10552-011-9813-9
Serraino D, et al. (2017). Dietary fiber intake and PMID:21785948
head and neck cancer risk: a pooled analysis in the Laprise C, Cahoon EK, Lynch CF, Kahn AR, Copeland
International Head and Neck Cancer Epidemiology G, Gonsalves L, et al. (2019). Risk of lip cancer after
solid organ transplantation in the United States.
200
Am J Transplant. 19(1):227–37. doi:10.1111/ajt.15052 Lockhart PB, Norris CM Jr, Pulliam C (1998). Dental
PMID:30074684 factors in the genesis of squamous cell carcinoma of
Lee CH, Ko YC, Huang HL, Chao YY, Tsai CC, Shieh TY, the oral cavity. Oral Oncol. 34(2):133–9. doi:10.1016/
et al. (2003). The precancer risk of betel quid chewing, S1368-8375(97)00086-9 PMID:9682777
tobacco use and alcohol consumption in oral leuko- Lubin JH, Muscat J, Gaudet MM, Olshan AF, Curado
plakia and oral submucous fibrosis in southern Taiwan. MP, Dal Maso L, et al. (2011). An examination of male
Br J Cancer. 88(3):366–72. doi:10.1038/sj.bjc.6600727 and female odds ratios by BMI, cigarette smoking, and
PMID:12569378 alcohol consumption for cancers of the oral cavity,
Lee CH, Lee KW, Fang FM, Wu DC, Tsai SM, Chen PH, et pharynx, and larynx in pooled data from 15 case-con-
al. (2012). The neoplastic impact of tobacco-free betel- trol studies. Cancer Causes Control. 22(9):1217–31.
quid on the histological type and the anatomical site of doi:10.1007/s10552-011-9792-x PMID:21744095
aerodigestive tract cancers. Int J Cancer. 131(5):E733– Lubin JH, Purdue M, Kelsey K, Zhang ZF, Winn D, Wei
43. doi:10.1002/ijc.27401 PMID:22174014 Q, et al. (2009). Total exposure and exposure rate
Lee PN, Hamling J (2009). Systematic review of the relation effects for alcohol and smoking and risk of head and
between smokeless tobacco and cancer in Europe and neck cancer: a pooled analysis of case-control studies.
North America. BMC Med. 7(1):36. doi:10.1186/1741- Am J Epidemiol. 170(8):937–47. doi:10.1093/aje/kwp222
7015-7-36 PMID:19638245 PMID:19745021
Lee YC, Marron M, Benhamou S, Bouchardy C, Ahrens Luo J, Ye W, Zendehdel K, Adami J, Adami HO, Boffetta
W, Pohlabeln H, et al. (2009). Active and involun- P, et al. (2007). Oral use of Swedish moist snuff (snus)
tary tobacco smoking and upper aerodigestive tract and risk for cancer of the mouth, lung, and pancreas
cancer risks in a multicenter case-control study. in male construction workers: a retrospective cohort
Cancer Epidemiol Biomarkers Prev. 18(12):3353–61. study. Lancet. 369(9578):2015–20. doi:10.1016/S0140-
doi:10.1158/1055-9965.EPI-09-0910 PMID:19959682 6736(07)60678-3 PMID:17498797
Lei J, Ploner A, Elfström KM, Wang J, Roth A, Fang F, Maasland DH, van den Brandt PA, Kremer B, Goldbohm
et al. (2020). HPV vaccination and the risk of inva- RA, Schouten LJ (2014). Alcohol consumption, ciga-
sive cervical cancer. N Engl J Med. 383(14):1340–8. rette smoking and the risk of subtypes of head-neck
doi:10.1056/NEJMoa1917338 PMID:32997908 cancer: results from the Netherlands Cohort Study.
Lesseur C, Diergaarde B, Olshan AF, Wünsch-Filho V, BMC Cancer. 14(1):187. doi:10.1186/1471-2407-14-187
Ness AR, Liu G, et al. (2016). Genome-wide association PMID:24629046
analyses identify new susceptibility loci for oral cavity Maasland DH, van den Brandt PA, Kremer B, Goldbohm
and pharyngeal cancer. Nat Genet. 48(12):1544–50. RA, Schouten LJ (2015). Consumption of vegetables and
doi:10.1038/ng.3685 PMID:27749845 fruits and risk of subtypes of head-neck cancer in the
Lewin F, Norell SE, Johansson H, Gustavsson P, Wennerberg Netherlands Cohort Study. Int J Cancer. 136(5):E396–
J, Biörklund A, et al. (1998). Smoking tobacco, oral snuff, 409. doi:10.1002/ijc.29219 PMID:25220255
and alcohol in the etiology of squamous cell carcinoma Macigo FG, Mwaniki DL, Guthua SW (1995). The
of the head and neck: a population-based case-referent association between oral leukoplakia and use of
study in Sweden. Cancer. 82(7):1367–75. doi:10.1002/ tobacco, alcohol and khat based on relative risks
(S I C I )10 9 7- 014 2 (19 9 8 0 4 01) 8 2 :7<13 6 7: : A I D - assessment in Kenya. Eur J Oral Sci. 103(5):268–73.
CNCR21>3.0.CO;2-3 PMID:9529030 doi:10.1111/j.1600-0722.1995.tb00025.x PMID:8521116
Lewis MAO, Murray S (2006). Safety of alcohol-con- Macigo FG, Mwaniki DL, Guthua SW (1996). Influence of
taining mouthwashes. A review of the evidence. Dent dose and cessation of kiraiku, cigarettes and alcohol use
Health (London). 45:2–4. on the risk of developing oral leukoplakia. Eur J Oral
Li AC, Warnakulasuriya S, Thompson RP (2003). Neoplasia Sci. 104(5–6):498–502. doi:10.1111/j.1600-0722.1996.
of the tongue in a patient with Crohn’s disease treated tb00132.x PMID:9021316
with azathioprine: case report. Eur J Gastroenterol Madathil SA, Rousseau MC, Wynant W, Schlecht NF,
Hepatol. 15(2):185–7. doi:10.1097/00042737-200302000- Netuveli G, Franco EL, et al. (2016). Nonlinear asso-
00013 PMID:12560764 ciation between betel quid chewing and oral cancer:
Li L, Psoter WJ, Buxó CJ, Elias A, Cuadrado L, Morse DE implications for prevention. Oral Oncol. 60:25–31.
(2011). Smoking and drinking in relation to oral poten- doi:10.1016/j.oraloncology.2016.06.011 PMID:27531869
tially malignant disorders in Puerto Rico: a case-control Manoharan S, Nagaraja V, Eslick GD (2014). Ill-fitting
study. BMC Cancer. 11(1):324. doi:10.1186/1471-2407- dentures and oral cancer: a meta-analysis. Oral Oncol.
11-324 PMID:21801414 50(11):1058–61. doi:10.1016/j.oraloncology.2014.08.002
Li S, Lee YC, Li Q, Chen C-J, Hsu W-L, Lou P-J, et al. (2015). PMID:25169920
Oral lesions, chronic diseases and the risk of head and Mariano LC, Warnakulasuriya S, Straif K,
neck cancer. Oral Oncol. 51(12):1082–7. doi:10.1016/j. Monteiro L (2022). Secondhand smoke expo-
oraloncology.2015.10.014 PMID:26526128 sure and oral cancer risk: a systematic review and
201
meta-analysis. Tob Control. 31(5):597–607. doi:10.1136/ Murti PR, Gupta PC, Bhonsle RB, Daftary DK, Mehta
tobaccocontrol-2020-056393 PMID:33903278 FS, Pindborg JJ (1990). Effect on the incidence of
Marichalar-Mendia X, Rodriguez-Tojo MJ, Acha- oral submucous fibrosis of intervention in the areca
Sagredo A, Rey-Barja N, Aguirre-Urizar JM (2010). nut chewing habit. J Oral Pathol Med. 19(2):99–100.
Oral cancer and polymorphism of ethanol metab- doi:10.1111/j.1600-0714.1990.tb00805.x PMID:2341978
olising genes. Oral Oncol. 46(1):9–13. doi:10.1016/j. Muwonge R, Ramadas K, Sankila R, Thara S, Thomas
oraloncology.2009.09.005 PMID:19896412 G, Vinoda J, et al. (2008). Role of tobacco smoking,
Marks MA, Chaturvedi AK, Kelsey K, Straif K, Berthiller chewing and alcohol drinking in the risk of oral
J, Schwartz SM, et al. (2014). Association of mari- cancer in Trivandrum, India: a nested case-con-
juana smoking with oropharyngeal and oral tongue trol design using incident cancer cases. Oral Oncol.
cancers: pooled analysis from the INHANCE consor- 44(5):446–54. doi:10.1016/j.oraloncology.2007.06.002
tium. Cancer Epidemiol Biomarkers Prev. 23(1):160–71. PMID:17933578
doi:10.1158/1055-9965.EPI-13-0181 PMID:24351902 Nagao T, Ikeda N, Warnakulasuriya S, Fukano H, Yuasa
Marron M, Boffetta P, Zhang ZF, Zaridze D, Wünsch- H, Yano M, et al. (2000). Serum antioxidant micro-
Filho V, Winn DM, et al. (2010). Cessation of alcohol nutrients and the risk of oral leukoplakia among
drinking, tobacco smoking and the reversal of head Japanese. Oral Oncol. 36(5):466–70. doi:10.1016/S1368-
and neck cancer risk. Int J Epidemiol. 39(1):182–96. 8375(00)00037-3 PMID:10964055
doi:10.1093/ije/dyp291 PMID:19805488 Nagao T, Warnakulasuriya S, Ikeda N, Fukano H,
McCullough M, Jaber M, Barrett AW, Bain L, Speight Yamamoto S, Yano M, et al. (2001). Serum antiox-
PM, Porter SR (2002). Oral yeast carriage corre- idant micronutrient levels in oral lichen planus. J
lates with presence of oral epithelial dysplasia. Oral Oral Pathol Med. 30(5):264–7. doi:10.1034/j.1600-
Oncol. 38(4):391–3. doi:10.1016/S1368-8375(01)00079-3 0714.2001.300502.x PMID:11334461
PMID:12076705 Nagy KN, Sonkodi I, Szöke I, Nagy E, Newman HN
McCullough MJ, Farah CS (2008). The role of alcohol in (1998). The microflora associated with human oral
oral carcinogenesis with particular reference to alco- carcinomas. Oral Oncol. 34(4):304–8. doi:10.1016/
hol-containing mouthwashes. Aust Dent J. 53(4):302–5. S1368-8375(98)80012-2 PMID:9813727
doi:10.1111/j.1834-7819.2008.00070.x PMID:19133944 Nasher AT, Al-Hebshi NN, Al-Moayad EE, Suleiman AM
Mehanna H, Bryant TS, Babrah J, Louie K, Bryant JL, (2014). Viral infection and oral habits as risk factors for
Spruce RJ, et al. (2019). Human papillomavirus (HPV) oral squamous cell carcinoma in Yemen: a case-con-
vaccine effectiveness and potential herd immunity for trol study. Oral Surg Oral Med Oral Pathol Oral Radiol.
reducing oncogenic oropharyngeal HPV-16 prevalence 118(5):566–572.e1. doi:10.1016/j.oooo.2014.08.005
in the United Kingdom: a cross-sectional study. Clin PMID:25442493
Infect Dis. 69(8):1296–302. doi:10.1093/cid/ciy1081 NCI and CDC (2014). Smokeless tobacco and public
PMID:30590469 health: a global perspective (NIH Publication No.
Mello FW, Miguel AFP, Dutra KL, Porporatti AL, 14-7983). Bethesda (MD), USA: National Cancer
Warnakulasuriya S, Guerra ENS, et al. (2018a). Institute, National Institutes of Health and Centers for
Prevalence of oral potentially malignant disorders: a Disease Control and Prevention, U.S. Department of
systematic review and meta-analysis. J Oral Pathol Med. Health and Human Services. Available from: https://
47(7):633–40. doi:10.1111/jop.12726 PMID:29738071 stacks.cdc.gov/view/cdc/43373.
Mello FW, Scotti FM, Melo G, Warnakulasuriya S, Guerra Ndiaye C, Mena M, Alemany L, Arbyn M, Castellsagué
ENS, Rivero ERC (2018b). Maté consumption associa- X, Laporte L, et al. (2014). HPV DNA, E6/E7 mRNA,
tion with upper aerodigestive tract cancers: a system- and p16 INK4a detection in head and neck cancers: a
atic review and meta-analysis. Oral Oncol. 82:37–47. systematic review and meta-analysis. Lancet Oncol.
doi:10.1016/j.oraloncology.2018.04.023 PMID:29909900 15(12):1319–31. doi:10.1016/S1470-2045(14)70471-1
Meyer MS, Joshipura K, Giovannucci E, Michaud DS PMID:25439690
(2008). A review of the relationship between tooth Negri E, Boffetta P, Berthiller J, Castellsague X, Curado
loss, periodontal disease, and cancer. Cancer Causes MP, Dal Maso L, et al. (2009). Family history of
Control. 19(9):895–907. doi:10.1007/s10552-008-9163-4 cancer: pooled analysis in the International Head and
PMID:18478344 Neck Cancer Epidemiology Consortium. Int J Cancer.
Miranda J, Monteiro L, Albuquerque R, Pacheco JJ, 124(2):394–401. doi:10.1002/ijc.23848 PMID:18814262
Khan Z, Lopez-Lopez J, et al. (2017). Coffee is protec- Nomura T, Noma H, Shibahara T, Yokoyama A,
tive against oral and pharyngeal cancer: a system- Muramatusu T, Ohmori T (2000). Aldehyde dehydro-
atic review and meta-analysis. Med Oral Patol Oral genase 2 and glutathione S-transferase M 1 polymor-
Cir Bucal. 22(5):e554–61. doi:10.4317/medoral.21829 phisms in relation to the risk for oral cancer in Japanese
PMID:28809372 drinkers. Oral Oncol. 36(1):42–6. doi:10.1016/S1368-
8375(99)00048-2 PMID:10889918
202
Param K, Kanchan M, Manish P, Satish K (2018). 5(6):583–90. doi:10.1097/00001648-199411000-00005

Assessment of serum paraoxonase-1 enzyme activity, PMID:7841239
malondialdehyde and vitamin-C in oral premalignan- Pivovar A, Gonçalves Dos Santos ZFD, Torres-Pereira CC
cies. Int J Clin Biomed Res. 4(3):21–6. doi:10.31878/ (2017). Oral cancer screening for high-risk individ-
ijcbr.2018.43.05 uals in the primary healthcare setting using an active
Patel KR, Vajaria BN, Begum R, Shah FD, Patel JB, approach. J Oral Pathol Med. 46(9):786–91. doi:10.1111/
Shukla SN, et al. (2013). Association between p53 gene jop.12552 PMID:28122153
variants and oral cancer susceptibility in population Polesel J, Talamini R, La Vecchia C, Levi F, Barzan L,
from Gujarat, West India. Asian Pac J Cancer Prev. Serraino D, et al. (2008). Tobacco smoking and the
14(2):1093–100. doi:10.7314/APJCP.2013.14.2.1093 risk of upper aero-digestive tract cancers: a reanal-
PMID:23621193 ysis of case-control studies using spline models. Int
Pavia M, Pileggi C, Nobile CG, Angelillo IF (2006). J Cancer. 122(10):2398–402. doi:10.1002/ijc.23385
Association between fruit and vegetable consump- PMID:18224689
tion and oral cancer: a meta-analysis of observational Purdue MP, Hashibe M, Berthiller J, La Vecchia C, Dal
studies. Am J Clin Nutr. 83(5):1126–34. doi:10.1093/ Maso L, Herrero R, et al. (2009). Type of alcoholic
ajcn/83.5.1126 PMID:16685056 beverage and risk of head and neck cancer – a pooled
Perera M, Al-Hebshi NN, Speicher DJ, Perera I, Johnson analysis within the INHANCE Consortium. Am J
NW (2016). Emerging role of bacteria in oral carcino- Epidemiol. 169(2):132–4 2 . doi:10.1093/aje/ k w n306
genesis: a review with special reference to perio-patho- PM ID:19 06 4 6 4 4
genic bacteria. J Oral Microbiol. 8(1):32762. doi:10.3402/ Pushalkar S, Ji X, Li Y, Estilo C, Yegnanarayana R, Singh
jom.v8.32762 PMID:27677454 B, et al. (2012). Comparison of oral microbiota in
Petridou E, Zavras AI, Lefatzis D, Dessypris N, Laskaris tumor and non-tumor tissues of patients with oral
G, Dokianakis G, et al. (2002). The role of diet and squamous cell carcinoma. BMC Microbiol. 12(1):144.
specific micronutrients in the etiology of oral carci- doi:10.1186/1471-2180-12-144 PMID:22817758
noma. Cancer. 94(11):2981–8. doi:10.1002/cncr.10560 Quadri MF, Alharbi F, Bajonaid AM, Moafa IH, Sharwani
PMID:12115387 AA, Alamir AH (2015). Oral squamous cell carcinoma
Petti S, Masood M, Scully C (2013). The magnitude and associated risk factors in Jazan, Saudi Arabia: a
of tobacco smoking–betel quid chewing–alcohol hospital based case control study. Asian Pac J Cancer
drinking interaction effect on oral cancer in South-East Prev. 16(10):4335–8. doi:10.7314/APJCP.2015.16.10.4335
Asia. A meta-analysis of observational studies. PLoS PMID:26028095
One. 8(11):e78999. doi:10.1371/journal.pone.0078999 Radoï L, Menvielle G, Cyr D, Lapôtre-Ledoux B, Stücker
PMID:24260143 I, Luce D; ICARE Study Group (2015). Population
Petti S, Scully C (2006). Association between different attributable risks of oral cavity cancer to behavioral
alcoholic beverages and leukoplakia among non- to and medical risk factors in France: results of a large
moderate-drinking adults: a matched case-con- population-based case-control study, the ICARE study.
trol study. Eur J Cancer. 42(4):521–7. doi:10.1016/j. BMC Cancer. 15(1):827. doi:10.1186/s12885-015-1841-5
ejca.2005.07.038 PMID:16427777 PMID:26520570
Piemonte E, Lazos J, Belardinelli P, Secchi D, Brunotto Radoï L, Paget-Bailly S, Cyr D, Papadopoulos A, Guida
M, Lanfranchi-Tizeira H (2018). Oral cancer asso- F, Schmaus A, et al. (2013a). Tobacco smoking, alcohol
ciated with chronic mechanical irritation of the oral drinking and risk of oral cavity cancer by subsite:
mucosa. Med Oral Patol Oral Cir Bucal. 23(2):e151–60. results of a French population-based case-control study,
doi:10.4317/medoral.22017 PMID:29476673 the ICARE study. Eur J Cancer Prev. 22(3):268–76.
Piemonte ED, Lazos J (2018). Evidence for chronic doi:10.1097/CEJ.0b013e3283592cce PMID:22976386
mechanical irritation as a causative factor in oral Radoï L, Paget-Bailly S, Menvielle G, Cyr D, Schmaus
cancer: application of Bradford Hill criteria. Transl Res A, Carton M, et al. (2013b). Tea and coffee consump-
Oral Oncol. 3:1–10. doi:10.1177/2057178X18791034 tion and risk of oral cavity cancer: results of a large
Piemonte ED, Lazos JP, Brunotto M (2010). Relationship population-based case-control study, the ICARE
between chronic trauma of the oral mucosa, oral poten- study. Cancer Epidemiol. 37(3):284–9. doi:10.1016/j.
tially malignant disorders and oral cancer. J Oral Pathol canep.2013.02.001 PMID:23453554
Med. 39(7):513–7. doi:10.1111/j.1600-0714.2010.00901.x Rajkumar T, Sridhar H, Balaram P, Vaccarella S,
PMID:20456614 Gajalakshmi V, Nandakumar A, et al. (2003). Oral
Pintos J, Franco EL, Oliveira BV, Kowalski LP, cancer in Southern India: the influence of body size,
Curado MP, Dewar R (1994). Maté, coffee, and tea diet, infections and sexual practices. Eur J Cancer Prev.
consumption and risk of cancers of the upper aero- 12(2):135–43. doi:10.1097/00008469-200304000-00007
digestive tract in southern Brazil. Epidemiology. PMID:12671537
203
Ramaswamy G, Rao VR, Kumaraswamy SV, Anantha Bone Marrow Transplant. 56(5):1038–46. doi:10.1038/
N (1996). Serum vitamins’ status in oral leucopla- s41409-020-01147-z PMID:33235350
kias – a preliminary study. Eur J Cancer B Oral Schildt EB, Eriksson M, Hardell L, Magnuson A (1998).
Oncol. 32B(2):120–2. doi:10.1016/0964-1955(95)00076-3 Oral snuff, smoking habits and alcohol consumption in
PMID:8736174 relation to oral cancer in a Swedish case-control study.
Ranasinghe AW, Warnakulasuriya KA, Tennekoon GE, Int J Cancer. 77(3):341–6. doi:10.1002/(SICI)1097-
Seneviratna B (1983). Oral mucosal changes in iron 0215(19980729)77:3<341::AID-IJC6>3.0.CO;2-O
deficiency anemia in a Sri Lankan female popula- PMID:9663593
tion. Oral Surg Oral Med Oral Pathol. 55(1):29–32. Schiller JT, Lowy DR (2012). Understanding and learning
doi:10.1016/0030-4220(83)90302-X PMID:6572346 from the success of prophylactic human papilloma-
Rashidian H, Zendehdel K, Kamangar F, Malekzadeh R, virus vaccines. Nat Rev Microbiol. 10(10):681–92.
Haghdoost AA (2016). An ecological study of the asso- doi:10.1038/nrmicro2872 PMID:22961341
ciation between opiate use and incidence of cancers. Schlecht NF, Franco EL, Pintos J, Negassa A, Kowalski LP,
Addict Health. 8(4):252–60. PMID:28819556 Oliveira BV, et al. (1999). Interaction between tobacco
Razmpa E, Saedi B, Motiee-langroudi M, Garajei A, and alcohol consumption and the risk of cancers of the
Hoseinpor S, Motamedi MH (2014). Opium usage as upper aero-digestive tract in Brazil. Am J Epidemiol.
an etiologic factor of oral cavity cancer. J Craniofac 150(11):1129–37. doi:10.1093/oxfordjournals.aje.
Surg. 25(5):e505–7. doi:10.1097/SCS.0000000000001089 a009938 PMID:10588073
PMID:25148634 Schlecht NF, Masika M, Diaz A, Nucci-Sack A, Salandy
Ren JS, Freedman ND, Kamangar F, Dawsey SM, A, Pickering S, et al. (2019). Risk of oral human papil-
Hollenbeck AR, Schatzkin A, et al. (2010). Tea, coffee, lomavirus infection among sexually active female
carbonated soft drinks and upper gastrointestinal tract adolescents receiving the quadrivalent vaccine.
cancer risk in a large United States prospective cohort JAMA Netw Open. 2(10):e1914031. doi:10.1001/
study. Eur J Cancer. 46(10):1873–81. doi:10.1016/j. jamanetworkopen.2019.14031 PMID:31651968
ejca.2010.03.025 PMID:20395127 Schwartz SM, Daling JR, Doody DR, Wipf GC, Carter JJ,
Rezazadeh F, Haghighat S (2021). Serum vitamin Madeleine MM, et al. (1998). Oral cancer risk in rela-
profile in oral lichen planus patients in sout h- tion to sexual history and evidence of human papillo-
west of Iran. BioMed Res Int. 2021:8627435. mavirus infection. J Natl Cancer Inst. 90(21):1626–36.
doi:10.1155/2021/8627435 PMID:33708994 doi:10.1093/jnci/90.21.1626 PMID:9811312
Ribeiro KB, Levi JE, Pawlita M, Koifman S, Matos E, She Y, Nong X, Zhang M, Wang M (2017). Epstein-Barr
Eluf-Neto J, et al. (2011). Low human papillomavirus virus infection and oral squamous cell carcinoma risk: a
prevalence in head and neck cancer: results from two meta-analysis. PLoS One. 12(10):e0186860. doi:10.1371/
large case-control studies in high-incidence regions. journal.pone.0186860 PMID:29065191
Int J Epidemiol. 40(2):489–502. doi:10.1093/ije/dyq249 Shiu MN, Chen TH, Chang SH, Hahn LJ (2000). Risk
PMID:21224273 factors for leukoplakia and malignant transformation
Rosenquist K (2005). Risk factors in oral and oropharyn- to oral carcinoma: a leukoplakia cohort in Taiwan. Br
geal squamous cell carcinoma: a population-based J Cancer. 82(11):1871–4. doi:10.1054/bjoc.2000.1208
case-control study in southern Sweden. Swed Dent J PMID:10839305
Suppl. (179):1–66. PMID:16335030 Shukla K, Vun I, Lov I, Laparidis G, McCamley C,
Sankaranarayanan R, Duffy SW, Padmakumary G, Day Ariyawardana A (2019). Role of Candida infection in
NE, Padmanabhan TK (1989). Tobacco chewing, the malignant transformation of oral leukoplakia: a
alcohol and nasal snuff in cancer of the gingiva in systematic review of observational studies. Transl Res
Kerala, India. Br J Cancer. 60(4):638–43. doi:10.1038/ Oral Oncol. 4:1–10. doi:10.1177/2057178X19828229
bjc.1989.330 PMID:2803939 Shulman JD, Beach MM, Rivera-Hidalgo F (2004). The
Sankaranarayanan R, Mathew B, Jacob BJ, Thomas G, prevalence of oral mucosal lesions in U.S. adults:
Somanathan T, Pisani P, et al.; The Trivandrum Oral data from the Third National Health and Nutrition
Cancer Screening Study Group (2000). Early find- Examination Survey, 1988–1994. J Am Dent Assoc.
ings from a community-based, cluster-randomized, 135(9):1279–86. doi:10.14219/jada.archive.2004.0403
controlled oral cancer screening trial in Kerala, PMID:15493392
India. Cancer. 88(3):664–73. doi:10.1002/(SICI)1097- Siewchaisakul P, Wang ST, Peng SM, Sarakarn P, Chen LS,
0142(20000201)88:3<664::AID-CNCR25>3.0.CO;2-V Chen TH, et al. (2020). Effect of metabolic syndrome
PMID:10649262 on incidence of oral potentially malignant disorder:
Santarone S, Natale A, Angelini S, Papalinetti G, Vaddinelli a prospective cohort study in Taiwan. BMJ Open.
D, Di Bartolomeo A, et al. (2021). Secondary oral 10(10):e041971. doi:10.1136/bmjopen-2020-041971
cancer following hematopoietic cell transplantation. PMID:33055124
204
Sinusas K, Coroso JG, Sopher MD, Crabtree BF (1992). g i e . 75(12):759– 63. doi:10.1055/s-2007-997671
Smokeless tobacco use and oral pathology in a profes- PMID:9081282
sional baseball organization. J Fam Pract. 34(6):713–8. Tisch M, Maier H (1996). Squamous epithelial carci-
PMID:1593245 noma of the tongue after occupational exposure
Smith EM, Ritchie JM, Summersgill KF, Hoffman HT, to chromium VI compounds. Laryngorhinootologie.
Wang DH, Haugen TH, et al. (2004). Human papilloma- 75(8):455–8. doi:10.1055/s-2007-997614 PMID:8962606
virus in oral exfoliated cells and risk of head and neck Tomar SL, Winn DM, Swango PA, Giovino GA, Kleinman
cancer. J Natl Cancer Inst. 96(6):449–55. doi:10.1093/ DV (1997). Oral mucosal smokeless tobacco lesions
jnci/djh074 PMID:15026470 among adolescents in the United States. J Dent Res.
Stingone JA, Funkhouser WK, Weissler MC, Bell ME, 76(6):1277–86. doi:10.1177/00220345970760060701
Olshan AF (2013). Racial differences in the rela- PMID:9168861
tionship between tobacco, alcohol, and squamous Toporcov TN, Znaor A, Zhang ZF, Yu GP, Winn DM, Wei
cell carcinoma of the head and neck. Cancer Causes Q, et al. (2015). Risk factors for head and neck cancer
Control. 24(4):649–64. doi:10.1007/s10552-012-9999-5 in young adults: a pooled analysis in the INHANCE
PMID:22674225 consortium. Int J Epidemiol. 44(1):169–85. doi:10.1093/
Su CC, Lin YY, Chang TK, Chiang CT, Chung JA, Hsu ije/dyu255 PMID:25613428
YY, et al. (2010). Incidence of oral cancer in relation Tramacere I, Negri E, Bagnardi V, Garavello W, Rota
to nickel and arsenic concentrations in farm soils of M, Scotti L, et al. (2010). A meta-analysis of alcohol
patients’ residential areas in Taiwan. BMC Public Health. drinking and oral and pharyngeal cancers. Part 1:
10(1):67. doi:10.1186/1471-2458-10-67 PMID:20152030 overall results and dose-risk relation. Oral Oncol.
Subapriya R, Thangavelu A, Mathavan B, Ramachandran 46(7):497–503. doi:10.1016/j.oraloncology.2010.03.024
CR, Nagini S (2007). Assessment of risk factors for PMID:20444641
oral squamous cell carcinoma in Chidambaram, Trichopoulou A, Lagiou P (1997). Healthy tradi-
Southern India: a case-control study. Eur J Cancer Prev. tional Mediterranean diet: an expression of culture,
16(3):251–6. doi:10.1097/01.cej.0000228402.53106.9e history, and lifestyle. Nutr Rev. 55(11 Pt 1):383–9.
PMID:17415096 doi:10.1111/j.1753-4887.1997.tb01578.x PMID:9420448
Szymańska K, Hung RJ, Wünsch-Filho V, Eluf-Neto J, Tsai KY, Su CC, Chiang CT, Tseng YT, Lian IB (2017).
Curado MP, Koifman S, et al. (2011). Alcohol and Environmental heavy metal as a potential risk factor
tobacco, and the risk of cancers of the upper aerodi- for the progression of oral potentially malignant disor-
gestive tract in Latin America: a case-control study. ders in central Taiwan. Cancer Epidemiol. 47:118–24.
Cancer Causes Control. 22(7):1037–46. doi:10.1007/ doi:10.1016/j.canep.2017.02.003 PMID:28259083
s10552-011-9779-7 PMID:21607590 Tuovinen V, Väänänen M, Kullaa A, Karinpää A,
Takezaki T, Hirose K, Inoue M, Hamajima N, Kuroishi Markkanen H, Kumpusalo E (1992). Oral mucosal
T, Nakamura S, et al. (1996). Tobacco, alcohol and changes related to plasma ascorbic acid levels. Proc
dietary factors associated with the risk of oral cancer Finn Dent Soc. 88(3–4):117–22. PMID:1298961
among Japanese. Jpn J Cancer Res. 87(6):555–62. Turati F, Garavello W, Tramacere I, Pelucchi C, Galeone
doi:10.1111/j.1349-7006.1996.tb00259.x PMID:8766517 C, Bagnardi V, et al. (2013). A meta-analysis of alcohol
Thomas G, Hashibe M, Jacob BJ, Ramadas K, Mathew drinking and oral and pharyngeal cancers: results
B, Sankaranarayanan R, et al. (2003). Risk factors from subgroup analyses. Alcohol Alcohol. 48(1):107–18.
for multiple oral premalignant lesions. Int J Cancer. doi:10.1093/alcalc/ags100 PMID:22949102
107(2):285–91. doi:10.1002/ijc.11383 PMID:12949809 Vaccarezza GF, Antunes JL, Michaluart-Júnior P (2010).
Thomas SJ, Bain CJ, Battistutta D, Ness AR, Paissat D, Recurrent sores by ill-fitting dentures and intra-oral
Maclennan R (2007). Betel quid not containing tobacco squamous cell carcinoma in smokers. J Public Health
and oral cancer: a report on a case-control study in Dent. 70(1):52–7. doi:10.1111/j.1752-7325.2009.00143.x
Papua New Guinea and a meta-analysis of current PMID:19694933
evidence. Int J Cancer. 120(6):1318–23. doi:10.1002/ van der Waal I (2014). Oral potentially malignant disor-
ijc.22304 PMID:17163423 ders: is malignant transformation predictable and
Thomas SJ, Harris R, Ness AR, Taulo J, Maclennan preventable? Med Oral Patol Oral Cir Bucal. 19(4):e386–
R, Howes N, et al. (2008). Betel quid not containing 90. doi:10.4317/medoral.20205 PMID:24905952
tobacco and oral leukoplakia: a report on a cross-sec- van Leeuwen MT, Grulich AE, McDonald SP,
tional study in Papua New Guinea and a meta-anal- McCredie MR, Amin J, Stewart JH, et al. (2009).
ysis of current evidence. Int J Cancer. 123(8):1871–6. Immunosuppression and other risk factors for lip
doi:10.1002/ijc.23739 PMID:18688850 cancer after kidney transplantation. Cancer Epidemiol
Tisch M, Enderle G, Zöller J, Maier H (1996). Cancer of Biomarkers Prev. 18(2):561–9. doi:10.1158/1055-9965.
the oral cavity in machine workers. Laryngorhinootolo- EPI-08-0919 PMID:19190169
205
Varela-Lema L, Ruano-Ravina A, Juiz Crespo MA, Whitmore SE, Lamont RJ (2014). Oral bacteria and cancer.
Barros-Dios JM (2010). Tobacco consumption and PLoS Pathog. 10(3):e1003933. doi:10.1371/journal.
oral and pharyngeal cancer in a Spanish male popu- ppat.1003933 PMID:24676390
lation. Cancer Lett. 288(1):28–35. doi:10.1016/j. WHO (2019). Meeting of the Strategic Advisory Group
canlet.2009.06.015 PMID:19615813 of Experts on Immunization, October 2019: conclu-
Velly AM, Franco EL, Schlecht N, Pintos J, Kowalski sions and recommendations. Wkly Epidemiol Rec.
LP, Oliveira BV, et al. (1998). Relationship between 94(47):541–59. Available from: https://apps.who.int/
dental factors and risk of upper aerodigestive tract iris/handle/10665/329963.
cancer. Oral Oncol. 34(4):284–91. doi:10.1016/S1368- Winn DM, Blot WJ, Shy CM, Pickle LW, Toledo A, Fraumeni
8375(98)80009-2 PMID:9813724 JF Jr (1981). Snuff dipping and oral cancer among
Voltzke KJ, Lee YA, Zhang ZF, Zevallos JP, Yu GP, Winn women in the southern United States. N Engl J Med.
DM, et al. (2018). Racial differences in the relationship 304(13):745–9. doi:10.1056/NEJM198103263041301
between tobacco, alcohol, and the risk of head and neck PMID:7193288
cancer: pooled analysis of US studies in the INHANCE Woolley E, Magennis P, Shokar P, Lowe D, Edwards D,
Consortium. Cancer Causes Control. 29(7):619–30. Rogers SN (2006). The correlation between indices of
doi:10.1007/s10552-018-1026-z PMID:29761303 deprivation and health-related quality of life in patients
Warnakulasuriya S, Cronin-Fenton D, Jinot J, Kamangar with oral and oropharyngeal squamous cell carcinoma.
F, Malekzadeh R, Dar NA, et al.; IARC Monographs Vol Br J Oral Maxillofac Surg. 44(3):177–86. doi:10.1016/j.
126 group (2020). Carcinogenicity of opium consump- bjoms.2005.06.024 PMID:16105713
tion. Lancet Oncol. 21(11):1407–8. doi:10.1016/S1470- Worakhajit P, Fuangtharnthip P, Khovidhunkit SP,
2045(20)30611-2 PMID:33038940 Chiewwit P, Klongnoi B (2021). The relationship of
Warnakulasuriya S, Dietrich T, Bornstein MM, Casals tobacco, alcohol, and betel quid with the formation
Peidró E, Preshaw PM, Walter C, et al. (2010). Oral of oral potentially malignant disorders: a commu-
health risks of tobacco use and effects of cessation. Int nity-based study from northeastern Thailand. Int J
Dent J. 60(1):7–30. PMID:20361572 Environ Res Public Health. 18(16):8738. doi:10.3390/
WCRF (2016). The associations between food, nutrition ijerph18168738 PMID:34444487
and physical activity and the risk of mouth, pharynx and Wu Y-H, Wu YC, Chu FY, Cheng SJ, Sun A, Chen HM
larynx cancer. Systematic literature review. London, (2019). Significantly higher frequencies of hematinic
UK: World Cancer Research Fund. Available from: deficiencies and hyperhomocysteinemia in oral
https://www.wcrf.org/wp-content/uploads/2021/02/ precancer patients. J Formos Med Assoc. 118(9):1299–
mouth-pharynx-larynx-cancer-slr.pdf. 307. doi:10.1016/j.jfma.2019.05.016 PMID:31202497
WCRF (2018). Mouth, pharynx and larynx cancers. Diet, Wu YH, Yen CJ, Hsiao JR, Ou CY, Huang JS, Wong TY,
nutrition, physical activity and cancer: a global perspec- et al. (2016). A comprehensive analysis on the asso-
tive. London, UK: World Cancer Research Fund. ciation between tobacco-free betel quid and risk of
Available from: https://www.wcrf.org/dietandcancer/ head and neck cancer in Taiwanese men. PLoS One.
mouth-pharynx-larynx-cancers/. 11(10):e0164937. doi:10.1371/journal.pone.0164937
WCRF/AICR (2018). Diet, nutrition, physical activity PMID:27780233
and cancers of the mouth, pharynx and larynx. Wyss A, Hashibe M, Chuang SC, Lee YC, Zhang ZF, Yu
Continuous Update Project Expert Report. London, GP, et al. (2013). Cigarette, cigar, and pipe smoking and
UK and Arlington (VA), USA: World Cancer Research the risk of head and neck cancers: pooled analysis in
Fund and American Institute for Cancer Research. the International Head and Neck Cancer Epidemiology
Available from: https://www.wcrf.org/diet-activity- Consortium. Am J Epidemiol. 178(5):679–90.
a nd-ca ncer/g loba l-ca ncer-update-prog ra mme/ doi:10.1093/aje/kwt029 PMID:23817919
resources-and-toolkits/. Wyss AB, Hashibe M, Lee YA, Chuang SC, Muscat J,
Weitkunat R, Sanders E, Lee PN (2007). Meta-analysis of Chen C, et al. (2016). Smokeless tobacco use and
the relation between European and American smoke- the risk of head and neck cancer: pooled analysis
less tobacco and oral cancer. BMC Public Health. of US studies in the INHANCE consortium. Am J
7(1):334. doi:10.1186/1471-2458-7-334 PMID:18005437 Epidemiol. 184(10):703–16. doi:10.1093/aje/kw w075
Wen CP, Tsai MK, Chung WS, Hsu HL, Chang YC, PMID:27744388
Chan HT, et al. (2010). Cancer risks from betel quid Yang CM, Chen HC, Hou YY, Lee MC, Liou HH, Huang
chewing beyond oral cancer: a multiple-site carcin- SJ, et al. (2014). A high IL-4 production diplotype is
ogen when acting with smoking. Cancer Causes associated with an increased risk but better prognosis
Control. 21(9):1427–35. doi:10.1007/s10552-010-9570-1 of oral and pharyngeal carcinomas. Arch Oral Biol.
PMID:20458529 59(1):35–46. doi:10.1016/j.archoralbio.2013.09.010
PMID:24169152
206
Yang YH, Ho PS, Lu HM, Huang IY, Chen CH (2010). Zeng XT, Luo W, Geng PL, Guo Y, Niu YM, Leng WD
Comparing dose-response measurements of oral habits (2014). Association between the TP53 codon 72 poly-
on oral leukoplakia and oral submucous fibrosis from morphism and risk of oral squamous cell carcinoma
a community screening program. J Oral Pathol Med. in Asians: a meta-analysis. BMC Cancer. 14(1):469.
39(4):306–12. doi:10.1111/j.1600-0714.2009.00820.x doi:10.1186/1471-2407-14-469 PMID:24969046
PMID:20149061 Zeng XT, Luo W, Huang W, Wang Q, Guo Y, Leng
Yen AM, Chen SL, Chiu SY, Chen HH (2011). Association WD (2013b). Tooth loss and head and neck cancer:
between metabolic syndrome and oral pre-malig- a meta-analysis of observational studies. PLoS
nancy: a community- and population-based study One. 8(11):e79074. doi:10.1371/journal.pone.0079074
(KCIS No. 28). Oral Oncol. 47(7):625–30. doi:10.1016/j. PMID:24260154
oraloncology.2011.04.011 PMID:21592847 Zhang L, Liu Y, Zheng HJ, Zhang CP (2020). The oral
Yuan TH, Lian IeB, Tsai KY, Chang TK, Chiang CT, Su microbiota may have influence on oral cancer. Front Cell
CC, et al. (2011). Possible association between nickel Infect Microbiol. 9:476. doi:10.3389/fcimb.2019.00476
and chromium and oral cancer: a case-control study PMID:32010645
in central Taiwan. Sci Total Environ. 409(6):1046–52. Zhang Z-F, Morgenstern H, Spitz MR, Tashkin DP, Yu GP,
doi:10.1016/j.scitotenv.2010.11.038 PMID:21195455 Marshall JR, et al. (1999). Marijuana use and increased
Zarate AM, Don J, Secchi D, Carrica A, Galindez Costa risk of squamous cell carcinoma of the head and neck.
F, Panico R, et al. (2017). Study of the TP53 codon Cancer Epidemiol Biomarkers Prev. 8(12):1071–8.
72 polymorphism in oral cancer and oral poten- PMID:10613339
tially malignant disorders in Argentine patients. Zheng T, Holford T, Chen Y, Jiang P, Zhang B, Boyle P
Tumour Biol. 39(5):1010428317699113. doi:10.1016/j. (1997). Risk of tongue cancer associated with tobacco
scitotenv.2010.11.038 PMID:21195455 smoking and alcohol consumption: a case-control
Zaridze DG, Blettner M, Matiakin EG, Poljakov BP, Stich study. Oral Oncol. 33(2):82–5. doi:10.1016/S0964-
HF, Rosin MP, et al. (1986). The effect of nass use and 1955(96)00056-5 PMID:9231164
smoking on the risk of oral leukoplakia. Cancer Detect Zhou H, Wu W, Wang F, Qi H, Cheng Z (2018). Tea
Prev. 9(5–6):435–40. PMID:3779705 consumption is associated with decreased risk of oral
Zeng XT, Deng AP, Li C, Xia LY, Niu YM, Leng WD cancer: a comprehensive and dose-response meta-anal-
(2013a). Periodontal disease and risk of head and neck ysis based on 14 case-control studies (MOOSE
cancer: a meta-analysis of observational studies. PLoS compliant). Medicine (Baltimore). 97(51):e13611.
One. 8(10):e79017. doi:10.1371/journal.pone.0079017 doi:10.1097/MD.0000000000013611 PMID:30572470
PMID:24194957 Znaor A, Brennan P, Gajalakshmi V, Mathew A, Shanta V,
Zeng XT, Leng WD, Zhang C, Liu J, Cao SY, Huang W Varghese C, et al. (2003). Independent and combined
(2015). Meta-analysis on the association between effects of tobacco smoking, chewing and alcohol
toothbrushing and head and neck cancer. Oral Oncol. drinking on the risk of oral, pharyngeal and esopha-
51(5):446–51. doi:10.1016/j.oraloncology.2015.02.095 geal cancers in Indian men. Int J Cancer. 105(5):681–6.
PMID:25753558 doi:10.1002/ijc.11114 PMID:12740918
207
208
3. CESSATION OF SMOKELESS TOBACCO
AND/OR ARECA NUT USE
3.1 Product definition and and Mexico, and as shammah in the Arabian
description Peninsula. In Sweden and some other Nordic
countries, the use of snus, a particular type of
The term “smokeless tobacco” refers to a large moist snuff, is still prevalent (Siddiqi et al., 2020;
variety of commercially available or non-com- WHO FCTC and ICMR-NICPR, 2022).
mercially available products that contain tobacco Preparations of areca nut mixed with tobacco
as the principal constituent and that are used are widely available commercially, such as betel
either orally (chewing, sucking, placing in the quid and gutka. Areca nut may also be consumed
cheek or lip pouch, or drinking) or nasally, on its own, especially in South Asia in the form of
without combustion (IARC, 2007; Siddiqi et al., supari, paan masala, betel quid without tobacco,
2020). Areca nut is the seed of the fruit of the Areca binglang, or kili (IARC, 2004; Cruising Maldives,
catechu L. (Palmaceae) tree, a palm that is indig- 2016).
enous to South Asia (IARC, 2004). Smokeless Both SLT and areca nut have been classified as
tobacco and areca nut may be consumed sepa- carcinogenic to humans (Group 1) by the IARC
rately or combined (Mehrtash et al., 2017). Monographs programme (IARC, 2004, 2007,
Although in some publications the term 2012). Multiple carcinogens have been identified
“smokeless tobacco” may include products with in SLT, such as tobacco-specific N-nitrosamines,
tobacco and areca nut combined, this Handbook N-nitrosamino acids, volatile N-nitrosamines,
considers the following three product catego- and polycyclic aromatic hydrocarbons (IARC,
ries: (i) “smokeless tobacco”, defined as smoke- 2012; Hecht and Hatsukami, 2022). Areca nut
less tobacco not containing areca nut; (ii) “areca contains several alkaloids and tannins (polyphe-
nut without tobacco”; and (iii) “areca nut with nols). Arecoline, which has been classified as
tobacco” (Table 3.1). possibly carcinogenic to humans (Group 2B), is
Smokeless tobacco (SLT) is available as a the most abundant alkaloid and the key active
myriad of products. They vary substantially ingredient in areca nut (IARC, 2012, 2021).
in their names and their use in each region;
the greatest diversity is observed in South and
South-East Asia. For example, these products are
known as khaini, zarda, naswar, and gul in South-
East Asia, as chimó and rapé in South America,
as plug, snuff, and snus in the USA, Canada,
209
210

Table 3.1 Smokeless tobacco and areca nut products available in different regions
Product Alternative or colloquial names (if any) Major constituents Other features (mode of consumption,
name Location and processing and manufacturing)
Smokeless tobacco products (not containing areca nut)
Chimó WHO Region of the Americas (Venezuela, Colombia) Tobacco leaf, baking soda, brown sugar, Oral (sucked, held in mouth)
ashes from mamón tree Cottage industry or manufactured
commercially
Creamy Tobacco toothpaste Tobacco, clove oil, glycerin, spearmint, Oral (applied to teeth and gums)
snuff Commonly used in WHO South-East Asia Region (India) menthol, camphor Manufactured commercially
Dry snuff Scotch snuff, snuff (USA, Canada, Germany), taaba Tobacco (fire-cured or air-cured, fermented, Oral (sucked, held in mouth) or nasal
(Burkina Faso), snuif (South Africa), sneif (Botswana, powdered), flavourings Manufactured commercially
Lesotho, South Africa), azổ (Benin), simonte (Kalunga
community in Brazil), tapkeer, tapkir, bajjar (India)
Moist snuff Dip, spit tobacco (USA, Canada, Mexico) Tobacco (air-cured or fire-cured), Oral (sucked)
Shammah: el-shama, bajeli, haradi, sharaci, black shammah flavourings, inorganic salts, moisturizers, Manufactured commercially
(Yemen), al-shammah (Saudi Arabia), chemma (Algeria) slaked lime, ash, black pepper, oil
Toombak: saute, sute, ammari, saood Tobacco leaves (dried, fermented, ground, Oral (sucked, held in mouth) or nasal
Commonly used in WHO Eastern Mediterranean Region matured), sodium bicarbonate Cottage industry and custom-made
(Sudan), WHO African Region (Chad)
Dissolvable Dissolvables Tobacco, moisturizers, preservatives, Oral (sucked, held in mouth, dissolved)
tobacco Commonly used in WHO Region of the Americas (USA) flavourings Manufactured commercially
Tobacco- Gudaku Tobacco powder, molasses, red soil, lime, Oral (applied to teeth and gums, teeth
based Commonly used in WHO South-East Asia Region (India) water cleaning, held in mouth)
toothpaste Gul or gul manjan Tobacco (fire-cured, fermented, powdered), Manufactured commercially and
or tooth Commonly used in WHO South-East Asia Region (India, molasses, unknown ingredients custom-made
powder Bangladesh)
Mishri or masheri Tobacco (toasted on hot metal plate,
Commonly used in WHO South-East Asia Region (India) powdered)
Tapkeer, tapkir, bajjar
Commonly used in WHO South-East Asia Region (India)
Iqmik Blackbull, dediguss Tobacco (fire-cured), tree fungus ash or Oral (chewed)
Commonly used in WHO Region of the Americas (USA, other ash derived from wood or bush Custom-made
Alaska)
Khaini Chadha, sada, surti (Nepal and neighbouring parts of Tobacco leaves (coarsely cut, sun-dried, Oral (sucked, held in mouth)
India) fermented), slaked lime Manufactured commercially, cottage
Commonly used in WHO South-East Asia Region (India, industry, and custom-made
Bangladesh, Nepal, Bhutan)
Kiwam Qiwam, qimam, khiwam, kimam Paste of tobacco extract, spices (cardamom, Oral (chewed or held in mouth)
Commonly used in WHO South-East Asia Region, WHO saffron, aniseed), additives such as musk Manufactured commercially
Eastern Mediterranean Region (Pakistan)
Nass Naswar, niswar, nasway, nasvay Tobacco, ash, cotton or sesame oil, water, Oral (chewed, sucked, held in mouth)
Commonly used in WHO Eastern Mediterranean flavourings such as cardamom and menthol Cottage industry and custom-made
Region (Pakistan, Islamic Republic of Iran, Afghanistan,
United Arab Emirates), WHO African Region (South
Africa), WHO European Region (Armenia, Kazakhstan,
Kyrgyzstan, Uzbekistan, Poland, Slovakia)
Rapé Commonly used in WHO Region of the Americas (Brazil) Dried tobacco leaf, selected tree ashes, Nasal inhalation
flavourings such as tonka bean, clove, Cottage industry and custom-made
cinnamon powder, and camphor
Red tooth Lal dant manjan Fine red tobacco powder, herbs, flavourings; Oral (teeth brushing, cleaning)
powder Commonly used in WHO South-East Asia Region (India) in addition, ginger, pepper, and camphor Manufactured commercially
may be used
Snus Commonly used in Nordic countries and some other Tobacco, moisturizers, sodium carbonate, Oral (held in mouth)
European countries, WHO Region of the Americas (USA, salt, sweeteners, flavourings Manufactured commercially
Canada, Brazil), WHO African Region (South Africa)
Tobacco Sada pata, chadha Tobacco leaf Oral (chewed)
leaf Commonly used in WHO South-East Asia Region (India, Custom-made
Bangladesh, Myanmar, Bhutan)
Hsey or hsey wah (Myanmar) Dried raw tobacco leaves
Hsey me’ (Myanmar) Cured and roasted tobacco leaves
Hsey paung or hnut hsey (Myanmar) Tobacco leaves treated with alcohol and
honey
Tobacco Tuibur, hidakpha Tobacco smoke, water Oral (sipped or gargled)
water Commonly used in WHO South-East Asia Region (India) Cottage industry and custom-made
Hsey paung Myanmar Scented tobacco soaked in honey, lime juice,
yay or black and water
water
Zarda Dokta Tobacco, lime, vegetable dyes, aromatic Oral (chewed; sometimes chewed with
Commonly used in WHO South-East Asia Region (India, spices areca nut or silver flecks)
Oral cancer screening

Bangladesh, Myanmar, Nepal, Bhutan), WHO Eastern Manufactured commercially
Mediterranean Region (Yemen)
211
212

Chewing Loose leaf, chew, chaw, spit tobacco Tobacco leaf (air-cured), sugar, liquorice Oral (chewed or held in mouth)
tobacco Commonly used in WHO Region of the Americas (USA) Manufactured commercially
Plug, chew, chaw, spit tobacco Heavy-grade or cigar tobacco top leaves, Oral (chewed, sucked, held in mouth)
Commonly used in WHO Region of the Americas (USA, immersed in liquorice or sugar, and pressed Manufactured commercially
Canada) into a plug
Twist, chew, chaw, chewing tobacco Tobacco, tobacco leaf extract, sweetener, Oral (chewed, held in mouth)
Commonly used in WHO Region of the Americas (USA) flavourings
Paraky (rural Madagascar) Oral (chewed)
Manufactured mainly in cottage
industry
Hsey or hsey-ywet kyee (Myanmar) Raw and cured tobacco
Hsey hmwe (Myanmar) Other varieties of tobacco mixture with
added fragrances
Bush tobacco, pituri or mingkulpa (Indigenous people in Fresh or dry leaves of certain tobacco Oral (sucked)
Australia) species, mixed with burned wood ash and
chewed into a quid
Areca nut products without tobacco
Betel quid Southern China, Pacific Islands Areca nut (fresh, unripe) alone or with lime Oral (chewed)
without Hunan Province (China) Areca nut (dried, unripe) alone or with lime Cottage industry and custom-made:
tobacco South Asia Areca nut (cured, ripe) alone or with lime prepared by individual vendors for sale,
or assembled at home by individual
Taiwan (China), Hainan Island (China), Papua New Areca nut (fresh, unripe) with lime and
users
Guinea, Pacific islands betel leaves
Lao-hwa quid Areca nut (fresh, unripe) with lime and
Taiwan (China), Papua New Guinea betel inflorescence
Stem quid:
Taiwan (China) Areca nut (fresh, unripe) with lime and
betel stem
Guam (USA) Areca nut (fresh, unripe) with betel leaves
South Asia Areca nut (cured, ripe) with lime and betel
leaves
Paan or pan (South Asia) Areca nut (cured, ripe) with lime, an
additional source of catechins, flavourings,
betel leaves
Paan Pan masala Areca nut, slaked lime, catechu, flavourings, Oral (chewed)
masala Commonly used in WHO South-East Asia Region sweeteners Manufactured commercially and
cottage industry
Areca nut Supari (WHO South-East Asia Region, India), doma Areca nut Oral (chewed raw, fermented, or
khando (Bhutan), buah pinang (Indonesia), meeru ripened; held in mouth)
bileygan’du and heera panna (Maldives), pugua (Guam,
USA), binglang (China)
Federated States of Micronesia: bu (Yap), bua (Belau), poc
(Pohnpei), pu (Chuuk)
Buai, dak (Papua New Guinea), pinang (Malaysia), puwak
(Sri Lanka), gua (Bangladesh), mak (Thailand), kun-ywet
(Myanmar)
Kili Commonly used in Maldives Areca nut, betel, cloves, cardamom, sugar Oral
Cottage industry and custom-made:
produced by individual vendors for sale
in small homemade paper pouches
Areca nut products with tobacco
Betel Paan or pan (India), khilli pan (Bangladesh) Tobacco, areca nut, slaked lime (calcium Oral (chewed)
quid with Commonly used in WHO South-East Asia Region, WHO hydroxide), betel leaf, catechu (Acacia Cottage industry and custom-made:
tobacco Eastern Mediterranean Region, WHO Western Pacific catechu tree extract) prepared by individual vendors for sale,
Region or assembled at home by individual
users
Dohra Commonly used in WHO South-East Asia Region (India) Tobacco, areca nut, catechu, slaked lime, Oral (chewed)
peppermint, cardamom Custom-made: produced by individual
vendors for sale
Gutka Commonly used in WHO South-East Asia Region, WHO Tobacco (sun-dried, finely chopped), areca Oral (chewed)
Eastern Mediterranean Region nut, slaked lime, catechu, flavourings, Manufactured commercially and
sweeteners cottage industry
Mainpuri Kapoori Tobacco leaves (pieces), slaked lime, areca Oral (chewed or held in mouth)
Commonly used in WHO South-East Asia Region (Uttar nut, flavourings (camphor, cloves) Cottage industry and custom-made:
Pradesh, India) produced by individual vendors for sale
Mawa Kharra Crushed tobacco leaves (sun-dried), areca Oral (chewed)

Commonly used in WHO South-East Asia Region (India) nut (sun-cured), slaked lime Cottage industry and custom-made:
produced by individual vendors for sale
Tombol Commonly used in WHO Eastern Mediterranean Region Tobacco, areca nut, noura, slaked lime, Oral (chewed, held in mouth)
(Yemen) catechu, tombol leaf Custom-made
WHO, World Health Organization.
Compiled by the Working Group, with data from Atkinson et al. (1964); Ahluwalia and Duguid (1966); Gupta and Ray (2002); Gupta and Warnakulasuriya (2002); IARC (2004, 2012);
Lim (2012); Blecher et al. (2014); Moghbel et al. (2016); Novais (2017); Buente et al. (2020); Gunjal et al. (2020); Joo et al. (2020); Siddiqi et al. (2020); WHO (2021a, b, c); WHO FCTC and
ICMR-NICPR (2022).
213
3.2 Prevalence of consumption reported a slightly higher prevalence of SLT use

in girls (14.8%) than in boys (12.2%) (WHO,
3.2.1 WHO South-East Asia Region 2021b). Such averages hide wide variations, given
There are almost 266 million adult users of the cultural diversity of the region (Table 3.2).
SLT or areca nut with tobacco (184 million men In an extremely detailed global analysis of
and 83 million women) in the World Health the prevalence of “chewing tobacco” in 1990–
Organization (WHO) South-East Asia Region; it 2019, unlike the trend for tobacco smoking, no
is the WHO region with the highest prevalence significant decrease was noted in the trends of
of use of these products in adults (WHO, 2021a). prevalence of SLT use in male or female individ-
Estimates for all the countries in the WHO uals aged ≥ 15 years in countries in the WHO
South-East Asia Region are given in Table 3.2. South-East Asia Region: Bangladesh, Bhutan,
[Although several recent detailed publications India, Myanmar, Nepal, and Sri Lanka (GBD
are available on “smokeless tobacco” or “chewing 2019 Chewing Tobacco Collaborators, 2021).
tobacco” in the WHO South-East Asia Region, Some trends can also be interpolated from the
they have imprecise definitions of the products repeated WHO Global Adult Tobacco Survey
involved; also, the words “areca nut” or “betel (GATS), which now includes data on SLT but
quid” rarely appear. Therefore, it was not always not on use of areca nut products of any kind. For
possible to present quantitative information on instance, the GATS India reported a significant
the prevalence of use of the three important decrease (−17.4%) in the percentage of current
product categories, i.e. SLT alone, areca nut SLT users between 2009–2010 (25.9%) and 2016–
without tobacco, and areca nut with tobacco.] 2017 (21.4%) (TISS and MOHFW, 2017).
Most of the countries in the region have In the WHO South-East Asia Region, a
reported a high overall prevalence (≥ 5%) of SLT common way of using tobacco is as an ingredient
use, ranging from 15.8% in Sri Lanka to 27.5% in in betel quid (i.e. areca nut with tobacco) (see
Bangladesh, with a few exceptions, such as the Section 3.1 and Table 3.1). Use of betel quid is an
Democratic People’s Republic of Korea (0.0%) ancient practice; tobacco was added beginning in
and Thailand (2.1%). The prevalence of SLT use about 1600, and this is now done in many parts of
is generally high in both men and women in South-East Asia, such as India, Bhutan, Myanmar,
most of the countries (WHO, 2021b). However, Nepal, and Sri Lanka (NCI and CDC, 2014). The
in several countries (e.g. Bangladesh, Indonesia, largest variety of SLT and areca nut products are
and Thailand), the prevalence of SLT use is available in India, such as khaini, gutka, zarda,
slightly higher in women than in men (WHO, gul, gudaku, mishri, tobacco water, and snuff, to
2017, 2021b). Similar to the situation for adults, name a few. The GATS-2 reported the highest
the WHO South-East Asia Region is the WHO prevalence of use for khaini (11.2%), followed by
region with the highest prevalence of SLT use in gutka (6.8%), betel quid with tobacco (5.8%), and
young people, with 4.2 million users (2.7 million oral tobacco (gul, mishri, gudaku) (3.8%) (TISS
boys and 1.5 million girls). Nepal has the highest and MOHFW, 2017). Products such as gutka,
reported prevalence of SLT use in adolescents khaini, and paan masala have been manufac-
(16.2%), followed by Timor-Leste (13.9%), Bhutan tured commercially since 1975 (NCI and CDC,
(12.5%), Maldives (6.2%), and Myanmar (5.7%). 2014). Khaini and gutka are also commonly used
The prevalence of SLT use was higher in boys in in Bangladesh (known as khoinee), Nepal, and
all the countries, ranging from 1.4% in Indonesia Sri Lanka. In addition to chewable products, the
to 19.7% in Nepal, except in Timor-Leste, which above-mentioned SLT and areca nut with tobacco
products administered through oral application,
214
Table 3.2 Prevalence of smokeless tobacco and areca nut use in adults and adolescents in the WHO South-East Asia Region
Country Product type and/or most Prevalence of use (%) Reference

popular names
Bangladesh Sada pata a, zarda a, gul a, SLT: WHO (2021b)
khoineea, gutka c, guab Adults:
Overall: 27.5
Men: 26.9
Women: 28.1
Youth: WHO (2021b)
Overall: 4.5
Boys: 5.9
Girls: 2.0
AN: 31 Flora et al. (2012)
Three quarters of users chewed BQ with tobacco
Bhutan BQ (usually with tobaccoc, SLT: WHO (2021b)
AN (called doma khando), Adults:
khainia Overall: 19.7
Men: 26.5
Women: 11.0
Youth: WHO (2021b)
Overall: 12.5
Boys: 17.0
Girls: 8.1
India Khainia, BQ (with and without SLT: TISS and
tobacco)b,c, gutka c, suparib, Adults: MOHFW (2017);
mishria, gul a, gudaku a Overall: 21.4 (199.4 million) Singh et al.
Men: 29.6 (2021); WHO
Women: 12.8 (2021b)
AN with tobacco: 14.2 (95% CI, 13.5–14.9)
Various products containing SLT alone or AN with tobacco:
Khaini: 11.2
Gutka: 6.8
BQ with tobacco: 5.8

Oral tobacco (gul, mishri, gudaku): 3.8
Paan masala with tobacco: 2.8
215
216

popular names
India SLT: MOHFW and
(cont.) Youth: IIPS (2019)
Overall: 4.1
Boys: 4.6
Girls: 3.4
AN: Singh et al.
~23.9 (95% CI, 23.1–24.8) (223.79 million adults) (2021)
National prevalence of use of plain AN products; lowest and highest prevalence among
states
Average % (statewise variation %):
BQ without tobacco: 8.7 (0.3–64.9)
Paan masala without tobacco: 4.8 (0.2–11.5)
AN alone without tobacco: 8.0 (0.2–22.6)
Tribal/Indigenous people are at high risk. Of 2186 tribal households in South India, 47.6% Sadath et al.
reported daily use of BQ (with or without tobacco) (2022)
Indonesia Buah pinang b, zarda a SLT: WHO (2017)
Men: 3.9
Women: 4.8
SLT: WHO (2021b)
Youth:
Overall: 1.0
Boys: 1.4
Girls: 0.7
AN without tobacco: Lee et al. (2011)
Women: 15.0
Men: 1.6
AN with tobacco:
Women: 31.7
Men: 10.4
Maldives Chewing tobaccoa, snuffa, dipa, SLT: WHO (2020a,
suparib, meeru bileygan’dub, Youth: 2021b)
heera pannab Overall: 6.2
Boys: 9.2
Girls: 2.9
SLT: WHO (2021b)
Men: 8.5
Women: 4.2
popular names
Myanmar Hsey or hsey-ywet kyee, hsey SLT: WHO (2018,
or hsey wah, hsey me’, hsey Youth: 2021b, c)
paung or hnut hsey, hsey paung Overall: 5.7
ya or black water, hsey hmwea, Boys: 11.0
kun-yab Girls: 1.5
Also, imported commercial SLT: WHO (2021b)
products Men: 58.9
BQ with tobaccoc, such as Women: 18.2
tobacco leaf, hnut hsey, hsey AN with tobacco: Papke et al.
paung, chewing tobacco leaf, 84% of respondents in a survey in Yangon (2020)
kun-ywet b
Nepal Khainia, gutka c, zarda a, paan Lifetime BQ (with tobacco) chewing: Lee et al. (2011)
masalab, snuffa, gul a, BQ with Men: 43.6
tobaccoc Women: 34.9
SLT: WHO (2021b)
Adults:
Overall: 18.3
Men: 33.3
Women: 4.9
SLT: WHO (2021b)
Youth:
Overall: 16.2
Boys: 19.7
Girls: 12.9
Sri Lanka BQ with tobaccoc, paan SLT: WHO (2021b)
masalab, mawa a, red tooth Adults:
powdera, khainia, tobacco Overall: 15.8
powdera, zarda a, gutka c, Men: 26
puwak b Women: 5.3
SLT: WHO (2021b)

Youth:
Overall: 2.4
Boys: 4.2
Girls: 0.5
217
218

popular names
Sri Lanka AN without tobacco: Lee et al. (2011)
(cont.) Men: 11.6
Women: 10.4
AN with tobacco:
Men: 6.4
Women: 3.2
AN with or without tobacco: Amarasinghe
Varies by ethnicity and geography; in one province in 1029 subjects (64.6% Sinhalese, et al. (2018)
34.9% Tamil, 0.5% other) aged > 30 years, prevalence of daily BQ chewing was 53.8%:
15.7% without tobacco and 47.4% with tobacco
Thailand Zarda a, mak b SLT: WHO (2021b)
Adults:
Overall: 2.1
Men: 1.5
Women: 2.7
SLT: WHO (2021b)
Youth:
Overall: 2.7
Boys: 4.1
Girls: 1.3
Timor-Leste Buab SLT: WHO (2021b)
Youth:
Overall: 13.9
Boys: 12.2
Girls: 14.8
SLT: WHO (2021b)
Adults:
Men: 20.9
Women: 0.2
AN, areca nut; BQ, betel quid; CI, confidence interval; SLT, smokeless tobacco; WHO, World Health Organization.
a SLT alone.
b AN alone (without tobacco).
c AN with tobacco.
Compiled by the Working Group.

such as gul, gudaku, and mishri, are also widely across the region) or as a component of betel quid.
prevalent in Bangladesh and Nepal. Areca nut chewing is a very ancient custom in
Consumption of areca nut is deeply the Philippines, from where it gradually spread
embedded in the social and cultural history across the Western Pacific islands, as planting
of the entire WHO South-East Asia Region. of the Areca catechu palm increased (NCI and
The Areca catechu palm tree is indigenous CDC, 2014).
to the Malay Peninsula and Sri Lanka, and A significant geographical variation is noted
cultivation has been widespread across South- both within and among the countries in this
East and South Asia for millennia (Gupta and region; in and close to continental Asia, the
Warnakulasuriya, 2002). Areca nut or its prepa- habits overlap with those in the WHO South-
rations without tobacco are known by various East Asia Region, whereas further east, they tend
colloquial names across the region, such as doma to mimic the habits of Chinese origin. Both the
khando in Bhutan, supari in India and Maldives, nature of the habits and the subpopulations in
buah pinang in Indonesia, meeru bileyn’d and which particular constituents of a betel quid are
heera panna in Maldives, and bua in Timor- favoured vary widely, and these are not always
Leste (Table 3.1). Areca nut is the primary adequately described in the literature. Also, in
component of betel quid, which may also be the WHO and Global Burden of Disease anal-
consumed without tobacco. The GATS-2 India yses conducted for these subpopulations, SLT use
reported the prevalence of the various plain is frequently referred to as the sole habit distin-
areca nut products: betel quid (8.7%), areca nut guished from smoked tobacco use, with no or
(8%), and paan masala (4.8%) (Singh et al., 2021). rare mentions of areca nut (Siddiqi et al., 2020;
The multicountry Asian Betel-Quid Consortium GBD 2019 Chewing Tobacco Collaborators,
study, in 2009–2010, reported a high prevalence 2021). As an example of the cultural varia-
of chewing betel quid (without tobacco) in the tions, in Taiwan (China) and Palau, unripe nuts
adult population in Indonesia (15% in women are used in the betel quid, whereas in Guam
and only 1.6% in men) and Sri Lanka (11.6% in (USA), white immature or red mature nuts are
men and 10.4% in women) (Lee et al., 2011). The preferred. Unwrapped quid is preferred in Papua
prevalence of use of common SLT and areca nut New Guinea and the Solomon Islands, whereas
with tobacco products (paan and gutka) was wrapped betel quid (in betel leaf) is consumed
recently reviewed (Niaz et al., 2017). in Cambodia, Palau, and the Federated States of
In summary, the WHO South-East Asia Micronesia. Also, the use of tobacco with areca
Region has the highest prevalence of SLT and nut or in a betel quid is not seen in all cultures
areca nut use among all WHO regions, and a in the WHO Western Pacific Region. The multi-
large variety of both SLT and areca nut products country Asian Betel-Quid Consortium study,
are consumed in this region. in 2009–2010, reported a prevalence of chewing
betel quid (without tobacco) ranging from 3.6%
3.2.2 WHO Western Pacific Region in Malaysia to 23.9% in China in men and from
1.8% in China to 17.5% in Malaysia in women
Areca nut or betel quid with tobacco are the (Lee et al., 2011). Similarly, users in island coun-
main products consumed in the WHO Western tries of Melanesia are unlikely to add tobacco to
Pacific Region. Chewing of areca nut is deeply the quid. Certain specific subpopulations in a
embedded in the social and cultural history of few countries have a higher prevalence of use of
many parts of the region; it may be consumed areca nut and SLT products, such as South Asian
on its own (known by various colloquial names immigrants in Australia, Fiji, and Singapore, and
219
Indigenous people in Australia (Kuek et al., 1990; Cambodia to 42.5% in Kiribati, except in three
Nambiar et al., 2020; Greenhalgh et al., 2022). countries that reported a relatively higher prev-
There are about 13.3 million users of SLT alence of SLT use in girls – Palau (16.8%), Papua
(11 million male and 2.3 million female) in the New Guinea (13.6%), and Tuvalu (3.3%) – than in
WHO Western Pacific Region; it is the WHO boys (WHO, 2021b).
region with the lowest average prevalence of In the extremely detailed global analysis of
SLT use in adults (0.9% overall, 1.4% in men, the prevalence of “chewing tobacco” in 1990–
and 0.3% in women) (WHO, 2021a). The WHO 2019, unlike the trend for tobacco smoking, no
Western Pacific Region is socially, culturally, significant decrease was noted in the trends of
economically, politically, and ethnically diverse, prevalence of SLT use in male or female individ-
containing both the world’s most populous uals aged ≥ 15 years in countries in the WHO
country, China, and the smallest territory in the Western Pacific Region: Cambodia, the Marshall
world, Pitcairn Island (NCI and CDC, 2014). The Islands, and Palau (GBD 2019 Chewing Tobacco
prevalence of SLT use varies widely, ranging from Collaborators, 2021). Increases in the prevalence
0.1% in women in China to 48.8% in women in have been reported in specific communities,
Palau (WHO, 2020b, 2021b). such as South Asian immigrants in Australia
There are limited robust longitudinal epide- and non-Chamorros in Guam (USA), whereas
miological studies on the prevalence of use decreases have been seen in Indigenous people
of these products, although estimates from in Australia, and in a few other locations, such as
many countries in this region are presented in Papua New Guinea, Singapore, Taiwan (China),
Table 3.3. Based on the available information, 4 and Viet Nam.
countries in the WHO Western Pacific Region In summary, although the WHO Western
have reported a high overall prevalence (≥ 5%) Pacific Region has reported the lowest average
of SLT use; the prevalence was highest in Palau prevalence of SLT use of all WHO regions, the
(44.4%), followed by the Marshall Islands (21.6%), prevalence of consumption of areca nut products
the Federated States of Micronesia (11.4%), is high and this practice is spreading further
and Malaysia (10.9%) (WHO, 2020b, 2021b). across the region.
The prevalence of SLT use is generally higher
in men in most of the countries in the region 3.2.3 WHO European Region
(WHO, 2021b). However, countries such as Palau
(48.8%), Cambodia (8.6%), and the Lao People’s In recent years, mass migration patterns and
Democratic Republic (8.6%) have a significantly commercial integration have affected the histor-
higher prevalence of SLT use in women than in ical regional prevalence of use of SLT products,
men (WHO, 2020b, 2021b). The WHO Western which are now widely available in the WHO
Pacific Region is the WHO region with the European Region (IARC, 2007; NCI and CDC,
lowest prevalence of SLT use in adolescents (aged 2014; WHO, 2017, 2019).
13–15 years), with 0.9 million users (0.6 million Table 3.4 provides data for countries for
boys and 0.3 million girls), but the prevalence which the estimated prevalence of SLT use was
of use is significantly high in Kiribati (38.6%), ≥ 2% in adults. Overall, the prevalence of SLT
the Federated States of Micronesia (16.0%), the use is low in the WHO European Region, with
Marshall Islands (14.9%), Palau (14.7%), and diverse geographical and subregional trends that
Papua New Guinea (12.2%) (WHO, 2021a, b). are greatly influenced by cultural and migration
The prevalence of SLT use was higher in boys patterns.
in most of the countries, ranging from 1.3% in
220
Table 3.3 Prevalence of smokeless tobacco and areca nut use in adults and adolescents in the WHO Western Pacific Region
Country or territory Product type and/or most popular Prevalence of use (%) Trends of prevalence
names Reference
Australia Both tobacco and AN productsa,b,c SLT: Increasing in immigrants
(South Asian immigrants) Overall: 0.4 WHO (2021b)
Men: 0.6
Women: 0.3
No national data on AN products
Bush tobacco, pituri or mingkulpa Chewing tobacco prevalence in Indigenous people in the Northern Decreasing in Indigenous
(Indigenous people) Territory (central Australia) in 1986–1987: people
Women: 61 Greenhalgh et al. (2022)
Men: 20
Cambodia AN with tobaccoc AN with tobacco: Decreased slightly
Women: 12.8 Chher et al. (2018); Gunjal
Men: 1.7 et al. (2020); WHO (2021b)
SLT:
Overall: 4.9
Men: 0.8
Women: 8.6
China AN with or without quidb, binglang b AN prevalence in 11 046 individuals in Xiangtan City, Hunan Tang et al. (1997); Lee et al.
Province: (2011); WHO (2021b)
Overall: 1.2
Men: 0.6
Women: 0.6
AN without tobacco:
Men: 23.9
Women: 1.9
SLT:
Overall: 0.9
Men: 1.6
Women: 0.1
Cook Islands SLT, AN SLT: AN: although use is
Overall: 3 spreading rapidly, no data are

Boys: 3.8 available
Girls: 2.4 WHO (2021b)
Fiji Paan masalab and other imported SLT: 14.2 In Fijians of Indian descent
packaged ingredients (South Asians) AN or paan masala: 20 in Suva aged ≥ 18 yr
Nambiar et al. (2020)
Guam (USA) AN with or without tobaccoc, puguab Adults (AN with tobacco): 46 AN: increased (in non-
Youth (pugua): 48 Chamorros)
AN (5-yr prevalence): 11 Paulino et al. (2017a)
221
222

names Reference
Kiribati SLT: WHO (2021b)
Adults
Overall: 4.2
Men: 7.6
Women: 1.4
Youth:
Overall: 38.6
Boys: 42.5
Girls: 35.3
Lao People’s BQb,c, AN b SLT: WHO (2021b)
Democratic Republic Overall: 4.3
Men: 0.5
Women: 8.6
Malaysia BQ with or without tobaccob,c, SLT: Lee et al. (2011);
pinang b Adults: WHO (2020b, 2021b)
Overall: 10.9
Men: 20.4
Women: 0.8
Youth:
Overall: 6.3
Boys: 8.2
Girls: 4.3
AN with tobacco:
Women: 12.0
Men: 6.2
AN without tobacco:
Women: 17.5
Men: 3.6
Marshall Islands “Chewing tobacco”: Increasing
Men: 10.36 GBD 2019 Chewing Tobacco
Women: 4.06 Collaborators (2021); WHO
SLT: (2021b)
Adults:
Overall: 21.6
Youth:
Overall: 14.9
Boys: 18.9
Girls: 11.8
names Reference
Micronesia Bu, bua, poc, pub, BQb,c AN: Oakley et al. (2005); Paulino
(Federated States of) School students: 63.4 et al. (2017b); WHO (2021b)
In families: 42 (from 3 in the Marshall Islands to 94 in Yap)
AN with tobacco: 84
SLT:
Adults:
Overall: 11.4
Men: 22.4
Women: 3.0
Youth:
Overall: 16.0
Boys: 20.0
Girls: 12.7
Mongolia SLT: WHO (2021b)
Overall: 8.2
Boys: 11.8
Girls: 4.5
Palau BQ with or without tobacco b AN without tobacco: Ysaol et al. (1996);
Men: 70 WHO (2020b, 2021b)
Women: 80
AN with tobacco: 80
SLT:
Adults:
Overall: 44.4
Men: 40.2
Women: 48.8
Youth:
Overall: 14.7
Boys: 12.2
Girls: 16.8

Papua New Guinea Buaib, dak b Chewing tobacco: Decrease (slight)
Men: 40 WHO (2021b); GBD
Women: 18 2019 Chewing Tobacco
SLT: Collaborators (2021)
Youth:
Overall: 12.2
Boys: 10.9
Girls: 13.6
223
224

names Reference
Singapore Paanb, makan sirihb AN: 6.4 Decreased (AN); more
common in Indian
community
Kuek et al. (1990); Lim and
Pakiam (2020)
Solomon Islands AN b AN: 94 in a sample of 400 people aged 15–24 yr Increased
Quinn et al. (2017); Moore
(2020)
Taiwan (China) BQ with and without tobaccob,c, AN without tobacco (in the multicountry ABC study): Decreased
binglang b Men: 10.7 Lee et al. (2011); Tsou et al.
Women: 2.5 (2022)
AN: 0.3 in 429 108 participants from the Senior Citizen Health
Examination in Taiwan (China) over 10 yr (2001–2010)
Tonga SLT: GBD 2019 Chewing Tobacco
Men: 5 Collaborators (2021)
Women: 2
Vanuatu SLT: WHO (2021b)
Overall: 5.2
Boys: 5.9
Girls: 4.6
Viet Nam AN b Women: 6.7 (in Ho Chi Minh City) Decreased
Reichart and Nguyen (2008);
Gunjal et al. (2020)
ABC, Asian Betel-Quid Consortium; AN, areca nut; BQ, betel quid; SLT, smokeless tobacco; WHO, World Health Organization; yr, year or years.
a SLT alone.
b AN alone (without tobacco).
c AN with tobacco.

Table 3.4 Countries with high prevalence of smokeless tobacco and areca nut use in adults in the WHO European Regiona
Country or population Product name or colloquial Prevalence of use (%) Reference

name
Men Women Overall
Czechia Snuff and chewing tobaccob 5.9 2.5 4.2 NCI and CDC (2014); WHO (2021b)
Denmark Snus b 4.0 1.0 3.0 Siddiqi et al. (2020); WHO (2021b)
Estonia Not reported 9.2 2.3 5.1 WHO (2021b)
Finland Snus b 9.2 1.0 5.2 Siddiqi et al. (2020); WHO (2021b)
Germany Dry snuff b 3.4 3.4 2.0 Agaku et al. (2014); NCI and CDC (2014)
Iceland Snus b 8.7 3.5 6.6 Siddiqi et al. (2020); WHO (2021b)
Kyrgyzstan Naswar b 10.1 0.1 5.2 Siddiqi et al. (2020); WHO (2021b)
Malta Chewing tobaccob 5.5 1.5 3.5 Agaku et al. (2014); NCI and CDC (2014)
Norway Snus b 25.0 10.0 18.0 Siddiqi et al. (2020); WHO (2021b)
Portugal Not reported 4.4 1.1 2.7 Agaku et al. (2014)
Slovenia Not reported 3.1 1.2 2.2 WHO (2021b)
Spain Not reported 2.1 2.9 2.5 Leon et al. (2016)
Sweden Snus b 22.0 6.0 14.0 Siddiqi et al. (2020); WHO (2020b)
Switzerland Snuff and chewing tobaccob 4.2 1.2 2.7 NCI and CDC (2014); WHO (2017)
South Asian immigrants in the Paanc,d, gutkad, zardad, khainib, 7.0 6.0 7.0 ASH (2019)
United Kingdom naswar b
Uzbekistan Naswar b 19.8 0.4 9.9 Siddiqi et al. (2020); WHO (2021b)
WHO European Region 1.9 0.4 1.1 WHO (2021a)
a Countries with a prevalence of smokeless tobacco and areca nut use of ≥ 2% are included in the table; countries with a prevalence of < 2% (Armenia, Austria, France, Hungary, Ireland,
Italy, Kazakhstan, Poland, Slovakia, Turkmenistan, and the United Kingdom) have been excluded.
b Smokeless tobacco alone.
c Areca nut without tobacco.
d Areca nut with tobacco.

225
Population-specific studies describing the commercial and mixed-use preparations, or

patterns and prevalence of SLT use were not avail- their variation in terms of natural and chemical
able for several countries in the WHO European compositions (IARC, 2007; NCI and CDC, 2014;
Region in which isolated SLT use had previously WHO, 2017, 2019). Table 3.4 shows a limited vari-
been reported (Leon et al., 2016). However, 34 of ation in terms of the products and their use in
53 countries (64.1%) presented data on SLT use in the WHO European Region. Regulations for the
adults; the regional average prevalence was 1.1%, consumption of SLT vary widely within coun-
with a higher prevalence in men (1.9%) than in tries in this region (WHO, 2017); however, in the
women (0.4%) (WHO, 2021a). Prevalence of SLT European Union (EU), SLT is regulated under
use was high in Estonia (5.1%), Finland (5.2%), the scope of the EU Tobacco Products Direc-
Iceland (6.6%), Kyrgyzstan (5.2%), Norway (18%), tive 2014/40/EU (European Parliament, 2014),
Sweden (14%), and Uzbekistan (9.9%) and in which banned all tobacco products for oral use.
South Asian immigrants in the United Kingdom Although most SLT products were banned by the
(WHO, 2020b, 2021b). Four of these countries European Council Directive in 1989, in western
exceeded the global average prevalence of SLT Europe the use of snus, a particular type of moist
use (6%) (WHO, 2021a). In the countries where snuff (see Section 3.1), is still prevalent among
the practice is highly prevalent, hotspots of high Scandinavian people, living mostly in Norway
prevalence of SLT use by men are observed in and Sweden (which are exempted from the ban)
subregions, including the Nordic countries and as well as in other Nordic countries, such as
in populations in central Asia (Ansara et al., Denmark, Finland, and Iceland (Council of the
2013; WHO, 2020b, 2021b). European Communities, 1989; IARC, 2007; Leon
The WHO European Region is the WHO et al., 2016). Other SLT products such as chewing
region with the second-lowest prevalence of SLT tobacco and dry snuff are also allowed for sale
use in adolescents (aged 13–15 years), after the and marketing in the WHO European Region
WHO Western Pacific Region (WHO, 2021b). (Leon et al., 2016). Originally from India, gutka
Based on data from 12 countries, the prevalence and zarda (see Section 3.1) are the most consumed
of SLT use in adolescents was 1.5% (1.8% in boys products in the United Kingdom, where about
and 1.1% in girls) (WHO, 2021a). The lowest prev- 75% of Asian immigrants had already consumed
alence of SLT use in adolescents was observed in them. Similarly, areca nut products are also often
Belarus, Kazakhstan, and San Marino (0.6%), consumed within immigrant communities from
and the highest prevalence was observed in Pakistan and Bangladesh, among others, living
Poland (5.6%), followed by Latvia (5.3%), Czechia in other parts of the WHO European Region
(4.7%), and Georgia (4.4%) (WHO Regional (IARC, 2004; Lechner et al., 2019; Siddiqi et al.,
Office for Europe, 2020). These hotspots of high 2020).
prevalence of SLT use by adolescents, such as In summary, a relatively small range of SLT
Latvia, may be due to the geographical prox- products is currently consumed in nearly half
imity to Sweden, where the prevalence of SLT of the countries in the WHO European Region,
use is one of the highest among countries in the with large regional and cultural variations.
WHO European Region (Leon et al., 2016). In
the United Kingdom, evidence about SLT use in 3.2.4 WHO Region of the Americas
adolescents is limited (WHO Regional Office for
Europe, 2020). Despite the heritage of SLT as an early
Few specific data are available about the American product (Shafey et al., 2009), SLT
spectrum of products used, which encompass use is not heavily culturally embedded in
226
Table 3.5 Countries with high prevalence of smokeless tobacco use in adults in the WHO Region
of the Americasa
Country Product name or Prevalence of use (%)b Reference

colloquial name
Men Women Overall
Haiti Not reported N/A 3.1 N/A WHO (2021b)
Paraguay Not reported 3 1.6 2.3 WHO (2021b)
USA Snuff b, snus b, iqmik b, plugb 6.2 0.6 3.3 Siddiqi et al. (2020);
WHO (2021b)
Venezuela Chimó b 6.2 0.9 3.5 Siddiqi et al. (2020);
WHO (2021b)
WHO Region of the Americas 2.5 0.3 1.4 WHO (2021a)
N/A, not available; WHO, World Health Organization.
a Countries with a prevalence of smokeless tobacco use of ≥ 2% are included in the table; countries with a prevalence of < 2% (e.g. Argentina,
Barbados, Canada, Dominican Republic, and Grenada) have been excluded.

contemporary societies in the WHO Region in girls) (WHO, 2021a). A total of 27 countries
of the Americas, and only limited data are in the WHO Region of the Americas reported
available about the prevalence of SLT use in SLT use in adolescents, of which Saint Vincent
this region (IARC, 2007; NCI and CDC, 2014). and the Grenadines (6.3%), Venezuela (5.1%),
Recent evidence on the patterns and prevalence and Barbados (5.0%) had the highest prevalence
of SLT use was not found for several countries (PAHO, 2018).
in this region in which isolated SLT use had There is significant variation in terms of the
previously been reported (WHO, 2017, 2019). products used in the subregions (Siddiqi et al.,
Although countries in this region have a mark- 2020; Table 3.1). For example, chimó is the most
edly low overall prevalence of SLT use, there are widely consumed product in Venezuela and
several subregions, with wide population diver- Colombia, whereas rapé is more common in
sity and a potentially variable prevalence of SLT Brazil. In the USA, Canada, and Mexico, plug,
use (Ansara et al., 2013). Table 3.5 provides snuff, and snus are the major oral SLT products,
the overall prevalence of SLT use in some of the whereas iqmik is commonly consumed by Alaska
countries in the WHO Region of the Americas Natives (Siddiqi et al., 2020; Table 3.5). Areca nut
for which the estimated prevalence of SLT use consumption is reported among the residents of
was ≥ 2% in adults. The regional average prev- Hawaii, with a low prevalence in young people
alence was 1.4%, and overall the prevalence was (ever use of 3.1% in high school students; current
higher in men (2.5%) than in women (0.3%) in use of 1.3% in middle school students and 2%
this region (WHO, 2021a); however, in countries in high school students) compared with a much
such as Argentina (0.2%), Barbados (0.6%), and higher prevalence in immigrants from the
Haiti (3.1%), the prevalence was higher in women. Federated States of Micronesia (20.6%) (Pobutsky
Hotspots of high prevalence of SLT use by men and Neri, 2012).
were identified in the USA (6.2%), Venezuela In summary, a relatively small range of SLT
(6.2%), and Paraguay (3%) (WHO, 2017, 2021b). or areca nut products are currently consumed
The average prevalence of SLT use reported by nearly 1.5% of the population of the WHO
in adolescents was 2.6% (3.4% in boys and 1.7% Region of the Americas.
227
Table 3.6 Countries with high prevalence of smokeless tobacco use in the WHO African Regiona
Country Product name or Prevalence (%) Trends of prevalenceb Reference

colloquial name Overall (male; female)
Algeria Chemma or 8.9 (17.3; 0.4) Increasing (men), NCI and CDC (2014);
shammahc decreasing (women) Oudjehih et al. (2020); WHO
(2021b)
Benin Azổc 5.7 (8; 3.2) Decreasing Siddiqi et al. (2015); WHO
(2020b)
Burkina Faso Taaba c 8.9 (5.6; 11.7) Unknown NCI and CDC (2014); WHO
(2021b)
Central African Snuffc 16.3 (17.3; 15.5) Unknown NCI and CDC (2014); WHO
Republic (2021b)
Comoros Unknown 18.4 (19.5; 17.4) Unknown WHO (2021b)
Madagascar Parakyc 17.3 (24.6; 9.6) Unknown Blecher et al. (2014); WHO
(2021b)
Mozambique Unknown 5.6 (2.5; 7.9) Unknown WHO (2021b)
Sierra Leone Snuffc, chewing 7.8 (2.9; 12.1) Decreasing (slightly) Samai et al. (2011); WHO
tobaccoc (2021b); Drope et al. (2022)
Togo Unknown 3.6 (5.1; 2.2) Unknown WHO (2021b)
a Countries with a prevalence of smokeless tobacco and areca nut use of ≥ 5% (either overall or in males or females) are included in the table.
b Unknown: no comparable data over a time period to make a call on trend.
c Smokeless tobacco alone.
3.2.5 WHO African Region large number of residents of South Asian origin
(Mamudu et al., 2013; WHO, 2021b). In the
There are an estimated 15 million adult users Comoros, which has a high overall prevalence
of SLT (8 million men and 7 million women) in of SLT use, the prevalence of SLT use in women
the WHO African Region; it is the WHO region (17.4%) is the highest of the African countries
with the second-highest prevalence of SLT use in (WHO, 2021b). However, in some countries with
adults, after the WHO South-East Asia Region a relatively low overall prevalence (< 5%) of SLT
(WHO, 2021a). The prevalence of use varies use, use is reported predominantly in women
widely, ranging from 0.1% in women in Eritrea (prevalence > 5%), such as Botswana and Cabo
and Senegal to 24.6% in men in Madagascar Verde. The prevalence of SLT use is much higher
(WHO, 2021b). Of the 46 countries in the WHO in men than in women in countries such as
African Region, only 8 countries (Algeria, Benin, Algeria (17.3% vs 0.4%), Eritrea (11.6% vs 0.1%),
Burkina Faso, Central African Republic, the and Madagascar (24.6% vs 9.6%) (WHO, 2021b;
Comoros, Madagascar, Mozambique, and Sierra Table 3.6).
Leone) had a moderate to high (≥ 5%) overall Information about the trends in prevalence of
prevalence of SLT use (WHO, 2020b, 2021b). SLT use has been reported for few countries. The
The prevalence of SLT use is generally available data suggest a decreasing trend in SLT
high in both male and female individuals in use in women in Algeria, from a reported prev-
Burkina Faso, the Central African Republic, alence of 0.8% in 2010 to a prevalence of 0.4% in
the Comoros, and Madagascar (WHO, 2021b; 2017, whereas the estimated prevalence in men
Table 3.6). The high overall prevalence (17.3%) of increased, from 9.8% in 2010 to 17.3% in 2017
SLT use in Madagascar may be attributed to the (Oudjehih et al., 2020; WHO, 2021b). Recent data
228
from a 2016 survey in South Africa also show a possibly because there was little or no uptake by
marked decrease in the prevalence of SLT use most South Africans (Tobacco Control Research
in women, from 10.9% in 2003 to 1.3% in 2016, Group, 2021).
whereas the prevalence in men increased, from
2.4% in 2003 to 6.4% in 2016 (Siddiqi et al., 2015; 3.2.6 WHO Eastern Mediterranean Region
WHO, 2021b).
Youth surveys have suggested an increased There are an estimated 20.9 million adult
uptake of SLT use in adolescent boys and girls, users of SLT (17.7 million men and 3.2 million
even in countries with a relatively low prevalence women) in the WHO Eastern Mediterranean
of SLT use in adults, such as Botswana, Eswatini, Region (WHO, 2021a). The prevalence of SLT
Liberia, Malawi, Nigeria, Rwanda, South Africa, use varies widely, ranging from null in women
and Uganda (WHO, 2021b). In countries with a in Egypt, Iraq, and Kuwait and in both men and
moderate or high overall prevalence of SLT use, women in the Syrian Arab Republic to 33.7% in
such as Burkina Faso and Mozambique, SLT use men in Afghanistan (WHO, 2021b; Table 3.7).
is also common in adolescents, with a reported The prevalence of SLT use is generally high
prevalence of 10.2% in Burkina Faso and 7.5% in in adults in Afghanistan, Yemen, the Sudan,
Mozambique and not much difference between Pakistan, and Tunisia (WHO, 2020b, 2021b).
sexes (WHO, 2021b). In contrast, recent data In Afghanistan, the 2019 WHO STEPwise
from Madagascar suggest a very low prevalence Approach to Surveillance (STEPS) survey showed
of SLT use in adolescents (1.6%); this is an indian overall prevalence of SLT use of 19.3% (33.7%
cator that this practice is probably becoming in men and 3.7% in women); it is the country
unpopular there (WHO, 2021b). with the highest percentage of SLT users in the
The dominant SLT product type used in WHO Eastern Mediterranean Region (WHO,
the WHO African Region is snuff (moist and 2021b). Although in this region SLT is consumed
dry) (see Table 3.1) (NCI and CDC, 2014). It is predominantly by men, Yemen has reported a
also locally known as taaba in Burkina Faso, substantial prevalence of use (5.9%) in women
chemma or shammah (moist snuff) in Algeria, (WHO, 2021b; Table 3.7).
snuif in South Africa, Botswana, and Lesotho, Information about trends in prevalence of
and azổ in Benin (NCI and CDC, 2014; Oudjehih SLT use is available for some countries in this
et al., 2020). The use of chewing tobacco is less region (Table 3.7). In Pakistan, in adult men the
common. However, paraky is mostly used in prevalence of SLT use decreased from 16.3%
rural areas of Madagascar (Blecher et al., 2014), in 2012–2013 to 11.4% in 2014 and to 14.6% in
and use of betel quid without tobacco (areca nut) 2017–2018, but in women it increased from
is common in a minority population of South 2.44% in 2012–2013 to 3.7% in 2014 and to 3.4%
Asian descent in some parts of South Africa and in 2017–2018 (Siddiqi et al., 2015; WHO, 2020b,
the United Republic of Tanzania (Bissessur and 2021b). In the Sudan, the 2005 STEPS country
Naidoo, 2009; Bhat et al., 2010; NCI and CDC, report showed a prevalence of SLT use of 24.1%
2014). SLT use through the nasal route in the in men and 1% in women, but recent data from
form of dry snuff is still a common practice in the 2016 STEPS survey revealed a decreasing
some parts of the WHO African Region (Sinha trend in the prevalence of SLT use in both men
et al., 2018a), but oral application remains more (to 14.3%) and women (to 0.2%) (Siddiqi et al.,
popular (Table 3.1). In 2005, Scandinavian-type 2015; WHO, 2021b). In Yemen, when comparing
snus was also introduced to the South African the recent Demographic and Health Survey 2013
market, but data on its use have not been reported, data with the 2003 Individual Country Survey
229
Table 3.7 Countries with high prevalence of smokeless tobacco use in the WHO Eastern
Mediterranean Regiona
Country Product name or colloquial Prevalence (%) Trends of prevalence Reference

name Overall (male;
female)
Afghanistan Naswar or nass b 19.3 (33.7; 3.7) Unknown WHO (2021b)
Pakistan Gutka c, naswar b, chalia or 9e (14.6; 3.4) Decrease (males) Siddiqi et al. (2015);
suparid, paanc,d, zarda c Increase (females) WHO (2020b, 2021b)
Sudan Toombak b, saffab, saodb 7.9 (14.3; 0.2) Decrease Siddiqi et al. (2015);
Abakar et al. (2020);
WHO (2021b)
Yemen Shammahb, toombak b, tombol c 11.3 (17.0; 5.9) Increase (males) Siddiqi et al. (2015);
Al-Tayar et al. (2017);
WHO (2021b)
a Countries with a prevalence of smokeless tobacco and areca nut use of ≥ 5% (either overall or in males or females) are included in the table.
c Areca nut with tobacco.
d Areca nut and/or betel quid alone.
e Overall prevalence data were not provided in the data source; therefore, the estimate provided here was computed by the Working Group.
data, the percentage of male SLT users appears supari, and gutka. A study in Pakistan reported
to have increased slightly (from 15.1% in 2003 to that in a group of male and female users of SLT
17% in 2013), but the percentage of female SLT or areca nut products, the prevalence of use of
users seems to have remained almost stable (from naswar (4.1%) was the highest, followed by paan
6.2% in 2003 to 5.9% in 2013) (Siddiqi et al., 2015; (2.6%) (Abbas et al., 2014). The use of these prod-
WHO, 2021b). ucts is also culturally acceptable in Afghanistan,
In the WHO Eastern Mediterranean Region, predominantly a local product known as naswar
SLT use seems to be relatively less common in or nass. In the Sudan, SLT is referred to as
adolescents than in adults (WHO, 2021a). The toombak, saffa, or saod (Abakar et al., 2020). In
prevalence in adolescents is highest in Djibouti Yemen, some of the commonly used products are
(6.2%), followed by the occupied Palestinian shammah, tombol, and toombak (Al-Tayar et al.,
territory (6%), Pakistan (5.3%), and Yemen (5.1%) 2017; Table 3.7; see Section 3.1).
(WHO, 2021b). A relatively low prevalence of SLT
use in adolescents in the Sudan (4.9%) compared 3.2.7 Determinants of use
with that in adult men suggests that this practice
is becoming unpopular there, or that there is a Both SLT and areca nut contain addic-
cultural tendency towards uptake in adulthood tive substances; this explains their continued
(Idris et al., 1998; WHO, 2021b). use despite the proven adverse health effects,
The dominant product types used in the including oral cancer (Sumithrarachchi et al.,
WHO Eastern Mediterranean Region are plain 2021). Therefore, to effectively eliminate these
SLT, or areca nut mixed with tobacco (NCI and practices, it is imperative to understand the
CDC, 2014). A variety of products are available in reasons that influence the initiation and
the region, of which the most common forms are continued use of these products. Whereas ciga-
betel quid with tobacco (paan), naswar, chalia/ rette smoking has been widely studied because
230
it is the causative factor for many noncommu- Pakistan conducted in adolescent users of areca
nicable diseases (Bergen and Caporaso, 1999), nut and/or SLT reported that adolescents who
studies on determinants of use of SLT and areca had not attended the knowledge-based sessions
nut are fewer in comparison. on the harmful health effects of areca nut and/or
Multiple factors determine the initiation and SLT use were more likely to use these products
continued use of SLT and areca nut, with an inter- (Hussain et al., 2017). In another study in adult
play between some of the factors. These determi- chewers in Myanmar, use of areca nut was found
nants may be broadly grouped as (i) individual to be significantly associated with low knowledge
factors (knowledge and perceptions), (ii) social scores with respect to adverse health effects of
factors (sociodemographic, socioeconomic, and areca nut use (Myint et al., 2016).
sociocultural), and (iii) environmental factors The level of knowledge about the harmful
(Table 3.8) (Singh et al., 2016). effects of areca nut or SLT use may also depend
Identifying the individual, social, and envi- on the level of education, as reported in multiple
ronmental determinants of the initiation and studies, in which individuals with lower educa-
continuation of SLT and areca nut use is required tion levels had less awareness of the adverse
when planning programmes on awareness and effects of these substances (Khawaja et al., 2006;
cessation interventions for these established risk WHO Regional Office for South-East Asia, 2012;
factors. Myint et al., 2016; Bangladesh Bureau of Statistics
and National Tobacco Control Cell, 2019).
(a) Individual factors Beliefs or perceptions about substances
Inculcating appropriate knowledge or raising such as SLT or areca nut are another impor-
awareness has the ability to induce a desired tant factor determining their use. Some users
health-related behavioural change. Several believe that the use of SLT offers health benefits,
studies have shown that knowledge levels and such as improving sleep quality and relieving
perceptions are associated with the use of SLT toothaches, headaches, and tiredness (Solhi
and areca nut (Singh et al., 2016). A few selected et al., 2021). There is also a belief that SLT is less
studies are described here to illustrate this deter- harmful than smoked tobacco (Singh et al., 2016).
minant (Table 3.8). Certain perceived positive effects of chewing
A cross-sectional study conducted in adoles- areca nut have been proven to be important
cents in the USA reported a moderate level of determinants of its use; these include inducing
knowledge about the undesirable effects of SLT, relaxation, enhancing concentration and aiding
which had only little impact on male users (Lee decision-making, relieving boredom, improving
et al., 1994). In another cross-sectional study in stamina, curing cold, inducing a pleasant sensa-
school students in the USA, significant differ- tion, feeling energized, and conferring cosmetic
ences were observed in the knowledge level benefits (Changrani et al., 2006; Banerjee et al.,
and attitudes between SLT users and non-users; 2014; Myint et al., 2016; Lin et al., 2017; Hussain
students with higher knowledge and attitude et al., 2018; Do and Vu, 2020).
scores were less likely to use SLT (Goebel et al.,
2000). In contrast, a study in a sample of univer- (b) Social factors
sity students in the USA reported no influence of (i) Sociodemographic determinants
the observed high knowledge level on the prev- Multiple studies in India have ascertained
alence of SLT use, indicating a probable influ- the role of age at initiation for SLT use; younger
ence of multiple factors (Monson and Beaulieu, age at initiation is associated with a higher
2011). A school-based cross-sectional study in
231
232

Table 3.8 Determinants of use of smokeless tobacco and areca nut products
Determinants Facilitators Barriers Country or Reference

territory
Individual factors
Knowledge Higher tendency to use SLT or AN if lower India Singh et al. (2016)
knowledge level about their harmful effects Bangladesh Bangladesh Bureau of Statistics and
National Tobacco Control Cell (2019)
USA Lee et al. (1994); Goebel et al. (2000)
Myanmar Myint et al. (2016)
Pakistan Hussain et al. (2017)
Lower knowledge level about the harmful effects Pakistan Khawaja et al. (2006)
of SLT or AN was also due to low education level Myanmar Myint et al. (2016)
Bangladesh Bangladesh Bureau of Statistics and
Indonesia WHO Regional Office for South-East
Asia (2012)
Perceptions SLT use not as harmful as the other tobacco types India Singh et al. (2016); Shah et al. (2018)
(smoking) USA Goebel et al. (2000)
SLT perceived as suitable for dental health and
treatment of dental pain
Belief that SLT causes one or Indonesia WHO Regional Office for South-East
more of the following: serious Asia (2012)
illnesses, serious illnesses in Bangladesh Bangladesh Bureau of Statistics and
pregnancy, stroke, heart attack, National Tobacco Control Cell (2019)
oral cancer India TISS and MOHFW (2017)
Belief that SLT use has USA Lee et al. (1994); Goebel et al. (2000);
undesirable effects, such as oral Changrani et al. (2006); Monson and
diseases or hypertension, chest Beaulieu (2011)
pain or burning
Perceived positive effects of AN chewing: Taiwan Lin et al. (2017); Yang and Lin (2017)
considered cool in youth, as a cooling-off agent, (China)
improves work efficiency, improves stamina, USA Changrani et al. (2006); Banerjee
relieves tension, cures cold, provides relaxation, (migrants) et al. (2014)
relieves boredom, reduces stress, increases Pakistan Rozi and Akhtar (2007); Hussain
alertness, provides pleasant sensation, aids in et al. (2017, 2018); Saqib et al. (2018)
digestion, prevents bad breath, reduces appetite,
Sri Lanka Lee et al. (2011); Sinha et al. (2012)
cosmetic benefits (red teeth as a sign of beauty).
India Shah et al. (2018)
Guam (USA) Murphy and Herzog (2015)
territory
Social factors
1. Sociodemographic
Age Initiation of SLT and AN use at younger age India Singh et al. (2016); Sharapova et al.
USA (2020)
Continuation of AN and SLT use increases with Bangladesh Bangladesh Bureau of Statistics and
age National Tobacco Control Cell (2019)
Cambodia Sreeramareddy et al. (2014a)
Indonesia Lee et al. (2011)
India Rani et al. (2003); TISS and MOHFW
(2017)
Sri Lanka Sinha et al. (2012)
Thailand WHO Regional Office for South-East
Asia (2011)
Nepal Shrestha et al. (2019)
Malaysia IPH (2012)
United Arab Ali et al. (2020)
Emirates
(migrants)
Uganda Kabwama et al. (2016)
United Núñez-de la Mora et al. (2007)
Kingdom
(migrants)

233
234

territory
Sex Higher prevalence of SLT use in males (reported Sri Lanka Lee et al. (2011); Sinha et al. (2012)
that SLT includes all types of non-smoked India Sinha et al. (2012); TISS and
tobacco products and AN) MOHFW (2017)
Malaysia IPH (2012)
Higher prevalence of SLT use in females (reported Thailand WHO Regional Office for South-East
that SLT includes all types of non-smoked Asia (2011)
tobacco products and AN, BQ)
Higher prevalence of AN use in males Myanmar Myint et al. (2016)
Sri Lanka Lee et al. (2011)
Nepal Lee et al. (2011)
Taiwan Lee et al. (2011)
(China)
China Lee et al. (2011)
Higher prevalence of AN (BQ) use in females Malaysia Lee et al. (2011)
Indonesia Lee et al. (2011)
Ethnicity Higher initiation and continued use of SLT noted USA Ebbert et al. (2006); Chaffee et al.
in White people (2018)
territory
Residence Higher prevalence of SLT use in rural areas than Indonesia WHO Regional Office for South-East
in urban areas Asia (2012)
India MOHFW and IIPS (2019); Singh et al.
(2020)
Asia (2011)
Malaysia IPH (2012)
WHO African Kabwama et al. (2016); Bonnechère
Region et al. (2019); WHO FCTC and ICMR-
NICPR (2022)
WHO Eastern Al-Tayar et al. (2017); Alemi et al.
Mediterranean (2021)
Region
2. Socioeconomic
Income level Higher prevalence of SLT use in poorer groups/ Cambodia Sreeramareddy et al. (2014a)
lowest-income groups/lowest-wealth-index Bangladesh WHO Country Office for Bangladesh
groups (2018); Bangladesh Bureau of
Statistics and National Tobacco
Control Cell (2019)
India Thakur et al. (2015); Bhan et al.
(2016); Singh et al. (2016); Sinha et al.
(2018a)
Employment Higher prevalence of SLT use in unemployed Indonesia WHO Regional Office for South-East

status people and homemakers Asia (2012)
India Singh et al. (2016, 2020)
235
236

territory
Type of Higher prevalence of AN use in taxi drivers, Taiwan Yang and Lin (2017); Huang et al.
occupation three-wheel taxi drivers, transportation workers, (China) (2020)
security guards, labourers, construction workers, Sri Lanka Mahees et al. (2021)
agriculture workers, and plantation workers United Arab Ali et al. (2020)
Emirates
(migrants)
Higher prevalence of SLT use in military USA Lin et al. (2018)
personnel (higher percentage of users serving as
infantry and gun crew specialists, and enlisted
personnel)
Education level Higher prevalence of SLT and AN use with lower India Palipudi et al. (2012); Singh et al.
education levels (2016); TISS and MOHFW (2017)
Households with uneducated or less-educated Egypt Palipudi et al. (2012)
members tend to consume more SLT Philippines Palipudi et al. (2012)
Bangladesh WHO Country Office for Bangladesh
(2018); Bangladesh Bureau of
Statistics and National Tobacco
Control Cell (2019)
Nepal Lee et al. (2011); Sreeramareddy et al.
(2014a); Shrestha et al. (2019)
Asia (2011)
Cambodia Sreeramareddy et al. (2014a)
Malaysia Lee et al. (2011); IPH (2012)
Taiwan Lee et al. (2011)
(China)
Indonesia Lee et al. (2011); WHO Regional
Office for South-East Asia (2012)
Emirates
(migrants)
territory
3. Sociocultural
Family or peer One of the main determinants for initiation of Pakistan Rozi and Akhtar (2007); Hussain
pressure SLT or AN use et al. (2017)
Guam (USA) Murphy et al. (2019)
Considered rude and disrespectful to refuse Guam (USA) Murphy and Herzog (2015); Murphy
chewing of AN (BQ) if family members or peers Federated et al. (2019)
are chewing States of
Micronesia
Pakistan Hussain et al. (2017); Hussain et al.
(2018)
Social reasons During interactions with friends and peers and Taiwan Lin et al. (2017)
for social acceptability (China)
Guam (USA) Murphy and Herzog (2015)
Symbol of love and marriage Taiwan Ma et al. (2017)
(China)
India Ahuja and Ahuja (2011)
AN offered to visitors on special occasions Sri Lanka Wijesinghe (2018)
Cultural reasons An acceptable alternative to smoking in Indian India Singh et al. (2016); Shah et al. (2018)
culture
Ancestral practice of the Kalunga community Brazil Novais (2017)
(the largest quilombola community in Brazil)
Use of multiple Strong association between current smoking USA Ebbert et al. (2006)
substances practice and initiation of SLT use
Concurrent AN (BQ) chewing in people who Myanmar Myint et al. (2016)
consume alcohol and/or smoke Malaysia Lee et al. (2011)
Taiwan Lee et al. (2011); Lin et al. (2017);
(China) Yang and Lin (2017)

Emirates
(migrants)
237
238

territory
Environmental factors
Easy availability Around the house Guam (USA) Murphy and Herzog (2015)
In neighbourhood stores USA Banerjee et al. (2014); Do and Vu
From hawkers around educational institutions (migrants) (2020)
India Sinha et al. (2016)
Family Preparing the AN quid for elderly family Guam (USA) Murphy and Herzog (2015)
members
Strong influence from family members United Núñez-de la Mora et al. (2007)
Kingdom
(migrants)
School type Higher tendency to use by students attending Pakistan Rozi and Akhtar (2007); Hussain
government schools than those attending private et al. (2017)
schools
Advertisements Exposure to tobacco advertisements is a factor in USA Timberlake (2016)
SLT use, especially by young people India Arora et al. (2008)
Sudan Almahdi et al. (2020)
Not seeing anti-tobacco advertisements Pakistan Rozi and Akhtar (2007)
Sports figures SLT use by favourite professional baseball players USA Chaffee et al. (2018)
(determinant for initiation and continuation of
SLT use in youth)
Health messages Lack of anti-AN and anti-SLT public health USA Banerjee et al. (2014)
messages (migrants)
AN, areca nut; BQ, betel quid; SLT, smokeless tobacco; WHO, World Health Organization.
level of use and more prolonged use. In addi- TISS and MOHFW, 2017; Bangladesh Bureau of
tion, the GATS-1 India documented that female Statistics and National Tobacco Control Cell,
individuals and people living in rural areas had 2019). Similarly, in a cross-sectional study in
a younger age at initiation (Singh et al., 2016). Myanmar men were 3 times as likely as women to
In Bangladesh, the GATS also documented a chew areca nut (Myint et al., 2016). In Pakistan,
younger age at initiation of areca nut use in men were also found to have a higher prob-
women (Bangladesh Bureau of Statistics and ability than women of initiating use of areca
National Tobacco Control Cell, 2019). A study in nut (including betel quid) (Hussain et al., 2017).
middle school and high school students in the Furthermore, a multicountry study also docu-
USA reported similar findings; male students mented a higher prevalence of areca nut chewing
initiated SLT use at a slightly older age compared in men than in women in China, Nepal, Sri
with their female counterparts (Sharapova et al., Lanka, and Taiwan (China), whereas the oppo-
2020). site was observed in Indonesia and Malaysia (Lee
With regard to continuation of SLT or areca et al., 2011).
nut use, in a study in Pakistan conducted in Ethnicity was also reported to be a predictor
adolescent users of areca nut (including betel of the initiation and continuation of SLT use; in
quid) and/or SLT, age was positively associated the USA, a higher prevalence of initiation and
with continued use (Hussain et al., 2017). In continuation was found in White people than
India, an increased likelihood of SLT use with in individuals of other ethnicities (Ebbert et al.,
increasing age was also observed; men aged 2006; Chaffee et al., 2018).
≥ 60 years were 4 times as likely and women Evidence from some countries in the
aged ≥ 60 years were 8 times as likely to use WHO African Region and the WHO Eastern
SLT compared with younger individuals (aged Mediterranean Region has shown a higher
15–24 years) (Rani et al., 2003). The GATS-2 prevalence of SLT use in people living in
India further confirmed the increasing likeli- rural areas (Al Tayar et al., 2017; Bonnechère
hood of SLT use with increasing age (TISS and et al., 2019; Alemi et al., 2021; WHO FCTC
MOHFW, 2017). This finding has also been and ICMR-NICPR, 2022). In general, there are
noted in Afghanistan, Bangladesh, Malaysia, a higher percentage of adult SLT users in rural
Nepal, and Thailand (WHO Regional Office for areas than in urban areas, especially in the coun-
South-East Asia, 2011; IPH, 2012; Bangladesh tries in the WHO South-East Asia Region, such
Bureau of Statistics and National Tobacco as Bangladesh, India, Myanmar, Nepal, and
Control Cell, 2019; Shrestha et al., 2019; Alemi Thailand (WHO Regional Office for South-East
et al., 2021). Another study in adolescent male Asia, 2011; Myint et al., 2016; Bangladesh Bureau
SLT users in Pakistan reported a similar asso- of Statistics and National Tobacco Control
ciation; however, this weakened on multivariate Cell, 2019; MOHFW and IIPS, 2019; Shrestha
analysis (Rozi and Akhtar, 2007). et al., 2019). A recent report of the Global Youth
With regard to sex, a higher prevalence of Tobacco Survey (GYTS) India suggests a higher
SLT use has been noted in male individuals prevalence of SLT use in school-going adolescents
in many countries, such as India, Malaysia, in rural areas than in urban areas (MOHFW and
Pakistan, and Sri Lanka, whereas in Bangladesh IIPS, 2019; Table 3.8).
and Thailand the reported prevalence of SLT use
is higher in female individuals (Lee et al., 2011;
WHO Regional Office for South-East Asia, 2011;
IPH, 2012; Sinha et al., 2012; Hussain et al., 2017;
239
(ii) Socioeconomic determinants in military personnel (especially in the infantry

The socioeconomic determinants of use of or gun crew specialists) in the USA (Lin et al.,
SLT and areca nut are income level, employment, 2018). In Taiwan (China), drivers and construc-
and education level (Table 3.8). tion workers were reported to have a higher prev-
A sufficient amount of literature is available alence of use of areca nut (including betel quid)
on the role of these factors in India (Singh et al., (Huang et al., 2020). In Sri Lanka, three-wheel
2016). A clear trend has been observed of higher taxi drivers, transportation workers, secu-
prevalence of SLT use with lower income levels rity guards, construction workers, plantation
(Bhan et al., 2016). Thakur et al. (2015) showed workers, and fishers had a very high prevalence
that the probability of SLT use decreases with of use of commercially prepared SLT products
increasing income; wide economic inequalities (Mahees et al., 2021). It has been hypothesized
in the patterns of SLT use were observed in all the that individuals in such occupations that
states of India. The association between SLT use require long working hours or continuously
and low income levels was also observed in other repeated activities benefit from the perceived
countries in the WHO South-East Asia Region, positive effects of areca nut use, such as improving
such as Bangladesh, Cambodia, and Nepal concentration, reducing hunger, inducing a sense
(Sreeramareddy et al., 2014a; WHO Country of well-being, and relieving boredom (Winstock,
Office for Bangladesh, 2018; Bangladesh Bureau 2002; Yang and Lin, 2017).
of Statistics and National Tobacco Control Cell, With regard to education level, the GATS-1
2019; Shrestha et al., 2019), and in countries India reported clear educational gradients;
in sub-Saharan Africa (Sreeramareddy et al., individuals with no formal education or less
2014b). An analysis of 140 countries by Sinha than primary education were much more likely
et al. (2018a) showed that, in general, the burden to be users of SLT or areca nut compared with
of SLT use is greatest in the lowest-income individuals with secondary education or above
segments of the population. (MOHFW and IIPS, 2010). This pattern
Unemployment was found to be another persisted over time; the GATS-2 India reported
predictor of increased likelihood of SLT use in that despite the decreasing trend in SLT
India (Singh et al., 2020). Similarly, in Indonesia, use in all households, an association between
the largest proportion of SLT users are home- lower education levels and higher prevalence of
makers (WHO Regional Office for South-East SLT use remained (TISS and MOHFW, 2017).
Asia, 2012). In contrast, it has also been reported A large multicountry study involving 13 low-
that the expense incurred, especially for an and middle-income countries also reported
unemployed person, is an important reason for high prevalence of tobacco use (including SLT)
quitting this practice (Murphy and Herzog, 2015). in individuals in the lower educational attain-
This may be due to the wide differential pricing ment category in Egypt and the Philippines,
of the various SLT products or even different among other countries (Palipudi et al., 2012).
brands of the same product (Nargis et al., 2014). The association between prevalence of SLT use
Also, increases in the taxation of smoked tobacco and lower education levels was also observed in
products have led to comparatively lower prices Bangladesh, Cambodia, Malaysia, Nepal, and
of SLT. Thailand (WHO Regional Office for South-East
The type of occupation may also determine Asia, 2011; IPH, 2012; Sreeramareddy et al., 2014a;
the prevalence of use of SLT and areca nut. A Bangladesh Bureau of Statistics and National
high prevalence of SLT use has been reported Tobacco Control Cell, 2019). Similarly, the large
Asian Betel-Quid Consortium study, which
240
involved 8922 chewers of areca nut (betel quid it is deeply embedded in the Indian culture
with or without tobacco), reported that individ- (Singh et al., 2016; Shah et al., 2018). The Kalunga
uals with higher education levels in Indonesia, community, the largest quilombola commu-
Malaysia, Sri Lanka, and Taiwan (China) were nity in Brazil, still preserves ancestral practices
less likely to be users of areca nut (Lee et al., 2011). such as the use of a type of snuff called simonte
The outcomes of another study, in adult chewers (Novais, 2017).
of areca nut in Myanmar, further corroborated Current tobacco smoking was found to be
these findings (Myint et al., 2016). However, indi- a strong predictor of the initiation of SLT use
viduals with higher education levels in Hunan (Ebbert et al., 2006). In addition, concurrent use
(China) were slightly more likely to be users of of areca nut (including betel quid) was observed
areca nut, probably because of the influence of along with alcohol consumption and/or smoking
other factors (Lee et al., 2011). in various settings (Lee et al., 2011; Myint et al.,
2016; Lin et al., 2017; Yang and Lin, 2017; Ali
(iii) Sociocultural determinants
et al., 2020).
Many studies have reported an association
between various sociocultural factors and use of (c) Environmental factors
SLT and areca nut (Table 3.8). Some of the adult participants in a pilot study
Studies in Guam (USA), Myanmar, and conducted in Guam (USA) cited readily available
Pakistan have documented that use of SLT and/ areca nut, especially around the house, and the
or areca nut by family members and peer pres- practice of preparing the areca nut by softening it
sure are among the main determining factors orally to enable use by the toothless elders in the
for initiation of these practices (Rozi and Akhtar, family as the main reasons for initiation of areca
2007; Myint et al., 2016; Hussain et al., 2017; nut chewing (Murphy and Herzog, 2015).
Murphy et al., 2019). The effect of peer pressure Evidence from India and Pakistan has also
on SLT use was also reported in a review in India shown that the easy availability of SLT and areca
(Shah et al., 2018). Moreover, in a recent study nut from hawkers around educational institu-
in adolescent chewers in Pakistan, not chewing tions, such as schools, plays a major role in facil-
was considered rude if family members or itating their use in adolescents (Sinha et al., 2016;
friends were chewing (Hussain et al., 2018); this Hussain et al., 2017). The school environment
sentiment was shared by adults in Guam (USA)
may also play a role in determining use of SLT
(Murphy et al., 2019). and areca nut. A study conducted in adolescent
The practice of areca nut chewing reinforces male high school students in Pakistan observed
positive acceptance when socializing with friends a higher prevalence of SLT use in students
in Taiwan (China), because sharing of areca nut attending government schools than in those
is a usual practice during social gatherings (Lin attending private schools (Rozi and Akhtar,
et al., 2017; Ma et al., 2017). It is also considered 2007).
a symbol of love and marriage in China, India, Exposure to advertisements for SLT and areca
and Taiwan (China) (Ahuja and Ahuja, 2011; Ma nut products is another determining factor for
et al., 2017). In Sri Lanka, areca nut is also offered the use of these products. A cross-sectional study
to visitors on important occasions (Wijesinghe, in 11 462 adolescent students in India reported
2018). that greater exposure to tobacco advertisements
In India, tobacco smoking in the presence significantly increased the risk of initiating
of elders is a social taboo, whereas wide social tobacco use; a dose–response effect was noted
acceptance exists for tobacco chewing because for a subset of students (Arora et al., 2008). In
241
addition, a study in male high school students in (short-term abstinence assessment) and
Pakistan found a significantly higher tendency prolonged or continuous abstinence (long-term
to SLT use in those who did not see anti-to- abstinence assessment) confirmed by biochem-
bacco advertisements (Rozi and Akhtar, 2007). ical validation or self-reporting. Prolonged or
In the Sudan, exposure to the advertisement continuous abstinence was defined as a preferred
of toombak at point of sale is associated with measure, and point prevalence was defined as a
its increased perceived accessibility (Almahdi secondary measure recommended by Hughes
et al., 2020). Lack of anti-areca nut and et al. (2003).
anti-SLT public health messages was also cited
as a facilitator of areca nut and SLT use by a 3.3.1 Behavioural interventions
group of South Asian immigrants in the USA
(Banerjee et al., 2014). In the USA, perceived This section reviews studies assessing the
SLT use by favourite professional baseball effectiveness of behavioural interventions alone
players was shown to increase the susceptibility for cessation of SLT and/or areca nut use, both
to initiation and continuation of SLT use in in adults and in youth.
adolescent baseball players (Chaffee et al., 2018). (a) Behavioural interventions in adults
Nine studies (7 RCTs and 2 cohort studies)
3.3 Interventions for cessation of use using behavioural interventions for SLT cessa-
tion were conducted in adults. Two of the largest
The review included published intervention
studies were cohort studies conducted in India
studies with intervention and control groups,
(Gupta et al., 1992; Anantha et al., 1995); most
such as randomized controlled trials (RCTs) and
of the studies (6) were conducted in the USA
cohort studies. For those on behavioural inter-
(Stevens et al., 1995; Severson et al., 1998, 2007,
ventions alone, studies with follow-up from
2008, 2009; Walsh et al., 1999), and one study
the start of the intervention of ≥ 6 months were
was conducted in Sweden (Virtanen et al., 2015)
included, and for those on pharmacological inter-
(Table 3.9).
ventions (alone or in combination with behav-
The earliest interventions took place in India.
ioural interventions), studies with follow-up of
One quasi-experimental cohort trial was carried
≥ 6 weeks were included.
out for 10 years in Ernakulam District, Kerala,
The review excluded studies such as those not
India, in 7033 users of betel quid with tobacco,
targeted at SLT or areca nut use but at smoking
to reduce the incidence of oral mucosal lesions
cessation, those targeted at SLT or areca nut
by persuading participants to quit tobacco use.
use but non-cessation studies, non-randomized
Interventions were carried out through house-
intervention trials, and studies with SLT quit
to-house visits followed by an oral examina-
attempts, reduction, or withdrawal symptoms as
tion and an educational talk by a dentist and
the primary end-point.
social scientist, along with relevant information,
When the RR and 95% CI for cessation of
education, and communication materials such as
SLT use were not provided by the authors, they
films, radio broadcasts, posters, local newspaper
were calculated by the Working Group for each
articles, and lantern slides in local cinemas. [At
outcome with the longest follow-up period. In
10 years of follow-up, a statistically significant
most studies assessing the effectiveness of phar-
effect was noted for the cessation intervention:
macological interventions alone or in combina-
relative risk (RR), 2.81; 95% confidence interval
tion with behavioural interventions, abstinence
(CI), 2.38–3.32] (Gupta et al., 1992). The incidence
is defined by 7-day point-prevalence abstinence
242
Table 3.9 Behavioural interventions for cessation of smokeless tobacco and/or areca nut use in adults
Reference Study design Intervention arm Control arm Efficacy of intervention Comments/interpretation
Location Study population
Gupta et al. Cohort study (quasi- 4619 2414 At 120 months (10 Strengths: large sample size; long follow-
(1992) experimental) House-to-house survey/interview, Interview, oral years), quit rate: up on cessation and OPMDs
Ernakulam Men and women, oral examination, educational examination, Men: Limitations: ITT data absent; control
District, aged ≥ 15 years talk by dentist and social scientist, brief I: 15.1% group was not concurrent in time; not
Kerala, India Betel quid with tailored films, radio broadcasts, educational C: 2.3% an RCT; results were not confirmed
tobaccob posters, local newspaper articles, talk, advice to Women: biochemically
10-year follow-up exhibition of lantern slides in local quit tobacco by I: 18.4%
cinemas, dental camps dentist C: 7.8%
[RR (95% CI):
Men: 6.52 (3.96–10.76)
Women: 2.37 (1.98–2.83)
Overall: 2.81 (2.38–3.32)]
Anantha Cohort study (quasi- 6714 Two control At 60 months (5 years), Strengths: long-term intervention; large
et al. (1995) experimental) Anti-tobacco education through areas: quit rate in men: sample size
India Men and women, handbills, folders, cards, a photo Control area 1: I: 30.2% Limitations: age group of participants
all ages album, portable display boards, 12 152 Control area 1: 1.2% in intervention and control arms not
Chewed tobaccoa,b and audiovisual aids (films in local Control area 2: Control area 2: 1.1% mentioned; no randomization; quit
5-year follow-up languages) 8171 [RR (95% CI): rate in women not mentioned; ITT
No anti-tobacco 25.70 (13.26–49.84)] data absent; results were not confirmed
education biochemically
RR was calculated by the Working
Group comparing intervention with
combination of both control arms and
only for men
Stevens et al. RCT 245 273 At 12 months, quit rate: Strength: long-term follow-up
(1995) Men, aged ≥ 15 years Oral examination, prophylactic Oral I: 18.4% Limitations: this study was
USA Moist snuff and treatment, patient education examination, C: 12.5% contaminated with NRT use by 4.5% in
chewing tobaccoa (with feedback), advice to quit brief advice to RR (95% CI): the intervention arm and 6.4% in the
12-month follow-up SLT products by DH, follow-up by quit 1.47 (0.83–2.6) control arm; results were not confirmed
dentist, video and brief counselling biochemically

session, brief self-help booklet, RR calculated by Ebbert et al. (2015)
telephone number of a 24-hour
advice line, a quit kit and follow-up
call by DH after 1 week, tip sheets,
monthly newsletters
243
244

Severson RCT 394 293 At 12 months, quit rate: Strength: long-term follow-up
et al. (1998) Men and women, Oral examination, advice to quit Advice to quit I: 10.2% Limitation: results were not confirmed
USA aged ≥ 15 years SLT use, informative pamphlets, C: 3.3% biochemically
SLTa quit kit, setting of a quit date, RR (95% CI): RR calculated by Ebbert et al. (2015)
12-month follow-up motivational video, telephonic 3.03 (1.44–6.37)
follow-up within 2 weeks
Walsh et al. RCT 171 189 At 12 months, quit rate: Strength: long-term follow-up
(1999) Men Oral examination by dentist, No intervention I: 35.1% Limitations: nicotine gum was used by
USA College baseball and advice to quit tobacco use, self-help C: 15.9% only 10% in the intervention group; no
football athletes guide, brief counselling by DH RR (95% CI): biochemical confirmation; 4% (7 in the
SLTa (to SLT users, and also in groups 2.21 (1.50–3.25) intervention group and 5 in the control
12-month follow-up to non-users), nicotine gum, group) of the SLT users who were non-
telephone support smokers at baseline started smoking
cigarettes; of these athletes, only 1 in the
intervention group quit SLT use
Quit rates and RR calculated by Ebbert
et al. (2015)
Severson RCT 535 534 At 12 months, quit rate Strength: long-term follow-up
et al. (2007) Men (majority) Assisted self-help: SLT quitting SLT quitting (based on ITT): Limitation: results were not confirmed
USA and women, aged manual, video, telephone manual only I: 12.9% biochemically
17–82 years counselling C: 9.7% RR calculated by Ebbert et al. (2015)
SLTa RR (95% CI):
12-month follow-up 1.32 (0.94–1.86)
Severson RCT 1260 1263 At 6 months, quit rate: Limitation: results were not confirmed
et al. (2008) Men Enhanced website: guided, Basic website: [I: 21.4% biochemically
USA SLTa interactive programme to help printable C: 16.8%
6-month follow-up each user create a tailored plan for pocket guide RR (95% CI):
quitting and relapse prevention, and useful 1.28 (1.09–1.50)]
streaming video, broader range resources, links
of printable useful resources, to external
annotated links to external websites on SLT
websites, two web forums, two cessation and
modules (planning to quit and oral cancer
staying quit)
Severson RCT 392 393 At 6 months, quit rate: Limitation: results were not confirmed
et al. (2009) Male military SLT cessation manual, video Advice to quit I: 30.3% biochemically
USA personnel cessation guide tailored to military SLT use, referral C: 15.3% Quit rates and RR calculated by Ebbert
SLT personnel, telephone counselling to local military RR (95% CI): et al. (2015)
6-month follow-up installation 1.98 (1.50–2.61)
tobacco
cessation
programmes
Virtanen RCT 225 242 At 6 months, quit rate: Limitations: number of SLT users in
et al. (2015) (FRITT study) 94 SLT users 100 SLT users I: 7.5% intervention and control groups was
Sweden Men and women, Structured brief advice based on Usual care C: 2% limited; large loss to follow-up; results
aged 18–75 years the 5A modelc RR (95% CI): were not confirmed biochemically
Snusa 3.72 (0.79–17.47) Quit rates and RR calculated by Ebbert
6-month follow-up et al. (2015)
AN, areca nut; C, control; CI, confidence interval; DH, dental hygienist; FRITT, Free from Tobacco in Dentistry; I, intervention; ITT, intention-to-treat analysis; NRT, nicotine
replacement therapy; OPMDs, oral potentially malignant disorders; RCT, randomized controlled trial; RR, relative risk; SLT, smokeless tobacco.
a SLT only.
b AN with tobacco.
c 5A model: (1) Asking about tobacco use, (2) Advising to quit, (3) Assessing willingness to quit, (4) Assisting the tobacco user in quitting, for instance by providing information on
available counselling and medications, and (5) Arranging follow-up contacts.

245
rate of leukoplakia decreased significantly over the participants in the intervention arm, an oral
the 10-year study period, and much more so in the examination was conducted, followed by advice
intervention cohort than in the control cohort. to quit SLT use and the setting of a quit date,
[The results were based only on the response of informative pamphlets, a quit kit and a moti-
participants and were not confirmed biochem- vational video, and telephonic follow-up within
ically. RR with 95% CI and intention-to-treat 2 weeks. Participants in the control arm were
analysis were not provided by the authors, and provided with usual care and only advice to quit.
the control group was not concurrent in time.] At 12 months, the cessation rate of SLT use was
An anti-tobacco community education 10.2% in the intervention group compared with
programme was conducted through trained 3.3% in the usual-care group (RR, 3.03; 95% CI,
health workers in Kolar District, Karnataka, 1.44–6.37; Ebbert et al., 2015). [Results were not
India, in an intervention area (n = 6714) and two confirmed biochemically.]
control areas (n = 12 152 in control area 1 and Walsh et al. (1999) conducted an athletic team-
n = 8171 in control area 2) (Anantha et al., 1995). based SLT cessation programme based on cogni-
The intervention, which included anti-tobacco tive social learning theory in 360 male college
education through handbills, folders, cards, a baseball and football athletes. The intervention
photo album, portable display boards, and films included an oral examination by a dentist, advice
in local languages, was provided only to the inter- to quit SLT use, a self-help guide, individual or
vention group. After 5 years, the prevalence of group counselling by a dental hygienist, tele-
tobacco chewing in men decreased significantly phone support, and nicotine gum in some
in the intervention group, from 16.8% to 8.1%, participants. At 12 months, the cessation rate of
and remained almost unchanged in the control SLT use was 35.1% in the intervention colleges
group (6.9% vs 7.1% in control area 1, and 11.4% and 15.9% in the control colleges (RR, 2.21;
vs 11.4% in control area 2) [RR, 25.70; 95% CI, 95% CI, 1.50–3.25; Ebbert et al., 2015). [Results
13.26–49.84]. [The age group of the participants were not confirmed biochemically. Nicotine
was not mentioned. Results were not confirmed gum was used by 10% of participants at inter-
biochemically. RR with 95% CI and intention-to- vention colleges, and 4% of the SLT users who
treat analysis were not provided.] were non-smokers at baseline started smoking
More recently, the 7 RCTs conducted in the cigarettes.]
USA and Sweden assessed the impact of SLT Severson et al. (2008) assessed the impact of
cessation interventions. Most of these studies an interactive, tailored web-based intervention
used one or more of the following behavioural (enhanced condition) versus a more linear, text-
interventions: brief advice to quit SLT use, a based website (basic condition) in 2523 adult
self-help booklet or cessation manual, tip sheets, SLT users in the USA. At 6 months of follow-up,
monthly newsletter, pamphlets, and/or video the cessation rate of SLT use was 21.4% in the
and telephone calls for brief counselling and enhanced condition and 16.8% in the basic
follow-up (Table 3.9). Of the 7 RCTs, 4 studies condition [RR, 1.28; 95% CI, 1.09–1.50]. [Results
(Severson et al., 1998, 2008, 2009; Walsh et al., were not confirmed biochemically.]
1999) showed statistically significant effects and An RCT was conducted in 785 male military
are described below. personnel who used SLT, recruited from 24 mili-
Severson et al. (1998) conducted a brief dental tary dental clinics across the USA during annual
office-based intervention in 687 SLT users (n = 394 dental examinations (Severson et al., 2009). The
in the intervention arm and n = 293 in the control behavioural intervention included an SLT cessa-
arm) at 75 dental practices in Oregon (USA). For tion manual, a videotape cessation guide tailored
246
to military personnel, and three 15-minute tele- Walsh et al. (2010) conducted an SLT cessa-
phone counselling sessions using motivational tion intervention study in 246 male baseball
interviewing methods. The usual care provided players aged 14–18 years in rural high schools in
to the controls consisted of standard procedures the USA who used SLT, and showed a significant
of the annual dental examination, including effect at 12 months of follow-up. The intervention
advice to quit SLT use and referral to local involved peer-led educational sessions, an oral
tobacco cessation programmes. At 6 months, examination, brief advice to quit SLT use, a self-
the cessation rate of SLT use was 30.3% in the help guide, a follow-up oral examination, and
behavioural intervention arm and 15.3% in the group cessation counselling sessions led by the
usual-care arm (RR, 1.98; 95% CI, 1.5–2.61; school nurse. In SLT users who were non-smokers
Ebbert et al., 2015). [Results were not confirmed at baseline, at 12 months of follow-up the cessa-
biochemically.] tion rate of SLT use was 62% in the interven-
A recent meta-analysis (Nethan et al., 2020) tion arm and 36% in the control arm [RR, 1.70;
reported efficacy of behavioural interventions 95% CI, 1.50–1.86]. [Results were not confirmed
for SLT cessation in adults (RR, 1.63; 95% CI, biochemically.] In this study, the male students
1.32–1.94) in both developed countries (RR, who used SLT only were more likely to quit SLT
1.39; 95% CI, 1.16–1.63) and developing coun- use than those who also smoked (i.e. dual users).
tries (RR, 2.79; 95% CI, 2.32–3.25). Of the 16 Gansky et al. (2005) conducted a study in
studies included in the meta-analysis, 8 studies 637 collegiate baseball athletes aged 17–20 years
(Gupta et al., 1992; Stevens et al., 1995; Severson who used spit tobacco at 52 colleges in California
et al., 1998, 2007, 2008, 2009; Walsh et al., 1999; (USA). The participants in the intervention arm
Virtanen et al., 2015) are summarized above. received oral cancer screening with feedback
In addition, in a study conducted in and brief counselling during the pre-season
Minnesota (USA) among 210 adult male users health screenings, support from a certified
of spit tobacco, group behavioural interventions athletic trainer for SLT cessation, and a peer-led
alone provided a higher long-term abstinence educational team meeting. The participants
rate than the use of nicotine gum with minimal in the control arm received the usual anti-to-
contact (Hatsukami et al., 1996; for details, see bacco education, and the intervention materials
Section 3.3.3). were distributed only after the study ended. At
12 months, the cessation rate of SLT use in the
(b) Behavioural interventions in youth intervention group (36%) was not significantly
Interventions for cessation of SLT use in different from that in the control group (37%)
youth are different from those in adults, because (RR, 0.98; 95% CI, 0.80–1.20; Ebbert et al., 2015).
the related health risks are not a major concern In a larger cohort of 948 students from the same
for this age group. A total of 5 studies (4 RCTs colleges, a significant positive effect of the inter-
and 1 cohort study) were found that assessed vention on the prevention of initiation of SLT
behavioural interventions for SLT cessation in use was observed (RR, 0.58; 95% CI, 0.35–0.99).
youth. The 4 RCTs were conducted in the USA [Results were not confirmed biochemically.]
in schools and colleges, in baseball and football Danaher et al. (2013) assessed a web-based
players and other athletes (Gansky et al., 2002, intervention for SLT cessation, called the
2005; Walsh et al., 2010; Danaher et al., 2013). MyLastDip programme, in SLT users aged
The cohort study was conducted in India (Stigler 14–25 years in the USA; 857 SLT users were
et al., 2007) (Table 3.10). randomly assigned to receive the enhanced
website-based tailored intervention, and 859 SLT
247
248

Table 3.10 Behavioural interventions for cessation of smokeless tobacco and/or areca nut use in youth
Reference Study design Intervention arm Control arm Efficacy of Comments/interpretation

Location Study population intervention
Gansky RCT 355 375 At 24 months, Strength: long-term follow-up
et al. (2002) Males 141 users of spit tobacco 166 users of spit quit rate: Limitation: results were not confirmed
USA High school baseball Oral examination by dentist, brief tobacco I: 23% biochemically
athletes counselling, peer-led component No intervention C: 13% RR calculated by Carr and Ebbert (2012)
Spit tobaccoa (video, graphic slides, group RR (95% CI): Initiation of SLT use: 27% in
24-month follow-up discussion) 2.03 (0.89–4.60) intervention group, 28% in control
group (RR, 1.03; 95% CI, 0.75–1.41)
Gansky Cluster RCT 883 (27 colleges) 702 (25 colleges) At 12 months, Strength: long-term follow-up
et al. (2005) Males, aged 17– 285 SLT users 352 SLT users quit rate: Limitation: results were not confirmed
USA 20 years Dental component: oral cancer Usual anti-tobacco I: 36% biochemically
Collegiate baseball screening examination by dentist and/ education offered C: 37% RR calculated by Ebbert et al. (2015)
athletes or DH, brief advice to quit SLT use, at their colleges; RR (95% CI): Initiation of SLT use: 5.1% in
Spit tobaccoa self-help guide tailored to baseball all intervention 0.98 (0.80–1.20) intervention colleges, 8.4% in control
12-month follow-up athletes, brief counselling by DH, materials were colleges (RR, 0.58; 95% CI, 0.35–0.99)
follow-up by certified athletic trainer distributed at the Of the SLT-only users at baseline, 4%
(group sessions), referral to tobacco- end of the study reported at follow-up that they had
cessation counsellors on campus or in stopped SLT use but had initiated
the community (for athletes wanting smoking
more intensive support and problem- Of the dual users at baseline, 14%
solving) reported at follow-up that they had quit
Peer-led component: videos (one SLT use but continued to smoke
tailored to baseball athletes), slide
presentation, discussion
Stigler et al. Cohort study 4009 (16 schools) 4360 (16 schools) At 12 months, Strength: long-term follow-up
(2007) (Project MYTRI) Classroom activities (curriculum), Delayed quit rate: Limitation: results were not confirmed
India Male and female school posters, parent postcards, peer- intervention I: 1.1% biochemically
students, aged led health activism C: 0.9%
10–16 years [RR (95% CI):
School students in 1.23 (0.88–1.72)]
grade 6–9
Chewing tobaccoa,b
12-month follow-up
Walsh et al. Cluster RCT 2270 2461 At 12 months, Strength: long-term follow-up
(2010) Males, aged 14– 123 SLT users 123 SLT users quit rate in Limitations: results were not confirmed
USA 18 years Peer-led educational session (video, No intervention baseline non- biochemically; confounded by smoking
Baseball players slide presentation, discussion), oral smokers: in some participants
SLTa examination with feedback, brief I: 62% Prevalence of SLT initiation in baseline
12-month follow-up advice to quit SLT use, self-help guide, C: 36% non-SLT users:
follow-up oral examination by nurse, [RR (95% CI): Overall: 3% in intervention group, 3%
nurse-led group cessation counselling 1.70 (1.50–1.86)c] in control group (OR, 0.78; 95% CI,
sessions 0.49–1.23)
In baseline non-smokers: 2% in
intervention group, 3% in control group
(OR, 0.63; 95% CI, 0.36–1.13)
In baseline smokers: 9% in intervention
group, 7% in control group (OR, 0.88;
95% CI, 0.51–1.51)
SLT-only users at baseline (i.e. baseline
non-smokers) reported a significantly
higher percentage of smoking at follow-
up (19.4%)
Danaher RCT 857 859 At 6 months, quit Limitation: results were not confirmed
et al. (2013) (MyLastDip Enhanced condition: personalized Basic condition: rate: biochemically
USA programme) best-practices SLT cessation online version of I: 22.6% RR calculated by Ebbert et al. (2015)
Males (majority) programme with interactive and a self-help guide, C: 21.9%
and females, aged multimedia features, resource section resource section RR (95% CI):
14–25 years with informational materials with informational 1.07 (0.87–1.31)
SLTa materials, links
6-month follow-up to websites
with content on
SLT cessation
and relaxation

strategies
AN, areca nut; C, control; CI, confidence interval; DH, dental hygienist; GEE, generalized estimating equation; I, intervention; MYTRI, Mobilizing Youth for Tobacco-Related Initiatives
in India; OR, odds ratio; RCT, randomized controlled trial; RR, relative risk; SLT, smokeless tobacco.
a SLT only.
b AN with tobacco.
c The calculation did not adjust for the fact that the OR reported by the authors comes from a GEE model that adjusted for clustering at schools.
249
users (controls) used the simple website, which 3.3.2 Pharmacological interventions
provided guidelines in static text. At 6 months, in
the intention-to-treat analysis, the cessation rate This section reviews studies assessing the
of SLT use was 22.6% in the intervention group effectiveness of pharmacological interven-
and 21.9% in the control group (RR, 1.07; 95% CI, tions alone for cessation of SLT or areca nut
0.87–1.31; Ebbert et al., 2015). [Results were not use. Nicotine replacement therapy, such as
confirmed biochemically.] nicotine gums, lozenges, patches, and inhalers,
A 2-year school-based, multicomponent and non-nicotine agents such as bupropion
tobacco intervention, called Project MYTRI: and varenicline are used as pharmacological
Mobilizing Youth for Tobacco-Related Initiatives interventions for tobacco cessation (Aubin et al.,
in India, was conducted in two large cities (Delhi 2014). A total of 3 RCTs were considered; one
and Chennai) at 32 schools (16 schools in the was conducted in India (Raja et al., 2016), one
intervention arm and 16 in the control arm) with in the USA (Severson et al., 2015), and one in
two cohorts of students who were in grades 6 and Taiwan (China) (Hung et al., 2020) (Table 3.11).
8, aged 10–16 years, when the study began (Stigler (a) Nicotine replacement therapy
et al., 2007). Three surveys were conducted: the
first at baseline in 2004, the second at the midpoint A worksite-based RCT in India evaluated
in 2005, and the third at the end of the interven- and compared the effectiveness of nicotine gum
tion in 2006. The intervention was carried out (2 mg strength) and oral health education in
by trained field staff, teachers, and peer leaders 40 male users of SLT and areca nut (Raja et al.,
and consisted of four primary components: 2016). The tobacco abstinence rate (biochem-
(i) behavioural, (ii) awareness generation with ically confirmed by cotinine levels in urine) at
classroom activities and posters, (iii) parental 3 months was higher in the nicotine gum group
involvement, and (iv) peer leadership in health than in the group that received oral health
activism. The controls received a delayed inter- education, but the difference was not statisti-
vention. At 12 months, the cessation rate of SLT cally significant [RR, 1.33; 95% CI, 0.70–2.56].
use was 1.1% in the intervention arm and 0.9% [Limitations of this study are the small sample
in the control arm [RR, 1.23; 95% CI, 0.88–1.72]. size, the short follow-up period, no mention of
[Results were not confirmed biochemically, and nicotine gum treatment period or frequency of
RR with 95% CI and intention-to-treat analysis gum intake, and the presence of one bidi smoker
were not provided by the authors. The Working in the nicotine gum group.]
Group noted that in this study the term “SLT” A web-based study (Severson et al., 2015),
may include products with SLT only and prod- conducted in 1067 users of SLT in the USA,
ucts with areca nut and tobacco, because both are assessed the effectiveness of three separate inter-
predominant in India.] ventions for SLT cessation: (a) nicotine lozenge
In the recent meta-analysis by Nethan et al. (4 mg for 12 weeks) together with telephone
(2020), behavioural interventions for SLT cessa- counselling for 3 weeks (intervention arm),
tion did not prove effective in youth overall (RR, (b) nicotine lozenge (4 mg for 12 weeks) alone,
1.07; 95% CI, 0.73–1.41), in developed countries and (c) telephone counselling only. In the study,
(RR, 1.39; 95% CI, 0.58–2.21), or in developing groups (b) and (c) were considered as two control
countries (RR, 0.87; 95% CI, 0.68–1.07). Of arms. [The Working Group assessed the results
the 3 studies included in the meta-analysis, 2 of the lozenge-only arm (b), taking the telephone
studies (Stigler et al., 2007; Walsh et al., 2010) are counselling-only arm (c) as the control arm.
summarized above. There was no significant difference between the
250
Table 3.11 Pharmacological interventions for cessation of smokeless tobacco and/or areca nut use
Recruitment
Nicotine replacement therapy
Severson et al. (2015) RCT (a) Nicotine lozenge (c) 3 coaching 7-day repeated PP all-tobacco Strength: large sample size (> 100
USA Male (98%) users of (4 mg for 12 weeks) callsd (for 3 weeks) abstinence rate at 3-month for each arm)
SLTa only who were plus 3 coaching callsd (n = 354) and 6-month assessments Limitations: unstated allocation
ready to quit (for 3 weeks) (n = 357) (ITT) (self-reported): concealment; no biochemical
Web-based (b) Nicotine lozenge (a) 43.1% validation test
3-month and 6-month (4 mg for 12 weeks) (b) 32.6% Note: In this study, the two
follow-up (n = 356) (c) 31.6% groups (b) nicotine lozenge-
Lozenge alone (b) versus only group and (c) telephone
coaching calls alone (c): counselling-only group were
[RR (95% CI): considered as two control arms.
1.02 (0.87–1.19)] Based on the abstinence rates, the
Working Group could estimate
the effectiveness of the lozenge-
only intervention, by comparing
(b) versus (c)
Raja et al. (2016) Parallel RCT NRT group: nicotine Oral health Tobacco abstinence rate at Limitations: small sample size; no
India Male users of khainia gum (2 mg, depending education group 3 months (urinary cotinine): mention of NRT treatment period
and paan masala c on frequency of (n = 20) [I: 25% or frequency of gum intake; no
Worksite-based tobacco intake) C: 15%] information on mean age of study
3-month follow-up (n = 20) [RR (95% CI): participants; 1 smoker (bidi) was
1.33 (0.70–2.56)] included in the NRT group; loss
to follow-up was treated as non-
abstinent

251
252

Recruitment
Non-nicotine replacement therapy
Hung et al. (2020) Double-blind RCT Escitalopram (SSRI, Placebo (identical- Continuous abstinence rate Strength: this is a novel study of
Taiwan (China) Male users of AN 10 mg/day for 8 weeks) appearing) (n = 37) (ITT) for ≥ 6 weeks after prescribing antidepressants for
(e.g. BQ)c with (n = 38) 8-week treatment (urinary cessation of AN use
cigarette smoking Moclobemide arecoline): Limitations: the outcomes were
habits (except for 2 (reversible MAOI, Escitalopram: 34.2% confounded by the cigarette
participants assigned 150 mg/day for Moclobemide: 33.3% smoking habits; an assignment of
to moclobemide group) 8 weeks) (n = 36) Placebo: 5.4% behavioural therapy group as a
Health-care setting- Escitalopram versus placebo: control arm is lacking
based Adjusted proportion ratio:
6.3 (95% CI, 1.5–26.1)
[RR (95% CI):
6.33 (1.53–26.14)]
Moclobemide versus placebo:
Adjusted proportion ratio:
6.8 (95% CI, 1.6–28.0)
[RR (95% CI):
6.17 (1.48–25.64)]
AN, areca nut; BQ, betel quid; CI, confidence interval; ITT, intention-to-treat analysis; MAOI, monoamine oxidase inhibitor; NRT, nicotine replacement therapy; PP, point prevalence;
RCT, randomized controlled trial; RR, relative risk; SLT, smokeless tobacco; SSRI, selective serotonin reuptake inhibitor.
a SLT alone.
b AN with SLT.
c AN alone.
d Three planned proactive telephone counselling calls: 1 week after randomization for initial call, 2–3 days after the quit date, and 14–21 days after the second call.
two groups in the all-tobacco abstinence rate at 3.3.3 Combined pharmacological and
the 3-month and 6-month assessments of self-re- behavioural interventions
ported 7-day repeated point prevalence (RR,
1.02; 95% CI, 0.87–1.19). The main strength of This section reviews studies assessing the
this study is the large sample size. Limitations of effectiveness of pharmacological interventions in
this study are unstated allocation concealment combination with behavioural interventions for
and no biochemical validation test.] cessation of SLT or areca nut use. In addition to
the study selection criteria mentioned earlier, the
(b) Non-nicotine replacement therapy review excluded studies on a pharmacological
In Taiwan (China), areca nut (including intervention alone and studies with no placebo
betel quid) products are consumed without or behavioural intervention in the control group.
tobacco (Lee et al., 2011). A double-blind RCT A total of 16 RCTs were evaluated (Table 3.12),
was conducted in 111 male users of areca nut and of which 2 were on nicotine gum (Boyle, 1992;
betel quid, to assess the effectiveness of the anti- Hatsukami et al., 1996), 4 on nicotine patch
depressants escitalopram (a selective serotonin (Howard-Pitney et al., 1999; Hatsukami et al.,
reuptake inhibitor; 10 mg/day for 8 weeks) and 2000; Stotts et al., 2003; Ebbert et al., 2013), 4
moclobemide (a reversible monoamine oxidase on nicotine lozenge (Ebbert et al., 2009, 2010;
inhibitor; 150 mg/day for 8 weeks) in treating Danaher et al., 2015; Severson et al., 2015), 3 on
areca nut or betel quid use disorder or areca nut bupropion (Glover et al., 2002; Dale et al., 2002,
addiction (Lee et al., 2018; Hung et al., 2020). 2007), and 3 on varenicline (Fagerström et al.,
Follow-up was every 2 weeks for the 8-week trial. 2010; Ebbert et al., 2011; Jain et al., 2014). Most
The primary outcome was cessation of areca nut of the studies were conducted in the USA (Boyle,
chewing, which was defined as patients who had 1992; Hatsukami et al., 1996, 2000; Howard-
quit use of areca nut products continuously for Pitney et al., 1999; Dale et al., 2002, 2007; Glover
≥ 6 weeks. After 8 weeks of treatment, 34.2% of et al., 2002; Stotts et al., 2003; Ebbert et al., 2009,
participants in the escitalopram group, 33.3% in 2010, 2011, 2013; Danaher et al., 2015; Severson
the moclobemide group, and 5.4% in the placebo et al., 2015); one study was in Norway and Sweden
group quit use of areca nut products continu- (Fagerström et al., 2010), and one study was in
ously for ≥ 6 weeks. The adjusted proportion India (Jain et al., 2014). One study was conducted
ratio for areca nut chewing cessation (adjusted specifically in adolescents aged 14–19 years
for age, education level, cigarette smoking, and (Stotts et al., 2003). [The Working Group noted
the level of betel quid use disorder) was 6.3 (95% that in all the studies the same behavioural inter-
CI, 1.5–26.1) for escitalopram [RR, 6.33; 95% vention was given in both the intervention arm
CI, 1.53–26.14] and 6.8 (95% CI, 1.6–28.0) for and the control arm, which may limit the evalu-
moclobemide [RR, 6.17; 95% CI, 1.48–25.64], ation of the combined effect of pharmacological
compared with the placebo group. [This is an and behavioural interventions compared with no
innovative study prescribing antidepressants intervention.]
for cessation of use of areca nut or betel quid, (a) Nicotine replacement therapy
but it has limitations such as being confounded
by cigarette smoking and lack of behavioural (i) Nicotine gum
therapy in the control arm.] Boyle (1992) conducted the first RCT for SLT
cessation using nicotine replacement therapy in
100 users of moist snuff in the USA. The study
investigated the effectiveness for SLT cessation
253
254

Table 3.12 Combined pharmacological and behavioural interventions for cessation of smokeless tobacco and/or areca nut
use

Recruitment
Nicotine replacement therapy: nicotine gum
Boyle (1992) RCT Nicotine gum (2 mg, 12 pieces/ Placebo gum with Continuous all- Loss to follow-up was
USA Male users of SLT* only day) for 6 weeks with group group meeting and tobacco abstinence treated as non-abstinent
Mass/social media-based meeting and group social group social support rate at 6 weeks Different validation tests
*Moist snuff and chewing support for behavioural skills for behavioural skills (CO and tobacco were used at baseline and
tobacco training (20–60 minutes/week) training (20– alkaloids): during follow-up
6-week follow-up for 4 weeks (n = 50) 60 minutes/week) for Nicotine gum: 50% Limitations: control
4 weeks (n = 50) Placebo gum: 40% arm also received the
[RR (95% CI): behavioural intervention; in
1.25 (0.81–1.94)] biochemical validation tests
at baseline, saliva cotinine
levels were significantly
higher in the active gum
group; short follow-up
Hatsukami et al. RCT (a) 2 mg of nicotine gum (at (b) Placebo gum with 7-day PP SLT Loss to follow-up was
(1996) Male users of SLT* only least 6 pieces/day initially, group behavioural abstinence rate at treated as non-abstinent
USA who were motivated to quit then decrease) with group therapy** (n = 50) 12-month follow-up Limitations: control
(not regular users of other behavioural therapy** for (d) Placebo gum with (salivary cotinine): arms also received the
forms of tobacco products) 8 weeks (n = 55) minimal contact*** (a) 34.5% behavioural interventions;
Mass/social media-based (c) 2 mg of nicotine gum with (n = 54) (b) 26% not enough description
*Spit tobacco minimal contact*** for 8 weeks (c) 17.6% of the approach of group
12-month follow-up (n = 51) (d) 27.8% allocation of participants,
**Group behavioural therapy: Nicotine gum (a) although it was mentioned
8 sessions (45–60 minutes each versus placebo gum that they were randomized
over 10 weeks) (b) with group
***Minimal contact: 4 brief behavioural therapy:
sessions by nurse, self-help [RR (95% CI):
booklet 1.20 (0.83–1.74)]
Nicotine gum (c)
versus placebo gum
(d) with minimal
contact:
[RR (95% CI):
0.72 (0.41–1.26)]
Recruitment
Nicotine replacement therapy: nicotine patch
Howard-Pitney et al. Double-blind RCT 15 mg nicotine patch for Placebo patch plus 7-day PP SLT Loss to follow-up was
(1999) Male users of SLT* only 6 weeks plus minimal-contact minimal-contact abstinence rate treated as non-abstinent
USA (98% non-smokers) who behavioural therapy** for behavioural at 6 months after Strength: large sample size
were motivated to quit 6 weeks (n = 206) therapy** for 6 weeks treatment (salivary (≥ 100 in each arm)
Mass/social media-based **2 pharmacy visits (with (n = 204) cotinine): Limitations: control
*Chewing tobacco trained pharmacist), 2 support Active patch: 38% arm also received the
6-month follow-up calls (48 hours and 10 days Placebo patch: 34% behavioural intervention;
after the quit date), self-help [RR (95% CI): high relapse rate in both
materials 1.09 (0.90–1.33)] groups; at the 6-month
follow-up, the response
rate was low (74%) and the
distribution by group was
not described
Hatsukami et al. Double-blind RCT (a) Active nicotine patch (c) Placebo patch Continuous all- Loss to follow-up was
(2000) Male users of SLT* only (including tapering period of plus mint snuff for tobacco abstinence treated as non-abstinent
USA who were ready to quit 21 mg for 6 weeks, 14 mg for 10 weeks (n = 101) rate at 62-week No evidence of the effect
(not regular users of other 2 weeks, and 7 mg for 2 weeks) (d) Placebo patch assessment (saliva of mint snuff, and no
forms of tobacco products) plus mint snuff for 10 weeks and no mint snuff for cotinine): interaction with nicotine
Mass/social media-based (n = 100) 10 weeks (n = 101) (a) 33% patch (a versus b)
*Spit tobacco (b) Active nicotine patch (b) 29% Strength: large sample size
62-week follow-up (including tapering period, (c) 21% (≥ 100 in each arm)
same as group a) and no mint (d) 28% Limitations: control
snuff for 10 weeks (n = 100) Active patch (b) arms also received the
Individual brief behavioural versus placebo patch behavioural intervention;
interventions (10 minutes) with (d): not enough description
self-help manual were given for [RR (95% CI): of the approach of group
all groups at 8 visits 1.03 (0.76–1.39)] allocation of participants,
although it was mentioned

that they were randomized
255
256

Recruitment
Stotts et al. (2003) RCT (a) Nicotine patch with (b) Placebo patch 7-day PP spit tobacco Limitations: high dropout
USA Adolescent male users 50 minutes of behavioural with 50 minutes abstinence rate rate in the control group
of SLT* only who were intervention** for 6 weeks of behavioural (ITT) at 12 months as a result of knowing
motivated to quit (n = 98) intervention** for (salivary cotinine): that they had no chance
Youth-targeted (ages **Based on National Cancer 6 weeks (n = 100) (a) 17.3% of receiving NRT; a few
14–19 years) Institute educational materials, (c) Usual care: (b) 25.0% participants also smoked
*Spit tobacco (snuff and/or and invited for a free oral 5–10-minute (c) 11.4%
chewing tobacco) screening counselling with Active patch (a)
12-month follow-up follow-up telephone versus placebo (b):
call 2 weeks later [RR (95% CI):
(n = 105) 0.69 (0.40–1.20)]
Active patch (a)
versus usual care (c):
[RR (95% CI):
1.52 (0.77–3.01)]
Ebbert et al. (2013) Phase II RCT Nicotine patch (two 21 mg Identical-appearing Prolonged all- Loss to follow-up was
USA Male heavy users of SLT patches/day for 6 weeks and one placebo patch tobacco abstinence treated as non-abstinent
only (aged 18–55 years) 21 mg patch/day for 2 weeks) for 8 weeks with rate at 6 months Limitations: control
who use ≥ 3 cans or with behavioural intervention* behavioural (urinary anabasine): arm also received the
pouches per week and no (n = 25) intervention* Nicotine patch: 32% behavioural intervention;
other tobacco products *Individualized sessions (n = 27) Placebo patch: 19% the dropout rate was higher
Mass/social media-based (4 study visits during the [RR (95% CI): in the placebo group
6-month follow-up medication phase) with self- 1.41 (0.81–2.47)]
help manual, minimum of
10 minutes, delivered by trained
research staff
Recruitment
Nicotine replacement therapy: nicotine lozenge
Ebbert et al. (2009) RCT pilot study Nicotine lozenge (4 mg Placebo lozenge Prolonged SLT Loss to follow-up was
USA Adult (97% male) users of for 12 weeks) with brief (for 12 weeks) with abstinence rate treated as non-abstinent
SLT* only who were ready behavioural counselling** at brief behavioural at week 24 (no Strength: large sample size
to quit each visit (n = 136) counselling** at each verification by (≥ 100 in each arm)
Mass/social media-based **Including best-practice topics visit (n = 134) urinary cotinine): Limitations: control
*Snuff (10 minutes long, at week 2, 4, 6, Nicotine lozenge: arm also received the
6-month follow-up and 12), tailored to participant 30.2% behavioural intervention;
quitting status Placebo lozenge: a higher percentage
23.1% of the active group
[RR (95% CI): (18.3%) had biochemical
1.19 (0.93–1.52)] disconfirmation of the self-
reporting compared with
the placebo group (5.1%) in
week 12 (end of medication)
Ebbert et al. (2010) RCT pilot study Nicotine lozenge (4 mg for Placebo lozenge Prolonged SLT Loss to follow-up was
USA Adult (97% male) users of 12 weeks) with assisted self-help (for 12 weeks) with abstinence rate treated as non-abstinent
SLT* only who wanted to intervention** (n = 30) assisted self-help (self-reported) at Limitations: control
quit **Assisted self-help by a self- intervention** 6 months: arm also received the
Mass/social media-based help quitting guide, telephone (n = 30) Nicotine lozenge: behavioural intervention;
*Snuff support (5–15 minutes) by 27% small sample size; no
6-month follow-up trained study assistants Placebo lozenge: 38% biochemically confirmed
[RR (95% CI): abstinence; unstated
0.79 (0.43–1.43)] randomization method
Danaher et al. (2015) RCT Interactive web-based Web-based 7-day repeated PP Strength: large sample size
USA Adult (98% male) users of intervention* plus lozenge intervention* only SLT abstinence rate (≥ 100 in each arm)
SLT only who wanted to (4 mg) for 12 weeks (n = 205) (n = 202) (ITT) at 6 months Limitations: no placebo
quit *Automated email reminders (self-reported) lozenge was given to control
Web-based encouraged engagement with (primary outcome): arm; control arm also

6-month follow-up the programme before and after Intervention arm: received the behavioural
the quit date (supportive emails 45.9% intervention; unstated
sent 2 days, 1 week, and 2 weeks Control arm: 39.1% allocation concealment; no
after the quit date) [RR (95% CI): biochemical validation test
1.15 (0.95–1.39)]
257
258

Recruitment
Severson et al. (2015) RCT (a) Nicotine lozenge (4 mg for (b) Nicotine lozenge 7-day repeated Strengths: large sample size
USA Adult (98% male) users of 12 weeks) plus 3 coaching calls* (4 mg for 12 weeks) PP all-tobacco (≥ 100 in each arm)
SLT only who were ready (for 3 weeks) (n = 357) (n = 356) abstinence rate Limitations: unstated
to quit *3 planned proactive telephone (c) 3 coaching at 3-month and allocation concealment; no
Web-based counselling calls: 1 week after calls* (for 3 weeks) 6-month assessments biochemical validation test
3-month and 6-month randomization for initial call, (n = 354) (ITT) (self-reported):
follow-up 2–3 days after the quit date, and (a) 43.1%
14–21 days after the second call (b) 32.6%
(c) 31.6%
(a) versus (b):
[RR (95% CI):
1.24 (1.08–1.44)]
(a) versus (c):
[RR (95% CI):
1.27 (1.10–1.47)]
Non-nicotine replacement therapy: bupropion SR

Glover et al. (2002) Double-blind RCT Bupropion SR 150 mg (for Placebo (for 7 weeks: 7-day PP SLT Loss to follow-up was
USA Male users of SLT* only 7 weeks: 150 mg once a day 1 tablet once a abstinence rate treated as non-abstinent
who were motivated to quit for 3 days, 150 mg twice a day for 3 days, 1 at 12 weeks after Limitations: control
(smokers excluded) day for days 4–49) and brief tablet twice a day treatment (CO, arm also received the
Mass/social media-based counselling** (n = 35) for days 4–49) and cotinine, and behavioural intervention;
*Moist snuff **Trained clinician encouraged brief counselling** NicCheck): small sample size; short
12-week follow-up participants via telephone (n = 35) Intervention arm: follow-up; no mention of
(3 days after the quit date and 40% dropout or loss to follow-up
during the follow-up phase), Control arm: 26%
nurse or qualified staff provided [RR (95% CI):
brief individual counselling 1.36 (0.86–2.15)]
(< 5 minutes) at each visit
during the treatment phase
Recruitment
Dale et al. (2002) Double-blind pilot RCT Bupropion SR 150 mg (for Placebo (identical- Continuous all- Loss to follow-up was
USA Male users of SLT* only 12 weeks: 150 mg once a day appearing, same tobacco abstinence treated as non-abstinent
who were interested in for 3 days, 150 mg twice a day dosage schedule) rate at 24 weeks Limitations: control
quitting from day 4 onwards) with for 12 weeks (biochemically arm also received the
Mass/social media-based behavioural intervention** with behavioural confirmed): behavioural intervention;
*Snuff and/or chewing (n = 34) intervention** Intervention arm: small sample size; nearly
tobacco **10-minute behavioural (n = 34) 12% half of the participants (31
6-month follow-up intervention at each study visit Control arm: 12% of 68) withdrew or were
[RR (95% CI): lost to follow-up; unstated
1.00 (0.48–2.09)] randomization and double-
blinding method
Dale et al. (2007) Multicentre double-blind Bupropion 150 mg twice a day Placebo for 12 weeks Continuous all- Loss to follow-up was
USA RCT for 12 weeks plus behavioural plus behavioural tobacco abstinence treated as non-abstinent
Male users of SLT* only intervention** (n = 113) intervention** rate at week 52 Strengths: long follow-up;
who wanted to quit **All participants received (n = 112) (urinary cotinine): large sample size (≥ 100 in
Mass/social media-based an oral examination by a Intervention arm: each arm)
*Snuff and/or chewing periodontist and behavioural 18.6% Limitation: control
tobacco intervention with a manual Control arm: 21.4% arm also received the
12-month follow-up during the treatment and [RR (95% CI): behavioural intervention
follow-up phases 0.91 (0.65–1.29)]
Non-nicotine replacement therapy: varenicline
Fagerström et al. Multicentre, double- Varenicline 1 mg twice a day Placebo for Continuous SLT Loss to follow-up was
(2010) blind, placebo-controlled, (titrated up during the first 12 weeks with abstinence rate at treated as non-abstinent
Norway and Sweden parallel-group RCT week) for 12 weeks with brief brief behavioural weeks 9–26 (salivary Strength: large sample size
Adult (90% male) users of behavioural counselling* counselling* cotinine): (≥ 100 in each arm)
SLT* only (n = 213) (n = 218) Intervention arm: Limitation: control
Mass/social media-based *Simple advice and helpful 45% arm also received the
*Swedish snus tips at the discretion of the Control arm: 34% behavioural intervention
26-week follow-up investigator RR (95% CI):

1.42 (1.08–1.79)
259
260

Recruitment
Ebbert et al. (2011) Double-blind, placebo- Varenicline (0.5 mg once a Matching Prolonged SLT Loss to follow-up was
USA controlled, phase II RCT day for 3 days, then 0.5 mg placebo with abstinence rate at treated as non-abstinent
Male users of SLT only twice a day for days 4–7, then brief behavioural 6 months (urinary Limitations: control
Mass/social media-based 1.0 mg twice a day for a total of counselling** cotinine): arm also received the
6-month follow-up 12 weeks) with brief behavioural (n = 38) Intervention arm: behavioural intervention;
counselling** (n = 38) 44.7% small sample size
**Individualized programme Control arm: 31.6%
containing 4 sessions of [RR (95% CI):
counselling (10 minutes long) 1.31 (0.84–2.04)]
with an intervention manual
Jain et al. (2014) Double-blind RCT Varenicline (1 mg twice a day Matching placebo 7-day PP SLT Strength: large sample size
India Adult (97% male) SLT users for 12 weeks) with behavioural with behavioural abstinence rate at (≥ 100 in each arm)
(specific products used not counselling* (n = 119) counselling* 12 weeks (urinary Limitations: no mention of
mentioned) *6-session counselling with (n = 118) cotinine and CO) the type of SLT products
Population/community manual-based intervention (ITT): and whether they were
based Intervention arm: with or without areca nut;
12-week follow-up 25.2% the study did not observe
Control arm: 19.5% a long-term effect of the
[RR (95% CI): treatment; adherence to
1.18 (0.89–1.56)] varenicline use was low;
unstated randomization
and double-blinding
method; control arm also
received the behavioural
intervention
CI, confidence interval; CO, carbon monoxide; ITT, intention-to-treat analysis; NRT, nicotine replacement therapy; PP, point prevalence; RCT, randomized controlled trial; RR, relative
risk; SLT, smokeless tobacco; SR, sustained release.
of nicotine gum prescribed for 6 weeks along intervention. At 6 months after the treatment,
with behavioural support in the form of 4 weekly the biochemically confirmed 7-day point-prev-
group meetings and group social support (Boyle, alence SLT abstinence rate was slightly higher
1992). At the end of the 6-week study, there was in the nicotine patch group than in the placebo
no significant difference between the nicotine patch group, but the difference was not statisti-
gum arm and the placebo arm in the continuous cally significant [RR, 1.09; 95% CI, 0.90–1.33].
abstinence rate (verified by carbon monoxide [A limitation of this study is that at the 6-month
and tobacco alkaloid metabolites analysis) for follow-up, the response rate was low and the
use of any tobacco (including SLT) [RR, 1.25; 95% distribution by group was not described.]
CI, 0.81–1.94]. [Limitations of this study are the In another large trial, conducted in Minnesota
short follow-up period and that levels of salivary (USA) (Hatsukami et al., 2000), 402 adult partic-
cotinine were significantly higher at baseline in ipants (99% men) were randomly assigned to
the active nicotine gum group.] the following treatment groups for 10 weeks:
In a study conducted in Minnesota (USA), 210 (a) nicotine patch plus mint snuff, (b) nicotine
adult male users of spit tobacco were randomized patch and no mint snuff, (c) placebo patch plus
to determine the effectiveness of nicotine gum mint snuff, or (d) placebo patch and no mint
and behavioural therapy (Hatsukami et al.,1996). snuff. The participants were also given a self-
The participants were randomly assigned to the help manual, and individual brief behavioural
following groups: (a) group behavioural therapy interventions were conducted (10 minutes) at
and nicotine gum, (b) group behavioural therapy 8 visits. At the 62-week assessment (12 months
and placebo gum, (c) minimal contact and nico- after treatment), the continuous abstinence rate
tine gum, or (d) minimal contact and placebo was higher in the nicotine patch group (b) than
gum. At 12 months, there were no significant in the placebo patch group (d), but the difference
differences in 7-day point-prevalence SLT absti- was not statistically significant [RR, 1.03; 95%
nence rates between group a and group b [RR, CI, 0.76–1.39]. [A limitation of this study is not
1.20; 95% CI, 0.83–1.74] or between group c and enough description of the approach of group allo-
group d [RR, 0.72; 95% CI, 0.41–1.26]. [A limi- cation of participants, although it was mentioned
tation of this study is not enough description that they were randomized.]
of the approach of group allocation of partici- An RCT was conducted in adolescent male
pants, although it was mentioned that they were users of SLT in the USA to test the efficacy of
randomized.] nicotine patches in combination with behav-
In a systematic review of interventions for ioural intervention compared with the usual care
SLT cessation (Ebbert et al., 2015), based on these (Stotts et al., 2003). About 303 participants (aged
two trials (Boyle et al., 1992; Hatsukami et al., 14–19 years) were recruited from 41 high schools
1996), nicotine gum use did not increase absti- in Arkansas. Participants were provided with
nence compared with placebo (RR, 0.99; 95% CI, either a nicotine patch (group a) or a placebo
0.68–1.43). patch (group b) for 6 weeks along with a behav-
ioural intervention and were also invited for a
(ii) Nicotine patch
free oral screening, or were provided usual care
In a large double-blind RCT (Howard-Pitney (group c). At the 1-year follow-up, no signifi-
et al., 1999), 410 adult users of chewing tobacco cant difference was noted between the nicotine
(SLT) (99% men) received either a nicotine patch patch group and the placebo patch group [RR,
(15 mg) or a placebo patch treatment for 6 weeks 0.69; 95% CI, 0.40–1.20] or between the nicotine
combined with minimal-contact behavioural patch group and the usual-care group [RR, 1.52;
261
95% CI, 0.77–3.01]. [A limitation of this study is a A large, web-based intervention (the
high dropout rate in the control group as a result MyLastDip programme) was conducted in 407
of knowing that they had no chance of receiving adult (98% male) SLT users in the USA to evaluate
nicotine replacement therapy; a few partici- the benefits of the website and nicotine lozenge
pants also smoked. The study was conducted in (for 12 weeks) on SLT cessation (Danaher et al.,
adolescents.] 2015; Table 3.12). At 6 months, the 7-day repeated
Ebbert et al. (2013) conducted a phase II point-prevalence SLT abstinence rate for the
RCT in adult male heavy users of SLT (who used website plus lozenge group was not significantly
≥ 3 cans or pouches per week). The intervention higher than that for the website-only group [RR,
consisted of a 42 mg/day nicotine patch (for 1.15; 95% CI, 0.95–1.39]. [Limitations of this
6 weeks) followed by a 21 mg/day nicotine patch study are that no placebo lozenge was given to
(for 2 weeks) along with behavioural counselling the control arm, unstated allocation conceal-
for SLT cessation. At 6 months, the continuous ment, and no biochemical validation test.]
all-tobacco abstinence rate was higher in the Severson et al. (2015) conducted a large RCT
nicotine patch group than in the placebo patch in the USA using nicotine lozenge plus telephone
group, but the difference was not statistically counselling for SLT cessation in 1067 adult (98%
significant [RR, 1.41; 95% CI, 0.81–2.47]. [A limi- male) participants recruited through an online
tation of this study is that the dropout rate was marketing campaign. Participants were allo-
higher in the placebo group. The low power of the cated to one of three groups: (a) nicotine lozenge
test may be due to the small sample size.] (4 mg for 12 weeks) plus coaching calls (tele-
The systematic review by Ebbert et al. (2015) phone counselling), (b) nicotine lozenge (4 mg
also did not report significantly increased absti- for 12 weeks) alone, or (c) coaching calls alone.
nence with nicotine patch use (5 trials; RR, 1.13; For the telephone counselling, three planned
95% CI, 0.93–1.37). Of the 5 trials included in the proactive calls were made: 1 week after random-
systematic review, 4 trials (Howard-Pitney et al., ization for the initial call, 2–3 days after the quit
1999; Hatsukami et al., 2000; Stotts et al., 2003; date, and 14–21 days after the second call. At the
Ebbert et al., 2013) are summarized above. 3-month and 6-month assessments, the 7-day
repeated point-prevalence all-tobacco absti-
(iii) Nicotine lozenge
nence rate was higher for nicotine lozenge plus
Two randomized pilot studies were con- coaching calls (43.1%) than for nicotine lozenge
ducted in adult (mostly male) snuff users to alone (32.6%) or coaching calls alone (31.6%).
assess the effect of nicotine lozenge use (for The differences were statistically significant for
12 weeks) on SLT cessation (Ebbert et al., 2009, lozenge plus coaching calls versus lozenge only
2010; Table 3.12). Participants were randomly [RR, 1.24; 95% CI, 1.08–1.44] and for lozenge
allocated to either a nicotine lozenge group or plus coaching calls versus coaching calls only
a placebo lozenge group, combined with behav- [1.27; 95% CI, 1.10–1.47]. Overall, the all-tobacco
ioural intervention (i.e. brief behavioural coun- abstinence rates were relatively high in all three
selling) (Ebbert et al., 2009) or a self-help quitting groups. [A strength of this study is the large
guide and telephone support (Ebbert et al., 2010). sample size. Limitations are unstated allocation
At 6 months, neither of the studies showed signif- concealment and no biochemical validation test.]
icant differences in prolonged SLT abstinence In the meta-analysis by Ebbert et al. (2015),
rates: [RR, 1.19; 95% CI, 0.93–1.52] (Ebbert et al., nicotine lozenge intervention was effective in
2009) and [RR, 0.79; 95% CI, 0.43–1.43] (Ebbert helping people quit SLT use (5 trials; RR, 1.36;
et al., 2010). 95% CI, 1.17–1.59), but the quality of the evidence
262
was rated as low. Of the 5 studies included in the 0.54–1.44). Both of these trials (Dale et al., 2002,
systematic review, 4 trials (Ebbert et al., 2009, 2007) are summarized above.
2010; Danaher et al., 2015; Severson et al., 2015) (ii) Varenicline
are summarized above.
Varenicline, a partial agonist of the α4β2
(b) Non-nicotine replacement therapy nicotinic receptor (Coe et al., 2005), has been
used for smoking cessation (Gonzales et al.,
(i) Bupropion
2006; Jorenby et al., 2006; Wu et al., 2006; Cahill
Bupropion is a monocyclic antidepres- et al., 2012). Varenicline inhibits the activation of
sant that acts as a norepinephrine and dopa- dopaminergic activity caused by smoking while
mine reuptake inhibitor (Cooper et al., 1980). providing relief from the craving and withdrawal
Sustained-release bupropion has been used to symptoms associated with smoking cessation
treat nicotine dependence and for cessation in attempts (Coe et al., 2005).
cigarette smokers (Hurt et al., 1997; Jorenby A large, multicentre, double-blind, placebo-
et al., 1999; Cahill et al., 2013). controlled, parallel-group RCT was conducted in
A double-blind RCT was conducted in 70 Norway and Sweden to evaluate the efficacy of
adult male users of moist snuff in the USA, using varenicline for cessation of SLT (Swedish snus)
sustained-release bupropion (150–300 mg/day use in 431 adult (mostly male) users (Fagerström
for 7 weeks) or placebo, combined with brief et al., 2010). Participants were recruited through
counselling (< 5 minutes) (Glover et al., 2002). At newspaper advertisements and were given either
12 weeks, the 7-day point-prevalence SLT absti- varenicline (1 mg) twice daily (titrated during
nence rate was higher in the sustained-release the first week) with brief behavioural counsel-
bupropion plus brief counselling group than in ling, or placebo with brief behavioural counsel-
the placebo plus brief counselling group, but the ling, for 12 weeks with follow-up to 14 weeks
difference was not statistically significant [RR, after treatment. All participants received brief
1.36; 95% CI, 0.86–2.15]. [Limitations of this advice and helpful tips at the discretion of the
study are the small sample size and no mention investigator, together with discussion of any
of dropout or loss to follow-up rates.] topics or concerns they raised. The continuous
A double-blind pilot RCT (Dale et al., 2002) SLT abstinence rate at weeks 9–26 was signifi-
and a double-blind multicentre RCT (Dale et al., cantly higher in the varenicline plus behavioural
2007) were conducted in adult male SLT users in counselling group than in the placebo plus
the USA to assess the effectiveness of bupropion behavioural counselling group (RR, 1.42; 95%
150 mg or placebo along with behavioural inter- CI, 1.08–1.79). [A strength of this study is the
vention over a period of 12 weeks, with long-term large sample size.]
follow-up (at 6 months and 12 months, respec- Another double-blind, placebo-controlled,
tively). Neither of the studies found significant phase II RCT was conducted in the USA, using
differences in the continuous all-tobacco absti- varenicline (for 12 weeks) with brief behavioural
nence rates between the two groups. [A strength counselling for the treatment of SLT use (Ebbert
of the Dale et al. (2007) study is the large sample et al., 2011). At 6 months, the prolonged SLT
size.] abstinence rate was not significantly higher in
In the meta-analysis by Ebbert et al. (2015), the varenicline plus behavioural counselling
based on two trials, bupropion did not show group than in the placebo plus behavioural coun-
a benefit in SLT cessation (RR, 0.89; 95% CI, selling group [RR, 1.31; 95% CI, 0.84–2.04]. [A
limitation of this study is the small sample size.]
263
Another large double-blind RCT was and distribution patterns in different geograph-
conducted in 237 adult (mostly male) users of SLT ical regions. Other factors that make SLT control
in India, using varenicline (1 mg twice per day challenging include manufacturing, storage,
for 12 weeks) with behavioural counselling as the and consumption patterns, inadequate regu-
intervention (Jain et al., 2014). The end-of-treat- latory processes, and illegal trade routes, but
ment 7-day point-prevalence SLT abstinence rate SLT control is an indispensable component of
was higher in the varenicline group than in the tobacco control efforts (Sinha et al., 2018b).
placebo group, but the difference was not statis- The WHO FCTC has been acceded to by 182
tically significant [RR, 1.18; 95% CI, 0.89–1.56]. Parties as of May 2020 (WHO FCTC, 2021), and
[A strength of this study is the large sample size. progress in its implementation is at an early inter-
Limitations are that there was no mention of mediate stage for SLT (WHO, 2008). Table 3.13
the type of SLT products and whether they were gives the number of countries in which the indi-
with or without areca nut, that the study did not vidual policies have been implemented for SLT
observe a long-term effect of the treatment, and control (Mehrotra et al., 2019; WHO, 2021b).
that adherence to varenicline use was low.] The WHO MPOWER package for tobacco
In the meta-analysis by Ebbert et al. (2015), control (WHO, 2008) includes six evidence-based
pooled results from two trials of varenicline measures: monitoring tobacco use and preven-
reported a benefit in SLT cessation (RR, 1.34; tion policies (M); protecting people from tobacco
95% CI, 1.08–1.68). Both of these trials smoke (P); offering help to quit tobacco use (O);
(Fagerström et al., 2010; Ebbert et al., 2011) are warning people about the harms of tobacco (W);
summarized above. enforcing bans on tobacco advertising, promo-
tion, and sponsorship (E); and raising taxes on
tobacco (R). Two thirds of the countries monitor
3.4 Policies and their impacts SLT use. Just less than half of the countries offer
3.4.1 Control policies for smokeless tobacco help to quit SLT use, and more than one third
have a quitline. Most countries have required the
(a) Introduction placement of pictorial health warnings on SLT
The burden and the health effects of SLT use packages, but many of these are small relative to
have shown that it poses a global public health the package size. At least half of the countries
challenge, like tobacco smoking (NCI and enforce bans on advertising and promotion of
CDC, 2014). The WHO Framework Convention SLT products. Very few countries have provided
on Tobacco Control (FCTC) aims to reduce data on raising taxes on SLT (Mehrotra et al.,
consumption of all forms of tobacco (as stated 2019; WHO, 2021b).
in Article 4.4) (WHO, 2003). The sixth session of This section presents studies on the impact of
the Conference of the Parties to the WHO FCTC the above-mentioned policies in terms of reduc-
reviewed the challenges related to SLT products tion in prevalence of SLT use, increased cessa-
and recommended that the countries apply rele- tion of SLT use, thinking about quitting SLT
vant policy interventions for SLT products with use, reduction in frequency of SLT use, decrease
the same rigour as those for smoked tobacco in initiation of SLT use, or decrease in sales of
products (WHO FCTC, 2014). SLT to youth, mainly as reported in successive
However, it is difficult to have globally national surveys (after 2011) for countries with
uniform regulations and guidelines pertaining a medium to high prevalence of SLT use (i.e.
to SLT products, because of the wide variations Bangladesh, India, the Sudan, Thailand, and
in the use, type of products, tobacco markets, the USA), or from a few other resources. The
264
Table 3.13 Tobacco control policies applicable to smokeless tobacco, and number of countries where they have been
implemented
WHO FCTC policies Specific policy Data Number of Reference

applicable to SLT year countries
(%)a
Article 6: Price and tax Data on price and taxation of SLT products 2018 34 (19%) Mehrotra et al. (2019)
measures on SLT Two-point data on SLT taxation 2018 11 (6%) Mehrotra et al. (2019)
Data on price elasticity and affordability of SLT 2018 2 (1%) Mehrotra et al. (2019)
Article 9: Regulation of Ban on the display of quantitative information on relevant constituents or emissions 2021 43 (22%) WHO (2021b)
contents of SLT products of SLT
Article 10: Regulation of Mandate the display of qualitative information on relevant constituents or 2021 26 (13%) WHO (2021b)
disclosures of contents of emissions of SLT
SLT product Data on pH and free nicotine in different SLT tobacco products 2018 6 (3%) Mehrotra et al. (2019)
Article 11: Packaging and Pictorial health warnings on SLT products 2020 47 (24%) WHO (2021b)
labelling of SLT products Pictorial health warnings ≥ 50% of package size 2020 41 (21%) WHO (2021b)
Text warnings ≥ 50% of package size 2020 23 (12%) WHO (2021b)
Article 12: Education, Anti-tobacco mass media campaign 2018 65 (36%) Mehrotra et al. (2019)
communication, training, Data on adults who believe that using SLT causes serious illness 2018 19 (11%) Mehrotra et al. (2019)
and public awareness on Data on adults who noticed information about the dangers of using SLT 2018 1 (1%) Mehrotra et al. (2019)
SLT
Data on SLT users who noticed health warnings on SLT packages 2018 1 (1%) Mehrotra et al. (2019)
Tobacco use prevention is included in the school curriculum 2018 30 (17%) Mehrotra et al. (2019)
Training to prevent tobacco use in young people 2018 30 (17%) Mehrotra et al. (2019)
Non-classroom programmes or activities to teach tobacco use prevention to 2018 29 (16%) Mehrotra et al. (2019)
students
Access to teaching and learning materials about preventing tobacco use in young 2018 28 (16%) Mehrotra et al. (2019)
people
Article 13: Ban on SLT Ban on promotion on national television and radio 2020 166 (85%) WHO (2021b)
advertising, promotion, Ban on promotion in local magazines and newspapers 2020 155 (80%) WHO (2021b)
and sponsorship (TAPS) Ban on billboard and outdoor advertising 2020 158 (81%) WHO (2021b)
Ban on advertising at point of sale 2020 111 (57%) WHO (2021b)

Ban on free distribution in mail or through other means 2020 134 (69%) WHO (2021b)
Ban on promotional discounts 2020 126 (65%) WHO (2021b)
Ban on tobacco brands (product placement) on television or in films 2020 130 (67%) WHO (2021b)
Ban on tobacco products on television or in films 2020 49 (25%) WHO (2021b)
Complete ban on sponsorship 2020 66 (34%) WHO (2021b)
Fines for violations of bans on promotion and sponsorship 2020 151 (77%) WHO (2021b)
265
266

WHO FCTC policies Specific policy Data Number of Reference
applicable to SLT year countries
(%)a
Article 14: Demand Quitline available 2020 72 (37%) WHO (2021b)
reduction measures Nicotine replacement therapy available 2020 117 (60%) WHO (2021b)
concerning SLT Nicotine replacement therapy available as essential medicine 2020 47 (24%) WHO (2021b)
dependence and cessation
Nicotine replacement therapy available (cost covered) 2020 57 (29%) WHO (2021b)
Cessation support available in health facilities and/or in hospitals 2020 125 (64%) WHO (2021b)
Cessation support available in offices of health professionals 2020 78 (40%) WHO (2021b)
Cessation support available in the community 2020 80 (41%) WHO (2021b)
Article 16: Access to and Warning signboards at point of sale 2018 75 (42%) Mehrotra et al. (2019)
availability of SLT to Ban on display of tobacco products at point of sale 2020 50 (26%) WHO (2021b)
minors Ban on tobacco products in the form of sweets, toys, candies, etc. 2020 103 (52%) WHO (2021b)
Prohibition of vending machines that contain tobacco products 2020 113 (58%) WHO (2021b)
Ban on free distribution of tobacco products to minors 2018 72 (40%) Mehrotra et al. (2019)
Ban on sale of loose SLT products 2018 18 (10%) Mehrotra et al. (2019)
Penalty against sellers for violations 2018 113 (63%) Mehrotra et al. (2019)
Article 20: Research, Data on SLT use in adults 2020 125 (64%) WHO (2021b)
surveillance, and Data on recent SLT use in adults 2018 55 (31%) Mehrotra et al. (2019)
exchange of information Data on SLT use in adolescents 2020 117 (60%) WHO (2021b)
on SLT
Data on recent SLT use in adolescents 2018 70 (39%) Mehrotra et al. (2019)
Prevalence of SLT use > 10% in adults 2020 14 (7%) WHO (2021b)
Prevalence of SLT use > 10% in adolescents 2020 16 (8%) WHO (2021b)
Data on SLT-attributable major diseases risk factors 2018 10 (6%) Mehrotra et al. (2019)
SLT, smokeless tobacco; WHO FCTC, World Health Organization Framework Convention on Tobacco Control.
a 180 countries for Mehrotra et al. (2019); 195 countries for WHO (2021b).
studies are described in the order of relevance to teachers of the dangers of tobacco use, to assist
the WHO FCTC articles for which considerable them to quit tobacco use (Sorensen et al., 2013),
progress has been shown (Articles 4–6, 11–14, 16, and to assess the implementation of the tobacco
and 20, and bans on SLT products). control polices (Mathur et al., 2016). The inter-
vention, called the Tobacco-Free Teachers/
(b) Articles 4 and 5: Prevention of initiation of Tobacco-Free Society Program, focused on
smokeless tobacco use in youth tobacco control policies, educational efforts, and
The Global School Personnel Survey (GSPS) cessation support. The control group received
conducted in 2000 in the state of Bihar in India delayed intervention. At baseline, about one
reported that nearly 78% of school personnel, third of teachers and staff used SLT and 7%
including teachers, used tobacco (Sinha et al., were smokers. At 30 days after the intervention,
2002; Sorensen et al., 2005). the self-reported adjusted cessation rate of SLT
In the GYTS conducted in students in grades use was 49.6% in the intervention cohort and
8, 9, and 10 (generally aged 13–15 years) in 50 15.4% in the control cohort (P < 0.05), whereas
state government schools and 50 federal (central at 6 months, the adjusted cessation rate was
government) schools in Bihar, a significantly 18.5% in the intervention cohort and 7.3% in
higher prevalence of ever and current tobacco the control cohort (P = 0.06). When the analysis
use (for both smoking and SLT use) was found was restricted to teachers who were employed at
in students in state government schools without the school for the entire intervention period, the
tobacco-free policies than in students in federal adjusted 6-month cessation rate was 20% in the
schools with tobacco-free policies. Classroom intervention cohort and 6% in the control cohort
teaching about the harmfulness of tobacco use (P = 0.04) (Sorensen et al., 2013). About 97% of
to health was also much more common in federal the intervention schools posted “no tobacco”
schools. Students in state schools were much signboards. Also, 84.5% of the intervention
more likely to have friends who used tobacco schools adopted the recommended tobacco
compared with students in federal schools (Sinha control policy; this percentage was much higher
et al., 2004a). When the school personnel were than that in the control schools (odds ratio [OR],
surveyed (Sinha et al., 2004b), a significantly 7.54; 95% CI, 4.92–11.60). The percentage of
higher prevalence of smoking and SLT use was schools where tobacco was sold within 100 yards
found in state schools than in federal schools. [~91 m] of the school decreased from 32.0% to
More than half of the personnel in the federal 24.9% in the intervention schools and increased
schools knew about the policy prohibiting from 26.2% to 28.4% in the control schools (OR,
tobacco use by personnel and students and about 0.77; 95% CI, 0.54–1.11) (Mathur et al., 2016).
the means of enforcement. Teaching about the
health consequences of tobacco use was carried (c) Article 6: Price and tax measures on
out to some extent in the federal schools but not smokeless tobacco
in the state schools, and the federal schools had Price increases and/or increased taxation on
some access to teaching materials on this topic. SLT products have caused a decrease in the prev-
More than 90% of all personnel in both types of alence of SLT consumption, just like for smoked
schools supported a policy prohibiting tobacco tobacco products (Table 3.14).
use in schools. A study conducted in the USA (Huang and
An RCT was conducted in teachers and staff Chaloupka, 2012) assessed the impact of the
of grades 8–10 in 72 state government schools in 2009 federal tobacco excise tax increase (effective
Bihar, in 2009–2010 and 2010–2011, to inform on 1 April 2009, causing a price increase of 12%)
267
Table 3.14 Article 6: Effect of taxation and price increases on price elasticity for use of smokeless
tobacco products with or without areca nut
Reference Data source (dates) Estimated price elasticity

Location
Policy
John (2008) 55th round of the National Sample Survey For leaf tobacco consumption and expenditure for
India (1999–2000) purchase:
Price increase In rural areas: −0.871 (0.02)a
In urban areas: –0.874 (0.03)a
Huang and Chaloupka (2012) Monitoring the Future Surveys (2008 and For smokeless tobacco: −1.2 to −1.8
USA 2009)
Taxation and price increase
Joseph and Chaloupka (2014) Global Youth Tobacco Survey, India For gutka: −0.58
India (1999–2004)
Taxation and price increase
Nargis et al. (2014) International Tobacco Control- For zarda:
Bangladesh Bangladesh Wave 3 Survey (2011–2012) Lower-priced brands: −0.64
Higher-priced brands: −0.39
Cross-price elasticity with cigarettes: 0.35
Selvaraj et al. (2015) Consumer Expenditure Survey (2011– For leaf tobacco, by income group:
India 2012) Lowest income: −0.557
Price increase Middle income: −0.4537
Highest income: −0.0507
a Values in parentheses are bootstrapped standard errors (bidis are complements for leaf tobacco; users of one tend to also use the other).
on the use of SLT products in youth, by using aged 13–15 years, estimated the price elasticity of
two different models. The prevalence of SLT use gutka as −0.58. This implies that a 10% increase
in youth decreased from 6.06% before the tax in the price of a pouch of gutka would reduce the
increase to 4.22% 30 days after the tax increase likelihood of someone becoming a gutka chewer
– a relative decrease of 30.37% – in the first by 5.8%.
econometric model, which did not control for John (2008) estimated the price elasticity for
the other study variables. In the second model, tobacco products for urban and rural households
which controlled for all the variables (such as in India separately, using data from the 55th
individual, family, and school-level character- round of the National Sample Survey, conducted
istics, state-level tobacco control measures, and in 1999–2000 in 120 309 households in 10 140
state tobacco control funding), SLT use in youth villages, on tobacco consumption and expendi-
decreased by 16–24%. The study also reported ture incurred during the past 30 days. For both
a price elasticity of between −1.2 and −1.8 for urban and rural households, the values are close
the prevalence of SLT use; this implies that an to 1; this implies that a change in price (e.g. an
increase of 10% in the price of SLT products increase due to taxation) would have a large
would reduce the prevalence of SLT use in youth downward effect on demand.
by about 12–18% (Huang and Chaloupka, 2012). Another study in India (Selvaraj et al., 2015)
A study on tobacco taxation and price in examined the pattern of price elasticity of three
India (Joseph and Chaloupka, 2014), which used major tobacco products (bidi, cigarettes, and leaf
the GYTS data for 1999–2004 in 73 356 students tobacco) based on household monthly per capita
268
consumption expenditure using data from the showed that a 10% price increase would reduce
nationally representative Consumer Expenditure the demand for SLT by 2.1%. The price elasticity
Survey of 2011–2012 in 101 662 households. The estimates for SLT products in high-income coun-
price elasticity for leaf tobacco, estimated using tries and low- and middle-income countries were
a simulation model, was highest in the lowest similar (coefficient, −0.2) (Jawad et al., 2018). Of
income group (−0.557), followed by the middle the 17 studies included in the meta-analysis, 2
income group (−0.4537) and the highest income studies (Joseph and Chaloupka, 2014; Nargis
group (−0.0507). This implies that a 10% increase et al., 2014) are summarized above.
in tax would reduce the consumption by about
5% in the lowest income group, by about 4% in (d) Article 11: Packaging and labelling of
the middle income group, and by 0.5% in the smokeless tobacco products
highest income group (Selvaraj et al., 2015). A study in the USA (Adkison et al., 2014)
Nargis et al. (2014) used data from the third evaluated the association of three elements of
wave of the International Tobacco Control Survey SLT packaging (health warning labels, descrip-
in Bangladesh in 2011–2012 to estimate the price tive characteristics, and corporate branding)
elasticity of the most commonly used SLT product with knowledge of health risks and perceptions
in Bangladesh, zarda, and the cross-price elas- of novelty and appeal, by using a web-based
ticity for zarda with respect to cigarettes. The esti- survey in 1000 individuals. Perception of health
mated price elasticity was −0.64 for lower-priced risks was higher with a graphic or pictorial
brands and −0.39 for higher-priced brands. This health warning than with a text warning on SLT
implies that a 10% increase in the price would packaging for both adults and young respond-
cause a reduction in the prevalence of zarda use ents (Table 3.15).
by about 6% for the lower-priced brands and by In India, pictorial health warnings have
4% for the higher-priced brands. The estimated changed substantially in content, size, and
cross-price elasticity for zarda with respect to coverage during the past decade. The first picto-
the price of cigarettes was 0.35. This implies rial health warning on SLT packages (a symbol
that a 10% increase in the price of cigarettes of a scorpion), which covered < 30% of the front
with the price of zarda remaining unchanged of the package, was released in May 2009, just
would result in an increase of about 3.5% in before the GATS-1 (in 2009–2010) in India
the consumption of zarda. Taken together, these (MOHFW and IIPS, 2010). In a study analysing
estimates signify that only if the prices of both the GATS-1 India data, SLT users who thought
cigarettes and zarda were increased by 10%, a about quitting after seeing a health warning in
reduction of 2.5% (−6% + 3.5%) would be seen in the past 30 days were significantly more likely to
the consumption of zarda. make attempts to quit compared with those who
An evaluation of the effect of the goods did not see a health warning (OR, 3.41; 95% CI,
and services tax in India on the affordability 3.12–3.73) (Singh et al., 2018). In 2011, the picto-
of tobacco products revealed that all tobacco rial health warnings consisted of photographs
products, including SLT products, had become of patients with oral cancer, which covered 40%
increasingly affordable over the previous 10 years of the front of all SLT packages; by 2016, these
and that the goods and services tax had accentu- were enlarged to cover 85% of the front and the
ated the increase in the affordability of SLT prod- back of the package. As a result, the percentage
ucts (John and Dauchy, 2021). of SLT users who noticed these health warnings
A meta-analysis of 17 studies on the price elas- increased from 62.9% in the GATS-1 to 71.6% in
ticity of demand for SLT products in 5 countries the GATS-2 (in 2016–2017), and the percentage
269
270

Table 3.15 Effects of text and graphic or pictorial health warning labels on smokeless tobacco packaging on perceptions of
health risks
Reference Study description Perceptions of health risks (%)

Location
Adkison et al. (2014) Cross-sectional web survey Text HWL Graphic HWL
USA Participants (n = 1000): Reduce health risks 23.2 10.8
Youth: 14–17 yr Consider health risks 5.4 63.6
Young adults: 18–25 yr
Most dangerous to health 4.8 28.3
Older adults: 26–65 yr
Deliver dangerous chemicals 3.9 31.8
Gravely et al. (2016) Tobacco Control Project India Survey Wave 1 Wave 2
India from 4 states (2010–2011) (2012–2013)
Adult SLT users (n = 4733) Symbolic HWL Graphic HWL
Policy assessed: change of HWLs % (95% CI) % (95% CI)
from symbol to graphic images on Among all respondents (n = 4733):
SLT packages in 2011 Aware that SLT packages contain HWLs 72.7 (67.1–77.7) 73.0 (67.3–78.1)
Respondents who noticed HWLs
Noticed HWLs at least once in a while 34.3 (28.5–40.6) 28.1 (21.8–35.4)
(n = 2154)
Among respondents who noticed HWLs
(n = 2154):
Read HWLs at least once in a while 49.4 (42.0–56.9) 50.1 (40.4–59.9)
HWLs made you think about risks of SLT at 15.0 (11.9–18.8) 17.5 (12.1–24.6)
least a little
HWLs made you think about quitting SLT at 16.8 (13.0–21.4) 19.3 (13.6–26.6)
least a little
Avoided looking at HWLs 8.1 (5.5–11.8) 11.6 (7.8–17.0)
Gave up SLT use at least once because of 31.3 (24.3–39.3) 36.7 (27.2–47.5)
HWLs
Have any intentions to quit SLT use 19.8 (14.6–26.4) 20.5 (15.2–27.0)
MOHFW and IIPS Cross-sectional national survey: GATS-1 GATS-2
(2010); TISS and Global Adult Tobacco Survey (2009–2010) (2016–2017)
MOHFW (2017) Photograph warning Photograph warning
India covering 40% of front covering 85% of front and
of package back of package
Noticed HWLs 62.9 71.6
Thought of quitting because of the HWLs 33.8 46.2
CI, confidence interval; GATS, Global Adult Tobacco Survey; HWL, health warning label; SLT, smokeless tobacco; yr, year or years.
of SLT users who thought of quitting also compared with those who did not notice these
increased, from 33.8% in the GATS-1 to 46.2% messages (OR, 1.42; 95% CI, 1.30–1.56) (Singh
in the GATS-2 (TISS and MOHFW, 2017). et al., 2018).
Gravely et al. (2016) evaluated the impact of In 2009, a national mass media communica-
the change in the health warning labels on SLT tion campaign on the dangers of SLT use, called
packaging from a single symbol (a scorpion) in the Surgeon campaign, aired on television and
2009 to four new graphic images in 2011, using radio in India for 6 weeks (Murukutla et al., 2012)
data from the Tobacco Control Project India in three languages (Hindi, English, and Sindhi)
Survey (wave 1 in 2010–2011 and wave 2 in 2012– (Vital Strategies, 2010). A nationally representa-
2013) from 4 states (Bihar, West Bengal, Madhya tive survey was subsequently conducted to eval-
Pradesh, and Maharashtra) in 4733 individuals uate the impact of the campaign in SLT users
aged ≥ 15 years. The change from a symbol to aged 16–50 years who had access to television or
graphic images did not significantly increase radio during the previous 2 months. The survey
any of the health warning label indicators or was administered to 2108 users of SLT only and
intentions to quit SLT use. However, people who 790 SLT users who also smoked (dual users).
quit using SLT were significantly more aware of Of these, 1323 users of SLT only (62.8%) and
health warning labels compared with people who 565 dual users (71.5%), or a total of 1888 users
continued to use SLT. (65.1%), were aware of the campaign. Of the
A study in 99 tobacco users (smokers and SLT respondents who were aware of the campaign,
users) in Chennai, India, assessed the impact of 62% recalled the campaign on television only,
the pictorial health warnings (photographs of 21% on both television and radio, and 16% on
throat cancer on cigarette packages and of oral radio only. Of the campaign-aware respondents,
cancer on SLT packages, covering 85% of the 72% said that the campaign made them stop and
front and back of the package) on the motiva- think. Almost 75% of the users of SLT only and
tion of tobacco users to quit. Most (84.8%) of the 77% of the dual users said that the campaign
tobacco users noticed the health warning labels made them feel concerned about the effects of
(including the text warning); 21.2% of SLT users using SLT on their health. In a logistic regression
were able to identify the picture correctly, and analysis, users of SLT only who were aware of
55.5% of tobacco users could relate the pictures the campaign were 2.4 times as likely to say that
to health problems. Including pictorial health SLT causes mouth cancer (P < 0.001) compared
warnings made 52.5% of users think about quit- with those who were not aware of the campaign,
ting, and 72.7% said that these warnings would and they were more likely to agree that quitting
motivate them to quit tobacco use. Because the SLT use would improve their health. Dual users
text warning was only in English, not everyone who were aware of the campaign were 2.3 times
could read it, but those who could not read the as likely to say that SLT causes throat cancer
text understood the pictorial warning (Bincy (P < 0.001). When respondents were asked about
et al., 2018). non-campaign-relevant statements (e.g. “SLT
use by pregnant women causes low-birth-weight
(e) Article 12: Education, communication, babies”), there was little or no difference in the
training, and public awareness on responses between those who were aware of the
smokeless tobacco campaign and those who were not. Users of SLT
In a study analysing the GATS-1 India data, only who were aware of the campaign were more
SLT users who noticed anti-SLT messages were likely to have seriously considered quitting SLT
significantly more likely to make attempts to quit use in the previous 2 months (OR, 1.6; P < 0.001)
271
and were more likely to have attempted to quit at least one People Behind the Packs campaign
in the previous 2 months (OR, 1.9; P < 0.001) message, 83.6% had seen the new graphic
compared with those who were not aware of the warning labels on tobacco packaging, and 38.1%
campaign. had made an attempt to quit. Attempts to quit
After the Surgeon campaign, a new campaign were significantly associated with having seen
was developed based on the story of Mukesh, a the new graphic warning labels on tobacco pack-
young patient (age 24 years) who died of oral aging (P < 0.001), recalling at least one People
cancer. The campaign consisted of a 30-second Behind the Packs campaign message (P < 0.001),
video message of Mukesh speaking to the public and recalling a greater number of adverse effects
from his hospital bed, after an introduction by of using tobacco products (P < 0.001). However,
the surgeon. Subtitles were used in different attempts to quit were less likely in users of SLT
languages. The video was aired for 4 weeks in 2011 only (P < 0.001) and in dual users (P < 0.01) than
by the Government of India. Apart from public in smokers (Turk et al., 2018).
awareness, the Mukesh campaign also provided
a face and a story for advocacy and policy efforts (f) Article 13: Ban on smokeless tobacco
about the harms of SLT use (including the request advertising, promotion, and sponsorship
for a ban on gutka, as part of the Voice of Tobacco (TAPS)
Victims campaign spearheaded by surgeons There is a dearth of studies on the impact of
from Tata Memorial Hospital in Mumbai, India). policies to prohibit advertising and sponsorship
The Mukesh campaign was evaluated using of SLT on quitting or attempts to quit SLT use.
street intercept interviews of tobacco users in 5 A cross-sectional study in Mumbai, India,
states representing 5 zones of India. The find- in 1373 high school students and 436 tobacco
ings showed that 71% of SLT users recalled the vendors close to their schools reported a lower
campaign, 80% rated it as believable, 79% found risk of current SLT use in students at schools in
it personally relevant, and 77% said it made them areas with higher compliance by vendors with
feel concerned about the health effects of their tobacco point-of-sale policies (OR, 0.40; 95% CI,
own SLT use (Vital Strategies, 2011; Gupta et al., 0.21–0.77) compared with students at schools in
2016a). areas with lower compliance, when controlling
In 2016, a mass communication campaign, for student-level and community-level tobacco
called the People Behind the Packs campaign, use risk factors (Mistry et al., 2019).
was started in Bangladesh, in Bengali and A cross-sectional study in 1670 students
English, to support the introduction of packaged 13–15 years was conducted in 28 randomly
based graphic warning labels and persuade selected schools in 7 areas of Khartoum in the
tobacco users (including SLT users) to heed the Sudan. The students completed a questionnaire
warnings in order to avoid the depicted tobac- about their exposure to toombak advertise-
co-related diseases. Two of the messages from ments at point of sale, the social acceptability
the communication campaign aired on 13 of toombak use, the perceived accessibility of
national television stations, and all 4 messages toombak, susceptibility to toombak, and toombak
were portrayed on billboards and community use. Despite a legal ban on advertisement at
health centre posters. A cross-sectional face-to- point of sale, 41.8% of students reported expo-
face survey was conducted within 14 days of the sure to toombak advertisements at point of sale.
television campaign in 1796 adult tobacco users Exposure to such advertisements was associated
(including SLT users) aged 16–55 years. The with male sex, older age, ever use, more social
results showed that 66.5% of users were aware of
272
acceptability, and direct accessibility of toombak quitters increased 3.6-fold at 6 months (Kumar
(Almahdi et al., 2020). et al., 2021).
(g) Article 14: Demand reduction measures (h) Article 16: Access to and availability of
concerning smokeless tobacco dependence smokeless tobacco to minors
and cessation Although 174 countries have restrictions in
In Oklahoma (USA), a state with a high prev- place to prevent minors from purchasing tobacco
alence of SLT use, a sample of 959 male users of (including SLT products) (WHO, 2021b), no
SLT only who had registered with the Oklahoma evidence is available about adequate enforcement
Tobacco Helpline in 2004–2012 were assessed for of this policy or its efficacy (Choi et al., 2014;
factors related to SLT abstinence (Mushtaq et al., Khan, 2016; Huque et al., 2017; Nyi Latt et al.,
2015). Of the 374 SLT users who completed the 2018; Cho et al., 2020).
7-month follow-up, 162 (43%) reported 30-day In July 1992, the United States Congress
abstinence, representing a 15% cessation rate. SLT enacted the Alcohol, Drug Abuse, and Mental
users with higher levels of motivation to quit at Health Administration Reorganization Act
baseline were twice as likely to be abstinent than (Public Law 102-321). Through the Synar
those with low or moderate levels of motivation Amendment to this law, the sale or distribu-
to quit (OR, 2.05; 95% CI, 1.25–3.35), and each tion of any form of tobacco to minors (aged
additional completed helpline call increased the < 18 years) was prohibited. The 2014 Annual
likelihood of tobacco cessation by 20%. Synar Report in 50 states and 8 jurisdictions
In Rajasthan, India, a quitline service was reported a decrease in the sales of all tobacco to
initiated in January 2013 as a voluntary activity minors (aged < 18 years), from 40.1% in 1997 to
(Gupta et al., 2016b) and later became a part of 9.6% in 2013 (national weighted averages). Also,
the 104 Information Helpline of the Medical the states that fined retailers for selling tobacco
and Health Department of the Government to minors had fewer violations of the Synar
of Rajasthan. Of the 1525 callers in 2013, 1105 Amendment (SAMHSA, 2014).
(72.5%) were SLT users. A self-reported cessation
rate of about 20% was observed in the SLT users (i) Bans on smokeless tobacco products
at the 18-month follow-up. This is > 11 times the This section discusses studies that reported
cessation rate of 1.6% for former daily users of the impact of the prohibition of sale, manufac-
SLT (and former daily smokers) in Rajasthan ture, and importation of SLT on its consumption
reported in the GATS-1. and the quit intentions of users, in some high-
A national tobacco quitline was started in May burden countries (i.e. those with > 1 million users
2016 in India. Of the 5179 callers who registered or a prevalence of ≥ 10% in males or females)
during the first year (Kumar et al., 2018), 3169 (Mehrotra et al., 2017).
(61.2%) were SLT users and 644 (12.4%) were dual Among the high-burden countries, Thailand
users. When the dual users were excluded, 41% of was the first to impose a ban on the importa-
SLT users successfully quit (complete abstinence tion of SLT, in 1992, and the country undertook
for ≥ 3–4 weeks). After the national quitline stringent measures for compliance with the ban.
number was included on tobacco packages, from The tobacco control programme in Thailand
September 2018, the percentage of callers who contributed to a decrease in the prevalence of
were SLT users increased from 51.1% to 70.7%, SLT use in adults from 3.9% (1.3% in men and
the number of tobacco users registering with the 6.3% in women) in 2009 (WHO Regional Office
quitline increased 3.3-fold, and the number of for South-East Asia, 2009b) to 3.2% (1.1% in men
273
Table 3.16 Impact of the gutka ban on the prevalence of gutka use in Indiaa
Reference Prevalence of gutka use (%) Relative change in

Location prevalence of use (%)
Before the ban: GATS-1b After the ban: GATS-2b
(2009–2010) (2016–2017)
Overall Men Women Overall Men Women Overall Men Women
MOHFW and IIPS (2010); 8.2 13.1 2.9 6.8 10.8 2.7 −17.1 −17.6 −6.9
TISS and MOHFW (2017)
India
GATS, Global Adult Tobacco Survey.
a The gutka ban was implemented in 2012.
b Repeated cross-sectional household survey of individuals aged ≥ 15 years, with a multistage, geographically clustered sample design.
and 5.2% in women) in 2011 (WHO Regional A study in Maharashtra, India, in 68 gutka
Office for South-East Asia, 2011) and 2.1% in 2017 users (Mishra et al., 2014) found that since the
(National Statistical Office of Thailand, 2017). ban, 24% had quit gutka use, 56% had reduced
In India, a central law in 2011 prohibited their consumption, and 21% had not changed
tobacco or nicotine from being used in any their consumption; none of the participants
food products (MOHFW, 2011c), which led to a reported an increase in their use of gutka. Some
subsequent statewise ban on the manufacture, respondents had turned to products that are
storage, and sale of gutka. A resultant decrease custom-made by vendors and contain similar
was observed in the prevalence of gutka use, ingredients (e.g. mawa, betel quid) or to another
from 8.2% in the GATS-1 to 6.8% in the GATS-2 commercially available SLT product (khaini).
(Table 3.16). However, gutka continued to be A study conducted in Andhra Pradesh, India,
available illegally, including near educational in 368 gutka users (Reddy et al., 2016) reported
institutions (Pimple et al., 2014). that most of the users (81.5%) had tried to quit
A study conducted in 2014 to assess the impact gutka use and 29.9% of the users had turned to
of the gutka ban in the Indian states of Assam, other forms of SLT products, most commonly
Bihar, Gujarat, Karnataka, Madhya Pradesh, mawa (51.8%). Also, 62.2% of the users reported
Maharashtra, Odisha, and Delhi (National that gutka was still available commercially in the
Capital Region) revealed that 92% of the popu- form of two separate sachets, one of paan masala
lation supported the ban and 99% agreed that and the other of tobacco.
it was good for the youth of the country (WHO In Bhutan, despite a comprehensive ban on
Regional Office for South-East Asia, 2014). the cultivation, manufacture, distribution, and
Interviews with 1001 current and former users sale of tobacco since 2004, the prevalence of use
of gutka revealed that 49% of current users had of tobacco, especially SLT, is high. A cross-sec-
reduced their consumption and the remaining tional analysis of the nationally representa-
51% had attempted to stop using gutka in the tive Noncommunicable Disease Risk Factors
previous year. About 41–88% of respondents Surveillance STEPS Survey 2014 in 2820 adults
across the different states reported quitting gutka in Bhutan showed a high prevalence of SLT use
use as a result of the ban. (19.7%; 95% CI, 16.5–22.9%), especially in males,
younger individuals, and people who consumed
274
alcohol (Gurung et al., 2016). An increase in SLT and in Thailand, from 3.9% in 2009 to 3.2%
use in adolescents was also noted in the GYTS in 2011 (Suliankatchi Abdulkader et al., 2019)
in Bhutan, from 18.8% in 2006 to 30.3% in 2013 (Table 3.17). After the GATS-1 in Bangladesh,
(WHO Regional Office for South-East Asia, pictorial health warnings were introduced that
2015). covered 50% of SLT packages, anti-SLT media
Since 1992, there has been a ban on the sale campaigns were conducted, direct marketing of
of tobacco for oral use (i.e. snus) in the EU except SLT was prohibited, and taxation of SLT prod-
in Sweden (Delhomme, 2019). From 2001, the ucts increased (Bangladesh Bureau of Statistics
European Commission reaffirmed that the EU and National Tobacco Control Cell, 2019). In
Member States were prohibited from placing India, the ban on the manufacture and sale of
tobacco for oral use on the market (Article 8 of gutka was implemented in 2012. In Thailand,
Directive 2001/37/EC) (European Parliament, since 2009 pictorial health warning labels are
2001). However, this ban has been evaded also required on packaging of shredded tobacco
through online sale and promotion of snus in the products (used as SLT) (WHO Regional Office
EU (Peeters and Gilmore, 2013). for South-East Asia, 2011).
The GYTS is a school-based survey of students
(j) Overall tobacco control policy and Article aged 13–15 years. Between 2007 and 2013, the
20: Research, surveillance, and exchange of prevalence of current SLT use did not change
information on smokeless tobacco significantly in Bangladesh, India, or Myanmar,
Standard, nationally representative surveys but the prevalence increased significantly in
designed to measure tobacco use and the impact Bhutan and Nepal. During this period, there was
of tobacco control policies in countries in an either an absence of effective policies focusing on
internationally comparable way were devel- SLT control or a lack of enforcement of policies in
oped jointly by the United States Centers for these countries. For instance, in India, where the
Disease Control and Prevention and WHO. Cigarettes and Other Tobacco Products Act was
These surveys include the GYTS, the GSPS, the enacted in 2004, a few court cases by the tobacco
Global Health Professions Student Survey, and industry prevented adequate implementation
the GATS, which together make up the Global of the legislation for several years. In Nepal, a
Tobacco Surveillance System. tobacco control policy was enacted in 2010, but
The GATS is a household survey that is litigation by the tobacco industry continued until
administered in male and female individuals 2014 (Sinha et al., 2014).
aged ≥ 15 years. A few of the countries with a From 2010, the Tobacco Control Act of
high SLT burden in the WHO South-East Asia Bhutan (Parliament of Bhutan, 2010) prohibited
Region, such as Bangladesh, India, and Thailand, the cultivation, manufacture, sale, and supply
have completed two rounds of the GATS since of tobacco products; it remained in effect until
2009 (WHO Regional Office for South-East 2020 (Wangdi and Gyeltshen, 2020). Awareness
Asia, 2009a, b, 2011; MOHFW and IIPS, 2010; programmes on the dangers of tobacco were also
TISS and MOHFW, 2017; Bangladesh Bureau undertaken in Bhutan (Tshering et al., 2021). In
of Statistics and National Tobacco Control Cell, Sri Lanka, from 2006, the tobacco control law
2019) (Table 3.17). In all three countries, the prev- prohibited the sale of tobacco to minors (aged
alence of SLT use decreased significantly between < 21 years) (Sinha et al., 2014). In Nepal, tobacco
the GATS-1 and the GATS-2: in Bangladesh, control laws in 2011 required graphic health
from 27.2% in 2009 to 20.6% in 2017; in India, warnings covering 75% of both the front and
from 25.9% in 2009–2010 to 21.4% in 2016–2017; the back of the package for all tobacco products;
275
276

Table 3.17 Reduction in prevalence of smokeless tobacco use in adults after policy interventions in selected countries
Reference GATS-1a GATS-2a Reduction in Policies and population-level

Location prevalence of SLT useb interventions
Year Prevalence of Year Prevalence of (relative change) (%)
No. of SLT useb (%) No. of SLT useb (%) Overall (men; women)
households Overall (men; households Overall (men;
surveyed women) surveyed women)
WHO Regional 2009 27.2 (26.4; 27.9) 2017 20.6 (16.2; 24.8) −24.1* (−38.6*; –11.3) Pictorial health warnings to cover
Office for South- 10 751 14 880 * P < 0.05 50% of SLT packages, anti-SLT media
East Asia (2009a); campaigns; marketing of SLT prohibited,
Bangladesh Bureau and increased taxation of SLT products,
of Statistics and verified by tax stamp
National Tobacco
Control Cell (2019)
Bangladesh
MOHFW and IIPS 2009–2010 25.9 (32.9; 18.4) 2016–2017 21.4 (29.6; 12.8) −17.4 (−10.0; –30.4) Manufacture and sale of gutka and
(2010); TISS and 69 296 77 170 P < 0.01 paan masala containing tobacco or
MOHFW (2017) nicotine banned by nearly all states by
India 2012 under national law; taxes on SLT
increased marginally; public awareness
campaigns on SLT in different media; in
2012, tobacco use in films was regulated;
in 2016, pictorial health warnings were
enlarged to 85% of both principal display
areas on packages
WHO Regional 2009 3.9 (1.3; 6.3) 2011 3.2 (1.1; 5.2) −17.2 (−18.0; −17.0) Pictorial health warnings and text
Office for South-East 22 768 20 922 P < 0.05 warnings on tobacco packages; taxation
Asia (2009b, 2011)
Thailand
GATS, Global Adult Tobacco Survey; SLT, smokeless tobacco.
a Repeated cross-sectional household survey of individuals aged ≥ 15 years, with a multistage, geographically clustered sample design.
b SLT use includes use of SLT only and SLT use plus smoking; prevalence of current use includes daily and occasional use.
this was implemented in 2014 (Sinha et al., 2014). tobacco use in the school survey, the prevalence
In Myanmar, tax rates for tobacco products, of tobacco use would increase to 53.9%, which
including SLT, increased in 2012 and again in is still substantially lower than the prevalence in
2015 (World Bank Group, 2020), and from 2016 the GSPS 2000 (Gupta et al., 2014a).
the size of health warnings on SLT and smoked
tobacco products was increased to 75% of both (k) Modelling the impact of a set of policies
the front and the back of the package (Tun et al., using available data
2017; Campaign for Tobacco-Free Kids, 2021), In a study conducted in Minnesota (USA),
After 2014, the prevalence of SLT use in youth Levy et al. (2019) estimated the effect of tobacco
decreased in four countries with a high SLT control policies implemented in 1993–2018 on
burden: in Bhutan, from 21.6% in 2013 (Sinha SLT use using a previous SimSmoke model,
et al., 2014; WHO Regional Office for South-East updated and extended to incorporate SLT use
Asia, 2015) to 12.5% in 2019 (WHO Regional (both use of SLT only and dual use) (Table 3.19).
Office for South-East Asia, 2020); in India, from The SimSmoke model projected that the prev-
14.0% in 2003 to 4.1% in 2019 (MOHFW and alence of SLT use in men would decrease from
IIPS, 2019); in Myanmar, from 9.8% in 2011 3.9% in 1993 to 2.6% in 2015 and to 2.5% in
(Sinha et al., 2014) to 5.7% in 2016 (Tun et al., 2018. In addition, compared with no new poli-
2017); and in Sri Lanka, from 8.5% in 2011 to cies implemented after 1993, the model projected
2.4% in 2015 (WHO Regional Office for South- that the prevalence of SLT use in men would
East Asia, 2016) (Table 3.18). In Bhutan in the decrease to 2.9% in 2040 (Levy et al., 2019). The
GYTS 2019, 87.1% of current SLT users wanted Minnesota Adult Tobacco Survey conducted in
to stop using it right away. In Bhutan, according 2014 reported only a slight decrease in the prev-
to law, tobacco cannot be cultivated and tobacco alence of SLT use, to 3.6% (Boyle et al., 2015); this
products cannot be produced. Although tobacco was contradictory to the decrease predicted by
products can be imported for personal consump- the model.
tion, there are limits on the amounts, and impor- The SimSmoke model was also used to assess
tation is prohibited for minors (aged < 18 years). the effect of past tobacco control policies and to
The advertisement, promotion, and sponsor- project the effect of future policies on the preva-
ship of tobacco are banned, except for brand lence of snus use (and smoking) in Sweden (Near
stretching (WHO Regional Office for South-East et al., 2014; Table 3.20). The model predicted that
Asia, 2020). if all of the policies were implemented, the prev-
A survey was conducted in two waves, in alence of use of snus only would decrease from
2009 and 2010, in 755 school personnel in 72 state 14.6% in 2010 to 10.4% in 2040 in men and from
government schools in Bihar, India (Gupta et al., 3.3% in 2010 to 2.8% in 2040 in women. Overall,
2014a). The reported prevalence of current use of the study showed that a combination of the poli-
tobacco (mainly SLT) was 35.5% (48.0% in men cies would have a greater impact on the preva-
and 11.3% in women), which was much lower lence of SLT use than a single policy. According
than the prevalence of 77.4% previously reported to a survey in 2010, the overall prevalence of SLT
in the GSPS 2000. Use of lal dant manjan (red use [SLT product not specified] in Sweden was
tooth powder) was considered as use of a tobacco 12.3% (20.7% in men and 3.5% in women) (Leon
product in the GSPS 2000 but not in this school et al., 2016).
study, because the inclusion of tobacco in any oral
hygiene products was prohibited by a government
order. If use of lal dant manjan was included as
277
278

Table 3.18 Reduction in prevalence of smokeless tobacco use in students aged 13–15 years after policy interventions in
selected countries
Reference Earlier GYTSa Later GYTSa Reduction in Policies and population-level interventions
Location prevalence of
Year Prevalence of SLT Year Prevalence of SLT useb (relative
useb (%) SLT useb (%) change (%)
Overall (boys; Overall (boys; Overall (boys;
girls) girls) girls)
MOHFW and IIPS 2003 14.0 (18.0; 7.9) 2019 4.1 −70.7 Cigarettes and Other Tobacco Products Act (COTPA)
(2019) in 2004; ban on the manufacture and sale of gutka in
India 2012
Sinha et al. (2014); 2013 21.6 (25.0; 18.9) 2019 12.5 (17.0; 8.1) −42.1 (−32.0; −57.1) Tobacco Control Amendment Act of 2012 to the
WHO Regional Office Tobacco Control Act of Bhutan of 2010; Tobacco
for South-East Asia Control Rules and Regulations 2013. The rules prohibit
(2020) minors (aged < 18 years) from importing tobacco or
Bhutan tobacco products, even for personal consumption.
However, SLT is available and accessible to youth
Sinha et al. (2014); 2011 9.8 (15.2; 4.0) 2016 5.7 (11.0; 1.5) −41.8 (−27.6; −62.5) From 2016, the size of health warnings on SLT and
Tun et al. (2017); smoked tobacco products was increased to 75% of the
Campaign for front and back of the package
Tobacco-Free Kids
(2021)
Myanmar
Sinha et al. (2014); 2011 8.5 (13.0; 4.1) 2015 2.4 (4.2; 0.5) −71.8 (−67.7; −87.8) The school curriculum has contained lessons on the
WHO Regional Office harmfulness of tobacco use (mainly smoking) for
for South-East Asia several years, before these surveys
(2016)
Sri Lanka
GYTS, Global Youth Tobacco Survey; SLT, smokeless tobacco.
a Repeated cross-sectional national school-based, self-administered survey of students aged 13–15 years, with a two-stage sample design.
b SLT use includes use of SLT only and SLT use plus smoking; prevalence of current use includes daily and occasional use.
Table 3.19 Modelling projections of the impact of tobacco control policies on prevalence of smokeless tobacco use in men in
Minnesota (USA) for 1993–2040
Reference Study design Tobacco control policies Prevalence of SLT use in mena (%)
Location
Actual Projection
Best (lower, upper)b
1993 2018 2040
Levy et al. (2019) SimSmoke modelling to estimate the Policies remaining at 1993 levels 3.9 3.2 (3.2, 3.2) 2.9 (2.9, 2.9)
Minnesota impact of policies on SLT use All policies (cumulative) 3.9 2.5 (2.8, 2.2) 2.1 (2.4, 1.8)
(USA) Period of policies included in model: Price policies 3.9 2.8 (2.9, 2.6) 2.5 (2.6, 2.3)
1993–2018
Smoke-free air policies 3.9 3.2 (3.2, 3.1) 2.8 (2.9, 2.8)
Used data from the 1993 Tobacco Use
Supplement and information on state Tobacco control expenditure by state 3.9 3.1 (3.2, 3.1) 2.8 (2.8, 2.8)
policies Cessation treatment 3.9 3.1 (3.1, 3.0) 2.8 (2.8, 2.7)
Health warnings policies 3.9 3.2 (3.2, 3.1) 2.8 (2.9, 2.8)
Youth access policies 3.9 3.1 (3.2, 3.1) 2.7 (2.8, 2.6)
SLT, smokeless tobacco.
a According to the model, projected prevalence rates for SLT use in women were not affected by the policies.
b Estimates are given in terms of the best estimate and the lower and upper bounds based on the policy evaluation literature.

279
Table 3.20 Modelling projections of the impact of tobacco control policies on prevalence of snus
use in Sweden
Reference Study design Tobacco control policies Prevalence of use of snus only (%)
Location Projections for 2010–2040
Men Women
2010 2011 2020 2040 2010 2011 2020 2040
Near et al. SimSmoke Status quo 14.6 14.5 14.4 13.5 3.3 3.3 3.5 3.6
(2014) modelling Newly implemented policiesa
Sweden to estimate Raise excise taxes to 70% of 14.6 13.4 13.1 11.9 3.3 3.0 3.1 3.1
the impact of retail price
policies on
Complete smoke-free 14.6 14.5 14.4 13.5 3.3 3.3 3.5 3.6
prevalence of
use of snus only Comprehensive marketing 14.6 14.4 14.2 13.3 3.3 3.3 3.4 3.5
Used data from ban
the Health on High-intensity tobacco 14.6 14.1 13.8 12.9 3.3 3.2 3.3 3.4
Equal Terms control campaign
of the National Strong health warnings 14.6 14.5 14.3 13.4 3.3 3.3 3.4 3.5
Public Health Strong youth access 14.6 14.5 14.1 12.8 3.3 3.3 3.4 3.4
Survey for enforcement
2004–2010 Cessation treatment policies 14.6 14.5 14.2 13.3 3.3 3.3 3.4 3.5
All of the above policies 14.6 12.7 12.0 10.4 3.3 2.9 2.9 2.8
implemented
aNew policies implemented at levels consistent with the World Health Organization Framework Convention on Tobacco Control (WHO FCTC)
in 2010 and maintained at the same level until 2040.
3.4.2 Control policies for areca nut products turn, has led to the adoption in several countries
(including betel quid) over the past several decades of policies designed
to control use of areca nut (Table 3.21).
Areca nut is cultivated and consumed mainly Areca nut control policies began in Thailand
in South and South-East Asia. In the past few in 1940 with a campaign promoted by the prime
decades (1994–2019), there have been increases minister to discourage betel quid chewing,
in the global production, which is highest showing that streets stained with red juice from
in India, followed by Bangladesh, Indonesia, spitting were dirty and unhygienic, and prohib-
Myanmar, and Taiwan (China), and in areca iting betel quid chewing on government prem-
nut consumption and trade (FAO, 2021). The ises (Thai Cultural Encyclopedia Foundation,
increase in consumption of areca nut in different 1999). Currently, the most common policy to
forms has led to high incidence rates of oral curb areca nut consumption as well as SLT use is
cancers and oral potentially malignant disorders, a ban on spitting in public places; this has been
especially in India (Gupta et al., 2014b), Hunan adopted by several countries, most recently in
(China) (Zhou et al., 2019), Taiwan (China) (Su India during the COVID-19 pandemic (Gunjal
et al., 2020), Bhutan, Myanmar, Nepal, Papua et al., 2020; The Economic Times, 2020; Yang
New Guinea, Pakistan, Sri Lanka, and various et al., 2020). The next most common policy is a
South Pacific islands such as Guam (USA) and ban on betel quid chewing in certain places, such
the Solomon Islands (Gunjal et al., 2020). This, in as government offices, schools, and hospitals,
280
Table 3.21 Major areca nut control policies and where they have been adopted
Policya Locations
Ban on spitting in public places Bhutan, Myanmar, Papua New Guinea, India (by the railways
only), Taiwan (China), Hangzhou City (China)
Ban on chewing betel quid in certain places Myanmar (in or near government offices, schools, and hospitals),
Sri Lanka, Taiwan (China) (in the military and in some
workplaces)
Ban on advertising of areca nut products Hunan Province (China)
Ban on manufacture and/or sale of certain areca nut India, Sindh Province (Pakistan), Xiamen in Fujian Province
products (China), Myanmar
Text warnings on packages of areca nut products India
Betel quid cessation programmes Taiwan (China)
Mass media awareness programmes Myanmar, Taiwan (China)
Plantation programme Taiwan (China)
Oral mucosal screening programme Taiwan (China)
aIn most countries, betel quid usually also contains tobacco.
Compiled by the Working Group, with data from Vital Strategies (2017); Zhou et al. (2019); Gunjal et al. (2020); The Economic Times (2020);
Yang et al. (2020); Zhao and Davey (2020).
in the military, or in certain other workplaces the Environmental Protection Administration

(Gunjal et al., 2020). There have also been mass (Yang et al., 2020). Support for areca nut cessa-
media awareness programmes about the dangers tion has been implemented with culturally sensi-
of betel quid chewing in Taiwan (China) (Yang tive educational materials, especially in high-risk
et al., 2020) and in Myanmar (Vital Strategies, communities and workplaces. An oral mucosal
2017). Currently, Taiwan (China) has the highest screening programme is available for chewers,
number of areca nut control policies, followed by former chewers, and smokers (Yang et al., 2020).
Myanmar and India. Also, clothing restrictions have been introduced
In Taiwan (China), a national areca nut for the previously scantily clad young women
and betel quid cessation programme has been (called “betel quid beauties”) who sell areca nut at
implemented since the late 1990s (Yang et al., neon-lit stalls that are frequented by taxi drivers,
2020). From 1997, 3 December was declared truck drivers, and other workers (Nylander,
Areca Prevention Day, to raise awareness of 2016). In 2014, the Council of Agriculture intro-
the carcinogenicity of areca nut through mass duced a plantation programme that helped areca
media communication, school programmes, nut growers change to other cash crops; this led
and health-care providers. The government and to a reduction of 5% in the area cultivated and of
nongovernmental organizations have created 18% in production. Since the start of such areca
areca nut-free environments to promote healthy nut prevention efforts, the prevalence of betel
behaviour and support a reduction in use of betel quid chewing in adults (aged ≥ 18 years) has
quid and areca nut in the community. Beginning decreased steadily in all age groups, from about
in 2014 in Taipei (Hsu, 2014), spitting of betel 45% in 2007 to about 5% in 2018. Also, the annual
quid juice in public places has been prohibited incidence rate of oral cancer has plateaued since
under the Waste Disposal Act and enforced by 2009 at just more than 42 per 100 000 people,
281
after increasing for several decades (Yang et al., use among adults in 44 countries during 2008–2012:
2020). evidence for an integrative and comprehensive approach
in tobacco control. Drug Alcohol Depend. 139:60–70.
In India, there is a provision under the Food doi:10.1016/j.drugalcdep.2014.03.003 PMID:24685560
Safety and Standards Act, 2006 (Ministry of Ahluwalia HS, Duguid JB (1966). Malignant tumours in
Law and Justice, 2006) for the prohibition of Malaya. Br J Cancer. 20(1):12–5. doi:10.1038/bjc.1966.2
PMID:5949191
the manufacture, storage, distribution, or sale Ahuja SC, Ahuja U (2011). Betel leaf and betel nut in India:
of any article of food product for up to 1 year. history and uses. Asian Agri-hist. 15(l):13–35.
This has been used in some states to prohibit Al-Tayar BAS, Tin-Oo MM, Sinor MZ, Abdikarim M,
paan masala (primarily see table containing Al-Akhali M (2017). Oral melanosis and severe perio-
dontitis in toombak users: a case report. Eur J Med Case
areca nut). There is also a restriction on the Rep. 2(1):9–11. doi:10.24911/ejmcr/1/35
use of anticaking agents, such as magnesium Alemi Q, Stempel C, Montgomery S (2021). Prevalence
carbonate, in food products (MOHFW, 2011a). and social determinants of tobacco use in Afghanistan.
This restriction has been used in some states to Int Health. 13(1):3–12. doi:10.1093/inthealth/ihaa010
PMID:32304214
ban paan masala, which invariably contains Ali R, Loney T, Al-Houqani M, Blair I, Aziz F, Al Dhaheri S,
magnesium carbonate. Also, since 1990, pack- et al. (2020). Cigarette smoking and smokeless tobacco
ages of paan masala and supari have text health use among male south Asian migrants in the United
Arab Emirates: a cross-sectional study. BMC Public
warnings (MOHFW, 2011b; NIHFW, 2014). Health. 20(1):815. doi:10.1186/s12889-020-08942-9
Gutka, which consists of areca nut with tobacco, PMID:32473627
has been banned statewise in India since 2012 Almahdi HM, Ali RW, Nasir EF (2020). Exposure to the
(Gunjal et al., 2020). advertisement of toombak at the point-of-sale among
adolescents in Khartoum State, Sudan: a cross-sectional
In China, the first step towards regulating study. Tob Prev Cessat. 6:19. doi:10.18332/tpc/115800
areca nut was a 2019 ban on advertising of areca PMID:32548356
nut products by companies based in Hunan (Zhou Amarasinghe AAHK, Usgodaarachchi US, Johnson NW,
et al., 2019). Also, another city in China (Xiamen, Warnakulasuriya S (2018). High prevalence of life-
style factors attributable for oral cancer, and of oral
in Fujian Province) adopted a specific anti-areca potentially malignant disorders in rural Sri Lanka.
nut policy that banned the production, sale, and Asian Pac J Cancer Prev. 19(9):2485–92. doi:10.22034/
use of areca nut (Zhao and Davey, 2020). APJCP.2018.19.9.2485 PMID:30256041
Anantha N, Nandakumar A, Vishwanath N, Venkatesh
T, Pallad YG, Manjunath P, et al. (1995). Efficacy of
an anti-tobacco community education program in
References India. Cancer Causes Control. 6(2):119–29. doi:10.1007/
BF00052772 PMID:7749051
Ansara DL, Arnold F, Kishor S, Hsia J, Kaufmann R
Abakar MAA, Omer AAI, Yousif AMM (2020). The effect (2013). Tobacco use by men and women in 49 coun-
of Sudanese smokeless tobacco (toombak) using on oral tries with demographic and health surveys. DHS
microbiota. bioRxiv. doi:10.1101/2020.04.03.023408 Comparative Reports No. 31. Calverton (MD), USA:
Abbas SM, Alam AY, Usman M, Siddiqi K (2014). Smokeless ICF International. Available from: https://dhsprogram.
tobacco consumption in a multi-ethnic community com/publications/publication-cr31-comparative-
in Pakistan: a cross-sectional study. East Mediterr reports.cfm.
Health J. 20(6):385–90. doi:10.26719/2014.20.6.385 Arora M, Reddy KS, Stigler MH, Perry CL (2008).
PMID:24960515 Associations between tobacco marketing and use among
Adkison SE, Bansal-Travers M, Smith DM, O’Connor urban youth in India. Am J Health Behav. 32(3):283–94.
RJ, Hyland AJ (2014). Impact of smokeless tobacco doi:10.5993/AJHB.32.3.6 PMID:18067468
packaging on perceptions and beliefs among youth, ASH (2019). Tobacco and ethnic minorities. London,
young adults, and adults in the U.S: findings from an UK: Action on Smoking and Health. Available from:
internet-based cross-sectional survey. Harm Reduct J. https://ash.org.uk/wp-content/uploads/2019/08/
11(1):2. doi:10.1186/1477-7517-11-2 PMID:24433301 ASH-Factsheet_Ethnic-Minorities-Final-Final.pdf.
Agaku IT, Filippidis FT, Vardavas CI, Odukoya OO, Atkinson L, Chester IC, Smyth FG, Ten Seldam REJ
Awopegba AJ, Ayo-Yusuf OA, et al. (2014). Poly-tobacco (1964). Oral cancer in New Guinea: a study in
282
demography and etiology. Cancer. 17(10):1289–98. Buente W, Dalisay F, Pokhrel P, Kramer HK, Pagano I
doi:10.1002/1097-0142(196410)17:10<1289::AID-CN- (2020). An Instagram-based study to understand betel
CR2820171011>3.0.CO;2-E PMID:14236762 nut use culture in Micronesia: exploratory content anal-
Aubin HJ, Luquiens A, Berlin I (2014). Pharmacotherapy ysis. J Med Internet Res. 22(7):e13954. doi:10.2196/13954
for smoking cessation: pharmacological principles and PMID:32673220
clinical practice. Br J Clin Pharmacol. 77(2):324–36. Cahill K, Stead LF, Lancaster T (2012). Nicotine
doi:10.1111/bcp.12116 PMID:23488726 receptor partial agonists for smoking cessation.
Banerjee SC, Ostroff JS, Bari S, D’Agostino TA, Khera M, Cochrane Database Syst Rev. 18(4):CD006103.
Acharya S, et al. (2014). Gutka and tambaku paan use doi:10.1002/14651858.CD006103.pub6 PMID:22513936
among South Asian immigrants: a focus group study. J Cahill K, Stevens S, Perera R, Lancaster T (2013).
Immigr Minor Health. 16(3):531–9. doi:10.1007/s10903- Pharmacological interventions for smoking cessa-
013-9826-4 PMID:23579964 tion: an overview and network meta-analysis.
Bangladesh Bureau of Statistics and National Tobacco Cochrane Database Syst Rev. 2013(5):CD009329.
Control Cell (2019). Global Adult Tobacco Survey doi:10.1002/14651858.CD009329.pub2 PMID:23728690
Bangladesh Report 2017. Dhaka, Bangladesh. Available Campaign for Tobacco-Free Kids (2021). Legislation
from: https://ntcc.gov.bd/ntcc/uploads/editor/files/ by country: Myanmar. Tobacco control laws.
GATS%20Report%20Final-2017_20%20MB.PDF. Washington (DC), USA. Available from: https://
Bergen AW, Caporaso N (1999). Cigarette smoking. J Natl www.tobaccocontrollaws.org/legislation/country/
Cancer Inst. 91(16):1365–75. doi:10.1093/jnci/91.16.1365 myanmar/pl-health-warnings.
PMID:10451441 Carr AB, Ebbert J (2012). Interventions for tobacco cessa-
Bhan N, Karan A, Srivastava S, Selvaraj S, Subramanian tion in the dental setting. Cochrane Database Syst Rev.
SV, Millett C (2016). Have socioeconomic inequalities 2012(6):CD005084. doi:10.1002/14651858.CD005084.
in tobacco use in India increased over time? Trends pub3 PMID:22696348
from the National Sample Surveys (2000–2012). Chaffee BW, Couch ET, Gansky SA (2018). Adolescents’
Nicotine Tob Res. 18(8):1711–8. doi:10.1093/ntr/ntw092 smokeless tobacco susceptibility by perceived profes-
PMID:27048274 sional baseball players’ use. J Public Health Dent.
Bhat SJS, Blank MD, Balster RL, Nichter M, Nichter 78(1):5–8. doi:10.1111/jphd.12228 PMID:28590522
M (2010). Areca nut dependence among chewers Changrani J, Gany FM, Cruz G, Kerr R, Katz R (2006).
in a South Indian community who do not also use Paan and gutka use in the United States: a pilot study
tobacco. Addiction. 105(7):1303–10. doi:10.1111/j.1360- in Bangladeshi and Indian-Gujarati immigrants in
0443.2010.02952.x PMID:20642513 New York City. J Immigr Refug Stud. 4(1):99–110.
Bincy M, Vidhubala E, Priyadharshini R (2018). Does 85% doi:10.1300/J500v04n01_07 PMID:17492057
pictorial health warning on the tobacco products have Chher T, Hak S, Kallarakkal TG, Durward C, Ramanathan
the impact among tobacco users in India? Tob Induc A, Ghani WMN, et al. (2018). Prevalence of oral cancer,
Dis. 16(Suppl 1):A758. doi:10.18332/tid/84498 oral potentially malignant disorders and other oral
Bissessur S, Naidoo S (2009). Areca nut and tobacco mucosal lesions in Cambodia. Ethn Health. 23(1):1–15.
chewing habits in Durban, KwaZulu Natal. SADJ. doi:10.1080/13557858.2016.1246431 PMID:27781495
64(10):460–3. PMID:20306864 Cho SM, Saw YM, Latt NN, Saw TN, Htet H, Khaing M,
Blecher E, Liber AC, Chaussard M, Fedewa S (2014). Market et al. (2020). Cross-sectional study on tobacco adver-
structures, socioeconomics, and tobacco usage patterns tising, promotion and sponsorship (TAPS) and viola-
in Madagascar. Nicotine Tob Res. 16(Suppl 1):S56–64. tions of tobacco sale regulations in Myanmar: do these
doi:10.1093/ntr/ntt020 PMID:23703733 factors affect current tobacco use among Myanmar high
Bonnechère B, Cissé K, Millogo T, Ouédraogo GH, school students? BMJ Open. 10(2):e031933. doi:10.1136/
Garanet F, Ouedraogo MA, et al. (2019). Tobacco use bmjopen-2019-031933 PMID:32054624
and associated risk factors in Burkina Faso: results Choi K, Fabian LE, Brock B, Engman KH, Jansen J,
from a population-based cross-sectional survey. BMC Forster JL (2014). Availability of snus and its sale
Public Health. 19(1):1466. doi:10.1186/s12889-019- to minors in a large Minnesota city. Tob Control.
7826-6 PMID:31694594 23(5):449–51. doi:10.1136/tobaccocontrol-2012-050719
Boyle RG (1992). Smokeless tobacco cessation with nico- PMID:23303287
tine replacement: a randomized clinical trial. Diss Coe JW, Brooks PR, Vetelino MG, Wirtz MC, Arnold
Abstr Int. 54(3):825. EP, Huang J, et al. (2005). Varenicline: an α4β2 nico-
Boyle RG, Amato MS, Rode P, Kinney AM, St Claire AW, tinic receptor partial agonist for smoking cessation.
Taylor K (2015). Tobacco use among Minnesota adults, J Med Chem. 48(10):3474–7. doi:10.1021/jm050069n
2014. Am J Health Behav. 39(5):674–9. doi:10.5993/ PMID:15887955
AJHB.39.5.9 PMID:26248177 Cooper BR, Hester TJ, Maxwell RA (1980). Behavioral
and biochemical effects of the antidepressant
283
bupropion (Wellbutrin): evidence for selective Ebbert JO, Croghan IT, Severson HH, Schroeder DR,
blockade of dopamine uptake in vivo. J Pharmacol Exp Hays JT (2011). A pilot study of the efficacy of varen-
Ther. 215(1):127–34. PMID:6778989 icline for the treatment of smokeless tobacco users
Council of the European Communities (1989). Council in Midwestern United States. Nicotine Tob Res.
Directive 89/622/EEC of 13 November 1989 on the 13(9):820–6. doi:10.1093/ntr/ntr078 PMID:21504885
approximation of the laws, regulations and adminis- Ebbert JO, Elrashidi MY, Stead LF (2015). Interventions for
trative provisions of the Member States concerning smokeless tobacco use cessation. Cochrane Database
the labelling of tobacco products. Off J Eur Comm. Syst Rev. 2015(10):CD004306. doi:10.1002/14651858.
L359:1-4. Available from: https://eur-lex.europa.eu/ CD004306.pub5 PMID:26501380
legal-content/EN/TXT/?uri=CELEX%3A31989L0622 Ebbert JO, Haddock CK, Vander Weg M, Klesges RC,
&qid=1629332557112. Poston WS, DeBon M (2006). Predictors of smokeless
Cruising Maldives (2016). The chewing of areca nuts in tobacco initiation in a young adult military cohort. Am
Maldives. Available from: https://cruisingmaldives. J Health Behav. 30(1):103–12. doi:10.5993/AJHB.30.1.10
blog/2016/10/29/t he-chew ing-of-areca-nuts-in- PMID:16430325
maldives/, accessed 10 March 2022. Ebbert JO, Severson HH, Croghan IT, Danaher BG,
Dale LC, Ebbert JO, Glover ED, Croghan IT, Schroeder DR, Schroeder DR (2009). A randomized clinical trial of
Severson HH, et al. (2007). Bupropion SR for the treat- nicotine lozenge for smokeless tobacco use. Nicotine
ment of smokeless tobacco use. Drug Alcohol Depend. Tob Res. 11(12):1415–23. doi:10.1093/ntr/ntp154
90(1):56–63. doi:10.1016/j.drugalcdep.2007.02.008 PMID:19880578
PMID:17353101 Ebbert JO, Severson HH, Croghan IT, Danaher BG,
Dale LC, Ebbert JO, Schroeder DR, Croghan IT, Schroeder DR (2010). A pilot study of mailed nicotine
Rasmussen DF, Trautman JA, et al. (2002). Bupropion lozenges with assisted self-help for the treatment of
for the treatment of nicotine dependence in spit tobacco smokeless tobacco users. Addict Behav. 35(5):522–5.
users: a pilot study. Nicotine Tob Res. 4(3):267–74. doi:10.1016/j.addbeh.2009.12.020 PMID:20060229
doi:10.1080/14622200210153821 PMID:12215235 European Parliament (2001). Directive 2001/37/EC
Danaher BG, Severson HH, Andrews JA, Tyler MS, of the European Parliament and of the Council
Lichtenstein E, Woolley TG, et al. (2013). Randomized of 5 June 2001 on the approximation of the laws,
controlled trial of MyLastDip: a web-based smokeless regulations and administrative provisions of the
tobacco cessation program for chewers ages 14–25. Member States concerning the manufacture,
Nicotine Tob Res. 15(9):1502–10. doi:10.1093/ntr/ntt006 presentation and sale of tobacco products. Off J Eur
PMID:23410803 Comm. L194:26–34. Available from: https://eur-lex.
Danaher BG, Severson HH, Crowley R, van Meter N, Tyler eu ropa .eu /re s ou rc e.html?uri=cel lar:248aee57-
MS, Widdop C, et al. (2015). Randomized controlled abd f- 4 d f 2-929e-30 c2fa7f 22e 0.0 0 0 4 .02/
trial examining the adjunctive use of nicotine lozenges DOC_1&format=PDF.
with MyLastDip: an eHealth smokeless tobacco European Parliament (2014). Directive 2014/40/EU of the
cessation intervention. Internet Interv. 2(1):69–76. European Parliament and of the Council of 3 April
doi:10.1016/j.invent.2014.12.004 PMID:25750862 2014 on the approximation of the laws, regulations
Delhomme V (2019). The ban on tobacco for oral use and administrative provisions of the Member States
upheld by the Court of Justice: on subsidiarity and concerning the manufacture, presentation and sale of
proportionality in EU lifestyle risks policy. Eur J Risk tobacco and related products and repealing Directive
Regul. 10(1):227–35. doi:10.1017/err.2019.11 2001/37/EC. Off J Eur Union. L127:1–38. Available from:
Do ETT, Vu M (2020). A qualitative study of gutka and https://eur-lex.europa.eu/legal-content/EN/TXT/?uri
paan masala use among Bhutanese and Burmese =CELEX%3A32014L0040&qid=1629329420874.
migrants in Georgia. PLoS One. 15(8):e0237266. Fagerström K, Gilljam H, Metcalfe M, Tonstad S, Messig
doi:10.1371/journal.pone.0237266 PMID:32764817 M (2010). Stopping smokeless tobacco with varenicline:
Drope J, Schluger N, Cahn Z, Drope J, Hamill S, Islami randomised double blind placebo controlled trial. BMJ.
F, et al. (2022). Sierra Leone. The tobacco atlas. Atlanta 341:c6549. doi:10.1136/bmj.c6549 PMID:21134997
(GA), USA: American Cancer Society and Vital FAO (2021). FAOSTAT. Food and agriculture data. Rome,
Strategies. Available from: https://tobaccoatlas.org/ Italy: Food and Agriculture Organization of the United
country/sierra-leone/, accessed 23 June 2021. Nations. Available from: https://www.fao.org/faostat/
Ebbert JO, Croghan IT, Schroeder DR, Hurt RD (2013). en/.
A randomized phase II clinical trial of high-dose Flora MS, Mascie-Taylor CGN, Rahman M (2012). Betel
nicotine patch therapy for smokeless tobacco users. quid chewing and its risk factors in Bangladeshi adults.
Nicotine Tob Res. 15(12):2037–44. doi:10.1093/ntr/ WHO South-East Asia J Public Health. 1(2):169–81.
ntt097 PMID:23873976 https://apps.who.int/iris/handle/10665/329825
284
Gansky SA, Ellison JA, Kavanagh C, Hilton JF, Walsh Gupta PC, Lando HA, Pednekar MS, Narake SS, Nagler
MM (2002). Oral screening and brief spit tobacco EM, Pawar PS, et al. (2014a). Improvement in prevalence
cessation counseling: a review and findings. J of tobacco use among teachers in Bihar after COTPA.
Dent Educ. 66(9):1088–98. doi:10.1002/j.0022- Indian J Cancer. 51 (Suppl 1):S19–23. doi:10.4103/0019-
0337.2002.66.9.tb03578.x PMID:12374270 509X.147438 PMID:25526243
Gansky SA, Ellison JA, Rudy D, Bergert N, Letendre Gupta PC, Mehta FS, Pindborg JJ, Bhonsle RB, Murti PR,
MA, Nelson L, et al. (2005). Cluster-randomized Daftary DK, et al. (1992). Primary prevention trial of
controlled trial of an athletic trainer-directed spit oral cancer in India: a 10-year follow-up study. J Oral
(smokeless) tobacco intervention for collegiate baseball Pathol Med. 21(10):433–9. doi:10.1111/j.1600-0714.1992.
athletes: results after 1 year. J Athl Train. 40(2):76–87. tb00970.x PMID:1460581
PMID:15970953 Gupta PC, Ray CS (2002). Tobacco and youth in the South
GBD 2019 Chewing Tobacco Collaborators (2021). Spatial, East Asian region. Indian J Cancer. 39(1):5–34.https://
temporal, and demographic patterns in prevalence pubmed.ncbi.nlm.nih.gov/12931709doi:10.4103/0019-
of chewing tobacco use in 204 countries and territo- 509X.58849 PMID:12931709
ries, 1990–2019: a systematic analysis from the Global Gupta PC, Ray CS, Murti PR, Sinha DN (2014b). Rising
Burden of Disease Study 2019. Lancet Public Health. incidence of oral cancer in Ahmedabad city. Indian
6(7):e482–99. doi:10.1016/S2468-2667(21)00065-7 J Cancer. 51 (Suppl 1):S67–72. doi:10.4103/0019-
PMID:34051920 509X.147476 PMID:25526252
Glover ED, Glover PN, Sullivan CR, Cerullo CL, Hobbs Gupta PC, Warnakulasuriya S (2002). Global epide-
G (2002). A comparison of sustained-release bupro- miology of areca nut usage. Addict Biol. 7(1):77–83.
pion and placebo for smokeless tobacco cessation. Am doi:10.1080/13556210020091437 PMID:11900626
J Health Behav. 26(5):386–93. doi:10.5993/AJHB.26.5.7 Gupta R, Verma V, Mathur P (2016b). Quitline activity
PMID:12206448 in Rajasthan, India. Asian Pac J Cancer Prev.
Goebel LJ, Crespo RD, Abraham RT, Masho SW, Glover 17(sup2) S2:19–24. doi:10.7314/APJCP.2016.17.S2.19
ED (2000). Correlates of youth smokeless tobacco use. PMID:27108749
Nicotine Tob Res. 2(4):319–25. doi:10.1080/713688153 Gurung MS, Pelzom D, Dorji T, Drukpa W, Wangdi C,
PMID:11197311 Chinnakali P, et al. (2016). Current tobacco use and
Gonzales D, Rennard SI, Nides M, Oncken C, Azoulay its associated factors among adults in a country with
S, Billing CB, et al.; Varenicline Phase 3 Study Group comprehensive ban on tobacco: findings from the
(2006). Varenicline, an α4β2 nicotinic acetylcholine nationally representative STEPS survey, Bhutan, 2014.
receptor partial agonist, vs sustained-release bupro- Popul Health Metr. 14(1):28. doi:10.1186/s12963-016-
pion and placebo for smoking cessation: a randomized 0098-9 PMID:27507928
controlled trial. JAMA. 296(1):47–55. doi:10.1001/ Hatsukami D, Jensen J, Allen S, Grillo M, Bliss R (1996).
jama.296.1.47 PMID:16820546 Effects of behavioral and pharmacological treat-
Gravely S, Fong GT, Driezen P, Xu S, Quah AC, Sansone ment on smokeless tobacco users. J Consult Clin
G, et al. (2016). An examination of the effectiveness of Psychol. 64(1):153–61. doi:10.1037/0022-006X.64.1.153
health warning labels on smokeless tobacco products in PMID:8907095
four states in India: findings from the TCP India cohort Hatsukami DK, Grillo M, Boyle R, Allen S, Jensen J,
survey. BMC Public Health. 16(1):1246. doi:10.1186/ Bliss R, et al. (2000). Treatment of spit tobacco users
s12889-016-3899-7 PMID:27964733 with transdermal nicotine system and mint snuff. J
Greenhalgh EM, Scollo MM, Winstanley MH (2022). Consult Clin Psychol. 68(2):241–9. doi:10.1037/0022-
Tobacco in Australia: facts and issues. Melbourne, 006X.68.2.241 PMID:10780124
Australia: Cancer Council Victoria. Available from: Hecht SS, Hatsukami DK (2022). Smokeless tobacco and
https://www.tobaccoinaustralia.org.au. cigarette smoking: chemical mechanisms and cancer
Gunjal S, Pateel DGS, Yang YH, Doss JG, Bilal S, Maling prevention. Nat Rev Cancer. 22(3):143–55. doi:10.1038/
TH, et al. (2020). An overview on betel quid and areca s41568-021-00423-4 PMID:34980891
nut practice and control in selected Asian and South Howard-Pitney B, Killen JD, Fortmann SP (1999).
East Asian countries. Subst Use Misuse. 55(9):1533–44. Quitting chew: results from a randomized trial
doi:10.1080/10826084.2019.1657149 PMID:32569533 using nicotine patches. Exp Clin Psychopharmacol.
Gupta PC, Arora M, Sinha DN, Asma S, Parascandola M, 7(4):362–71. https://pubmed.ncbi.nlm.nih.gov/106099
editors (2016a). Smokeless tobacco and public health 71doi:10.1037/1064-1297.7.4.362 PMID:10609971
in India. New Delhi, India: Ministry of Health and Hsu S (2014). Betel nut classes target curbing habit,
Family Welfare, Government of India; p. 309. Available preventing cancer. Taipei Times. Apr 19:3. Available
from: https://portal-uat.who.int/fctcapps/sites/default/ from: https://www.taipeitimes.com/News/taiwan/
files/kh-media/Final-Version-of-SLT-Monograph.pdf. archives/2014/04/19/2003588400.
285
Huang J, Chaloupka FJ (2012). The impact of the 2009 IARC (2021). Acrolein, crotonaldehyde, and arecoline.
federal tobacco excise tax increase on youth tobacco IARC Monogr Identif Carcinog Hazard Hum. 128:1–
use. Working Paper 18026. Cambridge (MA), USA: 335. Available from: https://publications.iarc.fr/602
National Bureau of Economic Research. Available PMID:36924508
from: https://www.nber.org/papers/w18026. Idris AM, Ibrahim YE, Warnakulasuriya KA, Cooper DJ,
Huang YT, Ho PS, Jou YY, Wu CY, Wang YW, Yang YH Johnson NW, Nilsen R (1998). Toombak use and ciga-
(2020). Determining high prevalence of betel-quid rette smoking in the Sudan: estimates of prevalence in
chewing and cigarette smoking by occupation using the Nile State. Prev Med. 27(4):597–603. doi:10.1006/
the Taiwan National Health Interview Survey. Subst pmed.1998.0331 PMID:9672954
Use Misuse. 55(9):1472–82. https://pubmed.ncbi.nlm. IPH (2012). Report of the Global Adult Tobacco Survey
nih.gov/32569535doi:10.1080/10826084.2020.1732421 (GATS), Malaysia 2011. Kuala Lumpur, Malaysia:
PMID:32569535 Institute for Public Health, National Institute of
Hughes JR, Keely JP, Niaura RS, Ossip-Klein DJ, Richmond Health, Ministry of Health Malaysia. Available from:
RL, Swan GE (2003). Measures of abstinence in clin- https://extranet.who.int/ncdsmicrodata/index.php/
ical trials: issues and recommendations. Nicotine Tob catalog/253/download/2133.
Res. 5(1):13–25. doi:10.1080/1462220031000070552 Jain R, Jhanjee S, Jain V, Gupta T, Mittal S, Goelz P, et al.
PMID:12745503 (2014). A double-blind placebo-controlled randomized
Hung CC, Lee CH, Ko AM, Lane HY, Lee CP, Ko YC (2020). trial of varenicline for smokeless tobacco dependence
Effect of antidepressants for cessation therapy in betel- in India. Nicotine Tob Res. 16(1):50–7. doi:10.1093/ntr/
quid use disorder: a randomised, double-blind, place- ntt115 PMID:23946326
bo-controlled trial. Epidemiol Psychiatr Sci. 29:e125. Jawad M, Lee JT, Glantz S, Millett C (2008). Price elas-
doi:10.1017/S2045796020000384 PMID:32370818 ticity of demand of non-cigarette tobacco products:
Huque R, Zaman MM, Huq SM, Sinha DN (2017). a systematic review and meta-analysis. Tob Control.
Smokeless tobacco and public health in Bangladesh. 27(6):689–95. doi:10.1136/tobaccocontrol-2017-054056
Indian J Public Health. 61(Suppl 1):S18–24. https:// John RM (2008). Price elasticity estimates for tobacco
pubmed.ncbi.nlm.nih.gov/28928314doi:10.4103/ijph. products in India. Health Policy Plan. 23(3):200–9.
IJPH_233_17 PMID:28928314 doi:10.1093/heapol/czn007
Hurt RD, Sachs DP, Glover ED, Offord KP, Johnston JA, John RM, Dauchy E (2021). Trends in affordability of
Dale LC, et al. (1997). A comparison of sustained-re- tobacco products before and after the transition to
lease bupropion and placebo for smoking cessa- GST in India. Tob Control. 30(2):155–9. doi:10.1136/
tion. N Engl J Med. 337(17):1195–202. doi:10.1056/ tobaccocontrol-2019-055525
NEJM199710233371703 PMID:9337378 Joo YJ, Newcombe D, Nosa V, Bullen C (2020). Investigating
Hussain A, Zaheer S, Shafique K (2017). Individual, betel nut use, antecedents and consequences: a review
social and environmental determinants of smokeless of literature. Subst Use Misuse. 55(9):1422–42. doi:10.1
tobacco and betel quid use amongst adolescents of 080/10826084.2019.1666144 PMID:31549570
Karachi: a school-based cross-sectional survey. BMC Jorenby DE, Hays JT, Rigotti NA, Azoulay S, Watsky EJ,
Public Health. 17(1):913. doi:10.1186/s12889-017-4916-1 Williams KE, et al. (2006). Varenicline Phase 3 Study
PMID:29183302 Group. Efficacy of varenicline, an α4β2 nicotinic
Hussain A, Zaheer S, Shafique K (2018). Reasons for acetylcholine receptor partial agonist, vs placebo or
betel quid chewing amongst dependent and non-de- sustained-release bupropion for smoking cessation:
pendent betel quid chewing adolescents: a school- a randomized controlled trial. JAMA. 296(1):56–63.
based cross-sectional survey. Subst Abuse Treat doi:10.1001/jama.296.1.56 PMID:16820547
Prev Policy. 13(1):16. doi:10.1186/s13011-018-0154-5 Jorenby DE, Leischow SJ, Nides MA, Rennard SI, Johnston
PMID:29743089 JA, Hughes AR, et al. (1999). A controlled trial of
IARC (2004). Betel-quid and areca-nut chewing and some sustained-release bupropion, a nicotine patch, or both
areca-nut derived nitrosamines. IARC Monogr Eval for smoking cessation. N Engl J Med. 340(9):685–91.
Carcinog Risks Hum. 85:1–334. Available from: https:// doi:10.1056/NEJM199903043400903 PMID:10053177
publications.iarc.fr/103 PMID:15635762 Joseph RA, Chaloupka FJ (2014). The influence of
IARC (2007). Smokeless tobacco and some tobacco-spe- prices on youth tobacco use in India. Nicotine
cific N-nitrosamines. IARC Monogr Eval Carcinog Risks Tob Res. 16(Suppl 1):S24–9. doi:10.1093/ntr/ntt041
Hum. 89:1–626. Available from: https://publications. PMID:23743096
iarc.fr/107 PMID:18335640 Kabwama SN, Ndyanabangi S, Mutungi G, Wesonga R,
IARC (2012). Personal habits and indoor combustions. Bahendeka SK, Guwatudde D (2016). Tobacco use and
IARC Monogr Eval Carcinog Risks Hum. 100E:1–575. associated factors among adults in Uganda: findings
Available from: https://publications.iarc.fr/122 from a nationwide survey. Tob Induc Dis. 14(1):27.
PMID:23193840 doi:10.1186/s12971-016-0093-8 PMID:27524959
286
Khan Z (2016). A gaping gap (smokeless tobacco control in Lim TK (2012). Edible medicinal and non-medicinal
Pakistan). Tob Induc Dis. 14(1):36. doi:10.1186/s12971- plants. Volume 1. Fruits. 1st ed. Dordrecht, Netherlands:
016-0102-y PMID:27895546 Springer. doi:10.1007/978-90-481-8661-7
Khawaja MR, Mazahir S, Majeed A, Malik F, Merchant Lin CC, Tamí-Maury I, Ma WF, Lam C, Tsai MH, Lin MT,
KA, Maqsood M, et al. (2006). Chewing of betel, areca et al. (2017). Social and cultural context of betel quid
and tobacco: perceptions and knowledge regarding consumption in Taiwan and implications for preven-
their role in head and neck cancers in an urban squatter tion and cessation interventions. Subst Use Misuse.
settlement in Pakistan. Asian Pac J Cancer Prev. 52(5):646–55. doi:10.1080/10826084.2016.1246572
7(1):95–100. PMID:16629524 PMID:28139147
Kuek LB, Chin TW, Fong KW (1990). Lip and intraoral Lin J, Zhu K, Soliván-Ortiz AM, Larsen SL, Schneid
tumours: a local perspective. Ann Acad Med Singap. TR, Shriver CD, et al. (2018). Smokeless tobacco use
19(2):178–81. PMID:2346292 and related factors: a study in the US military popu-
Kumar R, Jha AK, Munish VG, Pusp A, Sinha P, Gupta lation. Am J Health Behav. 42(4):102–17. doi:10.5993/
P, et al. (2018). National tobacco quitline: the prelimi- AJHB.42.4.10 PMID:29973315
nary Indian experience. Indian J Chest Dis Allied Sci. Ma WF, Li CI, Gritz ER, Tamí-Maury I, Lam C, Lin CC
60(1):7–12. (2017). A symbol of connectedness between the self and
Kumar R, Saroj SK, Kumar M (2021). Tobacco Quitline the tribal home: betel quid in the lives of Indigenous
toll-free number on tobacco packets in India: an Taiwanese. [in Chinese] Hu Li Za Zhi. 64(3):65–73.
analysis on outcome. Monaldi Arch Chest Dis. 91(2): doi:10.6224/JN.000041 PMID:28580560
doi:10.4081/monaldi.2021.1612 PMID:33942599 Mahees M, Amarasinghe HK, Usgodaararachchi U,
Lechner M, Breeze CE, Vaz F, Lund VJ, Kotecha B (2019). Ratnayake N, Tilakaratne WM, Shanmuganathan S,
Betel nut chewing in high-income countries – lack of et al. (2021). A sociological analysis and exploration
awareness and regulation. Lancet Oncol. 20(2):181–3. of factors associated with commercial preparations of
doi:10.1016/S1470-2045(18)30911-2 PMID:30723038 smokeless tobacco use in Sri Lanka. Asian Pac J Cancer
Lee CH, Ko AM, Warnakulasuriya S, Yin BL, Sunarjo, Prev. 22(6):1753–9. doi:10.31557/APJCP.2021.22.6.1753
Zain RB, et al. (2011). Intercountry prevalences and PMID:34181330
practices of betel-quid use in south, southeast and Mamudu HM, John RM, Veeranki SP, Ouma AE (2013).
eastern Asia regions and associated oral preneoplastic The odd man out in Sub-Saharan Africa: under-
disorders: an international collaborative study by standing the tobacco use prevalence in Madagascar.
Asian Betel-Quid Consortium of south and east Asia. BMC Public Health. 13(1):856. doi:10.1186/1471-2458-
Int J Cancer. 129(7):1741–51. doi:10.1002/ijc.25809 13-856 PMID:24044737
PMID:21128235 Mathur N, Pednekar M, Sorensen G, Nagler E, Stoddard
Lee CH, Ko AM, Yang FM, Hung CC, Warnakulasuriya A, Lando H, et al. (2016). Adoption and implemen-
S, Ibrahim SO, et al. (2018). Association of DSM-5 tation of tobacco control policies in schools in India:
betel-quid use disorder with oral potentially malig- results of the Bihar School Teachers Study. Asian Pac J
nant disorder in 6 betel-quid endemic Asian popu- Cancer Prev. 17(6):2821–6. PMID:27356696
lations. JAMA Psychiatry. 75(3):261–9. doi:10.1001/ Mehrotra R, Sinha DN, Szilagyi T, editors (2017).
jamapsychiatry.2017.4307 PMID:29417149 Global smokeless tobacco control policies and their
Lee S, Raker T, Chisick MC (1994). Psychosocial factors implementation. Noida, India: WHO FCTC Global
influencing smokeless tobacco use by teen-age mili- Knowledge Hub on Smokeless Tobacco and Indian
tary dependents. Mil Med. 159(2):112–7. https:// Council of Medical Research-National Institute of
pubmed.ncbi.nlm.ni h.gov/8202235doi:10.1093/ Cancer Prevention and Research. Available from:
milmed/159.2.112 PMID:8202235 https://portal-uat.who.int/fctcapps/sites/default/files/
Leon ME, Lugo A, Boffetta P, Gilmore A, Ross H, Schüz J, et kh-media/Global-smokeless-NICPR-19418-1.pdf.
al. (2016). Smokeless tobacco use in Sweden and other Mehrotra R, Yadav A, Sinha DN, Parascandola M, John
17 European countries. Eur J Public Health. 26(5):817– RM, Ayo-Yusuf O, et al. (2019). Smokeless tobacco
21. doi:10.1093/eurpub/ckw032 PMID:27048433 control in 180 countries across the globe: call to
Levy DT, Yuan Z, Li Y, St Claire AW, Schillo BA (2019). action for full implementation of WHO FCTC meas-
The Minnesota SimSmoke tobacco control policy ures. Lancet Oncol. 20(4):e208–17. doi:10.1016/S1470-
model of smokeless tobacco and cigarette use. Am J Prev 2045(19)30084-1 PMID:30942182
Med. 57(4):e103–15. doi:10.1016/j.amepre.2019.06.003 Mehrtash H, Duncan K, Parascandola M, David A,
PMID:31542143 Gritz ER, Gupta PC, et al. (2017). Defining a global
Lim F, Pakiam G (2020). A bite of history: betel chewing research and policy agenda for betel quid and areca
in Singapore. Biblioasia. 16(3):4–9. https://biblioasia. nut. Lancet Oncol. 18(12):e767–75. doi:10.1016/S1470-
nlb.gov.sg/vol-16/issue-3/oct-dec-2020/betel-chewing/ 2045(17)30460-6 PMID:29208442
287
Ministry of Law and Justice (2006). Food Safety and in: Toxicol Lett. 2016;259:156.] https://pubmed.ncbi.
Standards Act, 2006. No. 34 of 2006. New Delhi, India. nlm.nih.gov/27178269doi:10.1016/j.toxlet.2016.05.011
Available from: https://www.fssai.gov.in/cms/food- PMID:27178269
safety-and-standards-act-2006.php. Monson AL, Beaulieu JA (2011). Smokeless tobacco
Mishra GA, Gunjal SS, Pimple SA, Majmudar PV, Gupta use and knowledge among university students.
SD, Shastri SS (2014). Impact of ‘gutkha and pan masala Internet J Allied Health Sci Pract. 9(1):1–9.
ban’ in the state of Maharashtra on users and vendors. doi:10.46743/1540-580X/2011.1335
Indian J Cancer. 51(2):129–32. doi:10.4103/0019- Moore C (2020). Solomon Islands historical encyclo-
509X.138182 PMID:25104193 paedia, 1893–1978. Brisbane, Australia: The University
Mistry R, Pednekar MS, McCarthy WJ, Resnicow K, Pimple of Queensland. Available from: https://www.
SA, Hsieh HF, et al. (2019). Compliance with point-of- solomonencyclopaedia.net/introduction.html.
sale tobacco control policies and student tobacco use in Murphy KL, Herzog TA (2015). Sociocultural factors that
Mumbai, India. Tob Control. 28(2):220–6. doi:10.1136/ affect chewing behaviors among betel nut chewers and
tobaccocontrol-2018-054290 PMID:29743339 ex-chewers on Guam. Hawaii J Med Public Health.
MOHFW (2011a). Food Safety and Standards (Food 74(12):406–11. PMID:26668772
Products Standards and Food Additives) Regulations, Murphy KL, Liu M, Herzog TA (2019). Confirmatory
2011, 3.1.7.1: Restriction on use of anticaking agents. factor analysis and structural equation modeling
The Gazette of India: Extraordinary. 34 (Part III, of socio-cultural constructs among Chamorro and
Section 4):430. New Delhi, India: Ministry of Health non-Chamorro Micronesian betel nut chewers. Ethn
and Family Welfare, Government of India. Available Health. 24(6):724–35. doi:10.1080/13557858.2017.1346
from: https://www.fssai.gov.in/upload/uploadfiles/ 177 PMID:28678534
files/Food_Additives_Regulations.pdf. Murukutla N, Turk T, Prasad CV, Saradhi R, Kaur J,
MOHFW (2011b). Food Safety and Standards (Packaging Gupta S, et al. (2012). Results of a national mass
and Labelling) Regulations, 2011, 2.4.5: Subsections media campaign in India to warn against the dangers
30, 31. The Gazette of India: Extraordinary (Part III, of smokeless tobacco consumption. Tob Control.
Section 4):49. New Delhi, India: Ministry of Health 21(1):12–7. doi:10.1136/tc.2010.039438 PMID:21508418
and Family Welfare, Government of India. Available Mushtaq N, Boeckman LM, Beebe LA (2015). Predictors
from: https://www.fssai.gov.in/upload/uploadfiles/ of smokeless tobacco cessation among telephone quit-
files/Packaging_Labelling_Regulations.pdf. line participants. Am J Prev Med. 48(Suppl 1):S54–60.
MOHFW (2011c). Food Safety and Standards (Prohibition doi:10.1016/j.amepre.2014.09.028 PMID:25528708
and Restrictions on Sales) Regulations, 2011. Myint SK, Narksawat K, Sillabutra J (2016). Prevalence
Regulation No. 2.3.4: Prohibition of use of tobacco and factors influencing betel nut chewing among
and nicotine in any food product. The Gazette of India: adults in West Insein Township, Yangon, Myanmar.
Extraordinary (Part III, Section 4). New Delhi, India: Southeast Asian J Trop Med Public Health. 47(5):1089–
Ministry of Health and Family Welfare, Government 97. PMID:29620822
of India. Available from: https://www.fssai.gov.in/ Nambiar V, Osea Gavidi R, Bhai K (2020). Tobacco and
upload/uploadfiles/files/Prohibition_Regulations.pdf. betel nut chewing habits amongst Fijians of Indian
MOHFW and IIPS (2010). Global Adult Tobacco Survey descent living in Suva, Fiji. Int J Advert. 8(5):638–50.
(GATS), India 2009–2010. New Delhi, Mumbai, India: doi:10.21474/IJAR01/10970
Ministry of Health and Family Welfare, Government Nargis N, Hussain AKMG, Fong GT (2014). Smokeless
of India, and International Institute for Population tobacco product prices and taxation in Bangladesh:
Sciences. Available from: https://ntcp.nhp.gov.in/assets/ findings from the International Tobacco Control Survey.
document/surveys-reports-publications/Global-Adult- Indian J Cancer. 51(Suppl 1):S33–8. doi:10.4103/0019-
Tobacco-Survey-India-2009-2010-Report.pdf. 509X.147452 PMID:25526246
MOHFW and IIPS (2019). GYTS-4. Global Youth Tobacco National Statistical Office of Thailand (2017). The Smoking
Survey. Fact sheet. India 2019. New Delhi, Mumbai, and Drinking Behaviour Survey 2017. [in Thai]
India: Ministry of Health and Family Welfare, Available from: https://www.nso.go.th/sites/2014en/
Government of India, and International Institute for Pages/sur vey/Social/Health/The-Smoking-And-
Population Sciences. Available from: https://ntcp.nhp. Drinking-Behavior-Survey-2011.aspx.
gov.in/assets/document/National_Fact_Sheet_of_ NCI and CDC (2014). Smokeless tobacco and public
fourth_round_of_Global_Youth_Tobacco_Survey_ health: a global perspective (NIH Publication No.
GYTS-4.pdf. 14-7983). Bethesda (MD): USA: National Cancer
Moghbel N, Ryu B, Cabot PJ, Ratsch A, Steadman KJ (2016). Institute, National Institutes of Health and Centers for
In vitro cytotoxicity of Nicotiana gossei leaves, used in Disease Control and Prevention, U.S. Department of
the Australian Aboriginal smokeless tobacco known as Health and Human Services. Available from: https://
pituri or mingkulpa. Toxicol Lett. 254:45–51. [Erratum stacks.cdc.gov/view/cdc/43373.
288
Near AM, Blackman K, Currie LM, Levy DT (2014). Palipudi KM, Gupta PC, Sinha DN, Andes LJ, Asma S,
Sweden SimSmoke: the effect of tobacco control poli- McAfee T; GATS Collaborative Group (2012). Social
cies on smoking and snus prevalence and attributable determinants of health and tobacco use in thirteen
deaths. Eur J Public Health. 24(3):451–8. doi:10.1093/ low and middle income countries: evidence from
eurpub/ckt178 PMID:24287030 Global Adult Tobacco Survey. PLoS One. 7(3):e33466.
Nethan ST, Sinha DN, Sharma S, Mehrotra R (2020). doi:10.1371/journal.pone.0033466 PMID:22438937
Behavioral interventions for smokeless tobacco cessa- Papke RL, Bhattacharyya I, Hatsukami DK, Moe I,
tion. Nicotine Tob Res. 22(4):588–93. doi:10.1093/ntr/ Glatman S (2020). Betel nut (areca) and smokeless
ntz107 PMID:31251347 tobacco use in Myanmar. Subst Use Misuse. 55(9):1385–
Niaz K, Maqbool F, Khan F, Bahadar H, Ismail Hassan 94. [Erratum in: Subst Use Misuse. 2019;54(14):2425.]
F, Abdollahi M (2017). Smokeless tobacco (paan and doi:10.1080/10826084.2019.1624774 PMID:31220988
gutkha) consumption, prevalence, and contribution to Parliament of Bhutan (2010). Tobacco Control Act
oral cancer. Epidemiol Health. 39:e2017009. doi:10.4178/ of Bhutan, 2010. Thimphu, Bhutan: Parliament
epih.e2017009 PMID:28292008 of Bhutan. Available from: https://web.archive.
NIHFW (2014). The Prevention of Food Adulteration org/web/20120316175424/http://www.nab.gov.bt/
Act, 1954 (amended in 1964, 1976, 1986). New Delhi, downloadsact/Dzo76.pdf.
India: National Institute of Health and Family Welfare. Paulino YC, Ettienne R, Novotny R, Wilkens LR, Shomour
Available from: http://www.nihfw.org/Legislations/ M, Sigrah C, et al. (2017b). Areca (betel) nut chewing
THEPREVENTIONOFFOODADUTERATION.html. practices of adults and health behaviors of their chil-
Novais TO (2017). The use of alcohol and other drugs in dren in the Freely Associated States, Micronesia:
the Kalunga quilombola rural community in Goiás. findings from the Children’s Healthy Living (CHL)
[in Portuguese] Com Ciências Saúde. 28(3/4):379–88. Program. Cancer Epidemiol. 50(Pt B):234–40.
https://bvsms.saude.gov.br/bvs/periodicos/ccs _ doi:10.1016/j.canep.2017.07.009 PMID:29120830
artigos/v28_3_us_alcool.pdf. Paulino YC, Hurwitz EL, Ogo JC, Paulino TC, Yamanaka
Núñez-de la Mora A, Jesmin F, Bentley GR (2007). AB, Novotny R, et al. (2017a). Epidemiology of areca
Betel nut use among first and second generation (betel) nut use in the Mariana Islands: findings from
Bangladeshi women in London, UK. J Immigr Minor the University of Guam/University of Hawai’i Cancer
Health. 9(4):299–306. doi:10.1007/s10903-007-9041-2 Center Partnership Program. Cancer Epidemiol.
PMID:17431783 50(Pt B):241–6. doi:10.1016/j.canep.2017.08.006
Nyi Latt N, Saw YM, Myat Cho S, Kariya T, Yamamoto PMID:29120831
E, Hamajima N (2018). Tobacco control law aware- Peeters S, Gilmore AB (2013). How online sales and promo-
ness, enforcement, and compliance among high school tion of snus contravenes current European Union
students in Myanmar. Nagoya J Med Sci. 80(3):379–89. legislation. Tob Control. 22(4):266–73. doi:10.1136/
doi:10.18999/nagjms.80.3.379 PMID:30214087 tobaccocontrol-2011-050209 PMID:22267980
Nylander J (2016). Hard times for ‘betel nut beauties’ Pimple S, Gunjal S, Mishra GA, Pednekar MS, Majmudar
as Taiwan tries to kick deadly addiction. CNN.com. P, Shastri SS (2014). Compliance to gutka ban and other
Available from: https://edition.cnn.com/2016/09/05/ provisons of COTPA in Mumbai. Indian J Cancer.
asia/taiwan-asia-betel-nuts/index.html. 51(Suppl 1):S60–6. doi:10.4103/0019-509X.147475
Oakley E, Demaine L, Warnakulasuriya S (2005). Areca PMID:25526251
(betel) nut chewing habit among high-school children Pobutsky AM, Neri EI (2012). Betel nut chewing in Hawai’i:
in the Commonwealth of the Northern Mariana Islands is it becoming a public health problem? Historical and
(Micronesia). Bull World Health Organ. 83(9):656– socio-cultural considerations. Hawaii J Med Public
60. https://apps.who.int/iris/handle/10665/269482 Health. 71(1):23–6. PMID:22413101
PMID:16211156 Quinn B, Peach E, Wright CJC, Lim MSC, Davidson L,
Oudjehih M, Deltour I, Bouhidel ML, Bouhidel A, Dietze P (2017). Alcohol and other substance use among
Marref A, Luzon V, et al. (2020). Smokeless tobacco a sample of young people in the Solomon Islands. Aust
use, cigarette smoking, and upper aerodigestive tract N Z J Public Health. 41(4):358–64. doi:10.1111/1753-
cancers: a case-control study in the Batna Region, 6405.12669 PMID:28616897
Algeria, 2008–2011. Tob Use Insights. 13:X20902239. Raja M, Saha S, Krishna-Reddy V, Mohd S, Narang R, Sood
doi:10.1177/1179173X20902239 PMID:32076371 P (2016). Effectiveness of oral health education versus
PAHO (2018). Youth and tobacco in the Region of the nicotine replacement therapy for tobacco cessation –
Americas. Results from the Global Youth Tobacco a parallel randomized clinical trial. J Clin Exp Dent.
Survey (2010–2017). Washington (DC), USA: Pan 8(1):e64–70. doi:10.4317/jced.52738 PMID:26855709
American Health Organization. Available from: Rani M, Bonu S, Jha P, Nguyen SN, Jamjoum L (2003).
https://iris.paho.org/handle/10665.2/51670. Tobacco use in India: prevalence and predictors of
smoking and chewing in a national cross sectional
289
household survey. Tob Control. 12(4):e4. doi:10.1136/ Severson HH, Gordon JS, Danaher BG, Akers L (2008).
tc.12.4.e4 PMID:14660785 ChewFree.com: evaluation of a web-based cessation
Reddy P, Anjum Sh, Monica M, Yadav Rao K, Akula program for smokeless tobacco users. Nicotine Tob
Sh, Sai Pravallika T (2016). Is there any impact of Res. 10(2):381–91. doi:10.1080/14622200701824984
the gutkha ban on users and vendors in Rangareddy PMID:18236303
district? A cross sectional study. Asian Pac J Cancer Severson HH, Peterson AL, Andrews JA, Gordon JS,
Prev. 17(11):5005–9. PMID:28032731 Cigrang JA, Danaher BG, et al. (2009). Smokeless
Reichart PA, Nguyen XH (2008). Betel quid chewing, oral tobacco cessation in military personnel: a rand-
cancer and other oral mucosal diseases in Vietnam: omized controlled trial. Nicotine Tob Res. 11(6):730–8.
a review. J Oral Pathol Med. 37(9):511–4. doi:10.1111/ doi:10.1093/ntr/ntp057 PMID:19395686
j.1600-0714.2008.00669.x PMID:18624933 Shafey O, Eriksen M, Ross H, Mackay J (2009) The tobacco
Rozi S, Akhtar S (2007). Prevalence and predictors of atlas. 3rd ed. Atlanta (GA), USA: American Cancer
smokeless tobacco use among high-school males in Society. Available from: https://www.afro.who.int/
Karachi, Pakistan. East Mediterr Health J. 13(4):916– publications/tobacco-atlas-3rd-edition.
24. PMID:17955775 Shah S, Dave B, Shah R, Mehta TR, Dave R (2018).
Sadath A, Jose K, Jiji KM, Mercy VT, Ragesh G, Arensman Socioeconomic and cultural impact of tobacco in India.
E (2022). Prevalence and determinants of substance J Family Med Prim Care. 7(6):1173–6. doi:10.4103/
use among Indigenous tribes in South India: findings jfmpc.jfmpc_36_18 PMID:30613493
from a tribal household survey. J Racial Ethn Health Sharapova S, Reyes-Guzman C, Singh T, Phillips E,
Disparities. 9(1):356–66. doi:10.1007/s40615-021- Marynak KL, Agaku I (2020). Age of tobacco use
00964-2 PMID:33495925 initiation and association with current use and nico-
Samai M, Bash-Taqi D, Samai HH, Edem-Hotah J, Daoh tine dependence among US middle and high school
SK, Alemu W (2011). Prevalence of tobacco use and students, 2014–2016. Tob Control. 29(1):49–54. https://
physical activity among adult Sierra Leonean popula- pubmed.ncbi.nlm.nih.gov/30498008doi:10.1136/
tion. Sierra Leone J Biomed Res. 3(1):49–59. doi:10.4314/ tobaccocontrol-2018-054593 PMID:30498008
sljbr.v3i1.66652 Shrestha N, Mehata S, Pradhan PMS, Joshi D, Mishra SR
SAMHSA (2014). FFY 2014 annual Synar reports. Tobacco (2019). A nationally representative study on socio-de-
sales to youth. Rockville (MD), USA: Substance Abuse mographic and geographic correlates, and trends in
and Mental Health Services Administration. Available tobacco use in Nepal. Sci Rep. 9(1):2682. doi:10.1038/
from: https://store.samhsa.gov/sites/default/files/d7/ s41598-019-39635-y PMID:30804493
priv/synar-15.pdf. Siddiqi K, Husain S, Vidyasagaran A, Readshaw A,
Saqib MAN, Rafique I, Qureshi H, Munir MA, Bashir R, Mishu MP, Sheikh A (2020). Global burden of disease
Arif BW, et al. (2018). Burden of tobacco in Pakistan: due to smokeless tobacco consumption in adults:
findings from Global Adult Tobacco Survey 2014. an updated analysis of data from 127 countries.
Nicotine Tob Res. 20(9):1138–43. doi:10.1093/ntr/ntx179 BMC Med. 18(1):222. doi:10.1186/s12916-020-01677-9
PMID:29059338 PMID:32782007
Selvaraj S, Srivastava S, Karan A (2015). Price elasticity Siddiqi K, Shah S, Abbas SM, Vidyasagaran A, Jawad M,
of tobacco products among economic classes in India, Dogar O, et al. (2015). Global burden of disease due
2011–2012. BMJ Open. 5(12):e008180. doi:10.1136/ to smokeless tobacco consumption in adults: anal-
bmjopen-2015-008180 PMID:26656009 ysis of data from 113 countries. BMC Med. 13(1):194.
Severson HH, Andrews JA, Lichtenstein E, Danaher doi:10.1186/s12916-015-0424-2 PMID:26278072
BG, Akers L (2007). Self-help cessation programs Singh A, Mathur MR, Palipudi KM, Arora M (2016).
for smokeless tobacco users: long-term follow-up Determinants of smokeless tobacco use in India. In:
of a randomized trial. Nicotine Tob Res. 9(2):281–9. Gupta PC, Arora M, Sinha DN, Asma S, Parascandola
doi:10.1080/14622200601080281 PMID:17365759 M, editors. Smokeless tobacco and public health
Severson HH, Andrews JA, Lichtenstein E, Gordon in India. New Delhi, India: Ministry of Health and
JS, Barckley MF (1998). Using the hygiene visit to Family Welfare, Government of India; pp. 97–114.
deliver a tobacco cessation program: results of a rand- Singh PK, Yadav A, Singh L, Mazumdar S, Sinha DN,
omized clinical trial. J Am Dent Assoc. 129(7):993–9. Straif K, et al. (2021). Areca nut consumption with
doi:10.14219/jada.archive.1998.0353 PMID:9685764 and without tobacco among the adult population: a
Severson HH, Danaher BG, Ebbert JO, van Meter N, nationally representative study from India. BMJ Open.
Lichtenstein E, Widdop C, et al. (2015). Randomized 11(6):e043987. doi:10.1136/bmjopen-2020-043987
trial of nicotine lozenges and phone counseling PMID:34130957
for smokeless tobacco cessation. Nicotine Tob Res. Singh PK, Yadav A, Singh L, Singh S, Mehrotra R (2020).
17(3):309–15. doi:10.1093/ntr/ntu145 PMID:25168034 Social determinants of dual tobacco use in India: an
analysis based on the two rounds of Global Adult
290
Tobacco Survey. Prev Med Rep. 18:101073. doi:10.1016/j. Med. 41(2):417–23. doi:10.1016/j.ypmed.2004.09.048
pmedr.2020.101073 PMID:32257776 PMID:15917036
Singh SK, Schensul JJ, Kashyap GC (2018). The reach Sorensen G, Pednekar MS, Sinha DN, Stoddard AM,
of media to smokers and smokeless tobacco users in Nagler E, Aghi MB, et al. (2013). Effects of a tobacco
India: evidence from the Global Adult Tobacco Survey control intervention for teachers in India: results of
(GATS). J Popul Soc Stud. 26(1):42–52. doi:10.25133/ the Bihar School Teachers Study. Am J Public Health.
JPSSv26n1.004 103(11):2035–40. doi:10.2105/AJPH.2013.301303
Sinha DN, Gupta PC, Kumar A, Bhartiya D, Agarwal N, PMID:24028234
Sharma S, et al. (2018a). The poorest of poor suffer the Sreeramareddy CT, Pradhan PM, Sin S (2014b). Prevalence,
greatest burden from smokeless tobacco use: a study distribution, and social determinants of tobacco use in
from 140 countries. Nicotine Tob Res. 20(12):1529–32. 30 sub-Saharan African countries. BMC Med. 12(1):243.
doi:10.1093/ntr/ntx276 PMID:29309692 doi:10.1186/s12916-014-0243-x PMID:25518855
Sinha DN, Gupta PC, Pednekar MS, Jones JT, Warren CW Sreeramareddy CT, Pradhan PMS, Mir IA, Sin S (2014a).
(2002). Tobacco use among school personnel in Bihar, Smoking and smokeless tobacco use in nine South and
India. Tob Control. 11(1):82–3. doi:10.1136/tc.11.1.82 Southeast Asian countries: prevalence estimates and
PMID:11891376 social determinants from Demographic and Health
Sinha DN, Gupta PC, Ray C, Singh PK (2012). Prevalence Surveys. Popul Health Metr. 12(1):22. doi:10.1186/
of smokeless tobacco use among adults in WHO s12963-014-0022-0 PMID:25183954
South-East Asia. Indian J Cancer. 49(4):342–6. https:// Stevens VJ, Severson H, Lichtenstein E, Little SJ, Leben
pubmed.ncbi.nlm.nih.gov/23442396doi:10.4103/0019- J (1995). Making the most of a teachable moment: a
509X.107726 PMID:23442396 smokeless-tobacco cessation intervention in the dental
Sinha DN, Gupta PC, Warren CW, Asma S (2004a). School office. Am J Public Health. 85(2):231–5. doi:10.2105/
policy and tobacco use by students in Bihar, India. AJPH.85.2.231 PMID:7856783
Indian J Public Health. 48(3):118–22. PMID:15709597 Stigler MH, Perry CL, Arora M, Shrivastav R, Mathur C,
Sinha DN, Gupta PC, Warren CW, Asma S (2004b). Effect Reddy KS (2007). Intermediate outcomes from Project
of school policy on tobacco use by school personnel in MYTRI: mobilizing youth for tobacco-related initi-
Bihar, India. J Sch Health. 74(1):3–5. doi:10.1111/j.1746 atives in India. Cancer Epidemiol Biomarkers Prev.
-1561.2004.tb06592.x PMID:15022368 16(6):1050–6. doi:10.1158/1055-9965.EPI-06-0929
Sinha DN, Mathur N, Nazar GP, Gajalakshmi Vendhan V, PMID:17548662
Palipudi KM, Asma S, et al. (2016). Smokeless tobacco Stotts RC, Roberson PK, Hanna EY, Jones SK, Smith CK
use among youth. In: Gupta PC, Arora M, Sinha DN, (2003). A randomised clinical trial of nicotine patches
Asma S, Parascandola M, editors. Smokeless tobacco for treatment of spit tobacco addiction among adoles-
and public health in India. New Delhi, India: Ministry cents. Tob Control. 12(Suppl 4):IV11–5. doi:10.1136/
of Health and Family Welfare, Government of India; tc.12.suppl_4.iv11 PMID:14645935
pp. 45–62. Su SY, Chen WT, Chiang CJ, Yang YW, Lee WC (2020).
Sinha DN, Palipudi KM, Jones CK, Khadka BB, Silva Oral cancer incidence rates from 1997 to 2016 among
PD, Mumthaz M, et al. (2014). Levels and trends of men in Taiwan: association between birth cohort
smokeless tobacco use among youth in countries of the trends and betel nut consumption. Oral Oncol.
World Health Organization South-East Asia Region. 107:104798. doi:10.1016/j.oraloncology.2020.104798
Indian J Cancer. 51(Suppl 1):S50–3. doi:10.4103/0019- PMID:32434121
509X.147472 PMID:25526249 Suliankatchi Abdulkader R, Sinha DN, Jeyashree K,
Sinha DN, Suliankatchi RA, Gupta PC, Thamarangsi Rath R, Gupta PC, Kannan S, et al. (2019). Trends in
T, Agarwal N, Parascandola M, et al. (2018b). Global tobacco consumption in India 1987–2016: impact of the
burden of all-cause and cause-specific mortality World Health Organization Framework Convention
due to smokeless tobacco use: systematic review and on Tobacco Control. Int J Public Health. 64(6):841–51.
meta-analysis. Tob Control. 27(1):35–42. doi:10.1136/ doi:10.1007/s00038-019-01252-x PMID:31134319
tobaccocontrol-2016-053302 PMID:27903956 Sumithrarachchi SR, Jayasinghe R, Warnakulasuriya S
Solhi M, Fattahi E, Manzari ZS, Gupta PC, Kargar M, (2021). Betel quid addiction: a review of its addiction
Kasmaei P, et al. (2021). The reasons for using smokeless mechanisms and pharmacological management as
tobacco: a review. Iran J Public Health. 50(3):492–501. an emerging modality for habit cessation. Subst Use
doi:10.18502/ijph.v50i3.5589 PMID:34178796 Misuse. 56(13):2017–25. doi:10.1080/10826084.2021.19
Sorensen G, Gupta PC, Sinha DN, Shastri S, Kamat 63990 PMID:34396897
M, Pednekar MS, et al. (2005). Teacher tobacco use Tang JG, Jian XF, Gao ML, Ling TY, Zhang KH
and tobacco use prevention in two regions in India: (1997). Epidemiological survey of oral submucous
results of the Global School Personnel Survey. Prev fibrosis in Xiangtan City, Hunan Province, China.
Community Dent Oral Epidemiol. 25(2):177–80.
291
https://pubmed.ncbi.nlm.nih.gov/9181294doi:10.11 Virtanen SE, Zeebari Z, Rohyo I, Galanti MR (2015).

11/j.1600-0528.1997.tb00918.x PMID:9181294 Evaluation of a brief counseling for tobacco cessa-
Thai Cultural Encyclopedia Foundation (1999). tion in dental clinics among Swedish smokers and
Suwanna chakkatam – cause of huar winas chiba. snus users. A cluster randomized controlled trial
The Encyclopedia of Thai Culture, the Northern (the FRITT study). Prev Med. 70:26–32. doi:10.1016/j.
Region. Volume 14. Bangkok, Thailand: Thai Cultural ypmed.2014.11.005 PMID:25445335
Encyclopedia Foundation, Siam Commercial Bank. Vital Strategies (2010). Campaigns, India. India – tobacco
Thakur JS, Prinja S, Bhatnagar N, Rana SK, Sinha DN, control – Surgeon (English). Video recording. Available
Singh PK (2015). Widespread inequalities in smoking from: https://www.vitalstrategies.org/resources/
& smokeless tobacco consumption across wealth quin- india-surgeon-english/.
tiles in States of India: need for targeted interventions. Vital Strategies (2011). Campaigns, India. India – tobacco
Indian J Med Res. 141(6):789–98. https://pubmed.ncbi. control – Mukesh (English). Video recording. Available
nlm.nih.gov/26205022doi:10.4103/0971-5916.160704 from: https://www.vitalstrategies.org/resources/
PMID:26205022 india-mukesh-english/.
The Economic Times (2020). Spitting in public now an Vital Strategies (2017). First ever campaign in Myanmar
offence under Disaster Management Act: MHA. 15 to warn about deadly betel chewing. Available from:
April 2020. Available from: https://economictimes. https://www.vitalstrategies.org/first-ever-campaign-
indiatimes.com/news/politics-and-nation/spitting-in- myanmar-warn-deadly-betel-chewing/.
public-now-an-offence-under-disaster-management- Walsh MM, Hilton JF, Masouredis CM, Gee L, Chesney
act-mha/articleshow/75161291.cms. MA, Ernster VL (1999). Smokeless tobacco cessa-
Timberlake DS (2016). Advertising receptivity and youth tion intervention for college athletes: results after 1
initiation of smokeless tobacco. Subst Use Misuse. year. Am J Public Health. 89(2):228–34. doi:10.2105/
51(9):1077–82. doi:10.3109/10826084.2016.1160115 AJPH.89.2.228 PMID:9949754
PMID:27159627 Walsh MM, Langer TJ, Kavanagh N, Mansell C, MacDougal
TISS and MOHFW (2017). Global Adult Tobacco Survey, W, Kavanagh C, et al. (2010). Smokeless tobacco cessa-
second round. India 2016–2017 report. Mumbai, New tion cluster randomized trial with rural high school
Delhi, India: Tata Institute of Social Sciences and males: intervention interaction with baseline smoking.
Ministry of Health and Family Welfare, Government Nicotine Tob Res. 12(6):543–50. doi:10.1093/ntr/ntq022
of India. Available from: https://ntcp.nhp.gov.in/assets/ PMID:20439384
document/surveys-reports-publications/Global-Adult- Wangdi P, Gyeltshen N (2020). Bhutan lifts tobacco ban
Tobacco-Survey-Second-Round-India-2016-2017.pdf. to block COVID spillover from India. Nikkei Asia.
Tobacco Control Research Group (2021). Cigarette Available from: https://asia.nikkei.com/Politics/
companies investing in snus. Tobacco tactics. Bath, Bhutan-lifts-tobacco-ban-to-block-COVID-spillover-
UK: Department of Health, University of Bath. from-India.
Available from: https://tobaccotactics.org/wiki/ WHO (2003). WHO Framework Convention on
cigarette-companies-investing-in-snus/. Tobacco Control. Geneva, Switzerland: World Health
Tshering PS, Doelkar S, Tshering S (2021). NA decides Organization. Available from: https://apps.who.int/
lifting ban on sale of tobacco products. Bhutan iris/bitstream/handle/10665/42811/9241591013.pdf.
Broadcasting Service. Available from: http://www.bbs. WHO (2008). WHO report on the global tobacco epidemic,
bt/news/?p=152230. 2008: the MPOWER package. Geneva, Switzerland:
Tsou HT, Wu TY, Kuo KL, Hsiao SH, Huang SJ, Tsai World Health Organization. Available from: https://
CY (2022). Ten-year trends in lifestyle habits among apps.who.int/iris/handle/10665/43818.
community-dwelling older people in Taiwan. Australas WHO (2017). WHO report on the global tobacco
J Ageing. 41(5): doi:10.1111/ajag.12969 PMID:34192403 epidemic, 2017: monitoring tobacco use and preven-
Tun NA, Chittin T, Agarwal N, New ML, Thaung Y, Phyo tion policies. Geneva, Switzerland: World Health
PP (2017). Tobacco use among young adolescents in Organization. Available from: https://apps.who.int/
Myanmar: findings from Global Youth Tobacco Survey. iris/handle/10665/255874.
Indian J Public Health. 61(Suppl 1):S54–9. doi:10.4103/ WHO (2018). Report of fifth Global Youth Tobacco
ijph.IJPH_236_17 PMID:28928320 Survey (GYTS), Myanmar, 2016. New Delhi, India:
Turk T, Newton F, Choudhury S, Islam MS (2018). World Health Organization Regional Office for South-
Predictors of quitting attempts among tobacco users in East Asia. Available from: https://apps.who.int/iris/
Bangladesh after a communication campaign to launch handle/10665/274351.
graphic warning labels on packaging. Health Educ WHO (2019). WHO global report on trends in prev-
Behav. 45(6):879–87. doi:10.1177/1090198118775486 alence of tobacco use 2000–2025. 3rd ed. Geneva,
PMID:29884073 Switzerland: World Health Organization. Available
from: https://www.who.int/publications/i/item/
292
who-globa l-report-on-trends-in-preva lence-of- WHO Regional Office for South-East Asia (2009a). Global
tobacco-use-2000-2025-third-edition. Adult Tobacco Survey: Bangladesh report 2009. New
WHO (2020a). Global Youth Tobacco Survey (GYTS), Delhi, India: World Health Organization Regional
Maldives 2019. New Delhi, India: World Health Office for South-East Asia. Available from: https://apps.
Organization Regional Office for South-East Asia. who.int/iris/handle/10665/205136.
Available from: https://www.who.int/publications/i/ WHO Regional Office for South-East Asia (2009b). Global
item/9789290228158. Adult Tobacco Survey: Thailand country report. New
WHO (2020b). Most recent nationally representative Delhi, India: World Health Organization Regional
survey reporting prevalence of smokeless tobacco Office for South-East Asia. Available from: https://apps.
use or e-cigarette use among adults. Geneva, who.int/iris/handle/10665/205138.
Switzerland: World Health Organization. Available WHO Regional Office for South-East Asia (2011). Global
from: https://www.who.int/data/gho/data/indicators/ Adult Tobacco Survey: Thailand report, 2011. New
indicator-details/GHO/gho-tobacco-control-monitor- Delhi, India: World Health Organization Regional
survey-reporting-prevalence-of-smokeless-tobacco- Office for South-East Asia. Available from: https://apps.
use-or-e-cigarette-use-among-adults. who.int/iris/handle/10665/205139.
WHO (2021a). WHO global report on trends in prevalence WHO Regional Office for South-East Asia (2012).
of tobacco use 2000–2025. 4th ed. Geneva, Switzerland: Global Adult Tobacco Survey: Indonesia report 2011.
World Health Organization. Available from: https:// New Delhi, India: WHO Regional Office for South-
www.who.int/publications/i/item/9789240039322. East Asia. Available from: https://apps.who.int/iris/
WHO (2021b). WHO report on the global tobacco epidemic handle/10665/205137.
2021: addressing new and emerging products. Geneva, WHO Regional Office for South-East Asia (2014). State-
Switzerland: World Health Organization. Available level laws banning gutka are impacting product
from: https://www.who.int/teams/health-promotion/ availability and use. [Press release.] New Delhi, India:
tobacco-control/global-tobacco-report-2021. World Health Organization Regional Office for South-
WHO (2021c). Country profiles on implementation of the East Asia. Available from: https://www.who.int/india/
WHO Framework Convention on Tobacco Control footer/quick-links/media/india-press-releases/state-
in the WHO South-East Asia Region. Available from: level-laws-banning-gutka-are-impacting-product-
https://apps.who.int/iris/handle/10665/204129. availability-and-use.
WHO Country Office for Bangladesh (2018). National WHO Regional Office for South-East Asia (2015). Global
STEPS survey for non-communicable diseases risk Youth Tobacco Survey (GYTS), Bhutan Report, 2013.
factors in Bangladesh 2018. Available from: https:// New Delhi, India: World Health Organization Regional
apps.who.int/iris/handle/10665/332886. Office for South East Asia. Available from: https://apps.
WHO FCTC (2014). Report of the sixth session of the who.int/iris/handle/10665/164336.
Conference of the Parties to the WHO Framework WHO Regional Office for South-East Asia (2016). Sri Lanka
Convention on Tobacco Control. Moscow, Russian 2015 country report: Global Youth Tobacco Survey
Federation, 13–18 October 2014. Available from: (GYTS). New Delhi: India: World Health Organization
https://apps.who.int/gb/fctc/PDF/cop6/FCTC_COP6_ Regional Office for South-East Asia. Available from:
Report-en.pdf. https://apps.who.int/iris/handle/10665/250251.
WHO FCTC (2021). Parties to the WHO Framework WHO Regional Office for South-East Asia (2020). Global
Convention on Tobacco Control. Available from: Youth Tobacco Survey (GYTS), Bhutan, 2019. New
https://fctc.who.int/who-fctc/overview/parties. Delhi, India: World Health Organization Regional
WHO FCTC and ICMR-NICPR (2022). Commonly Office for South-East Asia. Available from: https://
used smokeless tobacco products around the globe. www.who.int/publications/i/item/9789290227724.
Uttar Pradesh, India: WHO Framework Convention Wijesinghe M (2018). The leaf, synonymous with
on Tobacco Control and Indian Council of Medical Avurudu. Sunday Observer. Available from:
Research-National Institute of Cancer Prevention and ht t p s : // w w w. s u n d ay o b s e r v e r. l k / 2 018 /0 4 /0 8 /
Research. Available from: https://untobaccocontrol. leaf-synonymous-avurudu.
org/kh/smokeless-tobacco/paan-betel-quid-tobacco/, Winstock A (2002). Areca nut-abuse liability, depend-
accessed 10 March 2022. ence and public health. Addict Biol. 7(1):133–8.
WHO Regional Office for Europe (2020). Summary doi:10.1080/13556210120091509 PMID:11900633
results of the Global Youth Tobacco Survey in selected World Bank Group (2020). Myanmar – overview of
countries of the WHO European Region. Copenhagen, tobacco use, tobacco control legislation, and taxa-
Denmark: World Health Organization Regional Office tion. Country Brief. Washington (DC), USA:
for Europe. Available from: https://apps.who.int/iris/ World Bank Group. Global Tobacco Control
handle/10665/336752. Program. Available from: https://openknowledge.
worldba nk .org / bitstrea m/ha nd le/10986/31952/
293
M y a n m a r - O v e r v i e w - o f -To b a c c o - U s e - nut and betel quid use for control of oral cancer
To b a c c o - C o nt r o l -L e g i s l a t i o n - a n d -Ta x a t i o n . in Taiwan. Oral Oncol. 108:104915. doi:10.1016/j.
pdf?sequence=1&isAllowed=y. oraloncology.2020.104915 PMID:32712006
Wu P, Wilson K, Dimoulas P, Mills EJ (2006). Effectiveness Ysaol J, Chilton JI, Callaghan P (1996). A survey of betel
of smoking cessation therapies: a systematic review nut chewing in Palau. Isla J. Micrones Stud. 4:244–5.
and meta-analysis. BMC Public Health. 6(1):300. Zhao X, Davey G (2020). Areca nut use among a Chinese
doi:10.1186/1471-2458-6-300 PMID:17156479 ethnic minority, and its health implications. Subst Use
Yang TY, Lin HR (2017). Taking actions to quit chewing Misuse. 55(9):1519–24. doi:10.1080/10826084.2019.166
betel nuts and starting a new life: taxi drivers’ 0676 PMID:31519115
successful experiences of quitting betel nut chewing. Zhou B, Zhu W, Ren C (2019). First steps to regulate adver-
J Clin Nurs. 26(7–8):1031–41. doi:10.1111/jocn.13599 tising of areca nut in China. Lancet Oncol. 20(5):615–6.
PMID:27681205 doi:10.1016/S1470-2045(19)30231-1 PMID:31044700
Yang YH, Warnakulasuriya S, Yang HF, Lin LJ, Wang
YW (2020). Public health measures to reduce areca
294
4. SCREENING AND EARLY DIAGNOSIS OF
ORAL CANCER
4.1 Screening methods and cancer screening process for the general popula-
technologies tion in the USA. Abnormal oral mucosal findings
indicative of oral cancer or OPMDs will lead to
referral for further evaluation (Warnakulasuriya,
4.1.1 Clinical oral examination 2020).
The first-line approach to the identification
of oral cancer and oral potentially malignant (a) Necessary training
disorders (OPMDs) remains the standard clin- Expertise in the screening and diagnosis of
ical oral examination (COE). Traditionally, COE oral mucosal diseases varies substantially across
consists of a white-light visual examination and different clinicians and community workers
palpation of the oral cavity structures and the engaged in either organized or opportunistic
external facial and neck regions. screening activities, and these differences are
A routine can be established to examine linked to their different training backgrounds.
each oral anatomical subsite in a consistent way. A meta-analysis of eight studies comparing the
For example, one would first examine the lips/ knowledge, attitudes, and practices of dentists
labial mucosae, the buccal mucosae, and the and physicians related to oral cancer and
buccal aspects of the mandibular and maxillary OPMDs concluded that dentists were better
gingivae, and then the lingual aspects of the trained than physicians to perform COE and to
mandibular gingivae, followed by examination recognize white or red lesions (Coppola et al.,
of the palate (hard and soft), the floor of the 2021). Educational requirements for compe-
mouth, the tongue, and the tonsils. A detailed tence in performing oral cancer screening are
description of how to examine the oral cavity for not universal, but they have been formalized
cancer is available in Ramadas et al. (2008b). in some countries, including the USA, where
COE involves both a visual examination and the Commission on Dental Accreditation has
a tactile examination (i.e. digital palpation). The mandated that all graduating dentists be compe-
neck is examined to identify enlarged lymph tent in performing screening for oral cancer.
nodes or masses. There is no universally recog- Such competencies are not mandated for medical
nized, evidence-based determination of what school graduates, and the results from a survey
constitutes an appropriate oral cancer screening showed variable training across medical schools
examination. Li et al. (2013) described an expert in the United Kingdom (Carter et al., 2011).
consensus on what should be included in the
295
The need to improve training for medical the reference standard against which the results
providers to perform COE was suggested long of COE performed by the primary screener were
ago (Carter and Ogden, 2007; Shanks et al., 2011), compared (clinical diagnosis by an expert and/
and in one study most of the survey respondents or histological end-points). In all the studies,
indicated a desire for further education on the screeners were trained to perform COE. A
identification of oral cancer (Ni Riordain and negative COE finding was designated when the
McCreary, 2009). Interventions to train medical patients either had no discernible abnormality
practitioners have been associated with improve- or had an abnormality that was deemed to be
ments in knowledge, attitudes, and practices benign. Compared with the reference standard,
over the short term (Papadiochou et al., 2020). non-expert screeners who designated the COE
Web-based educational approaches seem feasible findings as negative performed very well (pooled
to facilitate teaching primary health-care workers specificity, 98%; 95% confidence interval [CI],
to perform COE (Wee et al., 2016). 97–100%) (Table 4.1). The small overall false-
In terms of allied clinicians, dental hygien- positive rate (1 − specificity) was attributed to
ists may play a primary role in performing the large number of true-negative examinations
opportunistic COE at recall visits in dental (linked to the low prevalence of disease in the
offices (Clarke et al., 2018). Similar to the situ- populations studied, which were mostly general
ation for medical education, nurses and nurse populations). The ability of the screener to
practitioners receive variable education on perform a risk assessment on detected abnormal-
oral cancer screening (Carter et al., 2009). The ities equated to the sensitivity of COE. A positive
perceived benefit of such education has been examination in patients with oral mucosal abnor-
recognized (Patton et al., 2006; Li et al., 2020). malities showed heterogeneous sensitivity across
In low-resource countries, there is evidence that studies, ranging from 50% (95% CI, 7–93%) to
community health-care workers can be success- 99% (95% CI, 97–100%); the heterogeneity of the
fully trained to perform oral cancer screening sensitivity prevented pooling of data. Compared
(Warnakulasuriya and Kerr, 2021). with false-positive rates, the higher and heteroge-
Even though dentists receive training on neous overall false-negative rate (1 − sensitivity)
performing COE and recognizing abnormalities, was attributed to the relatively small number of
there is evidence to suggest that they often lack patients with true-positive examinations in the
the skills to identify early lesions (Maybury et al., general populations studied. The sensitivity and
2012) and that they may lack the decision-making specificity outcomes were based on aggregate
skills to differentiate oral cancers and OPMDs data of both oral cancer and OPMDs.
from benign lesions (Kerr et al., 2020). In an attempt to explore the performance
of COE to detect oral cancer versus OPMDs, a
(b) Performance of COE re-analysis of the data was performed (Walsh
A recent analysis of nine studies (10 data et al., 2021b). In four of the data sets, no cancers
sets) assessed the accuracy of COE to detect oral were detected, and the performance of COE
cancer and OPMDs (Walsh et al., 2021b). These to detect OPMDs ranged from 60% to 81% for
studies varied widely in terms of the types of sensitivity and from 94% to 99% for specificity
primary screeners performing COE (non-ex- (Downer et al., 1995; Ikeda et al., 1995; Jullien
pert community health-care workers, dentists, et al., 1995). In one large data set in which only
physicians, or nurses), the settings in which the cancers were considered positive (i.e. OPMDs
studies were performed, the definition of what were considered negative) (Chang et al., 2011),
constitutes a positive or negative finding, and 3 cancers were missed (i.e. false-negatives) out of
296
Table 4.1 Performance of COE for detection of oral cancer and OPMDs
Outcome measured No. Sensitivity (%) Specificity (%) Reference

screened (95% CI) (95% CI)
Oral cancer and OPMDs 2140 59 (39–78) 98 (97–99) Mehta et al. (1986)
Oral cancer and OPMDs 1872 95 (92–97) 81 (79–83) Warnakulasuriya and Pindborg (1990)
Oral cancer and OPMDs 3522 97 (96–98) 75 (73–77) Warnakulasuriya and Nanayakkara (1991)
Oral cancer and OPMDs 2069 94 (90–97) 98 (98–99) Mathew et al. (1997)
OPMDs 309 71 (44–90) 99 (98–100) Downer et al. (1995)
OPMDs 985 61 (44–83) 99 (98–100) Jullien et al. (1995)
OPMDs 1042 81 (64–93) 99 (98–99) Jullien et al. (1995)
OPMDs 154 60 (32–84) 94 (88–97) Ikeda et al. (1995)
Oral cancer 13 606 99 (97–100) 99 (99–99) Chang et al. (2011)
Oral cancer 88 50 (7–93) 98 (92–100) Sweeny et al. (2011)
CI, confidence interval; COE, conventional oral examination; OPMDs, oral potentially malignant disorders.
Reproduced with permission from Walsh et al. (2021b). Copyright 2021, John Wiley & Sons.
a total of 285 cancers, yielding both sensitivity novel approach to oral cancer screening is using
and specificity of 99%. Four of the data sets mobile phone technology to transmit digital
comprised both oral cancers and OPMDs (Mehta images from the field for specialists to review
et al., 1986; Warnakulasuriya and Pindborg, remotely. Three preliminary studies (two in
1990; Warnakulasuriya and Nanayakkara, 1991; India and one in Brazil) (Gomes et al., 2017; Birur
Mathew et al., 1997), and among a combined et al., 2019; Vinayagamoorthy et al., 2019) were
total of more than 9000 people screened, only included in a recent systematic review exploring
1 cancer (out of 36; 2.8%) compared with 95 the accuracy of remote screening in low-resource
OPMDs (out of 2309; 4.1%) were falsely identi- settings (Walsh et al., 2021b). In data from 3600
fied as screen-negative. [There was no stratifica- remote screenings, the sensitivity ranged from
tion analysis of COE performance by outcome 82% to 94%, and the specificity ranged from
(cancer vs OPMDs). None of the studies specif- 72% to 100% (Table 4.2), although the overall
ically assessed whether health workers could certainty of the evidence was rated as very low.
adequately discriminate between oral cancers Subsequently, Haron et al. (2023) compared
and OPMDs; nonetheless, the high sensitivity the accuracy of COE and the decision to refer
and specificity of COE to detect cancer would (i.e. lesions suspicious for oral cancer or OPMDs)
indicate that such discrimination could be performed on site with those based on clinical
successfully done by trained health workers.] images sent via the Mobile Mouth Screening
The overall certainty of the evidence under- Anywhere (MeMoSA) smartphone applica-
lying the reported accuracy of COE to detect tion. Non-specialists were trained to capture
oral cancer and OPMDs was rated as low (Walsh the digital images. For remote assessment and
et al., 2021b). referral decision, the sensitivity was 94.0% and
the specificity was 95.5%.
(c) Mobile technology to improve the perfor- The feasibility of community health workers
mance of COE using a prototype mobile technology to perform
Over the past decade, advances in smart- oral cancer screening was evaluated in rural
phones have enabled their use in health care. A India (Bhatt et al., 2018). The screening process
297
Table 4.2 Performance of remote screening (with mobile phone technology) for detection
of oral cancer and OPMDs
Outcome measured No. Sensitivity (%) Specificity (%) Reference

screened (95% CI) (95% CI)
Oral cancer and OPMDs 55 82 (57–96) 100 (91–100) Gomes et al. (2017)
Oral cancer and OPMDs 3414 85 (81–88) 99 (99–100) Birur et al. (2019)
Oral cancer and OPMDs 131 94 (70–100) 72 (63–80) Vinayagamoorthy et al. (2019)
Reproduced with permission from Walsh et al. (2021b). Copyright 2021, John Wiley & Sons.
was well accepted by this high-risk local popu- (a) Description of the technique
lation, which traditionally had limited access To perform MSE, the person stands in front
to specialized health-care providers; it also had of a mirror under good light to visualize all
a positive impact on the social standing of the parts of the oral cavity and the visible parts of
community health workers using the prototype. the oropharynx. The procedure is a comprehen-
Collectively, these studies have shown sive examination, which is divided into eight
opportunities to develop oral cancer screening steps: facial symmetry, lips, gingivae, buccal
programmes using technology based on mobile mucosae, tongue and floor of the mouth, palate,
phone photography. oropharynx, and lateral aspect of the neck. This
is followed by digital palpation of these struc-
4.1.2 Mouth self-examination tures using the index finger in the same sequence
The oral cavity is easily accessible for exam- as COE.
ination, and most OPMDs and oral cancers are The main advantages of MSE are the low cost,
readily visible (see Section 1.3.1), which facilitates the possibility of performing the examination in
mouth self-examination (MSE). Almost 50 years remote, low-resource areas without diagnostic
ago, Glass et al. (1975) recommended teaching infrastructure, and increased awareness about
MSE as part of cancer prevention programmes; oral diseases. The disadvantages are the impact
after clinical examination, each patient was of overdiagnosis of oral diseases, unnecessary
taught the technique under supervision and referrals, and potential false-negative findings.
encouraged to repeat it every month. Worldwide,
(b) Compliance with and performance of MSE
MSE is being taught to apparently healthy popu-
for screening
lations as part of numerous public awareness
programmes to promote early detection of Mathew et al. (1995) were the first to assess
oral cancer, particularly in populations at high the feasibility and performance of MSE in a large
risk (tobacco smokers and/or alcohol drinkers) trial, in Trivandrum, Kerala, India. About 10 000
(Hussain and Sullivan, 2013; Jornet et al., 2015; copies of a brochure describing risk factors for
Mishra and Bhatt, 2017; Shrestha and Maharjan, oral cancer, the appearance of OMPDs and oral
2020). cancer, and the method for MSE were distrib-
uted to 9000 households by college students in
9 villages over a period of 10 days. In some situa-
tions, the students also demonstrated the proce-
dure to the villagers. One week later, a survey
298
was conducted. Of about 22 000 eligible individ- a suspicious lesion by MSE (of which 39 were
uals, only 8028 (36%) had read the brochure and confirmed as OPMDs), whereas 219 individuals
performed MSE, of whom 247 identified an oral had a suspicious lesion detected by the health
lesion and reported to a referral clinic. A benign workers. The sensitivity of MSE was 18.0%, and
lesion was diagnosed in 97 cases (39%), and 51 the specificity was 99.9%.
individuals (21%) had normal oral variations. In a study conducted in the Buksa tribal
[The accuracy of MSE against clinical diagnosis community in Dehradun District (India), out
was not reported.] of 539 participants, 220 (40.8%) practised MSE.
Scott et al. (2010) reported the results of a pilot The prevalence of oral mucosal lesions identified
study of diagnostic accuracy of MSE in smokers by COE performed by a health worker was 213
aged ≥ 45 years who were recruited from a list (39.5%), whereas only 69 lesions (12.8%) were
of general practitioners in south-eastern London, detected by MSE. The sensitivity was 24.6%,
United Kingdom. COE was performed by a and the specificity was 87.4%. The sensitivity
dentist in 53 participants and identified OPMDs varied from 10.2% for white lesions to 72.7% for
in 12 participants (22%). Without knowing the ulcers, and the specificity varied from 92.4%
results of the dentist’s examination, all of the for difficulty in mouth opening to 99.3% for red
participants received a leaflet on “how to spot lesions (Shah et al., 2020). In an MSE training
mouth cancer early”, with details of MSE, and programme conducted in this tribal community
were asked to proceed with self-examination (Singh et al., 2017), 85 participants attended a
in the room. Most of the participants (39; 74%) health education lecture on MSE and oral cancer.
found MSE easy to perform. A total of 23 partic- The participants were then asked to perform
ipants (43%) reported noticing one or more MSE and report the presence of any abnormali-
lesions. The sensitivity of MSE was 33%, and the ties or oral lesions. Of the 77 study participants
specificity was 54%. [The Working Group noted who performed MSE, 9 detected a lesion.
the poor performance of the test, leading to a The efficacy of MSE was also tested as an
risk of false reassurance for those with false-neg- alternative to follow-up hospital visits in treated
ative results and unnecessary anxiety for those patients with oral cancer (Vaishampayan et al.,
with false-positive results.] 2017). MSE is included in the contents of new
Elango et al. (2011) analysed the effectiveness technologies such as mobile apps for oral cancer
of MSE in coastal villages of Kerala, India, in a awareness (Deshpande et al., 2019).
high-risk population of 57 704 individuals. A
brochure was distributed with information on 4.1.3 Adjunctive techniques
risk factors for oral cancer and the MSE tech-
nique, and instructions to report to an oral cancer An adjunct is defined as a technique or test
screening clinic if any suspicious lesions were that if applied in a screening or diagnostic setting
identified. Four weeks after the brochure was would facilitate the detection or assessment of an
distributed, trained health workers performed abnormal lesion. A screening adjunct is not the
COE on 34 766 available individuals. A total of same as a diagnostic adjunct, and this distinction
30 342 individuals (87%) had practised MSE; is important. A screening adjunct is applied to
987 (3%) reported not knowing how to perform all apparently healthy individuals undergoing
MSE, 1751 (5%) reported disinterest, and 1580 oral cancer screening (as part of a population
(5%) did not report any reason. Of the available screening programme, or opportunistically
individuals, 791 (2%) refused to be examined by to patients attending dental offices) with the
a health worker. Only 54 individuals identified sole aim of improving the ability of a screener
299
to detect disease in a population. A diagnostic (i) Tissue autofluorescence

adjunct is typically applied only to patients with Tissue autofluorescence devices are hand-
abnormal mucosal findings after COE, to better held and generate violet-blue light (in the
characterize such findings and guide clinical 400–450 nm range). This light excites naturally
decisions. occurring tissue fluorophores, i.e. molecules
In the hands of primary care clinicians, the such as flavin adenine dinucleotide (FAD) and
distinction between a screening adjunct and a reduced nicotinamide adenine dinucleotide
diagnostic adjunct is subtle. Hypothetical differ- (NADH) in the epithelium and collagen or elastin
ences might be that occult or small lesions (i.e. cross-links in the submucosa. The result is visible
disease that is not readily visible during COE) fluorescence emission, which enables clinicians
would be more likely to be detected when the to visually scan the mucosa in a darkened envi-
technique is used as a screening adjunct (i.e. ronment to detect disruptions in natural tissue
when COE and the adjunctive technique are autofluorescence (Poh et al., 2010). Two early
used sequentially). In the hands of expert clini- case series of OPMDs harbouring carcinoma
cians, such adjunctive techniques might be used or high-grade dysplasia demonstrated that such
in a diagnostic way to facilitate selection of the lesions exhibited a characteristic loss of fluores-
site of biopsy to aid in mapping or assessing the cence visualization (fluorescence visualization
margins of disease for the purposes of excision. loss [FVL]), in contrast to normal tissue, which
In addition, these techniques might be used in shows normal fluorescence (fluorescence visual-
the surveillance setting to monitor patients with ization retained [FVR]) (Lane et al., 2006; Poh
OPMDs or with a history of oral cancer who are et al., 2007; Fig. 4.1).
at risk of malignant development or recurrence In a single, low-quality study, autofluores-
(Kerr, 2020). cence as a screening adjunct showed no differ-
The adjuncts used in a screening setting are ence compared with COE alone (Simonato
typically point-of-care technologies that provide et al., 2019). Autofluorescence has been evalu-
macroscopic or wide-field information about ated almost exclusively as a diagnostic adjunct
the entire mouth (i.e. when used as a screening in accuracy studies. A recent meta-analysis of
adjunct) or about specific abnormal areas (i.e. these studies reported a pooled sensitivity of
when used to examine a lesion or lesions detected 88% (95% CI, 80–93%) and a pooled specificity
by COE) (Kerr, 2020). Table 4.3 compares the of 61% (95% CI, 44–75%) compared with histo-
utility of adjunctive techniques. pathological outcomes, i.e. any grade of oral
epithelial dysplasia (OED), carcinoma in situ, or
(a) Visualization adjuncts
oral squamous cell carcinoma (OSCC) was rated
Visualization or optical adjuncts include as a positive reference outcome (Table 4.4; Walsh
devices or machines that expose tissues in vivo to et al., 2021b). The low specificity is attributed
various wavelengths of light, generating optical to the preponderance of benign lesions that
signals in real time. These adjuncts work on the demonstrate FVL (i.e. confounder lesions that
premise that the optical properties of diseased yield false-positive outcomes), predominantly
tissues differ from those of normal tissue (Kerr, inflammatory lesions (such as geographic tongue
2020). or erythematous candidiasis), non-inflamma-
tory vascular changes, or pigmented lesions, all
of which absorb blue light. Specificity may be
increased in primary dental settings through
300
Table 4.3 Comparison of adjunctive techniques for screening or diagnosis of oral cancer and OPMDs
Technique Inherent Inherent Sensitivity Specificity Benefits Disadvantages Costs for Costs for Relevance Current
advantages disadvantages for for screening screening assessment to state of
screening screening development
Autofluorescence Non- Requires High Low Minimal Challenging Single None, other Unclear Commercially
invasive, darkened for field purchase than time available
real-time, room; population of device; for clinician in some
hand-held infection- screening; purchase of if used in countries
control interpretation infection- opportunistic
supplies is challenging control setting
needed for non-experts supplies
Narrow-band Non- Large, High High Minimal Impossible Prohibitively None, other Not likely Commercially
imaging invasive, expensive (small (small for field high cost for than time available
real-time unit; number of number of population opportunistic for clinician in some
endoscope studies) studies) screening screening if used in countries
requires opportunistic
sterilization setting
between
patients
Tissue Non- Requires High Very low None Interpretation Single None, other Not Commercially
reflectance invasive, darkened is challenging purchase than time relevant available
real-time, room; for non- of device; for clinician in some
hand-held infection- experts; purchase of if used in countries
control significant infection- opportunistic
supplies overdiagnosis control setting
needed; supplies;
requires purchase of
consumable rinse
supplies;
requires
rinsing steps
Vital staining Non- Uses Intermediate Intermediate Minimal Interpretation Purchase of None, other Not likely Commercially
invasive, consumable is challenging kits than time available
real-time supplies; for non-experts for clinician in some

requires if used in countries, or
rinsing steps; opportunistic may be easily
can be messy setting prepared from
(stains skin/ raw materials
clothing)
301
Table 4.4 Performance of autofluorescence for detection of oral cancer and OPMDs
Reference No. of studies No. of Outcome measured Sensitivity (%) Specificity (%)
lesions (95% CI) (95% CI)
Walsh et al. (2021b) 16 2140 OED (any grade), CIS, OSCC 88 (80–93) 61 (44–75)
CI, confidence interval; CIS, carcinoma in situ; OED, oral epithelial dysplasia; OPMDs, oral potentially malignant disorders; OSCC, oral
squamous cell carcinoma.
adequate training and/or by reassessing patients screening or diagnostic adjuncts for OPMDs
with FVL lesions to rule out benign inflammatory (Lingen et al., 2017a).
lesions (Bhatia et al., 2014; Laronde et al., 2014). One issue that deserves consideration is the
False-negative outcomes may occur in patients mucosal changes associated with chewing of
with dysplastic OPMDs, largely in homogeneous smokeless tobacco or areca nut products. These
leukoplakias with histopathological evidence changes can cause substantial hyper-reflec-
of mild OED, but in rare cases even in OSCCs tance (i.e. a bright white signal) as a result of the
(Truelove et al., 2011). Occult lesions (i.e. lesions effect of surface debris on the mucosa (i.e. betel
not detected by COE) have been detected with chewers’ mucosa), keratosis (such as smokeless
autofluorescence, and a small fraction of them tobacco keratosis), or increased collagen deposi-
harboured OED (Truelove et al., 2011). These tion (i.e. oral submucous fibrosis). False-positives
results, coupled with the fact that most of the are also common due to the preponderance of
accuracy studies were not generalizable to a reactive pigmented lesions (i.e. melanosis) in
primary care dental setting, led an expert panel users of smokeless tobacco or areca nut prod-
to recommend against the use of tissue auto- ucts. Collectively, these findings can make inter-
fluorescence devices by frontline clinicians as pretation challenging, and there are no validated
Fig. 4.1 Oral squamous cell carcinoma involving the left retromolar trigone
The image on the left is under white light. The image on the right displays fluorescence visualization loss (FVL).
Courtesy of Alexander Ross Kerr.
302
Table 4.5 Performance of narrow-band imaging for detection of oral cancer and OPMDs
Reference No. of lesions Outcome measured Sensitivity (%) Specificity (%)

(95% CI) (95% CI)
Piazza et al. (2010) 97 Oral and oropharyngeal SCC 96a 98a
Yang et al. (2013) 317 OED (any grade), CIS, SCC 87 (78–96) 94 (91–97)
CI, confidence interval; CIS, carcinoma in situ; OED, oral epithelial dysplasia; OPMDs, oral potentially malignant disorders; SCC, squamous
cell carcinoma.
a 95% CI not reported.
objective measures to record or document auto- of a similar wavelength to the green light used in
fluorescence outcomes. NBI. The green light was incorporated into the
device to better identify changes in vascularity of
(ii) Narrow-band imaging
OPMDs. Two accuracy studies reported data on
Narrow-band imaging (NBI) is an endo- the green light compared with histopathological
scopic adjunctive technique that is used in outcomes. They demonstrated low sensitivity
the aerodigestive tract to evaluate the surface and specificity: a sensitivity of 40.0% (95% CI,
texture and vascular patterns of the mucosa. NBI 24.9–56.7%) and a specificity of 71.0% (95% CI,
units simultaneously emit two distinct narrow 63.8–78.0%) (Lalla et al., 2016) and a sensitivity
bands of light: one in the blue-green range of 78.4% (95% CI, 61.8–90.2%) and a specificity
(400–430 nm), which helps delineate superficial of 15.4% (95% CI, 4.4–34.9%) (Sharma et al.,
vasculature (blood vessels appear brown), and 2021). [The results showed wide heterogeneity,
the other in the green range (525–555 nm), which suggesting that this device is not a surrogate for
delineates thicker vessels in the submucosa (they an NBI unit.]
appear cyan). The endoscopic NBI unit also [An NBI unit is a sophisticated and expen-
facilitates the photographic capture of images. sive piece of equipment, unlikely to be used for
Compared with healthy tissues, OSCC and OED screening by frontline clinicians or in low-re-
may exhibit abnormal neovascular (angiogenic) source settings.]
patterns; this is the premise for the utility of NBI
in the oral cavity. (iii) Tissue reflectance
Based on two studies (Piazza et al., 2010; This diagnostic adjunct was first developed
Yang et al., 2013), the sensitivity and specificity for the evaluation of cervical neoplasia and then
compared with histopathological outcomes (i.e. adapted for use in the oral cavity (Kerr et al.,
any grade of OED, carcinoma in situ, or OSCC 2006). The proposed basis for its use in the oral
as a positive reference outcome) ranged from cavity is that OPMDs harbouring OSCC or OED
87% to 96% and from 94% to 98%, respectively have a differential tissue reflectance compared
(Table 4.5). In both studies, NBI was significantly with normal mucosa. The evaluation of OPMDs
more accurate than white-light evaluation alone. is performed in two steps: topical application of
[The studies were of low quality.] an acetic acid solution, followed by direct illumi-
A commercially available and comparatively nation using a low-wavelength (blue-white) light
inexpensive hand-held multimodal visualization source. In some of these platforms, the light is
adjunctive device sequentially uses three lights: a generated by a chemical reaction (hence the term
white light, a 405 nm violet light to detect auto- “chemiluminescence”), whereas in others the
fluorescence, and a 545 nm green light, which is source is a light-emitting diode (LED).
303
Table 4.6 Performance of tissue reflectance for detection of oral cancer and OPMDs
Reference No. of No. of Outcome measured Sensitivity (%) Specificity (%)

studies lesions (95% CI) (95% CI)
Lingen et al. (2017b) 4 307 Clinically evident, suspicious lesions 81 (71–89) 69 (63–75)
Walsh et al. (2021b) 6 432 OED (any grade), CIS, SCC 94 (35–99) 19 (3–67)
cell carcinoma.
A recent meta-analysis of five accuracy related to its affinity for nuclear material in
studies (Walsh et al., 2021b) yielded poor speci- the context of increased cellular permeability
ficity compared with histopathological outcomes in OSCC and high-grade OED. Toluidine blue
of OED or OSCC, with a pooled sensitivity of stain may be prepared as a 1% or 2% solution or
94% (95% CI, 35–99%) and a pooled specificity is available commercially in pre-prepared pack-
of 19% (95% CI, 3–67%) (Table 4.6). This tech- ages or bottles. It is used in conjunction with a
nology is currently marketed for use in combi- 1% acetic acid solution; acetic acid is applied first,
nation with toluidine blue vital staining. Based followed by toluidine blue, and then acetic acid
on four studies, the combined use of these two again (Kerr, 2020). A positive test is commensu-
adjuncts led to improvements in the pooled sensi- rate with dark blue staining (Fig. 4.2).
tivity to 81% (95% CI, 71–89%) and in the pooled Toluidine blue was tested as a screening
specificity to 69% (95% CI, 63–75%) (Lingen adjunct in a community-based randomized
et al., 2017b). The studies were considered to have controlled trial (RCT) in 7975 people at high risk
serious issues of risk of bias and indirectness of for oral cancer. Those identified as test-positive
evidence, which downgraded the quality level of (i.e. with positive toluidine blue staining) had
the evidence to very low. a 21% lower incidence rate of OSCC at 5 years
Collectively, these findings led an expert compared with the control group (COE only); this
panel to recommend against the use of tissue result was not statistically significant (Su et al.,
reflectance devices by general dentists (Lingen 2010). In a later systematic review, this study was
et al., 2017a). judged to have high concerns regarding applica-
bility, due to patient selection, and unclear risk of
(b) Vital staining differential verification bias related to the use of
Vital staining involves the topical application a national cancer registry as a reference standard
of a dye to the entire oral mucosa as a screening (Walsh et al., 2013).
adjunct, or more commonly as a diagnostic Most of the literature available for toluidine
adjunct to assess abnormal mucosal lesions. blue is about its use as a diagnostic adjunct. A
Most of the research on vital staining is related recent meta-analysis of 20 accuracy studies,
to the use of toluidine blue and Lugol’s iodine. predominantly using toluidine blue as a single
stain, reported a pooled sensitivity of 86% (95%
(i) Toluidine blue
CI, 79–90%) and a pooled specificity of 68% (95%
The use of toluidine blue vital staining as a CI, 58–77%) compared with histopathological
diagnostic adjunct for assessing OPMDs was end-points (i.e. any grade of OED or OSCC); the
first reported more than 50 years ago by Niebel certainty of the evidence was rated as low to very
and Chomet (1964). The mechanism of action of low (Table 4.7; Walsh et al., 2021b). There was
toluidine blue remains unclear, but it is probably broad heterogeneity in accuracy, which may be
304
Table 4.7 Performance of vital staining for detection of oral cancer and OPMDs

Walsh et al. (2021b) 21 1780 OED (any grade), CIS, SCC 86 (79–90) 68 (58–77)
cell carcinoma.
attributed to several factors, including the diver- leukoplakia). In most of the study populations,
sity of OPMDs tested (i.e. a higher percentage of there is a lower ratio of traumatic and inflam-
high-grade OED, carcinoma in situ, or OSCC matory oral lesions to OPMDs or OSCCs than
will lead to higher sensitivity) (Chainani-Wu would be expected in a general population.
et al., 2015) and variability both in the testing Given that primary care clinicians and health-
protocols and in the interpretation of light or care workers will encounter a blend of mucosal
equivocal staining patterns. abnormalities that reflects the general popula-
Vital staining has potential for both tion, even higher false-positive and false-neg-
false-positives and false-negatives, and the clini- ative rates may be anticipated. Training in the
cian’s experience is critical. False-positives may use of toluidine blue may reduce the number of
occur because toluidine binds to benign inflam- false-positive and false-negative outcomes (Li
matory, ulcerative, or regenerating tissues. In et al., 2019), and a follow-up visit for repeated
addition, the dye may be mechanically retained staining after allowing sufficient time for trau-
in the crevices of rough or fissured lesions and matic and inflammatory lesions to resolve has
the filiform papillae. False-negatives may be due long been recommended to improve specificity
to the inability of the dye to penetrate through (Mashberg, 1980).
thick hyperkeratotic lesions (e.g. homogeneous
Fig. 4.2 Oral squamous cell carcinoma involving the left lateral border of the tongue
The image on the left is under white light. The image on the right displays positive toluidine blue staining (royal blue). Note the small satellite of
blue staining superiorly.
Courtesy of Alexander Ross Kerr.
305
Collectively, these findings led an expert superficial, intermediate, and even basal cells
panel to recommend against the use of vital (i.e. transepithelial sampling). The malignant
staining as a diagnostic adjunct for OPMDs by or benign nature of the oral lesion is usually
general dentists (Lingen et al., 2017a). evaluated with computer-assisted analysis
(Sciubba, 1999; Acha et al., 2005). Epithelial
(ii) Lugol’s iodine and other vital stains
cells collected with a wooden or metallic spatula
Lugol’s iodine, named after the French physi- are usually scarce and can exhibit nuclear and
cian Lugol, stains for glycogen content. Therefore, cytoplasmic distortion (Ogden et al., 1992).
normal non-keratinized oral mucosa will prefer- Cytobrushes improve the capacity to harvest
entially retain the stain. oral mucosa cells and the quality of smears.
Given the contrasting staining effects of Although transepithelial sampling can cause
Lugol’s iodine and toluidine blue, the two agents some discomfort to the patient, the brush
have been tested in combination to improve the must penetrate deeper (indicated by pinpoint
specificity of toluidine blue staining in diag- bleeding) in order to collect basal cell layers. This
nostic accuracy studies for oral cancer and is necessary because dysplastic and early invasive
OPMDs (Epstein et al., 1992; Nagaraju et al., cancer cells are first detected in the basal cell
2010; Chaudhari et al., 2013). layer (Acha et al., 2005).
A few other vital stains, such as methylene Usually, slides are immediately fixed with
blue and rose bengal, have a similar staining 95% ethyl alcohol (96° GL), which enables further
profile and performance to toluidine blue (Chen staining with routine staining methods, such as
et al., 2007; Du et al., 2007). Papanicolaou, haematoxylin and eosin (H&E),
periodic acid–Schiff (PAS), or Feulgen tech-
4.1.4 Cytology and quantitative DNA niques, among others (Pérez-de-Oliveira et al.,
cytometry 2020).
(a) Cytology Subsequent laboratory processing methods
include simple centrifugation, cytocentrifuge
The use of cytology was introduced by preparation, or cell blocks. The cytocentrifuge
Papanicolaou and Traut (1943) to detect cervical approach, which was developed to overcome
cancer. Since the 1950s, exfoliative cytology and the issues of insufficient material when using
then brush biopsy cytology were increasingly simple centrifugation, enables better results
used as practical, low-risk, and low-cost diag- in processing specimens. Fresh samples are
nostic tools for the initial evaluation of OPMDs collected in anticoagulant vials, loaded into
and oral cancer (Silverman, 1959; Sciubba, an automated cytospin machine, and centri-
1999; Böcking et al., 2011; Koch et al., 2011; fuged. Slides containing smears prepared by the
Nanayakkara et al., 2016). cytospin technique are then fixed in 95% ethyl
Oral cavity samples are collected with a alcohol for 20–30 minutes and stained with
wooden or metallic spatula (scrape biopsy or H&E, Papanicolaou, or PAS techniques (Qamar
exfoliative biopsy), a curette, or a cytological et al., 2018). A modified Papanicolaou staining
brush (cytobrush biopsy), which is rubbed or procedure can be carried out in clinical settings
scraped (in the case of a spatula) or rotated that require faster decision-making processes
(in the case of the cytobrush) on the surface (Thakur and Guttikonda, 2017).
of the lesion and then spread onto a glass slide In liquid-based cytology, the cytobrush-col-
for analysis. Exfoliative cytology collects only lected specimen is placed into a vial containing
superficial cells, whereas cytobrushes can collect preservative fluid before transportation to the
306
laboratory where the specimen is processed, 2019); however, the higher cost can be a substan-
i.e. with cytospin and staining (modified tial problem in low-resource settings.
Papanicolaou, or Feulgen in the case of DNA The exfoliative and brush biopsy techniques
ploidy; see below) or for flow cytometry were compared in a prospective study of patients
(Hutchinson et al., 1994; Khandelwal and with leukoplakia (116 lesions) and lesions
Solomon, 2010; Olms et al., 2018). In the CDx with a suspicion of malignancy (76 lesions)
system, the cytology results are reported as (Nanayakkara et al., 2016). When only positive
positive (for dysplasia or carcinoma), atypical results were considered [“high-risk” lesions
(cellular changes of uncertain diagnosis), nega- defined as smears with any degree of dysplasia
tive (normal cells), or inappropriate (incomplete or malignant cells], compared with histopatho-
sample) (Sciubba, 1999; Mehrotra et al., 2011; logical end-points of OSCC, the brush technique
Nanayakkara et al., 2016). In other reporting had a sensitivity of 89.6% and a specificity of 100%,
systems, the categories may be different. and the exfoliative technique had a sensitivity
Cytology with exfoliative biopsy yields high of 60.4% and a specificity of 95.2%. When the
false-negative rates (up to 31%) (Folsom et al., histopathological end-points included moderate
1972). Modified liquid-based cytology with brush dysplasia or worse, the accuracy increased.
biopsy improves the diagnostic accuracy of Recent reviews of the performance of cytology
cytology for OPMDs and oral cancer (Delavarian for detection of oral cancer and OPMDs are
et al., 2010; Navone et al., 2011; Deuerling et al., presented in Table 4.8. In a review and meta-anal-
2019). When the preparation methods of conven- ysis of 16 studies (Lingen et al., 2017b), cytology
tional cytology (transfer procedure to glass in patients with OPMDs had the highest accu-
slides) and liquid-based cytology are compared, racy among all reviewed adjuncts, with a sensi-
liquid-based preparations show a more uniform tivity of 92% (95% CI, 86–98%) and a specificity
distribution and less cellular overlapping, of 94% (95% CI, 88–99%).
cellular deformation, mucus, microbial colonies, A recent review of 24 data sets compared the
and debris compared with those of conventional accuracy of cytology when using a cytobrush
cytology (Olms et al., 2018). Liquid-based plat- (n = 16) or scraping (n = 3) to harvest cells. The
forms also have technical advantages, including overall sensitivity was 90% (95% CI, 82–94%),
(i) enabling immediate fixation of cells while and the specificity was 94% (95% CI, 88–97%).
removing unwanted harvested material (e.g. For cytobrush, the sensitivity was 91% (95% CI,
mucus and debris), (ii) producing thin layers with 81–96%) and the specificity was 94% (95% CI,
a clear background and producing more homo- 87–97%); for scraping, the sensitivity was 93%
geneous samples than conventional smears, and (95% CI, 87–96%) and the specificity was 92%
(iii) reducing the proportion of unsatisfactory (95% CI, 81–97%) (Walsh et al., 2021a).
samples (Hayama et al., 2005; Deuerling et al.,
Table 4.8 Performance of cytology for detection of oral cancer and OPMDs

(data sets)
Lingen et al. (2017b) 16 2148 Clinically evident, suspicious lesions 92 (86–98) 94 (88–99)
Walsh et al. (2021a) 24 1950 Oral cancer and OPMDs 90 (82–94) 94 (88–97)
307
Table 4.9 Performance of DNA cytometry for detection of oral cancer and OPMDs
Reference No. of No. of lesions Outcome measured Sensitivity (%) Specificity (%)
patients (95% CI) (95% CI)
Maraki et al. (2004) 98 98 Oral cancer 100a 97.4a
Ng et al. (2012) 171 199 Oral cancer and OPMDs 89.3a 96.5a
Walsh et al. (2021a)b 216 525 Oral cancer and OPMDs 76 (68–82) 98 (72–99)
a 95% CI not reported.
b Meta-analysis with 5 studies.
In a prospective trial, Sciubba (1999) analysed DNA cytometry has been used in the context of
the accuracy of brush biopsy with computer-as- early diagnosis of oral cancer and OPMDs (Tong
sisted sample analysis. Of the 298 cases with et al., 2009).
lesions judged to be clinically suspicious that A recent review included 24 data sets, of
underwent brush and scalpel biopsy [excisional which 5 used DNA cytometry. The pooled sensi-
biopsy], 102 were malignant. The sensitivity of tivity was 76% (95% CI, 68–82%), and the pooled
brush biopsy was 100%, and the specificity was specificity was 98% (95% CI, 72–99%) (Walsh
100% for positive results [definitive cellular et al., 2021a) (Table 4.9).
evidence of epithelial dysplasia or carcinoma] In a series of 98 cytobrush and scalpel biop-
and 92.9% for atypical results [abnormal epithe- sies of clinically evident lesions, 75 samples were
lial changes of uncertain diagnostic significance]. cytologically and histologically negative (the
To evaluate the feasibility of oral brush biopsy cut-off for true positive was severe dysplasia or
in resource-constrained settings, Mehrotra et al. carcinoma). The remaining 23 samples, which
(2008) evaluated 94 patients with OPMDs or had positive (15 cases), suspicious (4 cases), or
oral cancer using a baby toothbrush followed by doubtful (4 cases) cytological results, underwent
scalpel biopsy, and the specimens were analysed DNA cytometry, and 19 of the 23 cases showed
without computer-assisted analysis. The speci- aneuploidy (a sensitivity of 100% and a specificity
mens were adequate in 74 cases, with a sensitivity of 97.4%) (Maraki et al., 2004).
of 76.8% and a specificity of 93.3%. In a retrospective review of 171 patients
Experts from the American Dental Associa- with 199 suspicious oral lesions who underwent
tion recommend the use of cytology as a triage biopsy and quantitative cytology, 28 patients
tool in primary care settings or if the patient had OPMDs with OED or OSCC, of whom 25
refuses a tissue biopsy (Lingen et al., 2017a). had positive quantitative cytology. False-positive
quantitative cytology was observed in 5 of the
(b) Quantitative DNA cytometry 143 patients with negative histology; the sensi-
DNA cytometry, which is used to detect the tivity was 89.3%, and the specificity was 96.5%
cytometric equivalent of chromosomal aneu- (Ng et al., 2012).
ploidy, was developed as an adjunctive technique
to improve the accuracy of cytology. Aneuploidy 4.1.5 Liquid biopsy
is defined as an alteration of the chromosome
Liquid biopsy is a non-invasive, conven-
number that is not a multiple of the haploid
ient, and low-cost method, and it is easy to
complement (Williams and Amon, 2009).
collect liquid samples (Mali and Dahivelkar,
Because aneuploidy is frequent in cancer cells,
2021). Tumour DNA was detected in 100% of
308
saliva samples from patients with oral cancer, tumour necrosis factor α (TNF-α), are present at
suggesting that saliva is preferentially enriched significantly higher concentrations in the saliva
with tumour DNA from tumours at this site of patients with oral cancer compared with that
(Wang et al., 2015). The diagnostic and prognostic of healthy people (Rezaei et al., 2019; Ferrari et al.,
applications of “salivaomics” (Wong, 2012) for 2021). Another systematic review also identified
oral cancer have been extensively explored, with IL-8 mRNA as a potential candidate (Gaba et al.,
the identification of many potential biomarkers: 2021).
minerals, peptides, proteins, DNA, messenger The most recent systematic review of sali-
RNA (mRNA), microRNA (miRNA), long vary diagnostic biomarkers for oral cancer and
coding RNA, oxidative stress-related molecules, OPMDs, which included 295 articles (Piyarathne
glucocorticoids, glycosylation-related molecules, et al., 2021), included proteomic biomarkers,
telomerase activity, and the microbiome (Li et al., cytokines, growth factors, angiogenic factors,
2004; Jou et al., 2011; Cheng et al., 2014; Yu et al., antigens, cytokeratin, cell surface receptors,
2016; Amer et al., 2017; Kaczor-Urbanowicz et al., enzymes, and silencing of tumour suppressor
2017; van Ginkel et al., 2017; Payne et al., 2018, genes via promoter hypermethylation. From
2019; Chen and Zhao, 2019; Rapado-González the reported data, IL-1β, IL-6, and IL-8 were
et al., 2019; Hofmann et al., 2020). However, selected as the most suitable salivary biomarkers
saliva testing has not yet been incorporated for early detection of OSCC and OPMDs. [Most
into commercial products or clinical practice of the studies were graded with fair quality and
(Masthan et al., 2012; Walsh et al., 2021a). moderate risk of bias.]
The role of cytokines and other proteins as Matrix metalloproteinases are also prom-
promising salivary biomarkers for oral cancer ising saliva biomarkers. Stott-Miller et al. (2011)
has been shown consistently in numerous observed that the concentrations of MMP-1 and
studies. In a large study that included five MMP-3 were higher in later stages of oral cancer
cohorts (169 cases and 226 controls), interleukin compared with controls, cases with dysplasia,
8 (IL-8) and SAT mRNA had the highest predic- and early-stage tumours, with an area under the
tive values (Elashoff et al., 2012). In a single study, curve (AUC) of the receiver operating character-
the combination of the three biomarkers IL-8, istic (ROC) curve of 0.845 for MMP-1 and 0.877
SAT, and H3F3A increased the sensitivity and for MMP-3. Chang et al. (2020) also identified
specificity to predict the presence of oral cancer MMP-1 as the most promising candidate from
compared with each of the biomarkers separately a panel of proteins, with a sensitivity of 76.6%
(Li et al., 2004). and a specificity of 86.8%. In a systematic review,
In one systematic review, high sensi- Hema Shree et al. (2019) observed a high sensi-
tivity and specificity were observed for IL-8, tivity for MMP-9 (95%; 95% CI, 88–100%) and
choline, pipecolinic acid, L-phenylalanine, chemerin (100%; 95% CI, 78–100%), with a spec-
and S-carboxymethyl-L-cysteine; however, the ificity of 100% for both MMP-9 and chemerin.
combination of different biomarkers did not In a systematic review of six studies (with a total
improve sensitivity or specificity (Guerra et al., of 775 participants), high performance rates
2015). In another systematic review, the proteins were reported for MMP-9 and for CYFRA 21-1
found most frequently were IL-8, CD44, matrix (Gualtero and Suarez Castillo, 2016; AlAli et al.,
metalloproteinase-1 (MMP-1), and MMP-3 2020).
(Gualtero and Suarez Castillo, 2016). Recent Several reviews and meta-analyses have
systematic reviews and a meta-analysis showed highlighted the diagnostic accuracy of miRNAs
that numerous cytokines, such as IL-6, IL-8, and in differentiating patients with oral cancer from
309
healthy controls (Tian et al., 2015; Arantes et al., most enriched species were Fusobacterium,
2018; Al Rawi et al., 2021; Liu et al., 2021). The Leptotrichia, Campylobacter, and Rothia species;
most recent meta-analysis, which included severe dysplasia was associated with specific
1106 patients and 732 controls, found a pooled microbial enrichments (Leptotrichia spp. and
sensitivity of salivary miRNAs of 70%, a pooled Campylobacter concisus) (Amer et al., 2017).
specificity of 82%, and an AUC of 0.80 (Liu [Despite the great potential of saliva
et al., 2021). A previous meta-analysis based on biomarkers in the diagnosis of OPMDs and oral
23 studies found a pooled sensitivity of 75.9%, a cancers, and the rapidly evolving knowledge in
pooled specificity of 77.3%, and an AUC of 0.83 the field and the consistently high accuracy of
(Tian et al., 2015). Among a panel of miRNAs some of the biomarkers in a research setting,
in saliva samples from patients with head and there is a lack of clinical validation regarding this
neck cancer (comprising cancers of the oral approach in oral cancer screening settings.]
cavity, oropharynx, larynx, and pharynx) and
from healthy controls, miR-9, miR-191, and 4.1.6 Use of emerging technologies in the
miR-154 had excellent discriminatory power, primary screening setting
with an AUC of 0.85, 0.74, and 0.98, respectively
(Salazar et al., 2014). Momen-Heravi et al. (2014) (a) Artificial intelligence for identification of
performed a genome-wide evaluation of miRNA OPMDs
patterns in saliva samples from patients with Artificial intelligence (AI) is defined as the
oral cancer, patients with oral lichen planus, and process by which a computer is able to learn by
healthy controls and observed that miR-27b had continuously incorporating new data into an
a sensitivity of 85.7% and a specificity of 100% existing statistical model (Deo, 2015). A prom-
for detection of oral cancer (AUC, 0.96; 95% CI, ising new approach to improve the detection and
0.88–1.05). diagnosis of OPMDs is to engage the interest of
Aberrant methylation of tumour suppressor mathematicians with expertise in AI or machine
genes is an important epigenetic mechanism of learning to apply these techniques to improve
carcinogenesis. Several genes have been found the clinical diagnosis of oral cancer and OPMDs
to be more frequently hypermethylated in saliva (Kar et al., 2020; García-Pola et al., 2021).
samples from patients with oral cancer than Several groups have investigated the use of
in those from controls (Carvalho et al., 2008; AI to improve the efficacy of COE (García-Pola
Arantes et al., 2018; Rapado-González et al., et al., 2021; Ilhan et al., 2021), and the prelimi-
2021a, b). In a meta-analysis of 18 studies, the nary findings have been promising.
frequency of methylation was higher in patients
with head and neck cancer (comprising mostly (b) Optical coherence tomography
cancers of the oral cavity) than in healthy controls Optical coherence tomography (OCT) is
(odds ratio, 8.34; 95% CI, 6.10–11.39); a signifi- an optical technology that uses back scattered
cant association between methylation of specific signals from different layers of tissue to construct
tumour-related genes and risk of head and neck in vivo cross-sectional images of tissue with high
cancer [not otherwise specified] was observed resolution (Huang et al., 1991; Machoy et al.,
for p16, MGMT, DAPK, TIMP3, and RASSF1A 2017). This technology is similar to that used in
(Rapado-González et al., 2021b). ultrasound, but whereas ultrasound uses sonic
Finally, changes in the microbiome have signals to generate tissue images, OCT uses
been associated with risk of oral cancer (Perera optical signals.
et al., 2016). In dysplastic leukoplakia, the
310
OCT has been used for many years for the (c) In vivo microscopy
evaluation and diagnosis of retinal lesions Whereas OCT provides a cross-sectional
(Fujimoto, 2003). Wilder-Smith et al. (2009) eval- image of the oral mucosa and submucosa, reflec-
uated the use of OCT for diagnosis of oral cancer tance microscopy and fluorescence microscopy
and OPMDs in 50 patients and found strong provide images of the oral mucosal surface
agreement between the diagnosis based on OCT (Muldoon et al., 2012). Emerging reflectance
images and that based on histology. Heidari et al. microscopy technologies, including those that
(2019) developed a portable OCT system and can analyse vascular patterns in the oral submu-
used it to evaluate oral lesions in 20 patients and cosa, are adequate to visualize oral tissue without
10 healthy individuals. Whereas previous studies use of contrast agents. However, most fluores-
had compared the qualitative evaluation of OCT cence microscopy approaches require the use of
images and histological images, in this small an optical contrast agent, either applied topically
study the researchers developed an objective or administered intravenously.
algorithm to differentiate between normal and Muldoon et al. (2012) described a new
abnormal oral mucosa based on the OCT images. high-resolution optical microscopy (high-res-
They reported a sensitivity and specificity of this olution microendoscope [HRME]), fluores-
algorithm for differentiating between healthy cence microscope (Yang et al., 2018b), which
and cancerous or dysplastic mucosa of 95% and could provide real-time images of the nuclear
100%, respectively, and a sensitivity and spec- morphology of the oral mucosa. To enable visu-
ificity for differentiating between cancer and alization of the nuclei, topical application of the
dysplasia of 91% and 100%, respectively. fluorescent dye proflavine was required. The
James et al. (2021) provided validation of a images obtained could be saved for further anal-
point-of-care OCT diagnostic device based on an ysis of the size and shape of the nuclei by an auto-
automated algorithm, which was used to examine mated computer algorithm (Yang et al., 2018b).
232 individuals across a spectrum ranging from Autofluorescence (see Section 4.1.3) has low spec-
normal mucosa to OPMDs and oral cancer. ificity for identifying benign lesions. To boost the
The process included first imaging the lesion specificity, a multimodal approach was suggested
and then providing the image to the algorithm of merging autofluorescence with HRME tech-
for further interpretation. The algorithm score nology (Yang et al., 2018b). Subsequent studies
was compared with standard histopathological that used the HRME instrument, alone and in
diagnoses if biopsy was indicated. The algorithm combination with wide-field autofluorescence
score was unable to distinguish between the imaging devices, have documented the ability of
grades of dysplasia, but it accurately differenti- this technology to objectively identify abnormal
ated oral cancers (OSCC, with a sensitivity of and dysplastic mucosa with high sensitivity
93%) and OPMDs (with a sensitivity of 95%) from and specificity (Yang et al., 2018a, 2019, 2020).
benign lesions and normal mucosa. To provide However, this HRME technology is not yet avail-
the delineation of high-grade dysplastic lesions able for clinical use.
(moderate or severe dysplasia) from low-grade Nathan et al. (2014) reported on a prelimi-
lesions (mild dysplasia, benign, or normal), the nary study of 21 participants with oral cancer
research team implemented the use of an artifi- or OPMDs, who underwent imaging of lesions
cial neural network, which reached a sensitivity with confocal laser endomicroscopy for in vivo
of 83% (James et al., 2021). evaluation of the oral mucosa before resection
or excisional biopsy. To provide optical contrast,
311
the participants underwent intravenous injec- Although the reflected light is usually the
tion of fluorescein before imaging. Qualitative same wavelength as the illumination light
analysis of the images by experts familiar with source, in rare cases light is reflected at a different
this technology was compared with histolog- wavelength, due to inelastic scattering (Bigio
ical diagnosis. The overall sensitivity was 80% and Bown, 2004; Sahu and Krishna, 2017). These
for diagnosis of dysplasia versus non-dysplasia. inelastic reflectance signals, which are often
Despite these initial positive findings, this tech- called Raman signals, are very faint compared
nology has not yet been adopted for clinical with fluorescence and standard reflectance
evaluation of patients with oral mucosal lesions, signals. However, spectroscopic analysis of
possibly due to the need for intravenous injection Raman signals can provide objective documen-
of fluorescein before imaging. tation of chemical changes in biological tissues
(Bigio and Bown, 2004; Sahu and Krishna, 2017).
(d) Spectroscopy Raman spectroscopy is a technology that enables
In contrast to optical imaging technologies non-invasive, molecular interrogation of the
such as OCT and microscopy, optical spectros- chemical composition of biological tissues, using
copy involves the objective detection and analysis optical interrogation. Four biological compo-
of optical signals collected after tissue is exposed nents contribute to Raman signals: nucleic acids,
to light of various wavelengths. Basically, clinical lipids, proteins, and water (Bigio and Bown, 2004).
spectroscopy is the analysis of how light interacts Several studies have investigated the potential
with tissue (Sahu and Krishna, 2017). Alterations efficacy of Raman spectroscopy to discriminate
in spectroscopic signals can be used to detect between oral cancer or OPMDs and benign or
biochemical and architectural changes in oral normal oral mucosa. These studies refer to the
tissue that are associated with neoplastic progres- possible use of this technology both ex vivo, with
sion (Müller et al., 2003; Bigio and Bown, 2004). the use of formalin-embedded tissues (Ibrahim
Several different types of spectroscopic analysis et al., 2021) and biopsies (Matthies et al., 2021),
have been evaluated for use in the detection of and in vivo, with possible clinical use indicating
oral cancer, including Raman spectroscopy, fluo- a potential novel adjunctive diagnostic technique
rescence spectroscopy, reflectance spectroscopy, (Sahu et al., 2012).
elastic scattering spectroscopy, and time-resolved In contrast, elastic scattering spectroscopy
autofluorescence spectroscopy. The distinction relies on gradients in the optical index of refrac-
between these spectroscopic technologies is tion after the light is scattered by specific orga-
based on multiple factors, including the type of nelles inside the cell (e.g. nuclei or mitochondria).
light illumination delivered to the tissue and the This spectroscopic method depends on the differ-
type of optical signal detected after this illumina- ences in the densities of the organelles; the elastic
tion (Sahu and Krishna, 2017). These differences scattering spectrum may change in cells under-
arise as a result of how light interacts with tissue. going carcinogenesis (Bigio and Bown, 2004).
For example, fluorescence spectroscopy involves Fluorescence and reflectance spectroscopy
illumination of tissue at wavelengths that are technologies have been used to evaluate oral
known to stimulate autofluorescence by tissue mucosal lesions in vivo (Schwarz et al., 2008;
components such as collagen, and collection of Messadi et al., 2014). Although these prelimi-
the autofluorescence light emitted from the illu- nary studies have shown promise for the ability
minated tissue at specific wavelengths (Romano of these technologies to discriminate between
et al., 2021). Reflectance spectroscopy involves normal or benign oral tissue and dysplastic or
assessment of the light reflected from tissue.
312
cancerous oral tissue, they showed insufficient programme requires that dentists provide oral
sensitivity and specificity. visual inspection annually in community dental
clinics and refer suspicious cases to the regional
(e) Molecularly targeted optical imaging head and neck and maxillofacial surgical service
agents for further management. A formal evaluation
Given that the standard COE and radio- of the programme for the period 1984–1990
graphic imaging are insufficient to determine was carried out in collaboration with IARC
the extent of OSCC in many patients, several (Fernández Garrote et al., 1995). The programme
molecularly targeted optical imaging agents have covered 12–26% of the target population annu-
been developed over the past decades to improve ally, and less than 30% of the individuals with
the surgeon’s ability to delineate the anatom- suspicious lesions complied with referral to the
ical extent of malignant tissue and high-grade maxillofacial surgical service. The programme
dysplastic disease, before or during surgical identified about 16% of the 4412 incident oral
resection (Fakurnejad et al., 2019; van Keulen cancers in Cuba during 1984–1990. After the
et al., 2019; Steinkamp et al., 2021). formal evaluation of the programme, the age
[Although these clinical trials may offer threshold for the target group was increased
new techniques to improve surgical resection to ≥ 35 years as part of reorganization efforts
of oral cancer, it is unclear how these molec- (González, 2014). No further formal evaluation
ularly targeted optical imaging agents might of the reorganized programme has been done
improve the early detection and diagnosis of oral since 1995.
precancer and cancer in individuals at high risk, A nationwide population-based oral cancer
particularly in low-resource settings.] screening programme, which conducts oral
visual inspection every 2 years, has been running
in Taiwan (China) since 2004. It targets resi-
4.2 Organized and opportunistic dents aged ≥ 30 years with a history of ciga-
oral cancer screening activities rette smoking and/or betel quid chewing,
and Indigenous people aged ≥ 18 years.
Worldwide, there are very few large-scale
In 2004–2009, about 55% of invited individ-
population-based organized or non-organized
uals (n = 4.2 million) participated in screening
oral cancer screening programmes, and there
(Chuang et al., 2017). More than 4.6 million
is very little sporadic screening activity. This is
individuals with the exposure of betel quid
despite the fact that most patients with oral
chewing and/or cigarette smoking have attended
cancer present in advanced stages with poor
the biennial oral cancer screening. A nationwide
prognosis. Previous reviews of oral cancer
online information system for breast cancer,
screening have concluded that there is “insuffi-
colorectal cancer, and oral cancer screening was
cient evidence to recommend inclusion or exclu-
successfully developed to support health profes-
sion of oral cancer screening” in the general
sionals and health decision-makers for planning,
population, and that opportunistic screening
delivery, management, and evaluation in the
of populations at high risk might be effective
population-based cancer screening programme
and should be considered (Hawkins et al., 1999;
(Lin, 2018).
Kujan et al., 2005; Brocklehurst et al., 2013).
India accounts for the largest contribution
A large-scale population-based oral cancer
to the burden of oral cancer globally (Ferlay
screening programme in people aged ≥ 15 years
et al., 2020). Although the Government of India
has been under way in Cuba since 1982. The
has issued guidelines for oral cancer screening
313
for all individuals in the age group 30–65 years State, Brazil (Almeida et al., 2012). In 2001–2008,
(National Health Mission of India, 2021), these 2 229 273 individuals were screened, with an
have yet to be implemented systematically on a increase in coverage from 4.1% in 2001 to 16% in
large scale and have mostly resulted in sporadic 2008, a decrease in the percentage of suspicious
screening. The draft national oral health policy lesions from 9% in 2005 to 5% in 2008, and a
released in February 2021 by the Ministry of decrease in the rate of confirmed cases of oral
Health and Family Welfare of India (Ministry of cancer per 100 000 examinations from 20.9 in
Health and Family Welfare, 2021) also empha- 2001 to 10.4 in 2008.
sizes the need for screening, but it provides no No population-based oral cancer screening
clear direction or roadmap on how to achieve programmes have been reported in Europe,
this. The Government of Tamil Nadu State in North America, or Oceania.
India has organized an oral cancer screening
programme since 2016 through public health
services. This programme targets people aged
4.3 Determinants of participation in
≥ 18 years who are users of tobacco and/or screening for oral cancer
alcohol (National Health Mission Tamil Nadu, The World Health Assembly adopted the
2021). It is supported by an information system, first resolution related to oral cancer diagnosis
but no data have yet been published from this in 2007, and the World Health Organization
programme. In an opportunistic oral cancer has formally provided guidance for oral health
screening activity, 1 061 088 people in 265 272 (WHO, 2007, 2013, 2021). Despite this, most
houses were surveyed in Kannur District, Kerala, countries have not widely adopted or reported
India (Philip et al., 2018). oral cancer screening. In addition, the litera-
Sporadic oral cancer screening involving ture on the determinants of participation in oral
small numbers of individuals has been cancer screening is scarce.
conducted both in India and in Sri Lanka, It is critical to identify and monitor the factors
demonstrating the feasibility of MSE and/or that positively and negatively influence cancer
home-based screening by community health screening programmes and their outcomes, in
workers, but such activities do not resemble order to facilitate translation of the scientific
sustained programmatic efforts (Amarasinghe evidence of benefit to the clinical setting. The
et al., 2016; Philip et al., 2018; Basu et al., 2019). predictors of participation in cancer screening,
Guidelines have been developed by the National adherence to follow-up screening rounds, and
Cancer Control Programme of Sri Lanka for compliance with referrals for diagnosis and treat-
oral cancer screening and management of oral ment are well established in the literature (Solar
lesions, targeting users of tobacco and areca nut and Irwin, 2010). They consist of (a) drivers that
(National Cancer Control Programme, 2020); influence the process at the level of (i) the indi-
however, these have not resulted in a sustained vidual, (ii) health-care providers, (iii) health-
programmatic activity. care systems, and (iv) health-care policies, and
There has been very little oral cancer (b) interventions to increase participation in
screening activity in Central and South screening (Table 4.10).
America. Since 2001, the São Paulo State
Health Secretariat has coordinated oral cancer
screening with annual COE, combined with the
national campaign for influenza immunization
of the population aged ≥ 60 years in São Paolo
314
Table 4.10 Determinants of participation in screening for oral cancer
Category of Facilitator Barrier Reference Location

determinant
Individual level
Risk factors
Smoking Talamini et al. (1994) Italy
Smoking Chang et al. (2011) Taiwan (China)
Smoking Ramadas et al. (2008a) India
Betel quid chewing Chang et al. (2011) Taiwan (China)
Alcohol consumption Nagao and Warnakulasuriya (2003) Japan
Alcohol consumption Ramadas et al. (2008a) India
Age and sex
Age (45–54 years) Age (> 65 years) Ramadas et al. (2008a) India
Age (40–60 years) Age (< 40 years and > 60 years) Chang et al. (2011) Taiwan (China)
Middle-aged (55–64 years) Younger and elderly (< 55 years and Talamini et al. (1994) Italy
≥ 65 years)
Elderly women Young and middle-aged women Mishra et al (2021) India
Female sex Ramadas et al. (2008a) India
Female sex Talamini et al. (1994) Italy
Socioeconomic factors
Hindu religion Mishra et al. (2021) India
Marathi mother tongue Mishra et al. (2021) India
High secondary school education Mishra et al. (2021) India
Owning mass media devices Ramadas et al. (2008a) India
(television and/or radio)
Larger household size Ramadas et al. (2008a) India
Medical factors
Absence of symptoms Talamini et al. (1994) Italy
Family history of cancer Mishra et al. (2021) India

Health-care system level
Inadequate patient referral Warnakulasuriya et al. (1984) Sri Lanka
315
4.3.1 Individual level (c) Socioeconomic factors

(a) Risk factors Mishra et al. (2021) evaluated socioeconomic
determinants of participation in oral cancer
Populations at high risk (i.e. individuals with
screening by women with current smoking habits
risk factors for oral cancer, such as tobacco use,
or previous smoking habits (for ≥ 3 consecu-
areca nut use, and alcohol consumption) were
tive years) in an organized population-based
found to be more likely to adhere to oral cancer
screening programme in Mumbai, India. High
referral consultations and procedures, compared
secondary school education level, Hindu reli-
with individuals without these risk factors
gion, and Marathi mother tongue were all
(Nagao and Warnakulasuriya, 2003; Ramadas
positive factors associated with participation
et al., 2008a; Chang et al., 2011), except in one
in oral cancer screening. In addition, Ramadas
study (Talamini et al., 1994), in which smoking
et al. (2008a) identified larger household size and
habits were negatively associated with compli-
owning mass media devices (television and/or
ance with referral.
radio) as socioeconomic factors associated with
(b) Age and sex higher participation rates.
In three studies, middle-aged patients were (d) Medical factors
more likely to comply with screening proce-
In a study of screening for head and neck
dures, compared with elderly patients and
cancer (including oral cancer), patients with
younger patients (Talamini et al., 1994; Ramadas
upper aerodigestive tract symptoms (Talamini
et al., 2008a; Chang et al., 2011). In contrast,
et al., 1994) or a family history of cancer (Mishra
Mishra et al. (2021) reported that elderly women
et al., 2021) were more likely to attend screening
were more likely to participate in oral cancer
than those who were asymptomatic.
screening, followed by younger women and
middle-aged women.
Inconsistent findings describe female sex 4.3.2 Health-care provider level
as a positive predictor (Ramadas et al., 2008a) Trained health-care providers are more
and a negative predictor (Talamini et al., 1994) likely to promote oral cancer screening. For
of participation in oral cancer screening. In instance, in a study in Ernakulam District,
the study of Ramadas et al. (2008a), accrual of Kerala, India, about 53 basic health workers were
individuals was based on home visits in India, trained by dentists to examine the oral cavity of
and the authors argued that their finding may individuals at high risk and recognize suspicious
be explained by the fact that in the population cancerous and precancerous lesions. Within
evaluated, women are more likely to be at home a 1-year period, screening participation of the
during home visits than their male partners. [It target population increased to 33.5% (Mehta
is not clear whether women are more likely than et al., 1986). In addition, 45% of individuals were
men to attend screening.] correctly referred, with a sensitivity of 59%. Thus,
[Although age and sex were important the training of basic health workers specifically
predictors in the above-mentioned studies for oral cancer screening through active recruit-
(Talamini et al., 1994; Ramadas et al., 2008a; ment dramatically changed the participation
Mishra et al., 2021), these determinants were not rate.
consistent within and between the studies; this
may be explained by confounders, biases in anal-
ysis, and study design.]
316
4.3.3 Health-care system level Pivovar et al. (2017) described an e-health

strategy to increase the selection of individuals
Warnakulasuriya et al. (1984) recognized at high risk, followed by an active home-based
that in Sri Lanka only 50% of individuals with a invitation to schedule oral cancer screening.
suspicious lesion detected by primary health- Selecting individuals at high risk through an
care workers were re-examined by skilled electronic database enabled improved effi-
professionals at the university referral centre. ciency and reduced the percentage of potential
The authors concluded that the low compliance participants to 1.4% of the total population.
of the community with follow-up at the referral [The Working Group noted that no comparison
centre may have been due to lack of awareness arm was provided to evaluate the magnitude of
about oral cancer and the value of the screening the impact associated with such an interven-
programme, and possibly an inadequate under- tion. The study was also sex-biased, by excluding
standing between the individuals and the health women at high risk.]
workers about referral. Jedele and Ismail (2010) conducted a 2-year
oral cancer awareness and screening campaign
4.3.4 Health-care policies that targeted African-American men aged
Only a few countries, such as Cuba, India, ≥ 40 years. The number of billboards and radio
Malaysia, Sri Lanka, and Taiwan (China), have advertisements was positively correlated with the
adopted oral cancer screening on a large scale. number of calls received on the campaign’s toll-
However, the determinants of participation free hotline number. Also, the calls to the toll-
have most often not been reported. Overall, the free number resulted in scheduled appointments
above-mentioned countries promote distinct and screening of patients.
screening programmes in terms of target popu-
lation, coverage, design, and framework; these 4.4 Effectiveness of screening
differences pose challenges for harmonization
and comparison of data in terms of not only the In 2013, Brocklehurst et al. (2013) conducted
health impact but also the determinants of partic- the most recent systematic review of RCTs on
ipation in screening (Warnakulasuriya et al., screening for oral cancer or OPMDs using
1984; López Cruz et al., 2003; Sankaranarayanan COE, toluidine blue vital staining, fluorescence
et al., 2013; Moyer et al., 2014). imaging, or brush biopsy. Based on the only RCT
that met the inclusion criteria (Sankaranarayanan
4.3.5 Strategies to increase participation in et al., 2000, 2013), they concluded that as an alter-
oral cancer screening native to a national-based screening programme,
opportunistic oral cancer screening by visual
Studies have reported on various endeav- examination in a population at high risk might
ours to increase participation in oral cancer be effective in reducing oral cancer mortality.
screening, including individual invitations
through billboards, radio advertisements, news- 4.4.1 Preventive effects of screening
paper advertisements, toll-free hotlines, letters,
home visits, educational leaflets, and phone calls To ascertain the effect of screening on oral
(Jedele and Ismail, 2010; Pivovar et al., 2017). cancer incidence and/or mortality, a search was
However studies designed to specifically evaluate performed for experimental and observational
the efficacy of these interventions are scarce and studies that used “no screening” as the control
have biases. group and that reported incidence of advanced
317
and/or early oral cancer and mortality from oral After four rounds of screening in the inter-
cancer. The Working Group identified one experi- vention arm, there was a statistically non-sig-
mental study, from which the outcomes observed nificant (12%) overall reduction in oral cancer
with the longest follow-up were extracted. No mortality compared with the control arm
current experimental studies targeted at meas- (Table 4.12). However, in users of tobacco and/
uring incidence of advanced oral cancer and or alcohol, per-protocol analysis showed a statis-
mortality from oral cancer were identified. In tically significant (24%; 95% CI, 3–40%) reduc-
addition, three observational studies reporting tion in oral cancer mortality and a statistically
the primary end-points for performance of oral significant (21%; 95% CI, 5–35%) reduction in
cancer screening for the screening and control incidence of advanced oral cancer (clinical stages
groups were identified. III and IV). The reduction in both incidence of
advanced oral cancer and mortality from oral
(a) Randomized controlled trials cancer increased with the number of rounds of
In the Trivandrum Oral Cancer Screening screening (Sankaranarayanan et al., 2013). To
Study, in Kerala, India, healthy residents aged adjust an imbalance in risk of oral cancer between
≥ 35 years from 13 rural administrative units, the two arms in this study, an intention-to-treat
considered as clusters, were randomized into analysis was recently performed based on the
an intervention arm (n = 7) and a control arm 9-year follow-up; this analysis demonstrated a
(n = 6) (Sankaranarayanan et al., 2000, 2005, 27% reduction in oral cancer mortality due to
2013; Ramadas et al., 2003). Eligible individ- screening (hazard ratio, 0.73; 95% CI, 0.54–0.98)
uals were identified through interviews during (Cheung et al., 2021).
home visits; they provided information about [The Kerala trial has multiple limitations,
their demographic characteristics and indi- in particular related to a high non-compliance
vidual habits related to risk factors for oral rate in screen-positive individuals, i.e. only 59%
cancer (i.e. tobacco use and alcohol consump- of screen-positive individuals complied with the
tion). The longest reported follow-up of this clinical assessment by the physicians. The publi-
trial was 15 years (until December 2010) cation does not describe well whether and how
(Sankaranarayanan et al., 2013; Table 4.11). All the interval cancers were followed up. The cancers
intervention health workers were taught about that developed in the non-compliant individuals
cancer and trained in oral cancer screening. Of were included in the no-screening group, which
the 96 517 eligible individuals in the interven- assumes per-protocol analysis instead of inten-
tion arm, 25 144 (26.1%) underwent one round tion-to-treat analysis; however, the intention-to-
of screening, 22 382 (23.2%) underwent two treat analysis performed later reached a similar
rounds, 22 008 (22.8%) underwent three rounds, conclusion. No formal training certificate was
and 19 288 (20.0%) underwent four rounds. issued to the health workers; however, all the
Eligible individuals in the control arm received health workers underwent an examination
routine care in 1996–2005 and were offered at the end of the training to test their skills in
screening in 2006–2008, in which 43 992 (46.1%) completing the questionnaire and also in identi-
of 95 356 individuals participated. Participants fying the relevant lesions in the oral cavity. Those
with positive screening results were referred whose performance was poor were retrained. It is
for further clinical examination by a specialist possible that the health workers’ lack of a certif-
(either a dentist or an oncologist). Examinations icate was perceived as indicating a low qualifica-
for all invasive oral cancers included both COE tion and may have resulted in the low follow-up
and histological investigation. rate of screen-positive individuals.]
318
Table 4.11 Description of the cluster-randomized trial of the efficacy of oral cancer screening (Sankaranarayanan et al., 2013)
Location No. of Participation Accrual period for Age at Description Follow-up Follow-up Screening No. of
Randomization participants rate screening entry of the for screen- rate for interval rounds of
(years) intervention positive screen- (years) screening
Invited Control individuals positive Follow-up
group group individuals (years)
Kerala, India 191 872 Intervention 1996– Routine ≥ 35 Clinical oral Clinical 59% 3 4
Cluster- recruited; arm (at least 2008 care in Mean, examination examination 15
randomized (at 96 517 in the one screen): 1996– 49 (SD, by non- by a specialist
the municipal intervention 92%; at first 2005, 0.7) medical health (dentist or
level) group; 77% round: 79% screened in worker oncologist)
men Control arm: 2006–2008
46.1%
SD, standard deviation.
Table 4.12 Results of the cluster-randomized trial of the efficacy of oral cancer screening (Sankaranarayanan et al., 2013)
Outcome Population group No. of participants Outcome per 100 000 RR (95% CI)
(screened/control group) person-years
(screened/control group)
Incidence of oral cancer
General population 895 310/898 280 31.2/27.2 1.14 (0.91–1.44)
Users of tobacco and/or alcohol 429 620/377 350 59.2/61.6 0.97 (0.79–1.19)
Incidence of stages III and IV oral cancer
Mortality from oral cancer

CI, confidence interval; RR, relative risk.
319
(b) Observational studies settings. Enrolment of participants in hospitals

Two cohort studies, both based on a nation- is likely to increase selection bias. Selection bias
wide population-based biennial oral cancer also increases with the retrospective choice of
screening programme in Taiwan (China), and the controls related to the outcome of interest.
one case–control study, evaluating the national Second, the participation rate of < 60% means that
oral cancer screening programme in Cuba, there is a high risk of non-response bias. Third,
compared oral cancer screening attenders with because the nationwide oral cancer screening
non-attenders in terms of oral cancer incidence programme in Taiwan (China) included an initial
and/or mortality. survey on the risk factors, this could potentially
A cohort of 4 234 393 adults (≥ 18 years) lead to contamination of the control group,
who smoked cigarettes and/or chewed betel quid which would lead to an underestimation of the
underwent biennial oral screening by dentists benefits of screening.] A retrospective analysis
or physicians in 2004–2009 in Taiwan (China). of the at-risk cohorts invited to the oral cancer
The individuals were followed up until 2012, screening programme in Taiwan (China) was
with a median follow-up of 4.5 years (Table 4.13; subsequently conducted by Ho et al. (2019). The
Chuang et al., 2017). Screen-positive individuals study used the databases of the National Cancer
were referred to specialists in hospitals for histo- Registry, the Nationwide Oral Mucosal Screening
pathological examinations. The study was linked Program, and the National Death Registry. The
to the National Cancer Registry to enable precise duration of follow-up was calculated from the
recording of oral cancer cases in attenders and date of cancer diagnosis to the date of death or
non-attenders in the screening programme. to the end of the follow-up period (until 2017). A
The expected incidence and mortality rates of total of 18 625 patients with oral cancer were iden-
non-attenders were estimated based on previous tified from the National Cancer Registry during
findings that about 90% of oral cancer cases 2012–2015. The screened status was defined as
were attributed to cigarette smoking and/or betel having no records, records without a previous
quid chewing. The participation rate at the first positive result, or records with a previous posi-
screening in the invited population was 55.1%. tive result. Of this cohort, 8165 patients (43.8%)
There was a 21% (95% CI, 18–24%) reduction attended at least one screening round and had
in the incidence of advanced oral cancer and a a previous positive result, 3560 patients (19.1%)
26% (95% CI, 23–28%) reduction in oral cancer had a negative result on screening or no previous
mortality in the screened group compared with positive result, and 6900 patients (37.0%) had
the non-screened group (Table 4.13). [The lower no records of attending the screening. Among
incidence rate of oral cancer in the screened the patients with cancer, most of the screened
group compared with the non-screened group patients were diagnosed with cancer at earlier
may be due to an imbalance in risk of oral cancer stages compared with the non-screened patients
between attenders and non-attenders, consid- (Table 4.14). The 3-year survival rates were 71.4%
ering the low participation rate.] for screened patients with positive results, 68.7%
[To assess the transferability of the conclu- for screened patients with negative results, and
sions on effectiveness of oral cancer screening 63.5% in the non-screened group; this showed a
to other settings, the following biases should be survival benefit of screening.
considered. First, enrolment of the participants [The study of Ho et al. (2019) has several
was conducted in communities and in hospitals, limitations. Although the oral cancer screening
with an unclear distribution between these two programme included individuals aged ≥ 18 years,
this study limited the cohort to ages ≥ 30 years.
320
Table 4.13 Prospective cohort study of the effectiveness of oral cancer screening
Reference Description of the Description of the Accrual and Participation Detection rate Cancer incidence/ Comments
Location cohort controls follow-up rate and follow- mortality RR
periods up rate for (95% CI)
screen-positive
individuals
Chuang et 4 234 393 high-risk Non-attenders; 10.5 million Participation rate: First screening: Incidence Reports also by age
al. (2017) invitees (cigarette incidence and person-years 55.1% Screen-positive, Cancer: groups. The highest
Taiwan smokers and/or betel mortality rates of follow-up Referral follow- 18 116 (0.8%) 0.83 (0.81–0.86) detection rate for men
(China) quid chewers), followed were adjusted to up rate: first Precancer, Advanced cancer: was in the age group
up until the end of attribute 90% of screening, 91.1%; 11 051 (0.5%) 0.79 (0.76–0.82) 50–69 years and for
2012; median follow- cases to a high-risk subsequent Cancer, Mortality women was in the age
up, 4.5 years (National population; 86% screening, 92.6% 4110 (0.2%) 0.74 (0.72–0.77)ª group ≥ 70 years
Cancer Registry) men Subsequent
screening:
Screen-positive,
5825 (1.0%)
Precancer,
3782 (0.6%)
Cancer,
791 (0.1%)
CI, confidence interval; RR, relative risk.
a Adjusted for self-selection bias.

321
322

Table 4.14 Retrospective cohort and case–control studies of the effectiveness of oral cancer screening
Reference Description of the cohort/ Description of the Established programme: Oral cancer Proportion of Cancer incidence/
Location cases controls year of start, screening or precancer patients with events mortality RR
age, screening interval end-point (95% CI)
Ho et al. (2019) Retrospective cohort of Patients without Population-based biennial Early-stage Stage 0–I diagnosis: Mortality in
Taiwan (China) patients with oral cancer previous screening programme since 2004 diagnosis Screened positive, 3 years:
(2012–2015); high-risk records; 82.1% men targeting population aged Survival 34.3% HRª: 0.78
invitees (cigarette smokers ≥ 30 years Mortality Screened negative, Stage 0–I diagnosis:
and/or betel quid chewers); 34.3% HRª: 1.23
95.4% men Not screened, 27.8%
3-Year survival:
Screened positive,
71.4%
Screened negative,
68.7%
Not screened, 63.5%
Sankaranarayanan Cases: 200 patients with oral Controls: 3 per Population-based annual Incidence Screened cases: Incidence of
et al. (2002) cancer (77% men); median case, matched programme via oral of advanced 56.0% advanced cancer
Cuba age, 65 years on age, sex, and inspection since 1984 in cancer Screened controls: OR:
residence; 77% population aged ≥ 15 years; 49.7% Adjusted, 0.78
men screening is mainly (0.53–1.15)
opportunistic Not adjusted, 0.67
(0.46–0.95)
CI, confidence interval; HR, hazard ratio; OR, odds ratio; RR, relative risk.
a Calculated from the probability of having an event in 3 years in the screened positive and not screened groups.
The retrospective design carries a risk of misclas- without an impact on mortality rates (Fernández
sification and information bias. The screened Garrote et al., 1995).
cohorts included only a population at high risk, [The Working Group noted that the low
whereas the proportion of cigarette smokers coverage of the programme and the poor compli-
and/or betel quid chewers among the screening ance with referral contribute to selection bias.
non-attenders was unclear. The higher propor- Given the study design, there is also a possible
tion of women in the non-screened group (17.9%) risk of reporting bias. Another risk is recall bias
than in the screened group (4.6%) suggests a risk and differential reporting of exposure in cases
of bias. The comparison is done between five and controls due to the timing of the event.
groups, none of which included the “all screened” Furthermore, the definition of the intervention,
population (i.e. with either a positive or a nega- which was “any visit to a community dentist”,
tive screening result). The lower hazard ratio for may lead to a possible overestimation of exposure
oral cancer mortality in all the groups in the in the controls. Finally, the number of cases may
reported Cox regression analysis (e.g. in those be too small to detect a difference in outcomes
with a confirmed cancer and in those who had with an opportunistic screening programme.]
a positive screening result but did not complete Several studies have assessed the impact
confirmation of diagnosis) compared with those of oral cancer screening on oral cancer
who were not screened suggests a possible risk incidence (Fernández Garrote et al., 1995;
of bias.] Sankaranarayanan et al., 2013; Chuang et al.,
A case–control study was conducted to 2017; Morikawa et al., 2021). Chuang et al. (2017)
evaluate the effectiveness of the national reported a statistically significant decrease of
oral cancer screening programme in Cuba 17% in the oral cancer incidence rate. All other
(Sankaranarayanan et al., 2002). The cases were studies reported no impact.
200 individuals with incident oral cancer of
stages III and IV registered in 1994–1997. Three 4.4.2 Harms of screening
controls of apparently healthy individuals were
matched to each case on sex, age (± 5 years), and Although screening must by definition be
residence (within a 200 m radius of the house- beneficial, it may be associated with some harms.
hold of the case). A total of 462 (77%) males and The harms related to screening for cancer at
138 (23%) females provided data on socioeco- other sites have been reviewed extensively (e.g.
nomic factors and individual risk factors for oral Welch and Black, 2010; Woolf and Harris, 2012;
cancer. The proportion of screened individuals Marmot et al., 2013).
was higher in cases than in controls (56.0% vs The potential harms of screening include
49.7%). The odds ratio for advanced oral cancer factors associated with false-positive tests,
in cases screened 3 months before diagnosis was false-negative tests, overdiagnosis, and over-
0.67 (95% CI, 0.46–0.95). After adjustment for treatment. A false-positive test result is a positive
the frequency of cigarette smoking to address test result in an individual who does not have
selection bias, the odds ratio was 0.78 (95% CI, cancer in the further assessment. A false-positive
0.53–1.15) (Table 4.14). A time series analysis test result can lead to unnecessary psychological
compared incidence of early oral cancer and distress and anxiety, unnecessary additional
mortality from oral cancer in Cuba in 1983–1990 investigations to rule out disease, side-effects,
and concluded that the proportion of stage I unnecessary treatment, and additional costs.
cases increased from 24% in 1983 to 49% in 1990, A false-negative test result is a negative test
result in an individual who has the disease. A
323
false-negative test leads to false reassurance of not Restricting screening to only individuals at high
having disease and consequent increased risk of risk may improve the efficiency and effective-
advanced disease, with poor treatment outcome ness of screening while minimizing the harms.
and poor cosmesis and functional outcomes. A risk-based screening strategy has been tested
Overdiagnosis is the diagnosis of a cancer as in several model-based studies and cohorts
a result of screening that would not have been (Amarasinghe et al., 2010; Shieh et al., 2017;
diagnosed in the patient’s lifetime if screening Cheung et al., 2019; Willoughby et al., 2019; de
had not taken place. Although the concept of Koning et al., 2020; Harkness et al., 2020; Ten
overdiagnosis is often discussed in the context Haaf et al., 2021). Recently, several studies have
of screening asymptomatic people, there is no reported that incorporating genomic informa-
agreement on how to estimate overdiagnosis. tion along with other individual risk factors
Estimates of overdiagnosis are highly hetero- can help in screening for breast cancer, prostate
geneous and vary depending on the analytical cancer, and lung cancer (Torkamani et al., 2018;
approach. Overall, the harms are worse when the Callender et al., 2019; Roberts et al., 2021).
quality of the test is poor. The Trivandrum Oral Cancer Screening
No studies have reported on harms from Study showed that the benefit of screening
the oral cancer screening test itself (COE), from is limited to the individuals at high risk, i.e.
false-positive or false-negative screening test those who use tobacco and/or consume alcohol
results, or from overdiagnosis. However, several (Sankaranarayanan et al., 2005). A reanalysis
studies have reported the detection rates and of the Trivandrum study using a risk-based
screening performance in various oral cancer screening strategy showed that the absolute
screening programmes (see Section 4.1.1). benefits of screening increased significantly with
Diagnostic harms are primarily related to increasing model-predicted risk of oral cancer
the side-effects and complications of biopsy for (Cheung et al., 2021). The difference in the oral
suspected oral cancer or its potential precur- cancer mortality rate between the intervention
sors. Although oral cancer screening can detect arm and the control arm increased from 0.5 per
OPMDs, it is unclear which OPMDs regress spon- 100 000 in the lowest quartile of oral cancer
taneously and which lesions persist or progress risk to 13.4 per 100 000 for individuals in the
further to malignancy (see Section 1.3.1) (Moyer highest quartile. Similarly, among ever-users of
et al., 2014). The treatment of some screen-de- tobacco and/or alcohol, the difference in the oral
tected OPMDs is limited by a field cancerization cancer mortality rate between the intervention
effect due to the entire oral mucosa being exposed arm and the control arm increased from 1.0 per
to carcinogens. Moreover, surgical and ablative 100 000 in the lowest quartile of oral cancer risk
treatments of OPMDs may lead to unwanted to 22.5 per 100 000 for individuals in the highest
side-effects, such as severe pain, infection, and quartile. In a population similar to that in the
bleeding due to complications of treatment. Kerala trial, screening of 100% of eligible indi-
viduals (ages ≥ 35 years) would lead to a 27.1%
reduction in oral cancer mortality at a number
4.5 Risk-based model for screening needed to screen of 2043. Restricting screening
Cancer screening has historically been to ever-users of tobacco and/or alcohol with no
based on age and applied for all eligible indi- additional risk stratification (43.4% of the popu-
viduals without any assessment of their expo- lation) would substantially increase efficiency
sure to known risk factors. However, the risk (23.3% reduction in oral cancer mortality at a
of developing cancer varies among individuals. number needed to screen of 1029). Screening the
324
50% of ever-users of tobacco and/or alcohol at Brazil. Braz Dent J. 23(3):241–5. doi:10.1590/S0103-
highest risk based on the risk-prediction model 64402012000300010 PMID:22814693
Amarasinghe AAHK, Usgodaarachchi US, Johnson
(21.7% of the population) would further enhance NW (2016). Evaluation of the utilization of primary
efficiency with little loss in programme sensi- healthcare staff for control of oral cancer: a Sri
tivity (19.7% reduction in oral cancer mortality Lankan experience. Transl Res Oral Oncol. 1(2):1–6.
doi:10.1177/2057178X16682544
at a number needed to screen of 610) (Cheung Amarasinghe HK, Johnson NW, Lalloo R,
et al., 2021). Kumaraarachchi M, Warnakulasuriya S (2010).
[This study provided the first proof of prin- Derivation and validation of a risk-factor model
ciple that a risk-based tailored approach may for detection of oral potentially malignant disor-
ders in populations with high prevalence. Br J
enhance the efficiency of screening, reduce Cancer. 103(3):303–9. doi:10.1038/sj.bjc.6605778
harms, and be more cost-effective. However, the PMID:20628386
magnitude of risk associated with each risk factor Amer A, Galvin S, Healy CM, Moran GP (2017). The
may vary in different populations and countries microbiome of potentially malignant oral leukoplakia
exhibits enrichment for Fusobacterium, Leptotrichia,
(see Section 2.1) (Winn et al., 2015). This aspect Campylobacter, and Rothia species. Front Microbiol.
should be considered before implementing a 8:2391. doi:10.3389/fmicb.2017.02391 PMID:29250055
risk-based approach for a particular country. Arantes LMRB, De Carvalho AC, Melendez ME,
Lopes Carvalho A (2018). Serum, plasma and saliva
The risk-based approach may be appropriate for biomarkers for head and neck cancer. Expert Rev Mol
resource-limited countries with a high incidence Diagn. 18(1):85–112. doi:10.1080/14737159.2017.1404
of oral cancer (Cheung et al., 2021; D’Cruz and 906 PMID:29134827
Vaish, 2021). However, the implementation of a Basu P, Mahajan M, Patira N, Prasad S, Mogri S, Muwonge
R, et al. (2019). A pilot study to evaluate home-based
risk-based screening programme faces several screening for the common non-communicable diseases
challenges in selecting the high-risk group by a dedicated cadre of community health workers in
without negatively influencing the trade-off a rural setting in India. BMC Public Health. 19(1):14.
between individual benefits and harms.] doi:10.1186/s12889-018-6350-4 PMID:30606132
Bhatia N, Matias MA, Farah CS (2014). Assessment of a
decision making protocol to improve the efficacy of
VELscope™ in general dental practice: a prospective
References evaluation. Oral Oncol. 50(10):1012–9. doi:10.1016/j.
oraloncology.2014.07.002 PMID:25096826
Bhatt S, Isaac R, Finkel M, Evans J, Grant L, Paul B, et
Acha A, Ruesga MT, Rodríguez MJ, Martínez de al. (2018). Mobile technology and cancer screening:
Pancorbo MA, Aguirre JM (2005). Applications of the lessons from rural India. J Glob Health. 8(2):020421.
oral scraped (exfoliative) cytology in oral cancer and doi:10.7189/jogh.08.020421 PMID:30603075
precancer. Med Oral Patol Oral Cir Bucal. 10(2):95–102. Bigio IJ, Bown SG (2004). Spectroscopic sensing of cancer
PMID:15735540 and cancer therapy: current status of translational
Al Rawi N, Elmabrouk N, Abu Kou R, Mkadmi S, Rizvi Z, research. Cancer Biol Ther. 3(3):259–67. doi:10.4161/
Hamdoon Z (2021). The role of differentially expressed cbt.3.3.694 PMID:15107613
salivary microRNA in oral squamous cell carcinoma. Birur NP, Gurushanth K, Patrick S, Sunny SP, Raghavan
A systematic review. Arch Oral Biol. 125:105108. SA, Gurudath S, et al. (2019). Role of community
doi:10.1016/j.archoralbio.2021.105108 PMID:33756383 health worker in a mobile health program for early
AlAli AM, Walsh T, Maranzano M (2020). CYFRA 21-1 detection of oral cancer. Indian J Cancer. 56(2):107–13.
and MMP-9 as salivary biomarkers for the detection doi:10.4103/ijc.IJC_232_18 PMID:31062727
of oral squamous cell carcinoma: a systematic review Böcking A, Sproll C, Stöcklein N, Naujoks C, Depprich
of diagnostic test accuracy. Int J Oral Maxillofac R, Kübler NR, et al. (2011). Role of brush biopsy and
Implants. 49(8):973–83. doi:10.1016/j.ijom.2020.01.020 DNA cytometry for prevention, diagnosis, therapy,
PMID:32035907 and followup care of oral cancer. J Oncol. 2011:875959.
Almeida FC, Cazal C, Pucca Júnior GA, Silva DP, Frias doi:10.1155/2011/875959 PMID:21209723
AC, Araújo ME (2012). Reorganization of secondary Brocklehurst P, Kujan O, O’Malley LA, Ogden G,
and tertiary health care levels: impact on the outcomes Shepherd S, Glenny AM (2013). Screening programmes
of oral cancer screening in the São Paulo State, for the early detection and prevention of oral cancer.
325
Cochrane Database Syst Rev. 2013(11):CD004150. Br J Oral Maxillofac Surg. 45(7):590–1. doi:10.1016/j.
d o i :10 .10 0 2 /14 6 518 5 8 .C D 0 0 415 0 . pu b 4 bjoms.2006.08.017 PMID:17081665
PMID:24254989 Cheng YS, Rees T, Wright J (2014). A review of research
Callender T, Emberton M, Morris S, Eeles R, Kote-Jarai on salivary biomarkers for oral cancer detection.
Z, Pharoah PDP, et al. (2019). Polygenic risk-tai- Clin Transl Med. 3(1):3. doi:10.1186/2001-1326-3-3
lored screening for prostate cancer: a benefit-harm PMID:24564868
and cost-effectiveness modelling study. PLoS Med. Cheung LC, Berg CD, Castle PE, Katki HA, Chaturvedi
16(12):e1002998. doi:10.1371/journal.pmed.1002998 AK (2019). Life-gained-based versus risk-based
PMID:31860675 selection of smokers for lung cancer screening. Ann
Carter LM, Harris AT, Kavi VP, Johnson S, Kanatas Intern Med. 171(9):623–32. doi:10.7326/M19-1263
A (2009). Oral cancer awareness amongst hospital PMID:31634914
nursing staff: a pilot study. BMC Oral Health. 9(1):4. Cheung LC, Ramadas K, Muwonge R, Katki HA,
doi:10.1186/1472-6831-9-4 PMID:19175923 Thomas G, Graubard BI, et al. (2021). Risk-based
Carter LM, Ogden GR (2007). Oral cancer awareness selection of individuals for oral cancer screening. J
of general medical and general dental practitioners. Clin Oncol. 39(6):663–74. doi:10.1200/JCO.20.02855
Br Dent J. 203(5):E10, discussion 248–9. doi:10.1038/ PMID:33449824
bdj.2007.630 PMID:17632458 Chuang SL, Su WWY, Chen SLS, Yen AMF, Wang CP,
Carter LM, Parsonage-Grant S, Marshall A, Achal KS, Fann JCY, et al. (2017). Population-based screening
Kanatas A (2011). Oral cancer teaching of medical program for reducing oral cancer mortality in
students in the UK: time for a new approach? J Cancer 2,334,299 Taiwanese cigarette smokers and/or betel
Educ. 26(2):308–14. doi:10.1007/s13187-010-0156-0 quid chewers. Cancer. 123(9):1597–609. doi:10.1002/
PMID:20872262 cncr.30517 PMID:28055109
Carvalho AL, Jeronimo C, Kim MM, Henrique R, Clarke AK, Kobagi N, Yoon MN (2018). Oral cancer
Zhang Z, Hoque MO, et al. (2008). Evaluation of screening practices of Canadian dental hygienists.
promoter hypermethylation detection in body fluids Int J Dent Hyg. 16(2):e38–45. doi:10.1111/idh.12295
as a screening/diagnosis tool for head and neck squa- PMID:28643363
mous cell carcinoma. Clin Cancer Res. 14(1):97–107. Coppola N, Mignogna MD, Rivieccio I, Blasi A, Bizzoca
doi:10.1158/1078-0432.CCR-07-0722 PMID:18172258 ME, Sorrentino R, et al. (2021). Current knowledge,
Chainani-Wu N, Madden E, Cox D, Sroussi H, Epstein attitudes, and practice among health care providers in
J, Silverman S Jr (2015). Toluidine blue aids in detec- OSCC awareness: systematic review and meta-analysis.
tion of dysplasia and carcinoma in suspicious oral Int J Environ Res Public Health. 18(9):4506. doi:10.3390/
lesions. Oral Dis. 21(7):879–85. doi:10.1111/odi.12358 ijerph18094506 PMID:33922752
PMID:26173924 D’Cruz AK, Vaish R (2021). Risk-based oral cancer
Chang IH, Jiang RS, Wong YK, Wu SH, Chen FJ, Liu SA screening – lessons to be learnt. Nat Rev Clin
(2011). Visual screening of oral cavity cancer in a male Oncol. 18(8):471–2. doi:10.1038/s41571-021-00511-2
population: experience from a medical center. J Chin PMID:33883701
Med Assoc. 74(12):561–6. doi:10.1016/j.jcma.2011.09.014 de Koning HJ, van der Aalst CM, de Jong PA, Scholten ET,
PMID:22196472 Nackaerts K, Heuvelmans MA, et al. (2020). Reduced
Chang YT, Chu LJ, Liu YC, Chen CJ, Wu SF, Chen CH, lung-cancer mortality with volume CT screening
et al. (2020). Verification of saliva matrix metallo- in a randomized trial. N Engl J Med. 382(6):503–13.
proteinase-1 as a strong diagnostic marker of oral doi:10.1056/NEJMoa1911793 PMID:31995683
cavity cancer. Cancers (Basel). 12(8):2273. doi:10.3390/ Delavarian Z, Mohtasham N, Mosannen-Mozafari P,
cancers12082273 PMID:32823758 Pakfetrat A, Shakeri MT, Ghafoorian-Maddah R
Chaudhari A, Hegde-Shetiya S, Shirahatti R, Agrawal D (2010). Evaluation of the diagnostic value of a modified
(2013). Comparison of different screening methods liquid-based cytology using OralCDx brush in early
in estimating the prevalence of precancer and cancer detection of oral potentially malignant lesions and oral
amongst male inmates of a jail in Maharashtra, India. cancer. Med Oral Patol Oral Cir Bucal. 15(5):e671–6.
Asian Pac J Cancer Prev. 14(2):859–64. doi:10.7314/ doi:10.4317/medoral.15.e671 PMID:20383114
APJCP.2013.14.2.859 PMID:23621252 Deo RC (2015). Machine learning in medi-
Chen M, Zhao H (2019). Next-generation sequencing in cine. Circulation. 132(20):1920–30. doi:10.1161/
liquid biopsy: cancer screening and early detection. CIRCULATIONAHA.115.001593 PMID:26572668
Hum Genomics. 13(1):34. doi:10.1186/s40246-019- Deshpande S, Radke U, Karemore T, Mohril R, Rawlani
0220-8 PMID:31370908 S, Ingole P (2019). A novel mobile app for oral cancer
Chen YW, Lin JS, Fong JH, Wang IK, Chou SJ, Wu CH, et awareness amongst general population: development,
al. (2007). Use of methylene blue as a diagnostic aid in implementation, and evaluation. J Contemp Dent Pract.
early detection of oral cancer and precancerous lesions.
326
20(2):190–6. doi:10.5005/jp-journals-10024-2496 literature and report of a three-year study. Oral Surg

PMID:31058634 Oral Med Oral Pathol. 33(1):61–74. doi:10.1016/0030-
Deuerling L, Gaida K, Neumann H, Remmerbach TW 4220(72)90209-5 PMID:4550162
(2019). Evaluation of the accuracy of liquid-based oral Fujimoto JG (2003). Optical coherence tomography for
brush cytology in screening for oral squamous cell ultrahigh resolution in vivo imaging. Nat Biotechnol.
carcinoma. Cancers (Basel). 11(11):1813. doi:10.3390/ 21(11):1361–7. doi:10.1038/nbt892 PMID:14595364
cancers11111813 PMID:31752196 Gaba FI, Sheth CC, Veses V (2021). Salivary biomarkers
Downer MC, Evans AW, Hughes Hallet CM, Jullien JA, and their efficacies as diagnostic tools for oral squa-
Speight PM, Zakrzewska JM (1995). Evaluation of mous cell carcinoma: systematic review and meta-
screening for oral cancer and precancer in a company analysis. J Oral Pathol Med. 50(3):299–307. doi:10.1111/
headquarters. Community Dent Oral Epidemiol. jop.12791 PMID:30339289
23(2):84–8. doi:10.1111/j.1600-0528.1995.tb00206.x García-Pola M, Pons-Fuster E, Suárez-Fernández C,
PMID:7781305 Seoane-Romero J, Romero-Méndez A, López-Jornet
Du GF, Li CZ, Chen HZ, Chen XM, Xiao Q, Cao ZG, et P (2021). Role of artificial intelligence in the early
al. (2007). Rose bengal staining in detection of oral diagnosis of oral cancer. A scoping review. Cancers
precancerous and malignant lesions with colorimetric (Basel). 13(18):4600. doi:10.3390/cancers13184600
evaluation: a pilot study. Int J Cancer. 120(9):1958–63. PMID:34572831
doi:10.1002/ijc.22467 PMID:17245698 Glass RT, Abla M, Wheatley J (1975). Teaching self-ex-
Elango KJ, Anandkrishnan N, Suresh A, Iyer SK, amination of the head and neck: another aspect of
Ramaiyer SK, Kuriakose MA (2011). Mouth self-ex- preventive dentistry. J Am Dent Assoc. 90(6):1265–8.
amination to improve oral cancer awareness and doi:10.14219/jada.archive.1975.0254 PMID:1056392
early detection in a high-risk population. Oral Oncol. Gomes MS, Bonan PR, Ferreira VY, de Lucena Pereira
47(7):620–4. doi:10.1016/j.oraloncology.2011.05.001 L, Correia RJ, da Silva Teixeira HB, et al. (2017).
PMID:21646040 Development of a mobile application for oral cancer
Elashoff D, Zhou H, Reiss J, Wang J, Xiao H, Henson B, screening. Technol Health Care. 25(2):187–95.
et al. (2012). Prevalidation of salivary biomarkers for doi:10.3233/THC-161259 PMID:27689559
oral cancer detection. Cancer Epidemiol Biomarkers González RS (2014). Cancer screening: global debates
Prev. 21(4):664–72. doi:10.1158/1055-9965.EPI-11-1093 and Cuban experience. MEDICC Rev. 16(3–4):73–7.
PMID:22301830 doi:10.37757/MR2014.V16.N3-4.14 PMID:25208124
Epstein JB, Scully C, Spinelli J (1992). Toluidine blue Gualtero DF, Suarez Castillo A (2016). Biomarkers in
and Lugol’s iodine application in the assessment saliva for the detection of oral squamous cell carcinoma
of oral malignant disease and lesions at risk of and their potential use for early diagnosis: a systematic
malignancy. J Oral Pathol Med. 21(4):160–3. review. Acta Odontol Scand. 74(3):170–7. doi:10.3109/0
doi:10.1111/j.1600-0714.1992.tb00094.x PMID:1376363 0016357.2015.1110249 PMID:26577643
Fakurnejad S, van Keulen S, Nishio N, Engelen M, van Guerra EN, Acevedo AC, Leite AF, Gozal D, Chardin H, De
den Berg NS, Lu G, et al. (2019). Fluorescence molec- Luca Canto G (2015). Diagnostic capability of salivary
ular imaging for identification of high-grade dysplasia biomarkers in the assessment of head and neck cancer:
in patients with head and neck cancer. Oral Oncol. a systematic review and meta-analysis. Oral Oncol.
97:50–5. doi:10.1016/j.oraloncology.2019.08.008 51(9):805–18. doi:10.1016/j.oraloncology.2015.06.010
PMID:31421471 PMID:26170140
Ferlay J, Ervik M, Lam F, Colombet M, Mery L, Piñeros Harkness EF, Astley SM, Evans DG (2020). Risk-based
M, et al. (2020). Global Cancer Observatory: Cancer breast cancer screening strategies in women. Best Pract
Today. Lyon, France: International Agency for Research Res Clin Obstet Gynaecol. 65(65):3–17. doi:10.1016/j.
on Cancer. Available from: https://gco.iarc.fr/today. bpobgyn.2019.11.005 PMID:31848103
Fernández Garrote L, Sankaranarayanan R, Lence Anta Haron N, Rajendran S, Kallarakkal TG, Zain RB,
JJ, Rodriguez Salvá A, Maxwell Parkin D (1995). An Ramanathan A, Abraham MT, et al. (2023). High
evaluation of the oral cancer control program in Cuba. referral accuracy for oral cancers and oral potentially
Epidemiology. 6(4):428–31. doi:10.1097/00001648- malignant disorders using telemedicine. Oral Dis.
199507000-00019 PMID:7548355 29(2):380–9. doi:10.1111/odi.13892 PMID:33914993
Ferrari E, Pezzi ME, Cassi D, Pertinhez TA, Spisni A, Hawkins RJ, Wang EE, Leake JL (1999). Preventive health
Meleti M (2021). Salivary cytokines as biomarkers for care, 1999 update: prevention of oral cancer mortality.
oral squamous cell carcinoma: a systematic review. The Canadian Task Force on Preventive Health Care.
Int J Mol Sci. 22(13):6795. doi:10.3390/ijms22136795 J Can Dent Assoc. 65(11):617–27. PMID:10859725
PMID:34202728 Hayama FH, Motta AC, Silva AdeP, Migliari DA
Folsom TC, White CP, Bromer L, Canby HF, Garrington (2005). Liquid-based preparations versus conven-
GE (1972). Oral exfoliative study. Review of the tional cytology: specimen adequacy and diagnostic
327
agreement in oral lesions. Med Oral Patol Oral Cir Jedele JM, Ismail AI (2010). Evaluation of a multifaceted
Bucal. 10(2):115–22. PMID:15735543 social marketing campaign to increase awareness of
Heidari AE, Sunny SP, James BL, Lam TM, Tran AV, and screening for oral cancer in African Americans.
Yu J, et al. (2019). Optical coherence tomography as Community Dent Oral Epidemiol. 38(4):371–82.
an oral cancer screening adjunct in a low resource doi:10.1111/j.1600-0528.2010.00545.x PMID:20646014
settings. IEEE J Sel Top Quantum Electron. 25(1):1–8. Jornet PL, Garcia FJ, Berdugo ML, Perez FP, Lopez AP
doi:10.1109/JSTQE.2018.2869643 (2015). Mouth self-examination in a population at risk
Hema Shree K, Ramani P, Sherlin H, Sukumaran G, Jeyaraj of oral cancer. Aust Dent J. 60(1):59–64. doi:10.1111/
G, Don KR, et al. (2019). Saliva as a diagnostic tool in adj.12274 PMID:25721279
oral squamous cell carcinoma – a systematic review Jou YJ, Lin CD, Lai CH, Tang CH, Huang SH, Tsai MH,
with meta analysis. Pathol Oncol Res. 25(2):447–53. et al. (2011). Salivary zinc finger protein 510 peptide
doi:10.1007/s12253-019-00588-2 PMID:30712193 as a novel biomarker for detection of oral squa-
Ho PS, Wang WC, Huang YT, Yang YH (2019). Finding mous cell carcinoma in early stages. Clin Chim Acta.
an oral potentially malignant disorder in screening 412(15-16):1357–65. doi:10.1016/j.cca.2011.04.004
program is related to early diagnosis of oral cavity cancer PMID:21497587
– experience from real world evidence. Oral Oncol. Jullien JA, Downer MC, Zakrzewska JM, Speight PM
89:107–14. doi:10.1016/j.oraloncology.2018.12.007 (1995). Evaluation of a screening test for the early
PMID:30732947 detection of oral cancer and precancer. Community
Hofmann L, Ludwig S, Vahl JM, Brunner C, Hoffmann Dent Health. 12(1):3–7. PMID:7697560
TK, Theodoraki MN (2020). The emerging role of Kaczor-Urbanowicz KE, Martín Carreras-Presas C,
exosomes in diagnosis, prognosis, and therapy in head Kaczor T, Tu M, Wei F, Garcia-Godoy F, et al. (2017).
and neck cancer. Int J Mol Sci. 21(11):4072. doi:10.3390/ Emerging technologies for salivaomics in cancer
ijms21114072 PMID:32517240 detection. J Cell Mol Med. 21(4):640–7. doi:10.1111/
Huang D, Swanson EA, Lin CP, Schuman JS, Stinson jcmm.13007 PMID:27862926
WG, Chang W, et al. (1991). Optical coherence Kar A, Wreesmann VB, Shwetha V, Thakur S, Rao VUS,
tomography. Science. 254(5035):1178–81. doi:10.1126/ Arakeri G, et al. (2020). Improvement of oral cancer
science.1957169 PMID:1957169 screening quality and reach: the promise of artifi-
Hussain SA, Sullivan R (2013). Cancer control in cial intelligence. J Oral Pathol Med. 49(8):727–30.
Bangladesh. Jpn J Clin Oncol. 43(12):1159–69. doi:10.1111/jop.13013 PMID:32162398
doi:10.1093/jjco/hyt140 PMID:24163419 Kerr AR (2020). Diagnostic adjuncts for oral cavity squa-
Hutchinson ML, Isenstein LM, Goodman A, Hurley AA, mous cell carcinoma and oral potentially malignant
Douglass KL, Mui KK, et al. (1994). Homogeneous disorders. In: Warnakulasuriya KA, Greenspan JS,
sampling accounts for the increased diagnostic accu- editors. Textbook of oral cancer: prevention, diag-
racy using the ThinPrep processor. Am J Clin Pathol. nosis and management. Cham, Switzerland: Springer
101(2):215–9. doi:10.1093/ajcp/101.2.215 PMID:8116578 Nature; pp. 99–117. doi:10.1007/978-3-030-32316-5_9
Ibrahim O, Toner M, Flint S, Byrne HJ, Lyng FM (2021). Kerr AR, Robinson ME, Meyerowitz C, Morse DE,
The potential of Raman spectroscopy in the diagnosis Aguilar ML, Tomar SL, et al.; National Dental PBRN
of dysplastic and malignant oral lesions. Cancers Collaborative Group (2020). Cues used by dentists in
(Basel). 13(4):619. doi:10.3390/cancers13040619 the early detection of oral cancer and oral potentially
PMID:33557195 malignant lesions: findings from the National Dental
Ikeda N, Downer MC, Ishii T, Fukano H, Nagao T, Inoue K Practice-Based Research Network. Oral Surg Oral Med
(1995). Annual screening for oral cancer and precancer Oral Pathol Oral Radiol. 130(3):264–72. doi:10.1016/j.
by invitation to 60-year-old residents of a city in Japan. oooo.2020.05.007 PMID:32561250
Community Dent Health. 12(3):133–7. PMID:7584579 Kerr AR, Sirois DA, Epstein JB (2006). Clinical evalu-
Ilhan B, Guneri P, Wilder-Smith P (2021). The contribution ation of chemiluminescent lighting: an adjunct for
of artificial intelligence to reducing the diagnostic delay oral mucosal examinations. J Clin Dent. 17(3):59–63.
in oral cancer. Oral Oncol. 116:105254. doi:10.1016/j. PMID:17022366
oraloncology.2021.105254 PMID:33711582 Khandelwal S, Solomon MC (2010). Cytomorphological
James BL, Sunny SP, Heidari AE, Ramanjinappa RD, analysis of keratinocytes in oral smears from tobacco
Lam T, Tran AV, et al. (2021). Validation of a point- users and oral squamous cell carcinoma lesions – a
of-care optical coherence tomography device with histochemical approach. Int J Oral Sci. 2(1):45–52.
machine learning algorithm for detection of oral doi:10.4248/IJOS10011 PMID:20690418
potentially malignant and malignant lesions. Cancers Koch FP, Kunkel M, Biesterfeld S, Wagner W (2011).
(Basel). 13(14):3583. doi:10.3390/cancers13143583 Diagnostic efficiency of differentiating small cancerous
PMID:34298796 and precancerous lesions using mucosal brush smears
of the oral cavity – a prospective and blinded study.
328
Clin Oral Investig. 15(5):763–9. doi:10.1007/s00784- Liu D, Xin Z, Guo S, Li S, Cheng J, Jiang H (2021). Blood
010-0434-6 PMID:20593209 and salivary microRNAs for diagnosis of oral squamous
Kujan O, Glenny AM, Duxbury J, Thakker N, Sloan P cell carcinoma: a systematic review and meta-analysis.
(2005). Evaluation of screening strategies for improving J Oral Maxillofac Surg. 79(5):1082.e1–13. doi:10.1016/j.
oral cancer mortality: a Cochrane systematic review. joms.2020.12.043 PMID:33516682
J Dent Educ. 69(2):255–65. doi:10.1002/j.0022- López Cruz E, Marrero Fente A, Castells Zayas Bazán
0337.2005.69.2.tb03911.x PMID:15689610 S, Agüero Díaz A (2003). Evaluation of the national
Lalla Y, Matias MA, Farah CS (2016). Assessment of oral program for the early diagnosis of oral cancer during
mucosal lesions with autofluorescence imaging and seven years. Arch Méd Camagüey. 7(5):534–46.
reflectance spectroscopy. J Am Dent Assoc. 147(8):650– Machoy M, Seeliger J, Szyszka-Sommerfeld L, Koprowski
60. doi:10.1016/j.adaj.2016.03.013 PMID:27117325 R, Gedrange T, Woźniak K (2017). The use of optical
Lane PM, Gilhuly T, Whitehead P, Zeng H, Poh CF, Ng coherence tomography in dental diagnostics: a
S, et al. (2006). Simple device for the direct visualiza- state-of-the-art review. J Healthc Eng. 2017:7560645.
tion of oral-cavity tissue fluorescence. J Biomed Opt. doi:10.1155/2017/7560645 PMID:29065642
11(2):024006. doi:10.1117/1.2193157 PMID:16674196 Mali SB, Dahivelkar S (2021). Liquid biopsy = indi-
Laronde DM, Williams PM, Hislop TG, Poh C, Ng S, vidualized cancer management: diagnosis, moni-
Bajdik C, et al. (2014). Influence of fluorescence on toring treatment and checking recurrence and
screening decisions for oral mucosal lesions in commu- metastasis. Oral Oncol. 123:105588. doi:10.1016/j.
nity dental practices. J Oral Pathol Med. 43(1):7–13. oraloncology.2021.105588 PMID:34744021
doi:10.1111/jop.12090 PMID:23750637 Maraki D, Becker J, Boecking A (2004). Cytologic and
Li HW, Huang CH, Chou C, Wang TF (2020). Knowledge, DNA-cytometric very early diagnosis of oral cancer.
attitudes, practice and related factors of oral cancer J Oral Pathol Med. 33(7):398–404. doi:10.1111/j.1600-
prevention among public health nurses in Taiwan. Eur J 0714.2004.0235.x PMID:15250831
Cancer Care (Engl). 29(5):e13262. doi:10.1111/ecc.13262 Marmot MG, Altman DG, Cameron DA, Dewar JA,
PMID:32458509 Thompson SG, Wilcox M (2013). The benefits and harms
Li L, Morse DE, Katz RV (2013). What constitutes a proper of breast cancer screening: an independent review. Br
routine oral cancer examination for patients at low risk? J Cancer. 108(11):2205–40. doi:10.1038/bjc.2013.177
Findings from a Delphi survey. Oral Surg Oral Med PMID:23744281
Oral Pathol Oral Radiol. 116(5):e379–86. doi:10.1016/j. Mashberg A (1980). Reevaluation of toluidine blue appli-
oooo.2013.07.010 PMID:24119528 cation as a diagnostic adjunct in the detection of
Li Y, St John MAR, Zhou X, Kim Y, Sinha U, Jordan RCK, asymptomatic oral squamous carcinoma: a continuing
et al. (2004). Salivary transcriptome diagnostics for prospective study of oral cancer III. Cancer. 46(4):758–
oral cancer detection. Clin Cancer Res. 10(24):8442–50. 63. doi:10.1002/1097-0142(19800815)46:4<758::AID-
doi:10.1158/1078-0432.CCR-04-1167 PMID:15623624 CNCR2820460420>3.0.CO;2-8 PMID:7397638
Li YN, Lu R, Zhang J, Zhou G (2019). Inter-and intra-ob- Masthan KM, Babu NA, Dash KC, Elumalai M
server agreement on the judgment of toluidine blue (2012). Advanced diagnostic aids in oral cancer.
staining for screening of oral potentially malignant Asian Pac J Cancer Prev. 13(8):3573–6. doi:10.7314/
disorders and oral cancer. Clin Oral Investig. 23(4):1709– APJCP.2012.13.8.3573 PMID:23098535
14. doi:10.1007/s00784-018-2595-7 PMID:30155573 Mathew B, Sankaranarayanan R, Sunilkumar KB,
Lin JY (2018). Information system of nationwide popula- Kuruvila B, Pisani P, Nair MK (1997). Reproducibility
tion-based cancer screening in Taiwan. J Glob Oncol. and validity of oral visual inspection by trained health
4(Suppl 2):209s. doi:10.1200/jgo.18.84500 workers in the detection of oral precancer and cancer.
Lingen MW, Abt E, Agrawal N, Chaturvedi AK, Cohen Br J Cancer. 76(3):390–4. doi:10.1038/bjc.1997.396
E, D’Souza G, et al. (2017a). Evidence-based clinical PMID:9252209
practice guideline for the evaluation of potentially Mathew B, Sankaranarayanan R, Wesley R, Nair MK
malignant disorders in the oral cavity: a report of (1995). Evaluation of mouth self-examination in
the American Dental Association. J Am Dent Assoc. the control of oral cancer. Br J Cancer. 71(2):397–9.
148(10):712–727.e10. doi:10.1016/j.adaj.2017.07.032 doi:10.1038/bjc.1995.81 PMID:7841060
PMID:28958308 Matthies L, Gebrekidan MT, Tegtmeyer JF, Oetter N,
Lingen MW, Tampi MP, Urquhart O, Abt E, Agrawal N, Rohde M, Vollkommer T, et al. (2021). Optical diagnosis
Chaturvedi AK, et al. (2017b). Adjuncts for the eval- of oral cavity lesions by label-free Raman spectros-
uation of potentially malignant disorders in the oral copy. Biomed Opt Express. 12(2):836–51. doi:10.1364/
cavity: diagnostic test accuracy systematic review BOE.409456 PMID:33680545
and meta-analysis – a report of the American Dental Maybury C, Horowitz AM, Yan AF, Green KM, Wang MQ
Association. J Am Dent Assoc. 148(11):797–813.e52. (2012). Maryland dentists’ knowledge of oral cancer
doi:10.1016/j.adaj.2017.08.045 PMID:29080605 prevention and early detection. J Calif Dent Assoc.
329
40(4):341–50. doi:10.1080/19424396.2012.12220903 Noninvasive imaging of oral neoplasia with a high-res-

PMID:22679674 olution fiber-optic microendoscope. Head Neck.
Mehrotra R, Mishra S, Singh M, Singh M (2011). The effi- 34(3):305–12. doi:10.1002/hed.21735 PMID:21413101
cacy of oral brush biopsy with computer-assisted anal- Müller MG, Valdez TA, Georgakoudi I, Backman V,
ysis in identifying precancerous and cancerous lesions. Fuentes C, Kabani S, et al. (2003). Spectroscopic detec-
Head Neck Oncol. 3(1):39. doi:10.1186/1758-3284-3-39 tion and evaluation of morphologic and biochem-
PMID:21864339 ical changes in early human oral carcinoma. Cancer.
Mehrotra R, Singh MK, Pandya S, Singh M (2008). The 97(7):1681–92. doi:10.1002/cncr.11255 PMID:12655525
use of an oral brush biopsy without computer-assisted Nagao T, Warnakulasuriya S (2003). Annual screening for
analysis in the evaluation of oral lesions: a study of 94 oral cancer detection. Cancer Detect Prev. 27(5):333–7.
patients. Oral Surg Oral Med Oral Pathol Oral Radiol doi:10.1016/S0361-090X(03)00101-6 PMID:14585319
Endod. 106(2):246–53. doi:10.1016/j.tripleo.2008.02.030 Nagaraju K, Prasad S, Ashok L (2010). Diagnostic effi-
PMID:18644521 ciency of toluidine blue with Lugol’s iodine in oral
Mehta FS, Gupta PC, Bhonsle RB, Murti PR, Daftary premalignant and malignant lesions. Indian J Dent
DK, Pindborg JJ (1986). Detection of oral cancer using Res. 21(2):218–23. doi:10.4103/0970-9290.66633
basic health workers in an area of high oral cancer PMID:20657091
incidence in India. Cancer Detect Prev. 9(3–4):219–25. Nanayakkara PG, Dissanayaka WL, Nanayakkara BG,
PMID:3742502 Amaratunga EA, Tilakaratne WM (2016). Comparison
Messadi DV, Younai FS, Liu HH, Guo G, Wang CY of spatula and cytobrush cytological techniques in early
(2014). The clinical effectiveness of reflectance optical detection of oral malignant and premalignant lesions:
spectroscopy for the in vivo diagnosis of oral lesions. a prospective and blinded study. J Oral Pathol Med.
Int J Oral Sci. 6(3):162–7. doi:10.1038/ijos.2014.39 45(4):268–74. doi:10.1111/jop.12357 PMID:26403502
PMID:25059250 Nathan C-AO, Kaskas NM, Ma X, Chaudhery S, Lian T,
Ministry of Health and Family Welfare (2021). Draft Moore-Medlin T, et al. (2014). Confocal laser endomi-
national oral health policy. New Delhi, India: National croscopy in the detection of head and neck precancerous
Oral Health Program Division, Ministry of Health lesions. Otolaryngol Head Neck Surg. 151(1):73–80.
and Family Welfare, Government of India. Available doi:10.1177/0194599814528660 PMID:24699456
from: https://main.mohfw.gov.in/sites/default/ National Cancer Control Programme (2020). National
files/N_56820_1613385504626.pdf. guideline on early detection and referral pathways of
Mishra GA, Shaikh HA, Pimple SA, Awasthi AA, Kulkarni common cancers in Sri Lanka. Colombo, Sri Lanka:
VY (2021). Determinants of compliance to popula- National Cancer Control Programme, Ministry of
tion-based oral cancer screening program among low Health and Indigenous Medical Services. Available
socioeconomic women in Mumbai, India. Indian J from: https://www.nccp.health.gov.lk/storage/post/
Community Med. 46(2):210–5. PMID:34321728 pdfs/National%20guideline%20on%20cancer%20
Mishra GS, Bhatt SH (2017). Novel program of using early%20detection%20and%20referal%20pathways.
village health workers in early detection and aware- pdf.
ness of head and neck cancers: audit of a community National Health Mission of India (2021). National
screening program. Indian J Otolaryngol Head Neck programme for prevention and control of cancer,
Surg. 69(4):488–93. doi:10.1007/s12070-017-1231-1 diabetes, cardiovascular diseases and stroke (NPCDCS).
PMID:29238679 New Delhi, India: National Health Mission, Ministry
Momen-Heravi F, Trachtenberg AJ, Kuo WP, of Health and Family Welfare, Government of India.
Cheng YS (2014). Genomewide study of salivary Available from: https://nhm.gov.in/index1.php?lang=1
microRNAs for detection of oral cancer. J Dent Res. &level=2&sublinkid=1048&lid=604.
93(7 Suppl):86S–93S. doi:10.1177/0022034514531018 National Health Mission Tamil Nadu (2021). National
PMID:24718111 oral health programme (NOHP). Tamil Nadu, India:
Morikawa T, Shibahara T, Takano M, Iwamoto M, Takaki T, National Health Mission Tamil Nadu, Department
Kasahara K, et al. (2021). Countermeasure and oppor- of Health and Family Welfare, Government of Tamil
tunistic screening systems for oral cancer. Oral Oncol. Nadu. Available from: https://www.nhm.tn.gov.in/
112:105047. doi:10.1016/j.oraloncology.2020.105047 node/6110.
PMID:33129059 Navone R, Pentenero M, Gandolfo S (2011). Liquid-based
Moyer VA; U.S. Preventive Services Task Force (2014). cytology in oral cavity squamous cell cancer. Curr Opin
Screening for oral cancer: U.S. Preventive Services Task Otolaryngol Head Neck Surg. 19(2):77–81. doi:10.1097/
Force recommendation statement. Ann Intern Med. MOO.0b013e328343af10 PMID:21252668
160(1):55–60. doi:10.7326/M13-2568 PMID:24276469 Ng SP, Mann IS, Zed C, Doudkine A, Matisic J (2012). The
Muldoon TJ, Roblyer D, Williams MD, Stepanek use of quantitative cytology in identifying high-risk
VM, Richards-Kortum R, Gillenwater AM (2012). oral lesions in community practice. Oral Surg Oral Med
330
Oral Pathol Oral Radiol. 114(3):358–64. doi:10.1016/j. Philip PM, Nayak P, Philip S, Parambil NA, Duraisamy
oooo.2012.06.008 PMID:22862977 K, Balasubramanian S (2018). Population-based cancer
Ni Riordain R, McCreary C (2009). Oral cancer – screening through community participation: outcome
current knowledge, practices and implications for of a district wide oral cancer screening program
training among an Irish general medical practitioner from rural Kannur, Kerala, India. South Asian J
cohort. Oral Oncol. 45(11):958–62. doi:10.1016/j. Cancer. 7(4):244–8. doi:10.4103/sajc.sajc_104_17
oraloncology.2009.04.008 PMID:19502101 PMID:30430093
Niebel HH, Chomet B (1964). Staining test for delineation Piazza C, Cocco D, Del Bon F, Mangili S, Nicolai P,
of oral intraepithelial neoplasms: in vivo staining test Majorana A, et al. (2010). Narrow band imaging and
for delineation of oral intraepithelial neoplastic change: high definition television in evaluation of oral and
preliminary report. J Am Dent Assoc. 68(6):801–6. oropharyngeal squamous cell cancer: a prospec-
doi:10.14219/jada.archive.1964.0190 PMID:14141814 tive study. Oral Oncol. 46(4):307–10. doi:10.1016/j.
Ogden GR, Cowpe JG, Green M (1992). Cytobrush and oraloncology.2010.01.020 PMID:20189447
wooden spatula for oral exfoliative cytology. A compar- Pivovar A, Gonçalves Dos Santos ZFD, Torres-Pereira CC
ison. Acta Cytol. 36(5):706–10. PMID:1523928 (2017). Oral cancer screening for high-risk individ-
Olms C, Hix N, Neumann H, Yahiaoui-Doktor M, uals in the primary healthcare setting using an active
Remmerbach TW (2018). Clinical comparison of liquid- approach. J Oral Pathol Med. 46(9):786–91. doi:10.1111/
based and conventional cytology of oral brush biopsies: jop.12552 PMID:28122153
a randomized controlled trial. Head Face Med. 14(1):9. Piyarathne NS, Rasnayake RMSGK, Angammana R,
doi:10.1186/s13005-018-0166-4 PMID:29843756 Chandrasekera P, Ramachandra S, Weerasekera M,
Papadiochou S, Papadiochos I, Perisanidis C, Papado- et al. (2021). Diagnostic salivary biomarkers in oral
georgakis N (2020). Medical practitioners’ educa- cancer and oral potentially malignant disorders and
tional competence about oral and oropharyngeal their relationships to risk factors – a systematic review.
carcinoma: a systematic review and meta-analysis. Expert Rev Mol Diagn. 21(8):789–807. doi:10.1080/1473
Br J Oral Maxillofac Surg. 58(1):3–24. doi:10.1016/j. 7159.2021.1944106 PMID:34148471
bjoms.2019.08.007 PMID:31785865 Poh CF, Lane P, MacAulay C, Zhang L, Rosin MP
Papanicolaou GN, Traut HF (1943). Diagnosis of uterine (2010). The application of tissue autofluorescence
cancer by the vaginal smear. New York, NY, USA: The in detection and management of oral cancer and
Commonwealth Fund. premalignant lesions. In: Rosenthal E, Zinn KR,
Patton LL, Ashe TE, Elter JR, Southerland JH, Strauss RP editors. Optical imaging of cancer. New York,
(2006). Adequacy of training in oral cancer prevention Dordrecht, Heidelberg, London: Springer; pp. 101–118.
and screening as self-assessed by physicians, nurse doi:10.1007/978-0-387-93874-5_6
practitioners, and dental health professionals. Oral Surg Poh CF, Ng SP, Williams PM, Zhang L, Laronde DM,
Oral Med Oral Pathol Oral Radiol Endod. 102(6):758– Lane P, et al. (2007). Direct fluorescence visualization
64. doi:10.1016/j.tripleo.2005.10.072 PMID:17138178 of clinically occult high-risk oral premalignant disease
Payne K, Brooks J, Spruce R, Batis N, Taylor G, Nankivell using a simple hand-held device. Head Neck. 29(1):71–6.
P, et al. (2019). Circulating tumour cell biomarkers in doi:10.1002/hed.20468 PMID:16983693
head and neck cancer: current progress and future Qamar I, Rehman S, Mehdi G, Maheshwari V, Ansari
prospects. Cancers (Basel). 11(8):1115. doi:10.3390/ HA, Chauhan S (2018). Utility of cytospin and cell
cancers11081115 PMID:31387228 block technology in evaluation of body fluids and urine
Payne K, Spruce R, Beggs A, Sharma N, Kong A, Martin samples: a comparative study. J Cytol. 35(2):79–82.
T, et al. (2018). Circulating tumor DNA as a biomarker doi:10.4103/JOC.JOC_240_16 PMID:29643653
and liquid biopsy in head and neck squamous cell Ramadas K, Arrossi S, Thara S, Thomas G, Jissa V, Fayette
carcinoma. Head Neck. 40(7):1598–604. doi:10.1002/ JM, et al. (2008a). Which socio-demographic factors are
hed.25140 PMID:29542214 associated with participation in oral cancer screening
Perera M, Al-Hebshi NN, Speicher DJ, Perera I, Johnson in the developing world? Results from a popula-
NW (2016). Emerging role of bacteria in oral carcino- tion-based screening project in India. Cancer Detect
genesis: a review with special reference to perio-patho- Prev. 32(2):109–15. doi:10.1016/j.cdp.2008.02.008
genic bacteria. J Oral Microbiol. 8(1):32762. doi:10.3402/ PMID:18632218
jom.v8.32762 PMID:27677454 Ramadas K, Lucas E, Thomas G, Mathew B, Balan A,
Pérez-de-Oliveira ME, Petersen Wagner V, do Amaral- Sankaranarayanan R, editors (2008b). Physical exam-
Silva GK, Almeida Leite A, Ajudarte Lopes M, Santos- ination of the oral cavity. In: A digital manual for
Silva AR, et al. (2020). An audit of cytopathology in the the early diagnosis of oral neoplasia. Lyon, France:
oral and maxillofacial region: 18 years of experience. International Agency for Research on Cancer.
Cytopathology. 31(6):555–63. doi:10.1111/cyt.12891 Available from: https://screening.iarc.fr/atlasoral_list.
PMID:32681691 php?cat=H2&lang=1.doi:10.1007/978-3-030-63234-2.
331
Ramadas K, Sankaranarayanan R, Jacob BJ, Thomas cancers. Cell Oncol (Dordr). 37(5):331–8. doi:10.1007/
G, Somanathan T, Mahé C, et al. (2003). Interim s13402-014-0188-2 PMID:25156495
results from a cluster randomized controlled oral Sankaranarayanan R, Fernandez Garrote L, Lence Anta
cancer screening trial in Kerala, India. Oral Oncol. J, Pisani P, Rodriguez Salva A (2002). Visual inspec-
39(6):580–8. doi:10.1016/S1368-8375(03)00041-1 tion in oral cancer screening in Cuba: a case-control
PMID:12798401 study. Oral Oncol. 38(2):131–6. doi:10.1016/S1368-
Rapado-González Ó, López-López R, López-Cedrún 8375(01)00033-1 PMID:11854059
JL, Triana-Martínez G, Muinelo-Romay L, Suárez- Sankaranarayanan R, Mathew B, Jacob BJ, Thomas G,
Cunqueiro MM (2019). Cell-free microRNAs as Somanathan T, Pisani P, et al.; The Trivandrum Oral
potential oral cancer biomarkers: from diagnosis to Cancer Screening Study Group (2000). Early find-
therapy. Cells. 8(12):1653. doi:10.3390/cells8121653 ings from a community-based, cluster-randomized,
PMID:31861130 controlled oral cancer screening trial in Kerala,
Rapado-González Ó, Martínez-Reglero C, Salgado- India. Cancer. 88(3):664–73. doi:10.1002/(SICI)1097-
Barreira Á, Muinelo-Romay L, Muinelo-Lorenzo J, 0142(20000201)88:3<664::AID-CNCR25>3.0.CO;2-V
López-López R, et al. (2021a). Salivary DNA methyl- PMID:10649262
ation as an epigenetic biomarker for head and neck Sankaranarayanan R, Ramadas K, Thara S, Muwonge R,
cancer. Part I: a diagnostic accuracy meta-analysis. Thomas G, Anju G, et al. (2013). Long term effect of
J Pers Med. 11(6):568. doi:10.3390/jpm11060568 visual screening on oral cancer incidence and mortality
PMID:34204396 in a randomized trial in Kerala, India. Oral Oncol.
Rapado-González Ó, Martínez-Reglero C, Salgado- 49(4):314–21. doi:10.1016/j.oraloncology.2012.11.004
Barreira Á, Santos MA, López-López R, Díaz-Lagares PMID:23265945
Á, et al. (2021b). Salivary DNA methylation as an Sankaranarayanan R, Ramadas K, Thomas G, Muwonge
epigenetic biomarker for head and neck cancer. Part R, Thara S, Mathew B, et al.; Trivandrum Oral Cancer
II: a cancer risk meta-analysis. J Pers Med. 11(7):606. Screening Study Group (2005). Effect of screening on
doi:10.3390/jpm11070606 PMID:34206840 oral cancer mortality in Kerala, India: a cluster-ran-
Rezaei F, Mozaffari HR, Tavasoli J, Zavattaro E, Imani domised controlled trial. Lancet. 365(9475):1927–33.
MM, Sadeghi M (2019). Evaluation of serum and sali- doi:10.1016/S0140-6736(05)66658-5 PMID:15936419
vary interleukin-6 and interleukin-8 levels in oral squa- Schwarz RA, Gao W, Daye D, Williams MD, Richards-
mous cell carcinoma patients: systematic review and Kortum R, Gillenwater AM (2008). Autofluorescence
meta-analysis. J Interferon Cytokine Res. 39(12):727–39. and diffuse reflectance spectroscopy of oral epithe-
doi:10.1089/jir.2019.0070 PMID:31314647 lial tissue using a depth-sensitive fiber-optic probe.
Roberts MC, Fohner AE, Landry L, Olstad DL, Smit Appl Opt. 47(6):825–34. doi:10.1364/AO.47.000825
AK, Turbitt E, et al. (2021). Advancing precision PMID:18288232
public health using human genomics: examples from Sciubba JJ (1999). Improving detection of precancerous
the field and future research opportunities. Genome and cancerous oral lesions. Computer-assisted analysis
Med. 13(1):97–106. doi:10.1186/s13073-021-00911-0 of the oral brush biopsy. U.S. Collaborative OralCDx
PMID:34074326 Study Group. J Am Dent Assoc. 130(10):1445–57.
Romano A, Di Stasio D, Petruzzi M, Fiori F, Lajolo C, doi:10.14219/jada.archive.1999.0055 PMID:10570588
Santarelli A, et al. (2021). Noninvasive imaging methods Scott SE, Rizvi K, Grunfeld EA, McGurk M (2010). Pilot
to improve the diagnosis of oral carcinoma and its study to estimate the accuracy of mouth self-examina-
precursors: state of the art and proposal of a three- tion in an at-risk group. Head Neck. 32(10):1393–401.
step diagnostic process. Cancers (Basel). 13(12):2864. doi:10.1002/hed.21341 PMID:20146330
doi:10.3390/cancers13122864 PMID:34201237 Shah A, Bhushan B, Akhtar S, Singh PK, Garg M, Gupta
Sahu A, Deshmukh A, Ghanate AD, Singh SP, Chaturvedi M (2020). Effectiveness of mouth self-examination for
P, Krishna CM (2012). Raman spectroscopy of oral screening of oral premalignant/malignant diseases
buccal mucosa: a study on age-related physiological in tribal population of Dehradun district. J Family
changes and tobacco-related pathological changes. Med Prim Care. 9(8):4381–5. doi:10.4103/jfmpc.
Technol Cancer Res Treat. 11(6):529–41. doi:10.7785/ jfmpc_535_20 PMID:33110864
tcrt.2012.500304 PMID:23101756 Shanks LA, Walker TW, McCann PJ, Kerin MJ (2011).
Sahu A, Krishna CM (2017). Optical diagnostics in oral Oral cavity examination: beyond the core curriculum?
cancer: an update on Raman spectroscopic applica- Br J Oral Maxillofac Surg. 49(8):640–2. doi:10.1016/j.
tions. J Cancer Res Ther. 13(6):908–15. doi:10.4103/0973- bjoms.2010.10.021 PMID:21345553
1482.191032 PMID:29237951 Sharma D, Rimal J, Kumar Maharjan I, Shrestha A,
Salazar C, Nagadia R, Pandit P, Cooper-White J, Banerjee Shrestha A, Regmee P (2021). Evaluation of oral
N, Dimitrova N, et al. (2014). A novel saliva-based potentially malignant disorders with autoflorescence,
microRNA biomarker panel to detect head and neck reflectance spectroscopy and vital staining and their
332
correlation with histopathology – hospital based screening. Otolaryngol Head Neck Surg. 145(6):956–60.
prospective study. Oral Oncol. 118:105312. doi:10.1016/j. doi:10.1177/0194599811416773 PMID:21804026
oraloncology.2021.105312 PMID:34023743 Talamini R, Barzan L, Franceschi S, Caruso G, Gasparin
Shieh Y, Eklund M, Madlensky L, Sawyer SD, Thompson A, Comoretto R (1994). Determinants of compli-
CK, Stover Fiscalini A, et al.; Athena Breast Health ance with an early detection programme for cancer
Network Investigators (2017). Breast cancer of the head and neck in north-eastern Italy. Eur J
screening in the precision medicine era: risk- Cancer B Oral Oncol. 30B(6):415–8. doi:10.1016/0964-
based screening in a population-based trial. J Natl 1955(94)90022-1 PMID:7719224
Cancer Inst. 109(5):djw290. doi:10.1093/jnci/djw290 Ten Haaf K, van der Aalst CM, de Koning HJ, Kaaks R,
PMID:28130475 Tammemägi MC (2021). Personalising lung cancer
Shrestha G, Maharjan L (2020). Mouth self-examina- screening: an overview of risk-stratification oppor-
tion for prevention and control of oral cavity cancer. tunities and challenges. Int J Cancer. 149(2):250–63.
JNMA J Nepal Med Assoc. 58(225):360–2. doi:10.31729/ doi:10.1002/ijc.33578 PMID:33783822
jnma.4910 PMID:32538936 Thakur M, Guttikonda VR (2017). Modified ultrafast
Silverman S Jr (1959). Intraoral exfoliative cytology. CA Papanicolaou staining technique: a comparative study.
Cancer J Clin. 9(4):125–8. doi:10.3322/canjclin.9.4.125 J Cytol. 34(3):149–53. doi:10.4103/JOC.JOC_23_16
PMID:14446770 PMID:28701828
Simonato LE, Tomo S, Scarparo Navarro R, Balbin Tian X, Chen Z, Shi S, Wang X, Wang W, Li N, et al.
Villaverde AGJ (2019). Fluorescence visualization (2015). Clinical diagnostic implications of body fluid
improves the detection of oral, potentially malig- miRNA in oral squamous cell carcinoma: a meta-anal-
nant, disorders in population screening. Photodiagn ysis. Medicine (Baltimore). 94(37):e1324. doi:10.1097/
Photodyn Ther. 27:74–8. doi:10.1016/j.pdpdt.2019.05.017 MD.0000000000001324 PMID:26376377
PMID:31116999 Tong H, Shen R, Wang Z, Kan Y, Wang Y, Li F, et al.;
Singh K, Sharma D, Kaur M, Gauba K, Thakur JS, Kumar Mass Cervical Cancer Screening Regimen Group
R (2017). Effect of health education on awareness (2009). DNA ploidy cytometry testing for cervical
about oral cancer and oral self-examination. J Educ cancer screening in China (DNACIC Trial): a prospec-
Health Promot. 6(1):27. doi:10.4103/jehp.jehp_82_15 tive randomized, controlled trial. Clin Cancer Res.
PMID:28584827 15(20):6438–45. doi:10.1158/1078-0432.CCR-09-1689
Solar O, Irwin A (2010). A conceptual framework for PMID:19825960
action on the social determinants of health. Social Torkamani A, Wineinger NE, Topol EJ (2018). The
Determinants of Health Discussion Paper 2 (Policy personal and clinical utility of polygenic risk scores.
and Practice). Geneva, Switzerland: World Health Nat Rev Genet. 19(9):581–90. doi:10.1038/s41576-018-
Organization. Available from: https://apps.who.int/ 0018-x PMID:29789686
iris/bitstream/handle/10665/44489/9789241500852_ Truelove EL, Dean D, Maltby S, Griffith M, Huggins K,
eng.pdf?sequence=1&isAllowed=y. Griffith M, et al. (2011). Narrow band (light) imaging
Steinkamp PJ, Voskuil FJ, van der Vegt B, Doff JJ, of oral mucosa in routine dental patients. Part I: assess-
Schepman KP, de Visscher SAHJ, et al. (2021). A stand- ment of value in detection of mucosal changes. Gen
ardized framework for fluorescence-guided margin Dent. 59(4):281–9, quiz 290–1, 319–20. PMID:21903568
assessment for head and neck cancer using a tumor Vaishampayan S, Malik A, Pawar P, Arya K, Chaturvedi P
acidosis sensitive optical imaging agent. Mol Imaging (2017). Short message service prompted mouth self-ex-
Biol. 23(6):809–17. doi:10.1007/s11307-021-01614-z amination in oral cancer patients as an alternative
PMID:34031845 to frequent hospital-based surveillance. South Asian
Stott-Miller M, Houck JR, Lohavanichbutr P, Méndez J Cancer. 6(4):161–4. doi:10.4103/sajc.sajc_258_16
E, Upton MP, Futran ND, et al. (2011). Tumor and PMID:29404294
salivary matrix metalloproteinase levels are strong van Ginkel JH, Slieker FJB, de Bree R, van Es RJJ, Van
diagnostic markers of oral squamous cell carcinoma. Cann EM, Willems SM (2017). Cell-free nucleic
Cancer Epidemiol Biomarkers Prev. 20(12):2628–36. acids in body fluids as biomarkers for the predic-
doi:10.1158/1055-9965.EPI-11-0503 PMID:21960692 tion and early detection of recurrent head and neck
Su WW, Yen AM, Chiu SY, Chen TH (2010). A cancer: a systematic review of the literature. Oral
community-based RCT for oral cancer screening Oncol. 75:8–15. doi:10.1016/j.oraloncology.2017.10.007
with toluidine blue. J Dent Res. 89(9):933–7. PMID:29224828
doi:10.1177/0022034510373763 PMID:20525960 van Keulen S, Nishio N, Fakurnejad S, Birkeland A,
Sweeny L, Dean NR, Magnuson JS, Carroll WR, Clemons Martin BA, Lu G, et al. (2019). The clinical appli-
L, Rosenthal EL (2011). Assessment of tissue auto- cation of fluorescence-guided surgery in head and
fluorescence and reflectance for oral cavity cancer neck cancer. J Nucl Med. 60(6):758–63. doi:10.2967/
jnumed.118.222810 PMID:30733319
333
Vinayagamoorthy K, Acharya S, Kumar M, Pentapati KC, Welch HG, Black WC (2010). Overdiagnosis in cancer.
Acharya S (2019). Efficacy of a remote screening model J Natl Cancer Inst. 102(9):605–13. doi:10.1093/jnci/
for oral potentially malignant disorders using a free djq099 PMID:20413742
messaging application: a diagnostic test for accuracy WHO (2007). Cancer control: knowledge into action:
study. Aust J Rural Health. 27(2):170–6. doi:10.1111/ WHO guide for effective programmes. Geneva,
ajr.12496 PMID:30942518 Switzerland: World Health Organization.
Walsh T, Liu JL, Brocklehurst P, Glenny AM, Lingen M, WHO (2013). WHO guidelines for screening and treatment
Kerr AR, et al. (2013). Clinical assessment to screen of precancerous lesions for cervical cancer prevention.
for the detection of oral cavity cancer and potentially Geneva, Switzerland: World Health Organization.
malignant disorders in apparently healthy adults. Available from: https://apps.who.int/iris/bitstream/
Cochrane Database Syst Rev. 2013(11):CD010173. handle/10665/94830/9789241548694_eng.pdf.
doi:10.1002/14651858.CD010173.pub2 PMID:24258195 WHO (2021). Oral health. Seventy-fourth World Health
Walsh T, Macey R, Kerr AR, Lingen MW, Ogden Assembly. WHA 74.5. Agenda item 13.2, 31 May 2021.
GR, Warnakulasuriya S (2021a). Diagnostic tests Geneva, Switzerland: World Health Organization.
for oral cancer and potentially malignant disor- Available from: https://apps.who.int/gb/ebwha/pdf_
ders in patients presenting with clinically evident files/WHA74/A74_R5-en.pdf.
lesions. Cochrane Database Syst Rev. 7(7):CD010276. Wilder-Smith P, Lee K, Guo S, Zhang J, Osann K, Chen
doi:10.1002/14651858.CD010276.pub3 PMID:34282854 Z, et al. (2009). In vivo diagnosis of oral dysplasia and
Walsh T, Warnakulasuriya S, Lingen MW, Kerr AR, malignancy using optical coherence tomography:
Ogden GR, Glenny AM, et al. (2021b). Clinical assess- preliminary studies in 50 patients. Lasers Surg Med.
ment for the detection of oral cavity cancer and 41(5):353–7. doi:10.1002/lsm.20773 PMID:19533765
potentially malignant disorders in apparently healthy Williams BR, Amon A (2009). Aneuploidy: cancer’s fatal
adults. Cochrane Database Syst Rev. 12(12):CD010173. flaw? Cancer Res. 69(13):5289–91. doi:10.1158/0008-
doi:10.1002/14651858.CD010173.pub3 PMID:34891214 5472.CAN-09-0944 PMID:19549887
Wang Y, Springer S, Mulvey CL, Silliman N, Schaefer J, Willoughby A, Andreassen PR, Toland AE (2019). Genetic
Sausen M, et al. (2015). Detection of somatic mutations testing to guide risk-stratified screens for breast
and HPV in the saliva and plasma of patients with head cancer. J Pers Med. 9(1):15. doi:10.3390/jpm9010015
and neck squamous cell carcinomas. Sci Transl Med. PMID:30832243
7(293):293ra104. doi:10.1126/scitranslmed.aaa8507 Winn DM, Lee Y-CA, Hashibe M, Boffetta P; INHANCE
PMID:26109104 consortium (2015). The INHANCE consortium:
Warnakulasuriya KA, Ekanayake AN, Sivayoham S, toward a better understanding of the causes and mech-
Stjernswärd J, Pindborg JJ, Sobin LH, et al. (1984). anisms of head and neck cancer. Oral Dis. 21(6):685–93.
Utilization of primary health care workers for early doi:10.1111/odi.12342 PMID:25809224
detection of oral cancer and precancer cases in Sri Wong DT (2012). Salivaomics. J Am Dent Assoc. 143(10
Lanka. Bull World Health Organ. 62(2):243–50. Suppl):19S–24S. doi:10.14219/jada.archive.2012.0339
PMID:6610492 PMID:23034834
Warnakulasuriya KA, Nanayakkara BG (1991). Woolf SH, Harris R (2012). The harms of screening: new
Reproducibility of an oral cancer and precancer detec- attention to an old concern. JAMA. 307(6):565–6.
tion program using a primary health care model in Sri doi:10.1001/jama.2012.100 PMID:22318274
Lanka. Cancer Detect Prev. 15(5):331–4. PMID:1751941 Yang EC, Brenes DR, Vohra IS, Schwarz RA, Williams
Warnakulasuriya S (2020). Oral potentially malignant MD, Vigneswaran N, et al. (2020). Algorithm to quan-
disorders: a comprehensive review on clinical aspects tify nuclear features and confidence intervals for clas-
and management. Oral Oncol. 102:104550. doi:10.1016/j. sification of oral neoplasia from high-resolution optical
oraloncology.2019.104550 PMID:31981993 images. J Med Imaging (Bellingham). 7(5):054502.
Warnakulasuriya S, Kerr AR (2021). Oral cancer screening: doi:10.1117/1.JMI.7.5.054502 PMID:32999894
past, present, and future. J Dent Res. 100(12):1313–20. Yang EC, Schwarz RA, Lang AK, Bass N, Badaoui H, Vohra
doi:10.1177/00220345211014795 PMID:34036828 IS, et al. (2018a). In vivo multimodal optical imaging:
Warnakulasuriya S, Pindborg JJ (1990). Reliability of oral improved detection of oral dysplasia in low-risk oral
precancer screening by primary health care workers mucosal lesions. Cancer Prev Res (Phila). 11(8):465–76.
in Sri Lanka. Community Dent Health. 7(1):73–9. doi:10.1158/1940-6207.CAPR-18-0032 PMID:29903741
PMID:2357611 Yang EC, Tan MT, Schwarz RA, Richards-Kortum RR,
Wee AG, Zimmerman LM, Pullen CH, Allen CM, Lambert Gillenwater AM, Vigneswaran N (2018b). Noninvasive
PM, Paskett ED (2016). Evaluating a web-based educa- diagnostic adjuncts for the evaluation of potentially
tional module on oral cancer examination based on premalignant oral epithelial lesions: current limi-
a behavioral framework. J Cancer Educ. 31(1):158–65. tations and future directions. Oral Surg Oral Med
doi:10.1007/s13187-014-0776-x PMID:25572460
334
Oral Pathol Oral Radiol. 125(6):670–81. doi:10.1016/j. broadband white light versus narrowband imaging. Int
oooo.2018.02.020 PMID:29631985 J Oral Maxillofac Implants. 42(6):693–701. doi:10.1016/j.
Yang EC, Vohra IS, Badaoui H, Schwarz RA, Cherry KD, ijom.2012.10.039 PMID:23583318
Quang T, et al. (2019). Development of an integrated Yu JS, Chen YT, Chiang WF, Hsiao YC, Chu LJ, See LC,
multimodal optical imaging system with real-time et al. (2016). Saliva protein biomarkers to detect oral
image analysis for the evaluation of oral premalig- squamous cell carcinoma in a high-risk population
nant lesions. J Biomed Opt. 24(2):1–10. doi:10.1117/1. in Taiwan. Proc Natl Acad Sci USA. 113(41):11549–54.
JBO.24.2.025003 PMID:30793567 doi:10.1073/pnas.1612368113 PMID:27663741
Yang SW, Lee YS, Chang LC, Chien HP, Chen TA (2013).
Light sources used in evaluating oral leukoplakia:
335
5. SUMMARY
5.1 Oral cancer and oral potentially 5.1.2 Global burden

malignant disorders Cancer of the oral cavity is the most common
cancer type in the head and neck region of
5.1.1 Anatomy of the oral cavity and the the body, with about 380 000 new cases and
oropharynx 180 000 deaths worldwide in 2020. Cancer of the
oropharynx is less common, with an estimated
The oral cavity is the entrance to the gastro- 98 412 new cases and 48 143 deaths worldwide in
intestinal tract. It is bounded anteriorly by the 2020. The estimated age-standardized incidence
lips, posteriorly by the faucial arches anterior rates of oral cancer are highest in Melanesia and
to the tonsils, laterally by the cheeks (buccal South Asia. The incidence rate of oropharyngeal
mucosae), superiorly by the palate, and inferiorly cancer is highest in Europe. The incidence rate of
by the muscular floor. The tongue occupies the oral cancer is highest in countries with medium
floor of the oral cavity. The subsites of the oral levels of the Human Development Index, and
cavity include the lips (mucosal surface or labial the incidence rate of oropharyngeal cancer is
mucosae), oral commissures, buccal mucosae, highest in countries with very high levels of the
tongue, gingivae (gums), floor of the mouth, and Human Development Index; low socioeconomic
palate. status is associated with an increased risk of
The oropharynx is a tube-shaped fibromus- both cancer types. During the past two decades,
cular structure behind the oral cavity, contin- the observed incidence rates of oral cancers
uous with the nasopharynx superiorly and the have generally decreased, especially in North
hypopharynx inferiorly. It extends from the America, some countries in South-East Asia,
lower surface of the soft palate to the upper and some countries in Europe. The incidence
border of the epiglottis and communicates rates of oropharyngeal cancer appear to be
with the oral cavity anteriorly. The pala- increasing in most countries worldwide. Because
tine tonsils project from the lateral wall of the of population growth, the burden of oral cancer
oropharynx, and the lingual tonsils are found on and oropharyngeal cancer will increase in the
the posterior third (base) of the tongue. next two decades.
337
5.1.3 Oral neoplasia to health-care systems, associated comorbidi-

ties, and the quality of treatment planning also
(a) Classification and natural history of OPMDs affect the prognosis. The 5-year survival rate
and oral cancer of oral cancer ranges from 0% to 64% across
An oral potentially malignant disorder countries, with a median of 39%. The Union for
(OPMD) is defined as any oral mucosal abnor- International Cancer Control/American Joint
mality that is associated with a statistically Committee on Cancer (UICC/AJCC) staging
increased risk of developing oral cancer. system (eighth edition) has recently been updated
OPMDs include leukoplakia, proliferative verru- to enable improved staging and prediction of
cous leukoplakia, erythroplakia, oral submu- prognosis of patients with oral cancer.
cous fibrosis, oral lichen planus, oral lichenoid
lesions, oral graft-versus-host disease, oral lupus (c) Treatment and management of OPMDs
erythematosus, actinic keratosis, palatal lesions and oral cancer
in reverse smokers, and dyskeratosis congenita. OPMDs are a clinically diverse group of
OPMDs are a heterogeneous group of lesions, disorders, which require a careful clinico-
and the transformation rates to cancer vary pathological evaluation and monitoring over
from 1.4% to 49.5%; the presence of epithelial long periods, and their clinical management
dysplasia is the most significant predictor. The is challenging. Consensus guidelines for clin-
highest-risk OPMDs are proliferative verrucous ical management of patients with leukoplakia
leukoplakia and erythroplakia, and the risk is or erythroplakia, oral submucous fibrosis, and
lowest for oral lichen planus. The natural history oral lichen planus have recently been proposed.
of OPMDs is not always linear or predictable. After clinical identification, the current refer-
They may persist unchanged, progress towards ence standard is to perform a biopsy for histo-
cancer, or even regress. pathological diagnosis, treatment guidance, and
The clinical features of OPMDs vary prognostication based on the grade of epithe-
according to the type of OPMD. The clinical lial dysplasia. Preventive strategies (modifying
features of oral cancer vary depending on the known risk factors) can reduce risk of oral
site and the stage of presentation. Early cancers cancer. Both topical or systemic agents and
may present as erythroleukoplakic lesions with surgical excision or ablation of OPMDs have
red, white, or mixed red and white areas. As been proposed. No consensus exists about time
the disease advances, there is ulceration and/or intervals of serial follow-up, which remains
nodularity (exophytic or endophytic tumours). important for monitoring and surveillance.
Prognosis of oral cancer depends on multiple Management and treatment of oral cancer
factors, including tumour-, host-, and treat- are multidisciplinary. Treatment pillars include
ment-related factors. Spread of cancer to the combinations of surgery, chemotherapy, radio-
regional lymph nodes has a direct negative effect therapy, and, more recently, immunotherapy, for
on prognosis. early-stage and locally advanced disease.
(b) Stage at diagnosis and stage-related
survival
Stage at diagnosis is one of the main factors
that affects cancer prognosis. Most patients
with cancers of the oral cavity are diagnosed
with advanced disease. Besides staging, access
338
5.2 Reducing incidence of cancer or (b) Alcohol consumption

precancer Consumption of alcohol, particularly heavy
alcohol consumption, is associated with an
5.2.1 Established risk factors increased risk of oral cancer and oropharyn-
Tobacco smoking, alcohol consumption, geal cancer. For the same cumulative exposure,
smokeless tobacco use, chewing areca nut prod- higher alcohol intake for shorter duration confers
ucts (including betel quid) [hereafter referred a greater risk of oral cancer compared with
to as “areca nut”] with or without tobacco, and lower alcohol intake for longer duration. Alcohol
infection with human papillomavirus type 16 consumption also confers an elevated risk of
(HPV16) are classified by the International OPMDs, particularly leukoplakia. Alcohol
Agency for Research on Cancer (IARC) Mono- consumption increases synergistically in combi-
graphs programme as carcinogenic to humans nation with tobacco smoking and accounts for
(Group 1) and are established risk factors for 64% of the population attributable risk of oral
cancers of the oral cavity. Tobacco smoking, cancer.
alcohol consumption, and chewing areca nut
(c) Smokeless tobacco use
with tobacco are also established risk factors for
pharyngeal cancer. HPV16 infection is an estab- Use of smokeless tobacco (products contain-
lished cause of oropharyngeal cancer. ing tobacco but not including areca nut or betel
quid) increases the risk of oral cancer in a dose-
(a) Tobacco smoking dependent manner with increasing frequency of
In most countries, tobacco smoking is the use (times per day), duration of use (in months),
leading cause of oral cancer and oral cancer and amount of time the product is retained in the
death. The risk of oral cancer increases with mouth. The risk of oral cancer, after accounting
increasing frequency (number of cigarettes for concurrent smoking, appears to be higher
smoked per day), duration (in years), and cumu- in women than in men. The fraction of oral
lative pack-years of smoking. The duration of cancers attributable to smokeless tobacco use is
smoking has a stronger effect on risk of oral high (50–68%) in countries in South and South-
cancer than the frequency of smoking does, East Asia and in the Sudan, and is much lower
but the elevated risk becomes significant at a (1.6–6.6%) in North America. Smokeless tobacco
low number of cigarettes smoked per day. For use increases the risk of OPMDs, particularly
a given level of exposure, the risk of oral cancer leukoplakia.
conferred by tobacco smoking is similar by sex.
(d) Chewing areca nut products (including
Cigar, pipe, and bidi smoking are associated
betel quid) with added tobacco
with a significantly increased risk of oral cancer.
Tobacco smoking also causes oropharyngeal Chewers of areca nut with added tobacco are
cancer, with a dose–response increase in risk. at increased risk of oral cancer and oropharyn-
Worldwide, about one quarter to one third of oral geal cancer compared with never-chewers. The
cancers and 30–49% of oropharyngeal cancers effect is larger in women than in men. The risk
are attributable to tobacco smoking alone. The of oral cancer increases with frequency (times
main OPMDs caused by tobacco smoking are per day) and duration (in years) of chewing.
leukoplakia and erythroplakia. Chewers of areca nut with added tobacco also
have increased risk of OPMDs compared with
never-chewers, and the risk is highest for oral
339
submucous fibrosis, erythroplakia, and leuko- cancers, ranging from about 40–50% in North
plakia. The risk of OPMDs increases with the America, Europe, Australia and New Zealand,
frequency of chewing (times per day), the dura- Japan, and the Republic of Korea to < 15% in
tion of chewing (in years), and a younger age at most other parts of the world. This heterogeneity
the start of chewing. In India, the population may arise from a combination of differences in
attributable fraction of chewing betel quid with relevant sexual behaviours and the prevalence of
added tobacco for oral cancer was estimated exposure to other risk factors.
to be higher in women (63.2%) than in men
(44.7%). The population attributable risk for (g) Combined effects of established risk factors
OPMDs was estimated to be 84% in Sri Lanka. Combined exposure to more than one of the
risk factors confers a risk that is the sum of the
(e) Chewing areca nut products (including individual risks for each of these carcinogens,
betel quid) without tobacco and can confer a risk that exceeds the sum or
Chewing areca nut without tobacco is the multiplication product of the individual risk
predominant in Taiwan (China) but also occurs estimates.
in other countries in South-East Asia, including The relative risk of oral cancer in individuals
Bangladesh, India, Myanmar, Sri Lanka, and who both smoke tobacco and consume alcohol
Thailand. Chewers of areca nut without tobacco is greater than multiplicative (i.e. the joint effect
are at increased risk of oral cancer compared is greater than the multiplication product of the
with never-chewers. The risk of oral cancer individual effects). The relative risk of oral cancer
increases with the frequency (times per day) in individuals who smoke tobacco, consume
and the duration of chewing. Chewers of areca alcohol, and chew areca nut with or without
nut without tobacco also have increased risk tobacco is greater than additive (i.e. the joint
of OPMDs compared with never-chewers; the effect is greater than the sum of the individual
highest risks have been reported for oral submu- effects).
cous fibrosis, erythroplakia, and leukoplakia.
The risk of OPMDs increases with the frequency (h) Additional potential risk factors
of chewing (times per day), the duration of Other potential risk factors for oral cancer
chewing (in years), and a younger age at the include second-hand smoke, indoor air pollu-
start of chewing. In Taiwan (China), the popu- tion, low socioeconomic status (measured by
lation attributable fraction of chewing areca nut education level, income, occupation), chronic
without tobacco was estimated to be 57.3% for mechanical irritation, and drinking hot maté.
oral cancer, 85.4% for oral submucous fibrosis, Other potential risk factors are dysbiosis of the
and 73.2% for leukoplakia. oral microbiome; exogenous environmental,
occupational, and infectious exposures; and
(f) HPV16 infection poor oral hygiene or oral health, resulting in
HPV infections are acquired primarily persistent or chronic inflammation.
through sexual activity. HPV16 infection is
associated with a < 5-fold increased risk of oral 5.2.2 Impact upon quitting
cancer and with a 14-fold to > 100-fold increased
risk of oropharyngeal cancer. HPV16 infection (a) Tobacco smoking
causes ~2% of oral cancers worldwide. In contrast, Since the IARC Handbooks Volume 11 eval-
there is wide geographical heterogeneity in uation (in 2006), two cohort studies, one meta-
attributable fractions of HPV for oropharyngeal analysis of 17 case–control studies (including
340
3302 cases of oral cancer and 16 377 controls), an intervention study in India reported a 70%
and two additional case–control studies on inci- increase in risk of leukoplakia in former smokers
dent oral cancer consistently showed a progres- compared with never-smokers, in contrast to a
sive reduction in the relative risk of oral cancer > 3-fold increase in risk in current smokers.
with increasing time since quitting smoking, In case–control studies, estimates of the rela-
with a statistically significant trend in four of tive risk in former smokers compared with
these studies. The reduction in risk was evident never-smokers ranged from 0.5 to 4.9; the 95%
in all studies within 10 years of smoking cessa- confidence interval often included 1, and the
tion. In the meta-analysis, the risk of oral cancer magnitude was mostly, but not always, markedly
became significantly lower in former smokers lower than the relative risk in current smokers
compared with current smokers within 4 years (which ranged from 0.48 to 10.0).
after cessation (35% reduction), and the esti-
mated relative risk in former smokers reached (b) Alcohol consumption
the relative risk in never-smokers after ≥ 20 years Four studies were identified that reported
of smoking cessation, based on fully adjusted estimates of the relative risk of oral cancer by
risk estimates taking into account frequency of time since cessation of alcohol consumption: one
alcohol consumption and cumulative smoking. pooled analysis of case–control studies and three
The body of evidence on oral cancer and other, smaller case–control studies. In addition,
smoking cessation included populations with two cohort studies were identified that had data
a wide geographical distribution across North on former alcohol drinkers relative to never-
America, Central and South America, Europe, drinkers. The large international meta-analysis
and Asia, including both men and women in four from 2010, which pooled data on 3302 cases
of five studies. The reported risk estimates were of oral cancer from 13 case–control studies,
based mostly on former and current smokers found that the reduction in risk of oral cancer
of cigarettes but also included a minority of after alcohol cessation increases with time since
smokers of other tobacco products (cigars, pipes, cessation; the effects were more pronounced
and hand-rolled cigarettes). in former heavy drinkers (≥ 3 drinks per day),
Based on a single study that combined oral reducing the risk by > 50% by 20 years of quitting
cancer and pharyngeal cancer, quitting smoking (odds ratio, 0.43; 95% confidence interval [CI],
at any age was associated with a significant 0.28–0.67) for oral cancer, and were less clear
reduction in the risk of these cancers compared for oropharyngeal and hypopharyngeal cancers
with current smokers, with a progressive and combined. Of the three earlier smaller informa-
significant lowering of the risk with decreasing tive case–control studies, only the one in India
age at quitting. reported data comparing former versus current
Nine studies on smoking cessation and inci- drinkers, and found a tendency for reduction in
dence or prevalence of OPMDs were identified, the risk of oral cancer associated with ≥ 10 years
conducted in Brazil, India, Kenya, Puerto Rico, of quitting (odds ratio, 0.62; 95% CI, 0.19–2.05).
Sri Lanka, Taiwan (China), and the USA. In a No studies on the impact of duration of
large cohort study in India comparing former alcohol cessation on risk of OPMDs were iden-
versus current bidi smokers after 10 years tified. Based on data from seven case–control
of follow-up, the incidence of leukoplakia studies, risk estimates for OPMDs in former
decreased substantially (85% decrease) after drinkers relative to never-drinkers were gener-
smoking cessation. In addition, a large commu- ally higher than for current drinkers relative to
nity-based case–control study nested within
341
never-drinkers, particularly for leukoplakia and case–control studies, all in India), one published
erythroplakia. meta-analysis, and two primary analyses under-
taken by the Working Group (one cohort study
(c) Smokeless tobacco use and one case–control study, both in India). A
A total of six studies were available that primary intervention study in India assessed only
examined the association between former use OPMDs (leukoplakia) as the primary outcome.
compared with never use of smokeless tobacco Results from the published studies were
and risk of oral cancer. (None of the studies inconsistent. Three out of five studies reported
considered current users of smokeless tobacco a non-significantly lower relative risk of oral
as the reference group, and none provided risk cancer in former chewers compared with that
estimates by time since quitting use.) There were in current chewers. The other two studies
two large cohort studies, in Sweden and Norway, reported an increased risk, but the estimates
and four case–control studies, three in Sweden were not adjusted for tobacco smoking and
and one in Yemen. Neither of the two cohort alcohol consumption, and the results could
studies found an association between use of oral be due to reverse causation. Results from the
snuff (former use or current use) and risk of oral meta-analysis did not show any inverse associ-
cancer; they reported non-significant relative risk ation. Nevertheless, primary analyses from the
estimates for former users of 0.7–1.0. Although cohort study in India showed a reduction in risk
both cohort studies were well powered, exposure of oral cancer in former chewers of 3% (95% CI,
categories for smokeless tobacco use (as current, 1–4%) per year of cessation of chewing areca nut
former, and never use) were defined at study entry with added tobacco. In addition, results were
only, with no reassessment of status of snuff consistent across studies for a reduction in the
use. In addition, neither of the studies adjusted relative risk of OPMDs by duration of cessation
for alcohol consumption. These limitations are of chewing areca nut with added tobacco. The
particularly important given the long follow-up primary prevention study showed strong reduc-
period of 12–35 years. The registry-based case– tions in the incidence of leukoplakia 5 years after
control study in Sweden, which included 128 the intervention, by 49% (95% CI, 7–72%) in men
cases of oral cancer and 756 matched controls, and 81% (95% CI, 70–89%) in women.
reported a 1.8-fold non-statistically significant
increased risk of oral cancer in former oral snuff (e) Chewing areca nut products (including
users after adjustment for potential confounding betel quid) without tobacco
factors. Evidence for reduction in risk of oral cancers
Data from eight studies on the association or OPMDs with cessation of chewing areca nut
between former use of smokeless tobacco and without tobacco comes from four published case–
risk of OPMDs were inconsistent, and all except control studies (three in Taiwan [China] and one
one study lacked a definition of former users in Papua New Guinea), one published meta-
with regard to duration of cessation. analysis, and four primary analyses undertaken
by the Working Group (three cohort studies and
(d) Chewing areca nut products (including one case–control study, all in Taiwan [China]).
betel quid) with added tobacco Cessation of chewing areca nut without tobacco
Evidence for reduction in risk of oral was consistently associated with a reduction in
cancers or OPMDs with cessation of chewing the relative risk of oral cancer in former chewers
areca nut with added tobacco comes from five compared with current chewers; the inverse asso-
published studies (three cohort studies and two ciation was significant after long-term cessation
342
(≥ 15 years of quitting). Based on the primary a non-significant protective effect and others
data analyses, risk reductions per year of cessa- showing no benefit; it should also be noted that
tion ranged from 2.3% to 6.7%. Furthermore, risk drinking very hot beverages (at temperatures
reductions were generally larger for younger ages > 65 °C) may increase risks of oral cancer and
at cessation. In addition, results were consistent pharyngeal cancer. Studies examining consump-
across studies for a reduction in the relative risk tion of fruits and vegetables found a general
of OPMDs in former chewers compared with reduction in the relative risk of oral cancer asso-
current chewers, and by duration of cessation of ciated with increasing consumption of fruits or
chewing areca nut without tobacco. vegetables. Consumption of dietary fibre has also
been shown to have protective effects on develop-
(f) HPV16 infection ment of oral cancer.
Three types of vaccines against HPV infec- The effects of dietary agents on the develop-
tion are currently available: a bivalent vaccine, a ment of OPMDs were examined in several popu-
quadrivalent vaccine, and a nonavalent vaccine. lation-based studies conducted in India and Sri
All three target HPV16, the type that causes most Lanka, and a hospital-based study in Italy. In
HPV-associated oral and oropharyngeal cancers. general, consumption of foods and nutrients rich
HPV vaccines are prophylactic (i.e. vaccination in dietary fibre, vitamins A, C, E, and B12, β-car-
prevents future acquisition of infection) and otene, lycopene, folate, retinol, α-tocopherol, and
not therapeutic (i.e. vaccination does not enable antioxidant mineral zinc has been found to be
clearance of prevalent infection). Studies show protective against the development of OPMDs.
strong evidence of reduction in the prevalence Also, biochemical studies were conducted on
of oral and oropharyngeal HPV16 infection in serum or plasma samples from patients with
vaccinated individuals compared with unvacci- leukoplakia or oral submucous fibrosis. The
nated individuals. Because HPV vaccines were available data indicate that the consumption
only approved recently (in 2006 for women and of foods and nutrients rich in certain vitamins
in 2011 for men in most countries worldwide), and antioxidants may inhibit the development of
the impact of HPV vaccination will take several OPMDs.
years or even decades to result in a reduction
in incidence of oral cancer or oropharyngeal 5.3 Cessation of smokeless tobacco
cancer. However, the anticipated reductions in
the incidence of HPV-associated oral cancer and and/or areca nut use
oropharyngeal cancer will depend on the extent 5.3.1 Product definition and description
of vaccination coverage in the population in any
specific country. The term “smokeless tobacco” refers to
a large variety of commercially available or
5.2.3 Preventive dietary agents non-commercially available products that
contain tobacco as the principal constituent and
Several studies have examined the protective that are used either orally (chewing, sucking,
effects of consuming coffee, tea, fruits and vege- placing in the cheek or lip pouch, or drinking)
tables, and dietary fibre on the incidence of oral or nasally, without combustion. Areca nut is the
cancer. Several population studies have shown a seed of Areca catechu L. and is used as a chewing
significant inverse relationship between coffee substance, either alone or in combination with
intake and incidence of oral cancer. The effect of other substances. Areca nut is the primary
tea intake is unclear, with some studies showing component of betel quid, which may also be
343
consumed without tobacco. Smokeless tobacco decreasing in older people. Use of areca nut and
and areca nut may be consumed separately or betel quid has spread from the Philippines across
combined. the Western Pacific islands over the past century;
Smokeless tobacco or areca nut products are the prevalence of use is about 80% in Palau and
available as a myriad of products. The products the Solomon Islands. Smokeless tobacco use is
vary substantially in their names and their use in not common. Initiation of use of areca nut prod-
each region; the greatest diversity is observed in ucts by young people is increasing, for example
South and South-East Asia. in Guam (USA).
Both smokeless tobacco and areca nut contain
multiple carcinogens, and both have been classi- (c) WHO European Region
fied as carcinogenic to humans (Group 1) by the In the WHO European Region, the overall
IARC Monographs programme. prevalence of use of both areca nut and smoke-
less tobacco is low, with a prevalence of use in
5.3.2 Prevalence of consumption most countries of less than 2%. However, in four
countries the prevalence exceeded the global
(a) WHO South-East Asia Region average for smokeless tobacco use (6%). There
The World Health Organization (WHO) is a marked use of specific types of smokeless
South-East Asia Region has the highest preva- tobacco (snus) by men, in Nordic countries and
lence of use of areca nut and smokeless tobacco in populations in central Asia; also, people of
products in adults worldwide, ranging from Asian descent and immigrants from South Asia
2.1% in Thailand to 27.5% in Bangladesh. In may have use patterns from those regions. The
2019–2020, about 30% of men and about 13% of prevalence of use was highest in Sweden (14%)
women in India were daily users of smokeless and Norway (18%), with a higher prevalence of
tobacco or areca nut products. In several coun- use in men than in women. Areca nut, gutka, and
tries (e.g. Bangladesh, Indonesia, and Thailand), zarda are imported and are used only by immi-
the prevalence of use is higher in women than in grant communities from Bangladesh, India, and
men. The prevalence of use of smokeless tobacco Pakistan, in which a prevalence of use of about
or areca nut products is also high in young people 7% is reported.
in this region; Nepal has the highest reported
prevalence in adolescents (16%). South-East Asia (d) WHO Region of the Americas
is culturally very diverse, and the forms in which In the WHO Region of the Americas, the
these products are prepared and mixed are use of smokeless tobacco and areca nut is not
highly variable. Therefore, it is generally not culturally embedded, and only limited data are
possible to disaggregate data for areca nut and available about patterns of use in the general
for smokeless tobacco. Commonly used prod- population. A relatively small spectrum of
ucts include gutka, khaini, gul, betel quid (with smokeless tobacco products (e.g. snuff, snus,
or without tobacco), and supari. iqmik, chimó, plug) is currently used by about
1.4% of the population (ranging from 0.2% in
(b) WHO Western Pacific Region Argentina to 3.5% in Venezuela), with a higher
The WHO Western Pacific Region is cultur- prevalence of use in men (2.5%) than in women
ally extremely diverse. Consumption of areca (0.3%). Several factors are involved in the prev-
nut is common in Hunan Province (China) alence of smokeless tobacco use in this region,
and in Taiwan (China), where the prevalence of such as immigrants from South Asia, military
use is about 10% in men, but the prevalence is personnel, baseball players, middle and high
344
school students, and the quilombola community. effects such as relieving headaches, improving
Areca nut is not commonly used in this region, sleep quality, inducing relaxation, aiding deci-
except in scattered populations in Hawaii (USA). sion-making, reducing boredom, and inducing
a feeling of being energized are facilitators for
(e) WHO African Region use of smokeless tobacco and areca nut. Use of
The WHO African Region has the second- smokeless tobacco is associated with older age
highest prevalence of smokeless tobacco use in groups, and men generally have a higher likeli-
adults worldwide, with an estimated 15 million hood of use.
adult users. Most users of smokeless tobacco The socioeconomic determinants of use of
are men (8 million), but the prevalence of use in smokeless tobacco and areca nut are income
women is high in some countries. The prevalence level, employment, and education level. A large
of smokeless tobacco use ranges from 0.1% in proportion of users of smokeless tobacco have
women in Eritrea to 25% in men in Madagascar. low socioeconomic status, especially in unem-
Smokeless tobacco is commonly used without ployed people. Type of employment also deter-
areca nut, through nasal or oral application. Some mines use behaviour. The proportions of users
of the commonly used products are shammah of smokeless tobacco and areca nut are high in
(moist snuff), taaba (snuff), and paraky. Use of occupations that require long working hours
areca nut without tobacco by a minority popula- or continuously repeated activities, such as in
tion of South Asian descent has been reported in drivers and construction workers. With regard
some countries (e.g. South Africa). to education level, lower education levels are
consistently associated with increased preva-
(f) WHO Eastern Mediterranean Region lence of smokeless tobacco use. The relatively low
In the WHO Eastern Mediterranean Region, cost of the quid compared with smoked tobacco
there are about 21 million adult users of smoke- has been reported to be a socioeconomic deter-
less tobacco. The prevalence of use is much higher minant of areca nut use.
in men (~18 million) than in women (~3 million). Socioculturally, influence from family
The prevalence of use and the products used vary members and peer pressure are important deter-
across countries. The most common smoke- minants of both smokeless tobacco use and
less tobacco products used in the region are areca nut use. Sharing of areca nut is a usual
toombak (especially in the Sudan) and naswar (a practice during social gatherings and is a signif-
similar product that is common in the Arabian icant cultural identifier, which reinforces social
Peninsula). The prevalence of smokeless tobacco acceptance. It is also considered a symbol of
use ranges from 0.1% in women in Egypt to 34% love and marriage in many places, notably in
in men in Afghanistan. India and in Taiwan (China). Sociodemographic
factors for smokeless tobacco use include area of
(g) Determinants of use residence; there is a higher propensity for smoke-
Determinants of use of smokeless tobacco less tobacco use in rural areas than in urban
and areca nut can be categorized into individual, areas. Advertisements are another determinant
social, and environmental factors. of use of smokeless tobacco and areca nut.
Among individual factors, users’ level of
knowledge about the harmful health effects of
use is a determinant of use of smokeless tobacco
or areca nut. In addition, perceived positive
345
5.3.3 Interventions for cessation of use of intervention in the control arm. One study
reported a significant effect on the prevention of
(a) Behavioural interventions initiation of smokeless tobacco use in some of the
Nine intervention studies assessed the effects participants in the intervention and control arms
of behavioural interventions for cessation of use who were non-users at baseline.
of smokeless tobacco or areca nut products in
adults: six randomized controlled trials (RCTs) (b) Pharmacological interventions
in the USA, one RCT in Sweden, and two large Three RCTs assessed the effects of pharmaco-
cohort studies in India. Interventions included logical interventions for cessation of use of smoke-
dental examination; brief advice by physicians, less tobacco or areca nut products: two with
dentists, or behavioural scientists, together with nicotine replacement therapy (one with nicotine
a written manual or leaflet; audiovisual support; gum in India and one with nicotine lozenge in
follow-up telephone call; quitline support; and the USA) and one with antidepressants in Taiwan
other forms of support. Controls received brief (China). Compared with the behavioural inter-
advice as in usual care, a written manual, and/ vention received by the controls, neither nicotine
or delayed intervention. Of the nine studies, four gum nor nicotine lozenge had an effect on the
RCTs in the USA and the two cohort studies in cessation rates for use of smokeless tobacco and
India showed significant effects on the cessation areca nut products. In the third study, use of both
rates, with relative risk ranging from 1.28 (95% antidepressants showed significant effects on
CI, 1.09–1.50) to 25.70 (95% CI, 13.26–49.84) for cessation rates compared with placebo for use of
the intervention arm compared with the control areca nut products (including betel quid) without
arm. The remaining three RCTs (two in the USA tobacco (escitalopram: relative risk, 6.33; 95% CI,
and one in Sweden) did not show significant 1.53–26.14; moclobemide: relative risk, 6.17; 95%
effects; in addition, the numbers of smokeless CI, 1.48–25.64 at 2 months).
tobacco users in both the intervention group
and the control group were limited in the RCT (c) Combined pharmacological and
in Sweden. behavioural interventions
Five intervention studies assessed the effects A total of 16 RCTs were reviewed to assess
of behavioural interventions for cessation of the effects of pharmacological interventions in
use of smokeless tobacco or areca nut products combination with behavioural interventions for
in youth: four RCTs in the USA and one large cessation of use of smokeless tobacco or areca nut
cohort study in India. Interventions included products. Two studies were on nicotine gum, four
peer-led components, training by trained group on nicotine patch, four on nicotine lozenge, three
leaders or athletes, further supported with on bupropion (an antidepressant), and three on
tailored audiovisual meetings at periodic inter- varenicline (a nicotinic receptor partial agonist).
vals. Controls received either no intervention All of the studies were conducted in the USA,
or delayed intervention, general anti-tobacco except for two studies on varenicline, of which
education, or the general help of a support group. one study was in Norway and Sweden and one
One RCT in the USA showed significant effects study was in India. The study populations were
on the cessation rates in the intervention arm users of smokeless tobacco alone (in the USA and
compared with the control arm, which received in Norway and Sweden) or of smokeless tobacco
no intervention. The other four studies showed or areca nut with tobacco products (in India). All
effects that were not statistically significant; of studies performed intention-to-treat analyses or
note, three of the studies included some sort treated participants who withdrew and/or were
346
lost to follow-up as non-abstinent. Most studies product packages are noticed by users and lead to
(13 of 16) had a long follow-up (from 6 months to motivation to quit and thinking about quitting.
> 12 months). Nine of the 16 studies had at least Article 12: Education, communication,
100 participants each in the intervention and training, and public awareness on smokeless
control arms. In all studies except one, controls tobacco. Three studies in India and one study in
were provided with a combination of placebo plus Bangladesh showed that noticing anti-smokeless
behavioural therapy. Only two studies showed tobacco messages in the mass media is associated
significant positive effects on cessation rates, with intention to quit and attempts to quit.
one using varenicline (relative risk, 1.42; 95% CI, Article 13: Ban on smokeless tobacco adver-
1.08–1.79) and one using nicotine lozenge (rela- tising, promotion, and sponsorship (TAPS).
tive risk, 1.27; 95% CI, 1.10–1.47); in both studies, One study in India and one in the Sudan showed
the control arm also received the behavioural that restricting point-of-sale advertising near
intervention. Most of the other studies showed schools has an impact on the prevalence of
non-significant positive effects; non-significant smokeless tobacco use.
negative effects were found in three studies, and Article 14: Demand reduction measures
no effect was reported in one study. concerning smokeless tobacco dependence
and cessation. In one study in the USA and three
5.3.4 Policies and their impacts studies in India, quitlines reported high cessa-
tion rates in the callers.
(a) Control policies for smokeless tobacco Article 16: Access to and availability of
Implementation of the articles of the WHO smokeless tobacco to minors. In Sri Lanka,
Framework Convention on Tobacco Control to enforcement of the policy of no sales to minors
control smokeless tobacco use is at an interme- (aged < 18 years) led to a reduction in the prev-
diate stage in the 182 countries that have acceded alence of smokeless tobacco use by minors, as
to the WHO Framework Convention on Tobacco shown by two successive rounds of the Global
Control. Youth Tobacco Survey after this policy was
Articles 4 and 5: Prevention of initiation of adopted. However, in many places, adoption
smokeless tobacco use in youth. In two studies of a policy of no sales to minors has not been
in Bihar, India, school-based tobacco control successful.
policies showed positive effects in reducing the Bans on smokeless tobacco products. In
prevalence of smokeless tobacco use. three studies in India, bans have had some initial
Article 6: Price and tax measures on smoke- effect on prevalence of use. However, online sales
less tobacco. One study in the USA showed that and smuggling have been reported in Bhutan,
taxation had reduced the prevalence of smoke- India, Sri Lanka, and some European countries.
less tobacco use. Four other studies, three in One study in India showed that in view of the
India and one in Bangladesh, showed that higher gutka ban, former users of gutka were turning to
taxation would reduce the prevalence of use. A alternative, vendor-made mixtures (e.g. mawa)
large meta-analysis of studies in five countries containing similar ingredients to gutka.
showed that a 10% price increase would reduce Article 20: Research, surveillance, and
the demand for smokeless tobacco by 2.1%. exchange of information on smokeless to-
Article 11: Packaging and labelling of bacco. In Bangladesh, India, and Thailand, a
smokeless tobacco products. Three studies in reduction in smokeless tobacco use was found in
India showed that large pictorial warnings on the second round of the Global Adult Tobacco
347
Survey after several control policies were imple- 5.4 Screening and early diagnosis of
mented at the same time. oral cancer
Overall, applying multiple key tobacco con-
trol policies simultaneously has an amplifying 5.4.1 Screening methods and technologies
effect.
(a) Clinical oral examination
(b) Control policies for areca nut products Clinical oral examination is the only
(including betel quid) screening method for the detection of oral
Areca nut control policies are slowly emerging cancer and OPMDs that is routinely used. It
in countries where areca nut has traditionally consists of a visual inspection of the oral cavity
been used. Taiwan (China) is the only country and palpation of the neck to identify enlarged
with an areca nut control programme; the prev- lymph nodes or masses. The specificity of clinical
alence of betel quid use has continued to decrease oral examination ranges from 75% to 99%, based
for more than 10 years. on numerous studies conducted in Brazil, India,
Only a few countries have more than one Japan, Portugal, Sri Lanka, Taiwan (China), and
policy. Taiwan (China) has the highest number the United Kingdom. In contrast, the sensitivity
of policies (six policies), followed by Myanmar of clinical oral examination for OPMDs and oral
(four policies) and India (three policies). The cancer was more heterogeneous across studies,
most commonly adopted policy is a ban on spit- ranging from 50% to 99%. Correct risk strati-
ting in public places, as has been implemented fication of oral mucosal abnormalities detected
in Bhutan, Myanmar, Papua New Guinea, and by clinical oral examination is challenging,
Taiwan (China), as well as in India (by the rail- given the overlap in the signs and symptoms of
ways only) and in Hangzhou City (China). OPMDs and oral cancer with those of benign
The policies implemented in Taiwan (China) mucosal diseases; therefore, primary screeners
in 1997 are a ban on spitting in public places, a should be well trained. A limited number of
ban on chewing in certain workplaces and in the studies on dental care workers have assessed the
military, awareness programmes, a betel quid efficacy of training programmes. In low-resource
cessation programme, a plantation programme settings, community health-care workers can
to help areca nut growers change to other crops, be successfully trained to perform clinical oral
and an oral mucosal screening programme to examination.
monitor the effect of the policies. As a result, Mobile phone technology platforms for
the prevalence of areca nut use in adults (aged remote screening (i.e. via the transmission of
≥ 18 years) has decreased steadily, from about digital images for remote evaluation by an expert
45% in 2007 to about 5% in 2017. After having clinician) are currently being developed and
increased over several decades, the annual inci- tested.
dence rate of oral cancer has plateaued at about
42 per 100 000 people since 2009. (b) Mouth self-examination
The oral cavity is easily accessible for exam-
ination, and most OPMDs and oral cancers
are readily visible. In mouth self-examination,
individuals examine their own oral cavity to
identify OPMDs or cancerous lesions. The accu-
racy of mouth self-examination for detection of
OPMDs and early-stage cancer varies, from 8.6%
348
to 72.7% for sensitivity and from 54% to 99% for (d) Cytology and liquid biopsy
specificity; also, compliance with performing Brush biopsy cytology may be used as a prac-
mouth self-examination varies from 36% to 88%. tical, low-risk, and low-cost diagnostic adjunct
Several studies have assessed the detection rate in patients with OPMDs and oral cancer. Brush
of OPMDs. biopsy showed the highest accuracy among all
(c) Adjunctive techniques diagnostic adjuncts compared with histopatho-
logical end-points, with a sensitivity of 90% (95%
Visualization adjuncts and vital staining are CI, 82–94%) and a specificity of 94% (95% CI,
techniques that aim to improve the performance 88–97%).
of clinical oral examination to identify and/ Liquid biopsy is a minimally invasive diag-
or risk-stratify patients with OPMDs and oral nostic method that analyses biomarkers in
cancer. These adjuncts are rarely used in primary samples of circulating fluids, such as blood or
screening, and most of the evidence on their saliva (“salivaomics”). The use of saliva samples
performance comes from secondary or tertiary in the diagnosis of oral cancer and OPMDs
care settings, against the reference standard of has been extensively explored. Potential sali-
histopathology. vary biomarkers include minerals, peptides,
Visualization adjuncts include tissue auto- proteins, DNA, messenger RNA, microRNA,
fluorescence, narrow-band imaging, and tissue long coding RNA, oxidative stress-related
reflectance. Tissue autofluorescence shows the molecules, glucocorticoids, glycosylation-relat-
highest performance, with a pooled sensitivity of ed molecules, telomerase activity, and the micro-
88% (95% CI, 80–93%) and a pooled specificity biome. However, the diagnostic applications of
of 61% (95% CI, 44–75%). The low specificity salivaomics have been explored only recently.
is attributed to the preponderance of benign
lesions, which yield false-positive outcomes. (e) Emerging technologies
Consumption of smokeless tobacco or areca Artificial intelligence is used in medical
nut by the patient limits the interpretation of diagnostics, and its use has been proposed to
autofluorescence. improve the detection and diagnosis of OPMDs
Vital staining with toluidine blue or Lugol’s and to distinguish the oral mucosa at highest risk
iodine involves the topical application of a dye of malignant transformation from the normal
directly to the oral mucosa to detect or highlight mucosa.
abnormal mucosa. A recent meta-analysis of 20 Optical coherence tomography is an optical
data sets assessing the accuracy of toluidine blue diagnostic technique used for in vivo imaging of
staining reported a pooled sensitivity of 86% OPMDs. Also, in vivo microscopy can be adapted
(95% CI, 79–90%) and a pooled specificity of 68% for the same purpose to provide a cross-sec-
(95% CI, 58–77%). The potential for false-posi- tional image of the oral mucosa: techniques
tives is high because toluidine blue binds to such as reflectance microscopy and fluorescence
benign inflammatory, ulcerative, or regenerating microscopy provide the possibility of detailed
tissues. Toluidine blue staining has been tested mucosal diagnostics, sometimes with the aid of
as an adjunct to clinical oral examination in one an optical contrast agent (topical or intravenous).
screening study, showing a non-significant 21% When spectroscopy is used, contrast agents are
reduction in cancer incidence. not needed; instead, the light reflected at the
same wavelength (elastic scattering) or different
wavelengths (Raman spectroscopy) is measured.
349
Finally, molecularly targeted optical imaging follow-up visits or screening rounds).

agents have been developed to delineate the Determinants of participation can be identified
margins of oral cancer; however, it is unclear how mainly at three distinct levels: the individual,
they could be used in a screening setting. the health-care provider, and the health-care
system. Factors that are positively associated
5.4.2 Organized and opportunistic oral with screening participation and compliance
cancer screening with referral include the presence of symptoms, a
family history of cancer, higher education levels,
Worldwide, there are very few large-scale and exposure to risk factors such as tobacco
population-based organized or non-orga- smoking, alcohol consumption, and betel quid
nized oral cancer screening programmes, or chewing. In addition, training of primary
sporadic screening activities. A large-scale health-care workers for oral cancer screening
population-based annual oral cancer screening and adequate referral of patients are positively
programme has been under way in Cuba since linked to acceptance of oral cancer screening
1982, first targeting individuals aged ≥ 15 years from a health-care provider and a health-care
and later those aged ≥ 35 years. A nationwide system perspective, respectively.
population-based biennial oral cancer screening
programme has been running in Taiwan 5.4.4 Effectiveness of screening
(China) since 2004, targeting cigarette smokers
and/or betel quid chewers aged ≥ 30 years and (a) Preventive effects of screening
Indigenous people aged ≥ 18 years; more than One RCT (in India) and three observational
4.6 million people have participated in this studies – two based on the same screened cohort
screening programme. in Taiwan (China) and one in Cuba – have
In India and Sri Lanka, the governments assessed the effect of screening individuals at
have issued guidelines for oral cancer screening, high risk (defined in the studies as users of
but these have yet to be implemented systemat- tobacco, alcohol, and/or betel quid) with clinical
ically on a large scale. There has been very little oral examination on incidence of advanced oral
oral cancer screening activity in Central and cancer and on mortality from oral cancer. In
South America. Since 2001, the São Paulo State the RCT, after 15 years of follow-up, in users of
Health Secretariat has coordinated oral cancer tobacco and/or alcohol the relative risk of inci-
screening with annual clinical oral examination dence of advanced oral cancer was 0.79 (95% CI,
for the population aged ≥ 60 years in São Paulo 0.65–0.95) and the relative risk of oral cancer
State, Brazil. No population-based oral cancer mortality was 0.76 (95% CI, 0.60–0.97). The rela-
screening programmes have been reported in tive risks in the observational studies were very
Europe, North America, or Oceania. similar to those observed in the RCT, with reduc-
tions of 21–22% for incidence of advanced oral
5.4.3 Determinants of participation in cancer and 24–26% for oral cancer mortality,
screening for oral cancer and the findings were statistically significant in
the cohort studies.
Few studies have reported determinants
The studies were heterogeneous in the design
of participation in oral cancer screening, with
of the screening intervention. The observational
different outcome measurements for partici-
studies were based on evaluations of the perfor-
pation in screening (screening participation,
mance of the national screening programmes,
compliance with referral, and adherence to
which included mainly male participants. Most
350
of the studies did not identify an impact of oral

cancer screening on incidence of oral cancer
overall.
(b) Harms of screening

Screening programmes may be associ-
ated with some harms, and it is important to
consider the balance of benefits and harms for
any screening activity. The potential harms of
screening are associated with false-positive tests,
false-negative tests, overdiagnosis, and over-
treatment. Information about the harms of oral
cancer screening is lacking.
5.4.5 Risk-based model for screening

Cancer screening has historically been based
on age, without any assessment of exposure to
known risk factors. Selectively screening only
individuals at high risk may improve the effi-
ciency and effectiveness of screening while
minimizing the harms. A reanalysis of the find-
ings of the RCT in India showed that a tailored
approach based on alcohol consumption and use
of any tobacco by the individuals increased the
efficiency of oral cancer screening. A risk-based
model for screening has been considered to be an
appropriate approach for resource-limited coun-
tries with a high incidence of oral cancer.
351
6. EVALUATIONS, STATEMENTS, AND
CONSIDERATIONS
6.1 Impact of quitting exposure to non-malignant lung disease, other cancer types,
a risk factor on incidence of or and oral health problems.
Given that the joint effect of tobacco smoking
mortality from oral cancer and alcohol consumption is greater than multi-
6.1.1 Tobacco smoking plicative, quitting smoking reduces the large risk
of oral cancer in individuals who continue to
There is sufficient evidence that quitting drink alcohol; this is an indisputable additional
tobacco smoking reduces the risk of oral cancer. benefit of quitting smoking. Large reductions in
The risk decreases with increasing time since risk would also be expected after smoking cessa-
quitting smoking. tion in dual users of other agents known to be
Rationale. IARC Handbooks Volume 11, associated with oral cancer and correlated with
published in 2007, already concluded that the tobacco smoking for which greater-than-addi-
risk of oral cancer decreases with increasing tive or greater-than-multiplicative interactions
time since quitting smoking. Thus, in updating have been established (i.e. smokeless tobacco,
this evaluation, the Working Group restricted its areca nut).
review to recent studies that reported risk of oral
cancer by time since quitting smoking, adjusted 6.1.2 Alcohol consumption
for important confounders. Recent evidence
also reported a reduction in risk of oral cancer There is sufficient evidence that cessation of
within 10 years of quitting smoking; the relative alcohol consumption reduces the risk of oral
risk in former smokers reaches the relative risk cancer.
in never-smokers after ≥ 20 years of cessation. The risk decreases with increasing time since
The Working Group also reviewed the available cessation of alcohol consumption.
evidence on smoking cessation and risk of leuko- Rationale. In reaching this evaluation, the
plakia, which suggests that the risk of leukoplakia Working Group gave more weight to studies that
decreases after quitting smoking. reported risk estimates by time since cessation of
Additional considerations. Quitting smok- alcohol consumption; supporting evidence was
ing has additional benefits; it reduces the risk of provided by studies that reported risk estimates
other chronic diseases, such as vascular diseases in former drinkers or current drinkers relative
(coronary heart disease, cerebrovascular disease, to never-drinkers. Studies that adjusted for
aortic aneurysm, and peripheral arterial disease),
353
multiple potential confounders were also given world regions where use of smokeless tobacco is
more weight in the evaluation. highly prevalent.
The evidence comes mainly from one large The available studies had major limitations,
pooled analysis of data from 13 case–control including the absence of a clear period of absti-
studies, conducted in Asia, Europe, and North nence in the definition of former users of smoke-
and South America. Although the original less tobacco, sparse numerical representation,
case–control studies have limitations in terms and lack of sufficient adjustment for potential
of recall bias and selection bias and there is confounding factors.
significant heterogeneity in the pooled analysis, Data from eight studies on the association
robust methodologies were used for data harmo- between former use of smokeless tobacco and
nization and statistical analyses, and consistent risk of oral potentially malignant disorders
patterns of reduced risk after cessation of alcohol (OPMDs) were inconsistent.
consumption were observed. Additional considerations. The Working
Additional considerations. The reduction in Group noted the minimal geographical diversity
risk of oral cancer becomes more apparent after in the studies, particularly the absence of studies
10 years of cessation of alcohol consumption. from countries in South Asia, the world region
There is some evidence that the reduction in risk that has the highest prevalence of use of smoke-
of oral cancer is greater in former heavy drinkers less tobacco. The Working Group also noted the
(≥ 3 drinks per day). absence of studies for smokeless tobacco prod-
Increased reductions in risk would also be ucts other than moist snuff, except for one small
expected after cessation of alcohol consumption study on shammah use in Yemen.
in dual users of other agents known to be associ-
ated with oral cancer and correlated with alcohol 6.1.4 Chewing areca nut products (including
consumption for which greater-than-additive betel quid) with or without tobacco
or greater-than-multiplicative interactions have
been established (i.e. smoked tobacco, smokeless There is sufficient evidence that cessation of
tobacco, areca nut). use of areca nut products (including betel quid)
with or without tobacco reduces the risk of oral
cancer.
6.1.3 Smokeless tobacco use
Cessation of use of areca nut products
There is inadequate evidence that quitting (including betel quid) with or without tobacco
use of smokeless tobacco reduces the risk of also reduces the risk of OPMDs.
oral cancer. Rationale. The Working Group elected to
Rationale. In evaluating the body of evidence conduct a combined evaluation for chewing
on risk of oral cancer upon quitting exposure to areca nut products without tobacco and chewing
different risk factors, the Working Group gave the areca nut products with added tobacco, in view
most weight to cohort studies and case–control of several considerations. First, chewing behav-
studies that reported risk of oral cancer by time iours and use of areca nut products are very
since cessation. In the case of smokeless tobacco, heterogeneous between geographical regions
no studies were available based on this criterion. and subregions. Second, the available literature
The body of evidence supporting the evalu- does not enable a separate evaluation for each
ation consisted of two cohort studies and four product.
case–control studies, conducted predominantly The Working Group based the evaluation on
in Sweden and thus not providing data from other data from published studies and from primary
354
analyses, using principally evidence on time reduction in the incidence of HPV-associated

since cessation, supported by the comparison oral cancer or oropharyngeal cancer. HPV
of former users versus current users, and age vaccination has been shown to result in reduc-
at quitting. Particular attention was given to tion in the prevalence of oral HPV16 infection
adjustment for important confounders and to in vaccinated individuals and in populations
the precision of the risk estimates. Three cohort with high vaccination coverage. HPV vaccina-
studies had large sample sizes and long follow-up tion has also been shown to result in reduction
periods, which strengthened the temporal rela- in the incidence, prevalence, and persistence of
tionship between time since cessation and risk vaccine-type HPV infections, reduction in the
of oral cancer. incidence of associated precancers at the cervix,
Key observations that guided the Working vagina, vulva, penis, and anus, and reduction in
Group in making this evaluation include: the incidence of cervical cancer in vaccinated
• Three large cohort studies and two case– individuals and in populations with high vacci-
control studies consistently showed a statisti- nation coverage.
cally significant association and statistically There is a strong rationale and analogy, based
significant trend of reduced risk of oral on observations at other anatomical sites, that
cancer with increasing time since cessation of HPV vaccination would result in reduction in
chewing areca nut products without tobacco. the incidence of HPV-associated oral cancer and
• For cessation of chewing areca nut products oropharyngeal cancer in vaccinated individuals
with added tobacco, although the evidence and in the populations at large, depending on
was inconsistent across published studies, vaccination coverage.
one large cohort study showed reduced risk of
oral cancer with increasing time since cessa-
tion in former chewers.
6.2 Interventions for cessation of
• Risk reductions were also observed for smokeless tobacco or areca nut
OPMDs with increasing time since cessa- use
tion of chewing areca nut products without
tobacco, and for leukoplakia after cessation 6.2.1 Behavioural interventions in adults
of chewing areca nut products with added There is sufficient evidence that behavioural
tobacco. interventions in adults are effective in inducing
Additional considerations. Cessation of quitting use of smokeless tobacco.
chewing areca nut products with or without Rationale. Nine studies (seven randomized
tobacco would be broadly beneficial for a reduced controlled trials and two cohort studies),
global burden of oral cancer. In addition to oral including several high-quality studies, were
cancer and OPMDs, quitting chewing could available for evaluation. A positive effect of the
prevent other cancer types (e.g. cancers of the intervention on the quit rates was observed
pharynx and of the oesophagus) and other consistently in the body of evidence, and chance,
chronic diseases. bias, and confounding as causes of this associa-
tion were ruled out with reasonable confidence.
6.1.5 HPV16 infection Despite some limitations, all the studies showed
a positive association, and six of the studies
There are no studies to date on vaccination-re-
showed statistically significant effects.
lated reductions in oral infection with human
papillomavirus type 16 (HPV16) resulting in
355
6.2.2 Behavioural interventions in youth of some dual users (tobacco smoking and smoke-
less tobacco use).
There is limited evidence that behavioural Two studies showed an effect of the inter-
interventions in young people are effective in vention in inducing quitting; one was statisti-
inducing quitting use of smokeless tobacco. cally significant, and one was non-significant.
Rationale. Five studies (four randomized The third study showed no effect. However, in
controlled trials and one cohort study) were the latter two studies, the control groups were
available for evaluation. The body of evidence provided with behavioural intervention instead
provided apparently inconsistent results: one of placebo, thus reducing the potential effect size.
study showed significant effects, three studies
showed non-significant positive effects, and one
study showed a non-significant negative effect.
6.2.4 Combined pharmacological and
However, this could be explained as follows: behavioural interventions
• In the study that showed significant effects There is limited evidence that combined
on the quit rates, the control arm had no pharmacological and behavioural interven-
intervention. tions are effective in inducing quitting use of
• Three of the four studies that showed non-sig- smokeless tobacco.
nificant effects, two positive and one negative, Rationale. A large number of randomized
had some form of intervention in the control controlled trials (16) were available for evaluation.
arm, thus pulling the estimates towards the A positive effect of the intervention on the quit
null. rates was observed in some studies, but chance,
In addition, one study reported a significant bias, or confounding could not be ruled out with
positive effect of the intervention on the preven- reasonable confidence, for several reasons:
tion of initiation of smokeless tobacco use. • A positive effect of the intervention was
reported in most studies (13 of 16). However,
6.2.3 Pharmacological interventions in most of these studies (11 of 13), the associ-
ation was not statistically significant.
There is limited evidence that pharmacolog-
• Eight of the studies had large study popula-
ical interventions with nicotine replacement
tions (≥ 100 participants in each arm) and
therapy or antidepressants (escitalopram and
long follow-up periods. However, only two of
moclobemide) are effective in inducing quit-
these eight studies showed significant effects.
ting use of smokeless tobacco and areca nut
• Most studies had the same behavioural inter-
with tobacco.
vention in the control arm, thus pulling the
Rationale. Three randomized controlled
estimates towards the null.
trials were available for evaluation. Two studies
assessed the effectiveness of nicotine replacement
therapy (one with gum and one with lozenges),
and one study assessed the effectiveness of anti-
depressants (escitalopram and moclobemide).
All three studies followed a good methodology,
had adequate controls, and used proper outcome
measurements. However, the studies had several
limitations, including short follow-up periods
(< 12 months) and confounding by the presence
356
6.2.5 Policies • The impact of oral screening on oral cancer

mortality in the general population cannot
Few data are available on the effect of the be established on the basis of the current
individual World Health Organization Frame- evidence.
work Convention on Tobacco Control policies on • The limited number of studies of different
smokeless tobacco control. The strongest effect, designs (one randomized controlled trial,
despite limited evidence, was shown for taxation two cohort studies, and one case–control
in reducing the prevalence of smokeless tobacco study) in a few settings restricts generaliza-
use. One study in India showed a positive effect of tion of the outcomes.
school-based tobacco control policies. A combi- • The limitations of the included studies were
nation of policies was shown to be more effective likely to pull the effect of screening towards
than a single policy. the null:
There is a shortage of data with regard to the
effect of control policies for areca nut products, Ŝ In the randomized controlled trial, there
because these policies are new and have been was low compliance of screen-positive
implemented recently. The limited but positive cases with further assessment.
results from Taiwan (China) suggest that adop- Ŝ In the cohort studies, there was selection
tion of a comprehensive set of policies to control bias for screening, and possible contami-
areca nut use may lead to reductions in the prev- nation of the control group.
alence of areca nut use. Ŝ In the case–control study, there was lack
of power, possible overestimation of expo-
sure to the intervention (defined as “any
6.3 Screening for oral cancer and visit to a community dentist”), and a low
OPMDs coverage of the programme.
6.3.1 Effectiveness of screening by clinical • The included studies did not report whether
there were any primary prevention interven-
oral examination
tions within the studied population, which
Screening of individuals at high risk by could have an impact on the estimates.
clinical oral examination may reduce mortality • The included studies had other limitations
from cancer of the oral cavity (Group B). with an unclear effect on outcome:
Individuals at high risk are defined as those Ŝ The randomized controlled trial used a
with tobacco use, areca nut use, alcohol consump- small number of randomized units.
tion, or a combination of these, in any form.
Ŝ In the retrospective cohort study, the
Rationale. In reaching this evaluation, the
proportion of individuals at high risk in
Working Group noted the following:
the control group was unclear, possibly
• The randomized controlled trials and cohort
leading to information bias.
studies showed a statistically significant
Regimen to which the evaluation applies.
positive effect of oral screening on the inci-
The screening interval used in the included
dence of advanced oral cancer and on oral
studies was either 2 years or 3 years. The optimal
cancer mortality in individuals at high risk
age range could not be established.
(based on tobacco use, areca nut use, alcohol
consumption, or a combination of these, in
any form).
357
6.3.2 Additional considerations (c) Risk-based model for screening

(a) Adjunctive techniques to oral examination Assessment of risk, for example by question-
naire, has the potential to increase programme
Very few studies have evaluated adjunctive
efficiency and reduce the harms of overscreening,
techniques in population screening studies. Most
overdiagnosis, and overtreatment. However,
of the adjunctive techniques have been evalu-
implementation of screening programmes using
ated as diagnostic adjuncts in either prospec-
risk-based models for selection of participants is
tive accuracy studies or retrospective cohort or
a challenge from a programmatic perspective.
case–control studies. All of the available studies
report accuracy measures of test results against (d) Monitoring and evaluation of screening
histopathology as the reference standard. Given programmes
the unknown natural history of OPMDs in an
Assessment of determinants of partici-pation
individual patient, it is challenging to extrap-
at all steps of the screening pathway has been
olate accuracy data to important end-points
demonstrated to be critical for the optimization of
such as mortality or survival. The added value
cancer screening at other sites (e.g. cervix, breast,
of adjunctive techniques to clinical oral exam-
and colon). The existing oral cancer screening
ination remains unknown. There is a potential
programmes lack proper monitoring and eval-
for using adjunctive techniques and biomarkers
uation mechanisms, preventing evidence-based
in saliva for the diagnosis of OPMDs and oral
evaluation of their efficacy and health impact.
cancer. However, there is a lack of clinical vali-
It remains unclear whether the known risk
dation linked to important end-points as a stand-
factors for oral cancer, as well as age and sex, are
alone method in oral cancer screening settings.
positive or negative determinants of participa-
(b) Harms of screening tion in oral cancer screening. Identifying and
describing the predictors of participation in oral
A clear understanding of the harms linked
cancer screening, provider training, compliance
to false-positive screening test results and, more
with referral, the quality of available data, and
importantly, false-positive diagnostic findings
the interventions to improve these is crit-
leading to potential overtreatment is hampered
ical to increase the effectiveness of oral cancer
by a poor understanding of the natural history
screening programmes. Filling this gap may
of OPMDs. There is currently little evidence that
enable policy-makers and stakeholders to effi-
adjunctive techniques can reduce the propor-
ciently allocate human and financial resources
tion of false-positive results when screening by
to obtain higher benefits and reduce inequalities.
clinical oral examination. Adjunctive techniques
The screening trials have not provided a clear
or biomarkers that are predictive for cancer
understanding of the natural history of OPMDs.
progression in OPMDs are being investigated.
The impact of detection, treatment, and surveil-
Quality assurance of programme implementa-
lance of patients with OPMDs on oral cancer
tion is important to improve the performance
incidence and mortality has not been deter-
of screening programmes and reduce the harms
mined. Among the studies that assessed cancer
of screening. This issue has not been addressed
incidence, most did not observe an impact of oral
in the primary studies reviewed by the Working
cancer screening on oral cancer incidence.
Group.
The Working Group considers primary
prevention to be an integral part of a screening
programme.
358
A Working Group of 25 independent international experts, convened by the Interna
tional Agency for Research on Cancer (IARC) between September and December
2021, reviewed the scientific evidence on primary and secondary prevention of oral
cancer. Cancer of the lip and oral cavity ranks 16th in cancer incidence and mortality
worldwide and is a common cause of cancer death in men across much of South and
South-East Asia and the Western Pacific. The main causes of oral cancer worldwide
are tobacco smoking and alcohol consumption. Smokeless tobacco use and areca
nut use are the leading causes in those countries where their use is prevalent.
This is the first evaluation of oral cancer prevention by the IARC Handbooks
programme. The Working Group reviewed the body of evidence and provided
evidence-based evaluations of the impact of cessation of exposure to risk factors, of
cessation interventions for products containing smokeless tobacco or areca nut, and
of screening for oral cancer by clinical oral examination. In addition, this publication
presents background information related to oral cancer, such as the global burden of
oral cancer, the prevalence of use of products containing smokeless tobacco or areca
nut, and emerging technologies for oral cancer screening.
This volume of the IARC Handbooks brings together, for the first time, all the available
evidence related to primary and secondary prevention of oral cancer.
© University of Campinas, São Paulo

Handbook19 Book Final

Uploaded by

Copyright:

Available Formats

Handbook19 Book Final

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Handbook19 Book Final

Uploaded by

Copyright:

Available Formats

IARC HANDBOOKS

This publication represents the views and

LYON, FRANCE - 2023

How to cite this publication:

IARC Library Cataloging-in-Publication Data

NOTE TO THE READER. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

PREAMBLE – PRIMARY PREVENTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

PREAMBLE − SECONDARY PREVENTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

4. Current global screening practices. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

1. ORAL CANCER AND ORAL POTENTIALLY MALIGNANT DISORDERS. . . . . . . . . . . . . . . . . . . . . . 77

2. REDUCING INCIDENCE OF CANCER OR PRECANCER. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105

2.4 Preventive dietary agents. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191

3. CESSATION OF SMOKELESS TOBACCO AND/OR ARECA NUT USE. . . . . . . . . . . . . . . . . . . . 209

4. SCREENING AND EARLY DIAGNOSIS OF ORAL CANCER. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295

6. EVALUATIONS, STATEMENTS, AND CONSIDERATIONS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353

Devasena Anantharaman Anil Chaturvedi 2 (Subgroup Co-Chair,

David I. Conway Ravi Mehrotra (Meeting Co-Chair)

Luiz P. Kowalski Felipe Roitberg

Alan Roger Santos-Silva Invited Specialists

Patravoot Vatanasapt Frédéric de Bels

Observers 5 Rok Ho Kim, WHO headquarters

Véronique Bouvard, Evidence Synthesis and Administrative Assistance

A. GENERAL PRINCIPLES AND developing cancer. Secondary prevention entails

1997) reviewed the evidence on cancer-preven- 2. Objectives, scope, and

change in the intermediate outcome reduces have a significant impact on an interme-

Table 1 Roles of participants at IARC Handbooks meetings

Category of participant Role

by decades of experience documenting that Specialists are invited in limited numbers,

Approximate time frame Milestones

1. Intervention and outcome experimental and observational epidemiological

1.2 Outcome characterization In what follows, the term “cancer incidence”

population must have been identified in a confounding variables (including co-ex-

Results of studies that are judged to be of high to be modifiers of the intervention–outcome

(d) Overall evaluation intervention as commonly implemented in prac-

(mechanistic studies in humans) or sufficient evidence in experimental animals.

A. GENERAL PRINCIPLES AND developing cancer. Secondary prevention entails

1997) reviewed the evidence on cancer-preven- 2. Objectives, scope, and

2.2 Definition of interventions for (ii) subgroups with particular predisposing

of a randomized controlled trial”. Effectiveness potential harms is an important component of

In this framework, it is most commonly • The intervention is of putative preventive

Table 1 Roles of participants at IARC Handbooks meetings

Category of participant Role

Approximate time frame Milestones

• participate in review and discussion of B. SCIENTIFIC REVIEW AND

curable. Screening (organized or opportunistic) 3. Screening methods

and cross-contamination of the non-inter- including adjusting for matching when

(d) Evaluation of new technologies (e) Considerations in assessing the body of

the IARC Secretariat may commission or conduct 5.4 Cost–effectiveness

7.1 Evaluation A causal preventive association between use

In April 2016, an international group of sci- History of use of smokeless tobacco

References nut. Lancet Oncol. 18(12):e767–75. doi:10.1016/S1470-

ADH alcohol dehydrogenase

LOH loss of heterozygosity

1.1 Anatomy of the oral cavity and (a) Lips

Fig. 1.1 Anatomy of the oral cavity

(c) Gingivae (gums) (e) Hard palate

Fig. 1.2 Anatomy of the pharynx

From Ferlay et al. (2020).

From Ferlay et al. (2020).

From Ferlay et al. (2020).