• Share
  • Email
  • Embed
  • Like
  • Save
  • Private Content
IARC Monographs on the Evaluation of Carcinogenic Risks to Humans

IARC Monographs on the Evaluation of Carcinogenic Risks to Humans



"Riesgo cancerígeno" esta expresión de la serie Monografías de la IARC se entiende que un agente que es capaz de causar cáncer. EstasMonografías evaluan los riesgos de cáncer, a pesar de la ...

"Riesgo cancerígeno" esta expresión de la serie Monografías de la IARC se entiende que un agente que es capaz de causar cáncer. EstasMonografías evaluan los riesgos de cáncer, a pesar de la presencia histórica de los «riesgos» que figuran en el título.
La inclusión de un agente en las monografías no implica que se trata de un carcinógeno, sólo que los datos publicados han sido examinados. Igualmente, el hecho de que un agente aún no ha sido evaluado en una
Monografía no significa que no es cancerígeno. Del mismo modo, la identificación de los tipos de cáncer con pruebas suficientes o evidencia limitada en humanos no debe considerarse como excluyente de la posibilidad de que un agente puede causar cáncer en otros sitios.
Las evaluaciones de riesgo de cáncer son realizados por grupos de trabajo internacionales de científicos independientes y no son de naturaleza cualitativa. Ninguna recomendación se da para la regulación o legislación.
Cualquier persona que es consciente de los datos publicados que pueden alterar la evaluación del riesgo cancerígeno de un agente para el ser humano se le anima a hacer esta información disponible a la Sección de Monografías del IARC, Agencia Internacional para la Investigación del Cáncer, 150 cours Albert Thomas, 69372 Lyon Cedex 08 de Francia, con el fin de que el agente puede ser considerado para la re-evaluación de un futuro grupo de trabajo.
Aunque no se escatiman esfuerzos para preparar las monografías con la mayor precisión posible, los errores pueden ocurrir. Los lectores deben comunicar los errores a la Sección de Monografías del IARC, por lo que las correcciones pueden ser reportados en los volúmenes futuros.



Total Views
Views on SlideShare
Embed Views



0 Embeds 0

No embeds


Upload Details

Uploaded via as Adobe PDF

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
Post Comment
Edit your comment

    IARC Monographs on the Evaluation of Carcinogenic Risks to Humans IARC Monographs on the Evaluation of Carcinogenic Risks to Humans Document Transcript

    • WORLD HEALTH ORGANIZATION INTERNATIONAL AGENCY FOR RESEARCH ON CANCERIARC Monographs on the Evaluation of Carcinogenic Risks to Humans VOLUME 100 A Review of Human Carcinogens Part D: Radiation LYON, FRANCE
    • WORLD HEALTH ORGANIZATION INTERNATIONAL AGENCY FOR RESEARCH ON CANCERIARC Monographs on the Evaluation of Carcinogenic Risks to Humans Volume 100 A Review of Human Carcinogens Part D: Radiation This publication represents the views and expert opinions of an IARC Working Group on the Evaluation of Carcinogenic Risks to Humans, which met in Lyon, 2–9 June 2009
    • IARC MONOGRAPHS In 1969, the International Agency for Research on Cancer (IARC) initiated a programme on the evaluation of the carcinogenicrisk of chemicals to humans involving the production of critically evaluated monographs on individual chemicals. The programmewas subsequently expanded to include evaluations of carcinogenic risks associated with exposures to complex mixtures, lifestylefactors and biological and physical agents, as well as those in specific occupations. The objective of the programme is to elaborateand publish in the form of monographs critical reviews of data on carcinogenicity for agents to which humans are known to beexposed and on specific exposure situa­ ions; to evaluate these data in terms of human risk with the help of international working tgroups of experts in chemical carcinogenesis and related fields; and to indicate where additional research efforts are needed. Thelists of IARC evaluations are regularly updated and are available on the Internet at http://monographs.iarc.fr/. This programme has been supported since 1982 by Cooperative Agreement U01 CA33193 with the United States NationalCancer Institute, Department of Health and Human Services. Additional support has been provided since 1986 by the Health,Safety and Hygiene at Work Unit of the European Commission Directorate-General for Employment, Social Affairs and EqualOpportunities, and since 1992 by the United States National Institute of Environmental Health Sciences, Department of Healthand Human Services. The contents of this volume are solely the responsibility of the Working Group and do not necessarilyrepresent the official views of the U.S. National Cancer Institute, the U.S. National Institute of Environmental Health Sciences,the U.S. Department of Health and Human Services, or the European Commission Directorate-General for Employment, SocialAffairs and Equal Opportunities. This volume was made possible, in part, through Cooperative Agreement CR 834012 with the United States EnvironmentalProtection Agency, Office of Research and Development. The contents of this volume do not necessarily reflect the views orpolicies of the U.S. Environmental Protection Agency. Published by the International Agency for Research on Cancer, 150 cours Albert Thomas, 69372 Lyon Cedex 08, France © International Agency for Research on Cancer, 2012 Distributed by WHO Press, World Health Organization, 20 Avenue Appia, 1211 Geneva 27, Switzerland (tel.: +41 22 791 3264; fax: +41 22 791 4857; e-mail: bookorders@who.int). Publications of the World Health Organization enjoy copyright protection in accordance with the provisions of Protocol 2 of the Universal Copyright Convention. All rights reserved. The International Agency for Research on Cancer welcomes requests for permission to reproduce or translate its publications,in part or in full. Requests for permission to reproduce or translate IARC publications – whether for sale or for noncommercialdistribution – should be addressed to WHO Press, at the above address (fax: +41 22 791 4806; email: permissions@who.int). The designations employed and the presentation of the material in this publication do not imply the expression of anyopinion 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 orrecommended by the World Health Organization in preference to others of a similar nature that are not mentioned. Errors andomissions excepted, the names of proprietary products are distinguished by initial capital letters. The IARC Monographs Working Group alone is responsible for the views expressed in this publication.IARC Library Cataloguing in Publication Data A review of human carcinogens. Part D: Radiation / IARC Working Group on the Evaluation of Carcinogenic Risks to Humans (2009: Lyon, France) (IARC monographs on the evaluation of carcinogenic risks to humans ; v. 100D) 1. Carcinogens 2. Neoplasms – etiology 3. Radiation – adverse effects I. IARC Working Group on the Evaluation of Carcinogenic Risks to Humans II. Series ISBN 978 92 832 1321 5 (NLM Classification: W1) ISSN 1017-1606 PRINTED IN FRANCE
    • CONTENTSNOTE TO THE READER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1List of Participants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3PREAMBLE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 A. GENERAL PRINCIPLES AND PROCEDURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 1. Background. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2. Objective and scope. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8 3. Selection of agents for review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 4. Data for the Monographs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 5. Meeting participants. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 6. Working procedures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 B. SCIENTIFIC REVIEW AND EVALUATION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 1. Exposure data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 2. Studies of cancer in humans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 3. Studies of cancer in experimental animals. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 4. Mechanistic and other relevant data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 5. Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 6. Evaluation and rationale. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29General Remarks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31SOLAR AND UV RADIATION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 1. Exposure Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 1.1 Nomenclature and units. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 1.2 Methods for measuring UVR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 1.3 Sources and exposure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 2. Cancer in Humans. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 2.1 Natural sunlight. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 2.2 Artificial UV radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 2.3 UVA, UVB, and UVC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 2.4 Synthesis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 3. Cancer in Experimental Animals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 3.1 Non-melanoma skin cancer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 3.2 Melanoma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 v
    • IARC MONOGRAPHS - 100D 3.3 Synthesis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 4. Other Relevant Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 4.1 Transmission and absorption in biological tissues. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 4.2 Genetic and related effects: consequences of UVR exposure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 4.3 Genetic susceptibility: host factors modulating the response to UV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 4.4 Other effects. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 4.5 Synthesis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 5. Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90X- AND γ-RADIATION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 1. Exposure Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 1.1 Physical properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 1.2 Interactions with matter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 1.3 Exposure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 2. Cancer in Humans. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 2.1 Detonation of atomic bombs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 2.2 Fallout from nuclear weapons testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 2.3 Medical exposures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 2.4 Occupational studies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 2.5 Environmental studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 2.6 Synthesis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 3. Cancer in Experimental Animals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 3.1 Previous evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 3.2 Studies published since the previous IARC Monograph . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 3.3 Studies in adult animals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 3.4 Prenatal exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 3.5 Neonatal exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 3.6 Parental exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180 3.7 Synthesis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180 4. Other Relevant Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181 4.1 Radionuclides: determining the distribution of dose. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181 4.2 Mechanisms of carcinogenesis induced by all ionizing radiation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194 4.3 Mechanism of carcinogenesis of neutrons: an example of ionizing radiation. . . . . . . . . . . . . . . . . .207 4.4 Synthesis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209 5. Evaluation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210 References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210NEUTRON RADIATION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231 1. Exposure Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231 2. Cancer in Humans. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231 3. Cancer in Experimenal Animals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232 3.1 Previous IARC Monograph . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232 3.2 Carcinogenicity in adult animals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233 3.3 Prenatal exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237 3.4 Synthesis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237vi
    • Contents 4. Other Relevant Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237 5. Evaluation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237 References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237INTERNALIZED α-PARTICLE EMITTING RADIONUCLIDES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241 1. Exposure Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241 2. Cancer in Humans. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241 2.1 Radon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241 2.2 α-Particle emitters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250 3. Cancer in Experimental Animals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264 3.1 Previous IARC Monograph . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264 3.2 Studies published since the previous IARC Monograph . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265 3.3 Other studies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274 3.4 Synthesis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274 4. Other Relevant Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274 5. Evaluation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274 References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275INTERNALIZED β-PARTICLE EMITTING RADIONUCLIDES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285 1. Exposure Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285 2. Cancer in Humans. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285 2.1 Pure β-particle emitters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285 2.2 Mixed β-particle emitters–radioiodines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289 2.3 Synthesis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293 3. Cancer in Experimental Animals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294 3.1 Previous evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294 3.2 Pure β-particle-emitting radionuclides. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295 3.3 Mixed β-particle emitting radionuclides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296 3.4 Synthesis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297 4. Other Relevant Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297 5. Evaluation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297 References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298List of Abbreviations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305CUMULATIVE CROSS INDEX TO IARC MONOGRAPHS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309List of IARC Monographs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343 vii
    • NOTE TO THE READER The term ‘carcinogenic risk’ in the IARC Monographs series is taken to mean that an agent iscapable of causing cancer. The Monographs evaluate cancer hazards, despite the historical presenceof the word ‘risks’ in the title. Inclusion of an agent in the Monographs does not imply that it is a carcinogen, only that thepublished data have been examined. Equally, the fact that an agent has not yet been evaluated in aMonograph does not mean that it is not carcinogenic. Similarly, identification of cancer sites withsufficient evidence or limited evidence in humans should not be viewed as precluding the possibilitythat an agent may cause cancer at other sites. The evaluations of carcinogenic risk are made by international working groups of independentscientists 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 the carcinogenic riskof an agent to humans is encouraged to make this information available to the Section of IARCMonographs, International Agency for Research on Cancer, 150 cours Albert Thomas, 69372 LyonCedex 08, France, in order that the agent may be considered for re-evaluation by a future WorkingGroup. Although every effort is made to prepare the monographs as accurately as possible, mistakes mayoccur. Readers are requested to communicate any errors to the Section of IARC Monographs, so thatcorrections can be reported in future volumes. 1
    • List of ParticipantsMembers1Bruce Armstrong David J. Brenner (unable to attend) School of Public Health Center for Radiological Research University of Sydney Columbia University NSW 2006 New York, NY 10043 Australia USAKeith Baverstock Elisabeth CardisFaculty of Natural and Environmental Center for Research in Environmental Sciences Epidemiology (CREAL)University of Eastern Finland E-08003 BarcelonaFI-70211 Kuopio SpainFinland1 Working Group Members and Invited Specialists serve in their individual capacities as scientists and not as representa-tives of their government or any organization with which they are affiliated. Affiliations are provided for identificationpurposes only. Invited specialists are marked by an asterisk.Each participant was asked to disclose pertinent research, employment, and financial interests. Current financialinterests and research and employment interests during the past 3 years or anticipated in the future are identified here.Minor pertinent interests are not listed and include stock valued at no more than US$10 000 overall, grants that provideno more than 5% of the research budget of the expert’s organization and that do not support the expert’s research orposition, and consulting or speaking on matters not before a court or government agency that does not exceed 2% oftotal professional time or compensation. All grants that support the expert’s research or position and all consulting orspeaking on behalf of an interested party on matters before a court or government agency are listed as significant perti-nent interests. 3
    • IARC MONOGRAPHS – 100DAdele Green2 David Hoel4 Cancer & Population Studies Group College of Medicine Queensland Institute of Medical Research Medical University of South Carolina Queensland 4029 Charleston, SC 29401 Australia USARaymond A. Guilmette Daniel Krewski Lovelace Respiratory Research Institute Albuquerque, NM 87108-5127 McLaughlin Centre for Population USA Health Risk Assessment University of Ottawa Ottawa, Ontario K1N 6N5Janet Hall3 Canada Institut Curie Research Center INSERM Unit 612 University Centre Mark P. Little5 91405 Orsay Division of Epidemiology, Public Health France and Primary Care Imperial College Faculty of Medicine London W2 1PGMark A. Hill United KingdomGray Institute for Radiation Oncology & BiologyUniversity of OxfordOxford OX3 7DQUnited Kingdom2 Dr Green receives research funds (not exceeding 5% of total research support) from L’Oréal which makes productsintended to reduce the dose from solar radiation.3 Dr Hall’s research unit receives funds (not exceeding 5% of total research support) from Electricité de France, an elec-tric power company.4 Dr Hoel is providing assistance to Exxon Corp in court cases involving personal injury claimed to be related to radia-tion. He owns stock in Duke Energy Corp, an electric power company. His university salary is supported in part bygrants from the U.S. National Aeronautics and Space Administration (NASA) and the U.S. Department of Energy.5 Dr Little wrote software for the British Nuclear Group to calculate risks for workers in the nuclear industry. Thisrepresented less than 5% of his annual total professional compensation for 2006 and 2007, when the activity ceased. Healso advised the International Epidemiology Institute (USA) and Westlakes Research Institute (UK) on epidemiologicalmatters. This represented less than 5% of total professional income, and work ceased in 2007 for both contracts.New address: Radiation Epidemiology Branch, National Cancer Institute, Rockville MD, USA4
    • ParticipantsMichael Marshall (retired)6 David B. Richardson10 UK Atomic Energy Authority School of Public Health Blewbury, Didcot University of North Carolina at Chapel Hill Oxon OX11 9NW Chapel Hill, NC 27599-7435 United Kingdom USARonald E. J. Mitchel (retired)7 Anthony E. Riddell11 Atomic Energy of Canada Limited Westlakes Scientific Consulting Ltd Chalk River, Ontario K0J 1J0 University of Central Lancashire Canada Cumbria CA24 3JY United KingdomColin R. Muirhead8 Laure Sabatier Centre for Radiation, Chemical and Environmental Hazards Radiobiology and Oncology Unit Health Protection Agency French Alternative Energies and Atomic Chilton, Didcot Energy Commission Oxon OX11 0RQ 92265 Fontenay-aux-Roses United Kingdom FranceNicholas D. Priest9 Mikhail E. Sokolnikov Radiation Biology and Health Physics Southern Urals Biophysics Institute Atomic Energy of Canada Limited Ozyorsk, 456780 Chalk River, Ontario K0J 1P0 Russian Federation Canada6 Dr Marshall is retired from the United Kingdom Atomic Energy Authority (UKAEA).7 Dr Mitchel is retired from, and continues to consult for, Atomic Energy of Canada Ltd, a Crown corporation of Canadawhose mandate is to sustain and enhance nuclear technology, to manage nuclear wastes, and to maximize return oninvestment in nuclear technology. The corporation also produces more than half of the world’s medical isotopes.8 Dr Muirhead manages a section at the Health Protection Agency that receives partial funding from the UK Ministry ofDefence to maintain an epidemiological database of nuclear test veterans.New address: Institute of Health and Society, Newcastle University, UK.9 Dr Priest is a manager at Atomic Energy of Canada Ltd, a Crown corporation of Canada whose mandate is to sustainand enhance nuclear technology, to manage nuclear wastes, and to maximize return on investment in nuclear technol-ogy. The corporation also produces a significant fraction of the world’s medical isotopes.10 Dr Richardson provided written testimony on behalf of four persons seeking compensation for diseases claimed to berelated to X-rays. He reports receiving no compensation for this case.11 Dr Riddell is employed by Westlakes Scientific Consulting Ltd, a consulting firm specializing in the nuclear industry. 5
    • IARC MONOGRAPHS – 100DLadislav Tomasek Andrei Karotki National Radiation Protection Institute Ausra Kesminiene 140 00 Prague 4 Béatrice Lauby-Secretan (Rapporteur, Czech Republic Cancer in Humans) Ferid Shannoun (WHO geneva) Kurt Straif (Rapporteur, Cancer in Humans) Isabelle Thierry-ChefRobert L. Ullrich12 UTMB Cancer Center University of Texas Medical Branch Post-meeting Scientific Assistance Galveston, TX 77555-1048 USA Farhad Islami Administrative AssistanceIARC Secretariat Sandrine Egraz Philippe Autier Michel Javin Robert Baan (Co-Responsible Officer; Brigitte Kajo Rapporteur, Mechanistic and Other Relevant Helene Lorenzen-Augros Data) Karine Racinoux Lamia Benbrahim-Tallaa (Rapporteur, Cancer in Experimental Animals) Véronique Bouvard (Rapporteur, Reproduction of Graphics Mechanistic and Other Relevant Data) Rafael Carel (Visiting Scientist) Arthur Bouvard Vincent Cogliano (Head of Programme) Emilie van Deventer (WHO geneva) Jean-François Doré (Visiting Scientist) Fatiha El Ghissassi (Responsible Production Team Officer; Rapporteur, Mechanistic and Other Relevant Data) Elisabeth Elbers Crystal Freeman (Rapporteur, Cancer in Anne-Sophie Hameau Humans) Sylvia Moutinho Laurent Galichet (Editor) Dorothy Russell Yann Grosse (Rapporteur, Cancer in Experimental Animals) Neela Guha (Rapporteur, Cancer in Humans)12 Dr Ullrich provided assistance to Raytheon Co in a court case involving thyroid and kidney cancer claimed to berelated to X-rays.6
    • PREAMBLEThe Preamble to the IARC Monographs describes the objective and scope of the programme,the scientific principles and procedures used in developing a Monograph, the types ofevidence considered and the scientific criteria that guide the evaluations. The Preambleshould be consulted when reading a Monograph or list of evaluations.A. GENERAL PRINCIPLES AND risk of chemicals to man, which became the ini- PROCEDURES tial title of the series. In the succeeding years, the scope of the pro- gramme broadened as Monographs were devel-1. Background oped for groups of related chemicals, complex Soon after IARC was established in 1965, it mixtures, occupational exposures, physical andreceived frequent requests for advice on the car- biological agents and lifestyle factors. In 1988,cinogenic risk of chemicals, including requests the phrase ‘of chemicals’ was dropped fromfor lists of known and suspected human carcino- the title, which assumed its present form, IARCgens. It was clear that it would not be a simple Monographs on the Evaluation of Carcinogenictask to summarize adequately the complexity of Risks to Humans.the information that was available, and IARC Through the Monographs programme, IARCbegan to consider means of obtaining interna- seeks to identify the causes of human cancer. Thistional expert opinion on this topic. In 1970, the is the first step in cancer prevention, which isIARC Advisory Committee on Environmental needed as much today as when IARC was estab-Carcinogenesis recommended ‘...that a com- lished. The global burden of cancer is high andpendium on carcinogenic chemicals be pre- continues to increase: the annual number of newpared by experts. The biological activity and cases was estimated at 10.1 million in 2000 andevaluation of practical importance to public is expected to reach 15 million by 2020 (Stewarthealth should be referenced and documented.’ & Kleihues, 2003). With current trends in demo-The IARC Governing Council adopted a resolu- graphics and exposure, the cancer burden hastion concerning the role of IARC in providing been shifting from high-resource countries togovernment authorities with expert, independ- low- and medium-resource countries. As a resultent, scientific opinion on environmental carcino- of Monographs evaluations, national health agen-genesis. As one means to that end, the Governing cies have been able, on scientific grounds, to takeCouncil recommended that IARC should prepare measures to reduce human exposure to carcino-monographs on the evaluation of carcinogenic gens in the workplace and in the environment. 7
    • IARC MONOGRAPHS – 100D The criteria established in 1971 to evaluate causation of, and susceptibility to, malignantcarcinogenic risks to humans were adopted by the disease become more fully understood.Working Groups whose deliberations resulted in A cancer ‘hazard’ is an agent that is capablethe first 16 volumes of the Monographs series. of causing cancer under some circumstances,Those criteria were subsequently updated by fur- while a cancer ‘risk’ is an estimate of the carci-ther ad hoc Advisory Groups (IARC, 1977, 1978, nogenic effects expected from exposure to a can-1979, 1982, 1983, 1987, 1988, 1991; Vainio et al., cer hazard. The Monographs are an exercise in1992; IARC, 2005, 2006). evaluating cancer hazards, despite the historical The Preamble is primarily a statement of sci- presence of the word ‘risks’ in the title. The dis-entific principles, rather than a specification of tinction between hazard and risk is important,working procedures. The procedures through and the Monographs identify cancer hazardswhich a Working Group implements these prin- even when risks are very low at current exposureciples are not specified in detail. They usually levels, because new uses or unforeseen exposuresinvolve operations that have been established could engender risks that are significantly higher.as being effective during previous Monograph In the Monographs, an agent is termed ‘car-meetings but remain, predominantly, the pre- cinogenic’ if it is capable of increasing the inci-rogative of each individual Working Group. dence of malignant neoplasms, reducing their latency, or increasing their severity or multiplic- ity. The induction of benign neoplasms may in2. Objective and scope some circumstances (see Part B, Section 3a) con- The objective of the programme is to pre- tribute to the judgement that the agent is carci-pare, with the help of international Working nogenic. The terms ‘neoplasm’ and ‘tumour’ areGroups of experts, and to publish in the form of used interchangeably.Monographs, critical reviews and evaluations of The Preamble continues the previous usageevidence on the carcinogenicity of a wide range of the phrase ‘strength of evidence’ as a matterof human exposures. The Monographs repre- of historical continuity, although it should besent the first step in carcinogen risk assessment, understood that Monographs evaluations con-which involves examination of all relevant infor- sider studies that support a finding of a cancermation to assess the strength of the available evi- hazard as well as studies that do not.dence that an agent could alter the age-specific Some epidemiological and experimentalincidence of cancer in humans. The Monographs studies indicate that different agents may act atmay also indicate where additional research different stages in the carcinogenic process, andefforts are needed, specifically when data imme- several different mechanisms may be involved.diately relevant to an evaluation are not available. The aim of the Monographs has been, from their In this Preamble, the term ‘agent’ refers to inception, to evaluate evidence of carcinogenic-any entity or circumstance that is subject to ity at any stage in the carcinogenesis process,evaluation in a Monograph. As the scope of the independently of the underlying mechanisms.programme has broadened, categories of agents Information on mechanisms may, however, benow include specific chemicals, groups of related used in making the overall evaluation (IARC,chemicals, complex mixtures, occupational or 1991; Vainio et al., 1992; IARC, 2005, 2006; seeenvironmental exposures, cultural or behav- also Part B, Sections 4 and 6). As mechanismsioural practices, biological organisms and physi- of carcinogenesis are elucidated, IARC convenescal agents. This list of categories may expand as international scientific conferences to determine whether a broad-based consensus has emerged8
    • Preambleon how specific mechanistic data can be used exposure and (b) there is some evidence or sus-in an evaluation of human carcinogenicity. The picion of carcinogenicity. Mixed exposures mayresults of such conferences are reported in IARC occur in occupational and environmental set-Scientific Publications, which, as long as they still tings and as a result of individual and culturalreflect the current state of scientific knowledge, habits (such as tobacco smoking and dietarymay guide subsequent Working Groups. practices). Chemical analogues and compounds Although the Monographs have emphasized with biological or physical characteristics simi-hazard identification, important issues may also lar to those of suspected carcinogens may alsoinvolve dose–response assessment. In many be considered, even in the absence of data on acases, the same epidemiological and experimen- possible carcinogenic effect in humans or experi-tal studies used to evaluate a cancer hazard can mental animals.also be used to estimate a dose–response relation- The scientific literature is surveyed for pub-ship. A Monograph may undertake to estimate lished data relevant to an assessment of carci-dose–response relationships within the range nogenicity. Ad hoc Advisory Groups convenedof the available epidemiological data, or it may by IARC in 1984, 1989, 1991, 1993, 1998 andcompare the dose–response information from 2003 made recommendations as to whichexperimental and epidemiological studies. In agents should be evaluated in the Monographssome cases, a subsequent publication may be pre- series. Recent recommendations are avail-pared by a separate Working Group with exper- able on the Monographs programme web sitetise in quantitative dose–response assessment. (http://monographs.iarc.fr). IARC may schedule The Monographs are used by national and other agents for review as it becomes aware ofinternational authorities to make risk assess- new scientific information or as national healthments, formulate decisions concerning preventive agencies identify an urgent public health needmeasures, provide effective cancer control pro- related to cancer.grammes and decide among alternative options As significant new data become availablefor public health decisions. The evaluations of on an agent for which a Monograph exists, a re-IARC Working Groups are scientific, qualita- evaluation may be made at a subsequent meeting,tive judgements on the evidence for or against and a new Monograph published. In some cases itcarcinogenicity provided by the available data. may be appropriate to review only the data pub-These evaluations represent only one part of the lished since a prior evaluation. This can be usefulbody of information on which public health deci- for updating a database, reviewing new data tosions may be based. Public health options vary resolve a previously open question or identifyingfrom one situation to another and from country new tumour sites associated with a carcinogenicto country and relate to many factors, including agent. Major changes in an evaluation (e.g. a newdifferent socioeconomic and national priorities. classification in Group 1 or a determination that aTherefore, no recommendation is given with mechanism does not operate in humans, see Partregard to regulation or legislation, which are B, Section 6) are more appropriately addressed bythe responsibility of individual governments or a full review.other international organizations. 4. Data for the Monographs3. Selection of agents for review Each Monograph reviews all pertinent epi- Agents are selected for review on the basis of demiological studies and cancer bioassays intwo main criteria: (a) there is evidence of human experimental animals. Those judged inadequate 9
    • IARC MONOGRAPHS – 100Dor irrelevant to the evaluation may be cited but (a) The Working Groupnot summarized. If a group of similar studies isnot reviewed, the reasons are indicated. The Working Group is responsible for the crit- Mechanistic and other relevant data are also ical reviews and evaluations that are developedreviewed. A Monograph does not necessarily during the meeting. The tasks of Working Groupcite all the mechanistic literature concerning Members are: (i) to ascertain that all appropriatethe agent being evaluated (see Part B, Section data have been collected; (ii) to select the data rel-4). Only those data considered by the Working evant for the evaluation on the basis of scientificGroup to be relevant to making the evaluation merit; (iii) to prepare accurate summaries of theare included. data to enable the reader to follow the reasoning With regard to epidemiological studies, can- of the Working Group; (iv) to evaluate the resultscer bioassays, and mechanistic and other relevant of epidemiological and experimental studies ondata, only reports that have been published or cancer; (v) to evaluate data relevant to the under-accepted for publication in the openly available standing of mechanisms of carcinogenesis; andscientific literature are reviewed. The same publi- (vi) to make an overall evaluation of the carci-cation requirement applies to studies originating nogenicity of the exposure to humans. Workingfrom IARC, including meta-analyses or pooled Group Members generally have published sig-analyses commissioned by IARC in advance of a nificant research related to the carcinogenicity ofmeeting (see Part B, Section 2c). Data from gov- the agents being reviewed, and IARC uses litera-ernment agency reports that are publicly avail- ture searches to identify most experts. Workingable are also considered. Exceptionally, doctoral Group Members are selected on the basis of (a)theses and other material that are in their final knowledge and experience and (b) absence of realform and publicly available may be reviewed. or apparent conflicts of interests. Consideration Exposure data and other information on an is also given to demographic diversity and bal-agent under consideration are also reviewed. In ance of scientific findings and views.the sections on chemical and physical proper-ties, on analysis, on production and use and on (b) Invited Specialistsoccurrence, published and unpublished sources Invited Specialists are experts who also haveof information may be considered. critical knowledge and experience but have Inclusion of a study does not imply accept- a real or apparent conflict of interests. Theseance of the adequacy of the study design or of experts are invited when necessary to assist inthe analysis and interpretation of the results, and the Working Group by contributing their uniquelimitations are clearly outlined in square brack- knowledge and experience during subgroup andets at the end of each study description (see Part plenary discussions. They may also contributeB). The reasons for not giving further considera- text on non-influential issues in the section ontion to an individual study also are indicated in exposure, such as a general description of datathe square brackets. on production and use (see Part B, Section 1). Invited Specialists do not serve as meeting chair5. Meeting participants or subgroup chair, draft text that pertains to the description or interpretation of cancer data, or Five categories of participant can be present participate in the evaluations.at Monograph meetings.10
    • Preamble(c) Representatives of national and whether there is a conflict that warrants some international health agencies limitation on participation. The declarations are updated and reviewed again at the opening of Representatives of national and interna- the meeting. Interests related to the subject oftional health agencies often attend meetings the meeting are disclosed to the meeting par-because their agencies sponsor the programme ticipants and in the published volume (Coglianoor are interested in the subject of a meeting. et al., 2004).Representatives do not serve as meeting chair or The names and principal affiliations of par-subgroup chair, draft any part of a Monograph, ticipants are available on the Monographs pro-or participate in the evaluations. gramme web site (http://monographs.iarc.fr) approximately two months before each meeting.(d) Observers with relevant scientific It is not acceptable for Observers or third parties credentials to contact other participants before a meeting or Observers with relevant scientific credentials to lobby them at any time. Meeting participantsmay be admitted to a meeting by IARC in limited are asked to report all such contacts to IARCnumbers. Attention will be given to achieving a (Cogliano et al., 2005).balance of Observers from constituencies with All participants are listed, with their princi-differing perspectives. They are invited to observe pal affiliations, at the beginning of each volume.the meeting and should not attempt to influence Each participant who is a Member of a Workingit. Observers do not serve as meeting chair or Group serves as an individual scientist and not assubgroup chair, draft any part of a Monograph, a representative of any organization, governmentor participate in the evaluations. At the meeting, or industry.the meeting chair and subgroup chairs may grantObservers an opportunity to speak, generally 6. Working proceduresafter they have observed a discussion. Observersagree to respect the Guidelines for Observers A separate Working Group is responsible forat IARC Monographs meetings (available at developing each volume of Monographs. A vol-http://monographs.iarc.fr). ume contains one or more Monographs, which can cover either a single agent or several related(e) The IARC Secretariat agents. Approximately one year in advance of the The IARC Secretariat consists of scientists meeting of a Working Group, the agents to bewho are designated by IARC and who have rel- reviewed are announced on the Monographs pro-evant expertise. They serve as rapporteurs and gramme web site (http://monographs.iarc.fr) andparticipate in all discussions. When requested by participants are selected by IARC staff in consul-the meeting chair or subgroup chair, they may tation with other experts. Subsequently, relevantalso draft text or prepare tables and analyses. biological and epidemiological data are collected Before an invitation is extended, each poten- by IARC from recognized sources of informationtial participant, including the IARC Secretariat, on carcinogenesis, including data storage andcompletes the WHO Declaration of Interests to retrieval systems such as PubMed. Meeting par-report financial interests, employment and con- ticipants who are asked to prepare preliminarysulting, and individual and institutional research working papers for specific sections are expectedsupport related to the subject of the meeting. to supplement the IARC literature searches withIARC assesses these interests to determine their own searches. 11
    • IARC MONOGRAPHS – 100D For most chemicals and some complex mix- the entire volume is the joint product of thetures, the major collection of data and the prep- Working Group, and there are no individuallyaration of working papers for the sections on authored sections.chemical and physical properties, on analysis, on IARC Working Groups strive to achieve aproduction and use, and on occurrence are car- consensus evaluation. Consensus reflects broadried out under a separate contract funded by the agreement among Working Group Members, butUS National Cancer Institute. Industrial associ- not necessarily unanimity. The chair may electations, labour unions and other knowledgeable to poll Working Group Members to determineorganizations may be asked to provide input to the diversity of scientific opinion on issues wherethe sections on production and use, although consensus is not readily apparent.this involvement is not required as a general rule. After the meeting, the master copy is verifiedInformation on production and trade is obtained by consulting the original literature, edited andfrom governmental, trade and market research prepared for publication. The aim is to publishpublications and, in some cases, by direct con- the volume within six months of the Workingtact with industries. Separate production data Group meeting. A summary of the outcome ison some agents may not be available for a vari- available on the Monographs programme webety of reasons (e.g. not collected or made public site soon after the meeting.in all producing countries, production is small).Information on uses may be obtained from pub-lished sources but is often complemented by B. SCIENTIFIC REVIEW ANDdirect contact with manufacturers. Efforts are EVALUATIONmade to supplement this information with datafrom other national and international sources. The available studies are summarized by the Six months before the meeting, the mate- Working Group, with particular regard to therial obtained is sent to meeting participants to qualitative aspects discussed below. In general,prepare preliminary working papers. The work- numerical findings are indicated as they appearing papers are compiled by IARC staff and sent, in the original report; units are converted whenbefore the meeting, to Working Group Members necessary for easier comparison. The Workingand Invited Specialists for review. Group may conduct additional analyses of the The Working Group meets at IARC for seven published data and use them in their assessmentto eight days to discuss and finalize the texts of the evidence; the results of such supplemen-and to formulate the evaluations. The objectives tary analyses are given in square brackets. Whenof the meeting are peer review and consensus. an important aspect of a study that directlyDuring the first few days, four subgroups (cov- impinges on its interpretation should be broughtering exposure data, cancer in humans, cancer to the attention of the reader, a Working Groupin experimental animals, and mechanistic and comment is given in square brackets.other relevant data) review the working papers, The scope of the IARC Monographs pro-develop a joint subgroup draft and write sum- gramme has expanded beyond chemicals tomaries. Care is taken to ensure that each study include complex mixtures, occupational expo-summary is written or reviewed by someone sures, physical and biological agents, lifestylenot associated with the study being considered. factors and other potentially carcinogenic expo-During the last few days, the Working Group sures. Over time, the structure of a Monographmeets in plenary session to review the subgroup has evolved to include the following sections:drafts and develop the evaluations. As a result,12
    • Preamble Exposure data which the agent being evaluated is only one of Studies of cancer in humans the ingredients. Studies of cancer in experimental animals For biological agents, taxonomy, struc- Mechanistic and other relevant data ture and biology are described, and the degree Summary of variability is indicated. Mode of replication, Evaluation and rationale life cycle, target cells, persistence, latency, host In addition, a section of General Remarks at response and clinical disease other than cancerthe front of the volume discusses the reasons the are also presented.agents were scheduled for evaluation and some For physical agents that are forms of radia-key issues the Working Group encountered dur- tion, energy and range of the radiation areing the meeting. included. For foreign bodies, fibres and respir- This part of the Preamble discusses the types able particles, size range and relative dimensionsof evidence considered and summarized in each are indicated.section of a Monograph, followed by the scientific For agents such as mixtures, drugs or lifestylecriteria that guide the evaluations. factors, a description of the agent, including its composition, is given. Whenever appropriate, other information,1. Exposure data such as historical perspectives or the description Each Monograph includes general informa- of an industry or habit, may be included.tion on the agent: this information may vary sub-stantially between agents and must be adapted (b) Analysis and detectionaccordingly. Also included is information on An overview of methods of analysis andproduction and use (when appropriate), meth- detection of the agent is presented, includingods of analysis and detection, occurrence, and their sensitivity, specificity and reproducibility.sources and routes of human occupational and Methods widely used for regulatory purposesenvironmental exposures. Depending on the are emphasized. Methods for monitoring humanagent, regulations and guidelines for use may be exposure are also given. No critical evaluationpresented. or recommendation of any method is meant or implied.(a) General information on the agent For chemical agents, sections on chemical (c) Production and useand physical data are included: the Chemical The dates of first synthesis and of first com-Abstracts Service Registry Number, the latest pri- mercial production of a chemical, mixture ormary name and the IUPAC systematic name are other agent are provided when available; forrecorded; other synonyms are given, but the list agents that do not occur naturally, this informa-is not necessarily comprehensive. Information tion may allow a reasonable estimate to be madeon chemical and physical properties that are rel- of the date before which no human exposure toevant to identification, occurrence and biologi- the agent could have occurred. The dates of firstcal activity is included. A description of technical reported occurrence of an exposure are also pro-products of chemicals includes trade names, rel- vided when available. In addition, methods ofevant specifications and available information synthesis used in past and present commercialon composition and impurities. Some of the production and different methods of production,trade names given may be those of mixtures in 13
    • IARC MONOGRAPHS – 100Dwhich may give rise to different impurities, are place. For biological agents, the epidemiology ofdescribed. infection is described. The countries where companies report pro-duction of the agent, and the number of compa- (e) Regulations and guidelinesnies in each country, are identified. Available dataon production, international trade and uses are Statements concerning regulations andobtained for representative regions. It should not, guidelines (e.g. occupational exposure limits,however, be inferred that those areas or nations maximal levels permitted in foods and water,are necessarily the sole or major sources or users pesticide registrations) are included, but theyof the agent. Some identified uses may not be may not reflect the most recent situation, sincecurrent or major applications, and the coverage such limits are continuously reviewed and modi-is not necessarily comprehensive. In the case of fied. The absence of information on regulatorydrugs, mention of their therapeutic uses does not status for a country should not be taken to implynecessarily represent current practice nor does it that that country does not have regulations withimply judgement as to their therapeutic efficacy. regard to the exposure. For biological agents, leg- islation and control, including vaccination and therapy, are described.(d) Occurrence and exposure Information on the occurrence of an agent inthe environment is obtained from data derived 2. Studies of cancer in humansfrom the monitoring and surveillance of levels This section includes all pertinent epidemio-in occupational environments, air, water, soil, logical studies (see Part A, Section 4). Studies ofplants, foods and animal and human tissues. biomarkers are included when they are relevantWhen available, data on the generation, per- to an evaluation of carcinogenicity to humans.sistence and bioaccumulation of the agent arealso included. Such data may be available from (a) Types of study considerednational databases. Data that indicate the extent of past and pre- Several types of epidemiological study con-sent human exposure, the sources of exposure, tribute to the assessment of carcinogenicity inthe people most likely to be exposed and the fac- humans — cohort studies, case–control studies,tors that contribute to the exposure are reported. correlation (or ecological) studies and interven-Information is presented on the range of human tion studies. Rarely, results from randomized tri-exposure, including occupational and environ- als may be available. Case reports and case seriesmental exposures. This includes relevant findings of cancer in humans may also be reviewed.from both developed and developing countries. Cohort and case–control studies relate indi-Some of these data are not distributed widely and vidual exposures under study to the occurrence ofmay be available from government reports and cancer in individuals and provide an estimate ofother sources. In the case of mixtures, indus- effect (such as relative risk) as the main measuretries, occupations or processes, information is of association. Intervention studies may providegiven about all agents known to be present. For strong evidence for making causal inferences, asprocesses, industries and occupations, a histori- exemplified by cessation of smoking and the sub-cal description is also given, noting variations in sequent decrease in risk for lung cancer.chemical composition, physical properties and In correlation studies, the units of inves-levels of occupational exposure with date and tigation are usually whole populations (e.g. in14
    • Preambleparticular geographical areas or at particular Bias is the effect of factors in study design ortimes), and cancer frequency is related to a sum- execution that lead erroneously to a stronger ormary measure of the exposure of the population weaker association than in fact exists between anto the agent under study. In correlation studies, agent and disease. Confounding is a form of biasindividual exposure is not documented, which that occurs when the relationship with disease isrenders this kind of study more prone to con- made to appear stronger or weaker than it truly isfounding. In some circumstances, however, cor- as a result of an association between the apparentrelation studies may be more informative than causal factor and another factor that is associatedanalytical study designs (see, for example, the with either an increase or decrease in the inci-Monograph on arsenic in drinking-water; IARC, dence of the disease. The role of chance is related2004). to biological variability and the influence of sam- In some instances, case reports and case series ple size on the precision of estimates of effect.have provided important information about the In evaluating the extent to which these fac-carcinogenicity of an agent. These types of study tors have been minimized in an individual study,generally arise from a suspicion, based on clinical consideration is given to several aspects of designexperience, that the concurrence of two events — and analysis as described in the report of thethat is, a particular exposure and occurrence of study. For example, when suspicion of carcino-a cancer — has happened rather more frequently genicity arises largely from a single small study,than would be expected by chance. Case reports careful consideration is given when interpretingand case series usually lack complete ascertain- subsequent studies that included these data in anment of cases in any population, definition or enlarged population. Most of these considera-enumeration of the population at risk and esti- tions apply equally to case–control, cohort andmation of the expected number of cases in the correlation studies. Lack of clarity of any of theseabsence of exposure. aspects in the reporting of a study can decrease The uncertainties that surround the inter- its credibility and the weight given to it in thepretation of case reports, case series and corre- final evaluation of the exposure.lation studies make them inadequate, except in First, the study population, disease (or dis-rare instances, to form the sole basis for inferring eases) and exposure should have been wella causal relationship. When taken together with defined by the authors. Cases of disease in thecase–control and cohort studies, however, these study population should have been identified intypes of study may add materially to the judge- a way that was independent of the exposure ofment that a causal relationship exists. interest, and exposure should have been assessed Epidemiological studies of benign neo- in a way that was not related to disease status.plasms, presumed preneoplastic lesions and Second, the authors should have taken intoother end-points thought to be relevant to cancer account — in the study design and analysis —are also reviewed. They may, in some instances, other variables that can influence the risk of dis-strengthen inferences drawn from studies of ease and may have been related to the exposurecancer itself. of interest. Potential confounding by such vari- ables should have been dealt with either in the(b) Quality of studies considered design of the study, such as by matching, or in the analysis, by statistical adjustment. In cohort It is necessary to take into account the pos- studies, comparisons with local rates of diseasesible roles of bias, confounding and chance in may or may not be more appropriate than thosethe interpretation of epidemiological studies. with national rates. Internal comparisons of 15
    • IARC MONOGRAPHS – 100Dfrequency of disease among individuals at differ- The advantages of combined analyses areent levels of exposure are also desirable in cohort increased precision due to increased sample sizestudies, since they minimize the potential for and the opportunity to explore potential con-confounding related to the difference in risk fac- founders, interactions and modifying effectstors between an external reference group and the that may explain heterogeneity among studies instudy population. more detail. A disadvantage of combined analy- Third, the authors should have reported the ses is the possible lack of compatibility of databasic data on which the conclusions are founded, from various studies due to differences in sub-even if sophisticated statistical analyses were ject recruitment, procedures of data collection,employed. At the very least, they should have methods of measurement and effects of unmeas-given the numbers of exposed and unexposed ured co-variates that may differ among studies.cases and controls in a case–control study and Despite these limitations, well conducted com-the numbers of cases observed and expected in bined analyses may provide a firmer basis thana cohort study. Further tabulations by time since individual studies for drawing conclusions aboutexposure began and other temporal factors are the potential carcinogenicity of agents.also important. In a cohort study, data on all IARC may commission a meta-analysis orcancer sites and all causes of death should have pooled analysis that is pertinent to a particularbeen given, to reveal the possibility of reporting Monograph (see Part A, Section 4). Additionally,bias. In a case–control study, the effects of inves- as a means of gaining insight from the results oftigated factors other than the exposure of interest multiple individual studies, ad hoc calculationsshould have been reported. that combine data from different studies may Finally, the statistical methods used to obtain be conducted by the Working Group duringestimates of relative risk, absolute rates of can- the course of a Monograph meeting. The resultscer, confidence intervals and significance tests, of such original calculations, which would beand to adjust for confounding should have been specified in the text by presentation in squareclearly stated by the authors. These methods have brackets, might involve updates of previouslybeen reviewed for case–control studies (Breslow conducted analyses that incorporate the results& Day, 1980) and for cohort studies (Breslow & of more recent studies or de-novo analyses.Day, 1987). Irrespective of the source of data for the meta- analyses and pooled analyses, it is important that(c) Meta-analyses and pooled analyses the same criteria for data quality be applied as those that would be applied to individual studies Independent epidemiological studies of the and to ensure also that sources of heterogeneitysame agent may lead to results that are difficult between studies be taken into account.to interpret. Combined analyses of data frommultiple studies are a means of resolving this (d) Temporal effectsambiguity, and well conducted analyses can beconsidered. There are two types of combined Detailed analyses of both relative and abso-analysis. The first involves combining summary lute risks in relation to temporal variables, suchstatistics such as relative risks from individual as age at first exposure, time since first exposure,studies (meta-analysis) and the second involves a duration of exposure, cumulative exposure, peakpooled analysis of the raw data from the individ- exposure (when appropriate) and time sinceual studies (pooled analysis) (Greenland, 1998). cessation of exposure, are reviewed and sum- marized when available. Analyses of temporal16
    • Preamblerelationships may be useful in making causal of the agent being evaluated, data on this pheno-inferences. In addition, such analyses may sug- type may be useful in making causal inferences.gest whether a carcinogen acts early or late in theprocess of carcinogenesis, although, at best, they (f) Criteria for causalityallow only indirect inferences about mechanismsof carcinogenesis. After the quality of individual epidemiologi- cal studies of cancer has been summarized and assessed, a judgement is made concerning the(e) Use of biomarkers in epidemiological strength of evidence that the agent in question studies is carcinogenic to humans. In making its judge- Biomarkers indicate molecular, cellular or ment, the Working Group considers several crite-other biological changes and are increasingly ria for causality (Hill, 1965). A strong associationused in epidemiological studies for various pur- (e.g. a large relative risk) is more likely to indicateposes (IARC, 1991; Vainio et al., 1992; Toniolo causality than a weak association, although it iset al., 1997; Vineis et al., 1999; Buffler et al., 2004). recognized that estimates of effect of small mag-These may include evidence of exposure, of early nitude do not imply lack of causality and may beeffects, of cellular, tissue or organism responses, important if the disease or exposure is common.of individual susceptibility or host responses, Associations that are replicated in several studiesand inference of a mechanism (see Part B, Section of the same design or that use different epidemi-4b). This is a rapidly evolving field that encom- ological approaches or under different circum-passes developments in genomics, epigenomics stances of exposure are more likely to representand other emerging technologies. a causal relationship than isolated observations Molecular epidemiological data that identify from single studies. If there are inconsistentassociations between genetic polymorphisms results among investigations, possible reasonsand interindividual differences in susceptibility are sought (such as differences in exposure), andto the agent(s) being evaluated may contribute results of studies that are judged to be of highto the identification of carcinogenic hazards to quality are given more weight than those of stud-humans. If the polymorphism has been demon- ies that are judged to be methodologically lessstrated experimentally to modify the functional sound.activity of the gene product in a manner that is If the risk increases with the exposure, this isconsistent with increased susceptibility, these considered to be a strong indication of causality,data may be useful in making causal inferences. although the absence of a graded response is notSimilarly, molecular epidemiological studies that necessarily evidence against a causal relation-measure cell functions, enzymes or metabolites ship. The demonstration of a decline in risk afterthat are thought to be the basis of susceptibil- cessation of or reduction in exposure in indi-ity may provide evidence that reinforces biologi- viduals or in whole populations also supports acal plausibility. It should be noted, however, that causal interpretation of the findings.when data on genetic susceptibility originate Several scenarios may increase confidence infrom multiple comparisons that arise from sub- a causal relationship. On the one hand, an agentgroup analyses, this can generate false-positive may be specific in causing tumours at one site orresults and inconsistencies across studies, and of one morphological type. On the other, carci-such data therefore require careful evaluation. nogenicity may be evident through the causationIf the known phenotype of a genetic polymor- of multiple tumour types. Temporality, precisionphism can explain the carcinogenic mechanism of estimates of effect, biological plausibility and 17
    • IARC MONOGRAPHS – 100Dcoherence of the overall database are consid- 3. Studies of cancer in experimentalered. Data on biomarkers may be employed in animalsan assessment of the biological plausibility of epi-demiological observations. All known human carcinogens that have been Although rarely available, results from rand- studied adequately for carcinogenicity in experi-omized trials that show different rates of cancer mental animals have produced positive resultsamong exposed and unexposed individuals pro- in one or more animal species (Wilbourn et al.,vide particularly strong evidence for causality. 1986; Tomatis et al., 1989). For several agents When several epidemiological studies show (e.g. aflatoxins, diethylstilbestrol, solar radiation,little or no indication of an association between vinyl chloride), carcinogenicity in experimen-an exposure and cancer, a judgement may be made tal animals was established or highly suspectedthat, in the aggregate, they show evidence of lack before epidemiological studies confirmed theirof carcinogenicity. Such a judgement requires carcinogenicity in humans (Vainio et al., 1995).first that the studies meet, to a sufficient degree, Although this association cannot establish thatthe standards of design and analysis described all agents that cause cancer in experimental ani-above. Specifically, the possibility that bias, con- mals also cause cancer in humans, it is biologicallyfounding or misclassification of exposure or out- plausible that agents for which there is sufficientcome could explain the observed results should evidence of carcinogenicity in experimental ani-be considered and excluded with reasonable cer- mals (see Part B, Section 6b) also present a car-tainty. In addition, all studies that are judged to cinogenic hazard to humans. Accordingly, inbe methodologically sound should (a) be con- the absence of additional scientific information,sistent with an estimate of effect of unity for any these agents are considered to pose a carcinogenicobserved level of exposure, (b) when considered hazard to humans. Examples of additional scien-together, provide a pooled estimate of relative tific information are data that demonstrate thatrisk that is at or near to unity, and (c) have a nar- a given agent causes cancer in animals throughrow confidence interval, due to sufficient popula- a species-specific mechanism that does not oper-tion size. Moreover, no individual study nor the ate in humans or data that demonstrate that thepooled results of all the studies should show any mechanism in experimental animals also oper-consistent tendency that the relative risk of can- ates in humans (see Part B, Section 6).cer increases with increasing level of exposure. Consideration is given to all available long-It is important to note that evidence of lack of term studies of cancer in experimental animalscarcinogenicity obtained from several epidemio- with the agent under review (see Part A, Sectionlogical studies can apply only to the type(s) of 4). In all experimental settings, the nature andcancer studied, to the dose levels reported, and to extent of impurities or contaminants present inthe intervals between first exposure and disease the agent being evaluated are given when avail-onset observed in these studies. Experience with able. Animal species, strain (including genetichuman cancer indicates that the period from first background where applicable), sex, numbers perexposure to the development of clinical cancer is group, age at start of treatment, route of expo-sometimes longer than 20 years; latent periods sure, dose levels, duration of exposure, survivalsubstantially shorter than 30 years cannot pro- and information on tumours (incidence, latency,vide evidence for lack of carcinogenicity. severity or multiplicity of neoplasms or prene- oplastic lesions) are reported. Those studies in experimental animals that are judged to be irrel- evant to the evaluation or judged to be inadequate18
    • Preamble(e.g. too short a duration, too few animals, poor (a) Qualitative aspectssurvival; see below) may be omitted. Guidelinesfor conducting long-term carcinogenicity exper- An assessment of carcinogenicity involvesiments have been published (e.g. OECD, 2002). several considerations of qualitative impor- Other studies considered may include: exper- tance, including (i) the experimental conditionsiments in which the agent was administered in under which the test was performed, includingthe presence of factors that modify carcinogenic route, schedule and duration of exposure, spe-effects (e.g. initiation–promotion studies, co- cies, strain (including genetic background wherecarcinogenicity studies and studies in geneti- applicable), sex, age and duration of follow-up;cally modified animals); studies in which the (ii) the consistency of the results, for example,end-point was not cancer but a defined precan- across species and target organ(s); (iii) the spec-cerous lesion; experiments on the carcinogenic- trum of neoplastic response, from preneoplasticity of known metabolites and derivatives; and lesions and benign tumours to malignant neo-studies of cancer in non-laboratory animals (e.g. plasms; and (iv) the possible role of modifyinglivestock and companion animals) exposed to factors.the agent. Considerations of importance in the inter- For studies of mixtures, consideration is pretation and evaluation of a particular studygiven to the possibility that changes in the phys- include: (i) how clearly the agent was defined and,icochemical properties of the individual sub- in the case of mixtures, how adequately the sam-stances may occur during collection, storage, ple characterization was reported; (ii) whetherextraction, concentration and delivery. Another the dose was monitored adequately, particu-consideration is that chemical and toxicological larly in inhalation experiments; (iii) whether theinteractions of components in a mixture may doses, duration of treatment and route of expo-alter dose–response relationships. The relevance sure were appropriate; (iv) whether the survivalto human exposure of the test mixture adminis- of treated animals was similar to that of con-tered in the animal experiment is also assessed. trols; (v) whether there were adequate numbersThis may involve consideration of the following of animals per group; (vi) whether both male andaspects of the mixture tested: (i) physical and female animals were used; (vii) whether animalschemical characteristics, (ii) identified constitu- were allocated randomly to groups; (viii) whetherents that may indicate the presence of a class of the duration of observation was adequate; andsubstances and (iii) the results of genetic toxicity (ix) whether the data were reported and analysedand related tests. adequately. The relevance of results obtained with an When benign tumours (a) occur togetheragent that is analogous (e.g. similar in structure with and originate from the same cell type asor of a similar virus genus) to that being evalu- malignant tumours in an organ or tissue in aated is also considered. Such results may provide particular study and (b) appear to represent abiological and mechanistic information that is stage in the progression to malignancy, they arerelevant to the understanding of the process of usually combined in the assessment of tumourcarcinogenesis in humans and may strengthen incidence (Huff et al., 1989). The occurrence ofthe biological plausibility that the agent being lesions presumed to be preneoplastic may in cer-evaluated is carcinogenic to humans (see Part B, tain instances aid in assessing the biological plau-Section 2f). sibility of any neoplastic response observed. If an agent induces only benign neoplasms that appear to be end-points that do not readily undergo 19
    • IARC MONOGRAPHS – 100Dtransition to malignancy, the agent should nev- Gart et al., 1986; Portier & Bailer, 1989; Bieler &ertheless be suspected of being carcinogenic and Williams, 1993). The choice of the most appro-requires further investigation. priate statistical method requires consideration of whether or not there are differences in sur-(b) Quantitative aspects vival among the treatment groups; for example, The probability that tumours will occur may reduced survival because of non-tumour-relateddepend on the species, sex, strain, genetic back- mortality can preclude the occurrence ofground and age of the animal, and on the dose, tumours later in life. When detailed informa-route, timing and duration of the exposure. tion on survival is not available, comparisonsEvidence of an increased incidence of neoplasms of the proportions of tumour-bearing animalswith increasing levels of exposure strengthens among the effective number of animals (alive atthe inference of a causal association between the the time the first tumour was discovered) canexposure and the development of neoplasms. be useful when significant differences in sur- The form of the dose–response relation- vival occur before tumours appear. The lethal-ship can vary widely, depending on the par- ity of the tumour also requires consideration: forticular agent under study and the target organ. rapidly fatal tumours, the time of death providesMechanisms such as induction of DNA dam- an indication of the time of tumour onset andage or inhibition of repair, altered cell division can be assessed using life-table methods; non-and cell death rates and changes in intercellular fatal or incidental tumours that do not affectcommunication are important determinants of survival can be assessed using methods such asdose–response relationships for some carcino- the Mantel-Haenzel test for changes in tumourgens. Since many chemicals require metabolic prevalence. Because tumour lethality is often dif-activation before being converted to their reac- ficult to determine, methods such as the Poly-Ktive intermediates, both metabolic and toxicoki- test that do not require such information cannetic aspects are important in determining the also be used. When results are available on thedose–response pattern. Saturation of steps such number and size of tumours seen in experimen-as absorption, activation, inactivation and elim- tal animals (e.g. papillomas on mouse skin, liverination may produce nonlinearity in the dose– tumours observed through nuclear magneticresponse relationship (Hoel et al., 1983; Gart resonance tomography), other more complicatedet al., 1986), as could saturation of processes such statistical procedures may be needed (Shermanas DNA repair. The dose–response relationship et al., 1994; Dunson et al., 2003).can also be affected by differences in survival Formal statistical methods have been devel-among the treatment groups. oped to incorporate historical control data into the analysis of data from a given experiment. These methods assign an appropriate weight to(c) Statistical analyses historical and concurrent controls on the basis Factors considered include the adequacy of of the extent of between-study and within-studythe information given for each treatment group: variability: less weight is given to historical con-(i) number of animals studied and number exam- trols when they show a high degree of variability,ined histologically, (ii) number of animals with a and greater weight when they show little varia-given tumour type and (iii) length of survival. bility. It is generally not appropriate to discountThe statistical methods used should be clearly a tumour response that is significantly increasedstated and should be the generally accepted tech- compared with concurrent controls by arguingniques refined for this purpose (Peto et al., 1980; that it falls within the range of historical controls,20
    • Preambleparticularly when historical controls show high one subsection. For example, a mutation in abetween-study variability and are, thus, of little gene that codes for an enzyme that metabolizesrelevance to the current experiment. In analys- the agent under study could be discussed in theing results for uncommon tumours, however, the subsections on toxicokinetics, mechanisms andanalysis may be improved by considering histori- individual susceptibility if it also exists as ancal control data, particularly when between-study inherited polymorphism.variability is low. Historical controls should beselected to resemble the concurrent controls as (a) Toxicokinetic dataclosely as possible with respect to species, gen-der and strain, as well as other factors such as Toxicokinetics refers to the absorption, dis-basal diet and general laboratory environment, tribution, metabolism and elimination of agentswhich may affect tumour-response rates in con- in humans, experimental animals and, wheretrol animals (Haseman et al., 1984; Fung et al., relevant, cellular systems. Examples of kinetic1996; Greim et al., 2003). factors that may affect dose–response relation- Although meta-analyses and combined anal- ships include uptake, deposition, biopersis-yses are conducted less frequently for animal tence and half-life in tissues, protein binding,experiments than for epidemiological studies metabolic activation and detoxification. Studiesdue to differences in animal strains, they can be that indicate the metabolic fate of the agent inuseful aids in interpreting animal data when the humans and in experimental animals are sum-experimental protocols are sufficiently similar. marized briefly, and comparisons of data from humans and animals are made when possible. Comparative information on the relationship4. Mechanistic and other relevant between exposure and the dose that reaches the data target site may be important for the extrapola- tion of hazards between species and in clarifying Mechanistic and other relevant data may pro- the role of in-vitro findings.vide evidence of carcinogenicity and also help inassessing the relevance and importance of find- (b) Data on mechanisms of carcinogenesisings of cancer in animals and in humans. Thenature of the mechanistic and other relevant data To provide focus, the Working Groupdepends on the biological activity of the agent attempts to identify the possible mechanisms bybeing considered. The Working Group considers which the agent may increase the risk of cancer.representative studies to give a concise descrip- For each possible mechanism, a representativetion of the relevant data and issues that they con- selection of key data from humans and experi-sider to be important; thus, not every available mental systems is summarized. Attention isstudy is cited. Relevant topics may include toxi- given to gaps in the data and to data that suggestscokinetics, mechanisms of carcinogenesis, sus- that more than one mechanism may be operat-ceptible individuals, populations and life-stages, ing. The relevance of the mechanism to humansother relevant data and other adverse effects. is discussed, in particular, when mechanisticWhen data on biomarkers are informative about data are derived from experimental model sys-the mechanisms of carcinogenesis, they are tems. Changes in the affected organs, tissues orincluded in this section. cells can be divided into three non-exclusive lev- These topics are not mutually exclusive; thus, els as described below.the same studies may be discussed in more than 21
    • IARC MONOGRAPHS – 100D(i) Changes in physiology described for every possible level and mechanism Physiological changes refer to exposure- discussed above.related modifications to the physiology and/or Genotoxicity data are discussed here to illus-response of cells, tissues and organs. Examples trate the key issues involved in the evaluation ofof potentially adverse physiological changes mechanistic data.include mitogenesis, compensatory cell division, Tests for genetic and related effects areescape from apoptosis and/or senescence, pres- described in view of the relevance of gene muta-ence of inflammation, hyperplasia, metaplasia tion and chromosomal aberration/aneuploidyand/or preneoplasia, angiogenesis, alterations in to carcinogenesis (Vainio et al., 1992; McGregorcellular adhesion, changes in steroidal hormones et al., 1999). The adequacy of the reporting ofand changes in immune surveillance. sample characterization is considered and, when necessary, commented upon; with regard to(ii) Functional changes at the cellular level complex mixtures, such comments are similar to those described for animal carcinogenicity Functional changes refer to exposure-related tests. The available data are interpreted criticallyalterations in the signalling pathways used by according to the end-points detected, whichcells to manage critical processes that are related may include DNA damage, gene mutation, sisterto increased risk for cancer. Examples of func- chromatid exchange, micronucleus formation,tional changes include modified activities of chromosomal aberrations and aneuploidy. Theenzymes involved in the metabolism of xenobi- concentrations employed are given, and men-otics, alterations in the expression of key genes tion is made of whether the use of an exogenousthat regulate DNA repair, alterations in cyclin- metabolic system in vitro affected the test result.dependent kinases that govern cell cycle progres- These data are listed in tabular form by phyloge-sion, changes in the patterns of post-translational netic classification.modifications of proteins, changes in regula- Positive results in tests using prokary-tory factors that alter apoptotic rates, changes otes, lower eukaryotes, insects, plants and cul-in the secretion of factors related to the stimula- tured mammalian cells suggest that genetic andtion of DNA replication and transcription and related effects could occur in mammals. Resultschanges in gap–junction-mediated intercellular from such tests may also give information oncommunication. the types of genetic effect produced and on the(iii) Changes at the molecular level involvement of metabolic activation. Some end- points described are clearly genetic in nature Molecular changes refer to exposure-related (e.g. gene mutations), while others are associatedchanges in key cellular structures at the molec- with genetic effects (e.g. unscheduled DNA syn-ular level, including, in particular, genotoxicity. thesis). In-vitro tests for tumour promotion, cellExamples of molecular changes include forma- transformation and gap–junction intercellulartion of DNA adducts and DNA strand breaks, communication may be sensitive to changes thatmutations in genes, chromosomal aberrations, are not necessarily the result of genetic altera-aneuploidy and changes in DNA methylation tions but that may have specific relevance to thepatterns. Greater emphasis is given to irrevers- process of carcinogenesis. Critical appraisalsible effects. of these tests have been published (Montesano The use of mechanistic data in the identifica- et al., 1986; McGregor et al., 1999).tion of a carcinogenic hazard is specific to the Genetic or other activity manifest in humansmechanism being addressed and is not readily and experimental mammals is regarded to be of22
    • Preamblegreater relevance than that in other organisms. surgical implants of various kinds, and poorlyThe demonstration that an agent can induce soluble fibres, dusts and particles of variousgene and chromosomal mutations in mammals sizes, the pathogenic effects of which are a resultin vivo indicates that it may have carcinogenic of their physical presence in tissues or bodyactivity. Negative results in tests for mutagenicity cavities. Other relevant data for such materialsin selected tissues from animals treated in vivo may include characterization of cellular, tissueprovide less weight, partly because they do not and physiological reactions to these materi-exclude the possibility of an effect in tissues other als and descriptions of pathological conditionsthan those examined. Moreover, negative results other than neoplasia with which they may bein short-term tests with genetic end-points can- associated.not be considered to provide evidence that rulesout the carcinogenicity of agents that act through (c) Other data relevant to mechanismsother mechanisms (e.g. receptor-mediatedeffects, cellular toxicity with regenerative cell A description is provided of any structure–division, peroxisome proliferation) (Vainio et al., activity relationships that may be relevant to an1992). Factors that may give misleading results evaluation of the carcinogenicity of an agent, thein short-term tests have been discussed in detail toxicological implications of the physical andelsewhere (Montesano et al., 1986; McGregor chemical properties, and any other data relevantet al., 1999). to the evaluation that are not included elsewhere. When there is evidence that an agent acts by High-output data, such as those derived froma specific mechanism that does not involve gen- gene expression microarrays, and high-through-otoxicity (e.g. hormonal dysregulation, immune put data, such as those that result from testingsuppression, and formation of calculi and other hundreds of agents for a single end-point, pose adeposits that cause chronic irritation), that evi- unique problem for the use of mechanistic datadence is presented and reviewed critically in the in the evaluation of a carcinogenic hazard. Incontext of rigorous criteria for the operation of the case of high-output data, there is the possi-that mechanism in carcinogenesis (e.g. Capen bility to overinterpret changes in individual end-et al., 1999). points (e.g. changes in expression in one gene) For biological agents such as viruses, bacteria without considering the consistency of that find-and parasites, other data relevant to carcinogenic- ing in the broader context of the other end-pointsity may include descriptions of the pathology of (e.g. other genes with linked transcriptional con-infection, integration and expression of viruses, trol). High-output data can be used in assessingand genetic alterations seen in human tumours. mechanisms, but all end-points measured in aOther observations that might comprise cellu- single experiment need to be considered in thelar and tissue responses to infection, immune proper context. For high-throughput data, whereresponse and the presence of tumour markers the number of observations far exceeds the num-are also considered. ber of end-points measured, their utility for iden- For physical agents that are forms of radia- tifying common mechanisms across multipletion, other data relevant to carcinogenicity may agents is enhanced. These data can be used toinclude descriptions of damaging effects at the identify mechanisms that not only seem plausi-physiological, cellular and molecular level, as ble, but also have a consistent pattern of carci-for chemical agents, and descriptions of how nogenic response across entire classes of relatedthese effects occur. ‘Physical agents’ may also be compounds.considered to comprise foreign bodies, such as 23
    • IARC MONOGRAPHS – 100D(d) Susceptibility data found on the Monographs programme web site (http://monographs.iarc.fr). Individuals, populations and life-stages mayhave greater or lesser susceptibility to an agent, (a) Exposure databased on toxicokinetics, mechanisms of carcino-genesis and other factors. Examples of host and Data are summarized, as appropriate, on thegenetic factors that affect individual susceptibil- basis of elements such as production, use, occur-ity include sex, genetic polymorphisms of genes rence and exposure levels in the workplace andinvolved in the metabolism of the agent under environment and measurements in human tis-evaluation, differences in metabolic capacity due sues and body fluids. Quantitative data and timeto life-stage or the presence of disease, differ- trends are given to compare exposures in dif-ences in DNA repair capacity, competition for ferent occupations and environmental settings.or alteration of metabolic capacity by medica- Exposure to biological agents is described intions or other chemical exposures, pre-existing terms of transmission, prevalence and persis-hormonal imbalance that is exacerbated by a tence of infection.chemical exposure, a suppressed immune sys- (b) Cancer in humanstem, periods of higher-than-usual tissue growthor regeneration and genetic polymorphisms that Results of epidemiological studies pertinentlead to differences in behaviour (e.g. addiction). to an assessment of human carcinogenicity areSuch data can substantially increase the strength summarized. When relevant, case reports andof the evidence from epidemiological data and correlation studies are also summarized. The tar-enhance the linkage of in-vivo and in-vitro labo- get organ(s) or tissue(s) in which an increase inratory studies to humans. cancer was observed is identified. Dose–response and other quantitative data may be summarized when available.(e) Data on other adverse effects Data on acute, subchronic and chronic (c) Cancer in experimental animalsadverse effects relevant to the cancer evaluationare summarized. Adverse effects that confirm Data relevant to an evaluation of carcino-distribution and biological effects at the sites of genicity in animals are summarized. For eachtumour development, or alterations in physiol- animal species, study design and route of admin-ogy that could lead to tumour development, are istration, it is stated whether an increased inci-emphasized. Effects on reproduction, embryonic dence, reduced latency, or increased severityand fetal survival and development are summa- or multiplicity of neoplasms or preneoplasticrized briefly. The adequacy of epidemiological lesions were observed, and the tumour sites arestudies of reproductive outcome and genetic and indicated. If the agent produced tumours afterrelated effects in humans is judged by the same prenatal exposure or in single-dose experiments,criteria as those applied to epidemiological stud- this is also mentioned. Negative findings, inverseies of cancer, but fewer details are given. relationships, dose–response and other quantita- tive data are also summarized.5. Summary This section is a summary of data presented (d) Mechanistic and other relevant datain the preceding sections. Summaries can be Data relevant to the toxicokinetics (absorp- tion, distribution, metabolism, elimination) and24
    • Preamblethe possible mechanism(s) of carcinogenesis (e.g. relationship has been established between expo-genetic toxicity, epigenetic effects) are summa- sure to the agent and human cancer. That is, arized. In addition, information on susceptible positive relationship has been observed betweenindividuals, populations and life-stages is sum- the exposure and cancer in studies in whichmarized. This section also reports on other toxic chance, bias and confounding could be ruledeffects, including reproductive and developmen- out with reasonable confidence. A statement thattal effects, as well as additional relevant data that there is sufficient evidence is followed by a sepa-are considered to be important. rate sentence that identifies the target organ(s) or tissue(s) where an increased risk of cancer was observed in humans. Identification of a specific6. Evaluation and rationale target organ or tissue does not preclude the pos- Evaluations of the strength of the evidence for sibility that the agent may cause cancer at othercarcinogenicity arising from human and experi- sites.mental animal data are made, using standard Limited evidence of carcinogenicity:terms. The strength of the mechanistic evidence A positive association has been observedis also characterized. between exposure to the agent and cancer for It is recognized that the criteria for these which a causal interpretation is considered byevaluations, described below, cannot encompass the Working Group to be credible, but chance,all of the factors that may be relevant to an eval- bias or confounding could not be ruled out withuation of carcinogenicity. In considering all of reasonable confidence.the relevant scientific data, the Working Group Inadequate evidence of carcinogenicity: Themay assign the agent to a higher or lower cat- available studies are of insufficient quality, con-egory than a strict interpretation of these criteria sistency or statistical power to permit a conclu-would indicate. sion regarding the presence or absence of a causal These categories refer only to the strength of association between exposure and cancer, or nothe evidence that an exposure is carcinogenic data on cancer in humans are available.and not to the extent of its carcinogenic activ- Evidence suggesting lack of carcinogenicity:ity (potency). A classification may change as new There are several adequate studies covering theinformation becomes available. full range of levels of exposure that humans are An evaluation of the degree of evidence is lim- known to encounter, which are mutually consist-ited to the materials tested, as defined physically, ent in not showing a positive association betweenchemically or biologically. When the agents eval- exposure to the agent and any studied canceruated are considered by the Working Group to be at any observed level of exposure. The resultssufficiently closely related, they may be grouped from these studies alone or combined shouldtogether for the purpose of a single evaluation of have narrow confidence intervals with an upperthe degree of evidence. limit close to the null value (e.g. a relative risk of 1.0). Bias and confounding should be ruled(a) Carcinogenicity in humans out with reasonable confidence, and the studies should have an adequate length of follow-up. A The evidence relevant to carcinogenicity from conclusion of evidence suggesting lack of carcino-studies in humans is classified into one of the fol- genicity is inevitably limited to the cancer sites,lowing categories: conditions and levels of exposure, and length of Sufficient evidence of carcinogenicity: observation covered by the available studies. InThe Working Group considers that a causal 25
    • IARC MONOGRAPHS – 100Daddition, the possibility of a very small risk at the A single study in one species and sex might belevels of exposure studied can never be excluded. considered to provide sufficient evidence of carci- In some instances, the above categories may nogenicity when malignant neoplasms occur tobe used to classify the degree of evidence related an unusual degree with regard to incidence, site,to carcinogenicity in specific organs or tissues. type of tumour or age at onset, or when there are When the available epidemiological stud- strong findings of tumours at multiple sites.ies pertain to a mixture, process, occupation or Limited evidence of carcinogenicity:industry, the Working Group seeks to identify The data suggest a carcinogenic effect but arethe specific agent considered most likely to be limited for making a definitive evaluationresponsible for any excess risk. The evaluation because, e.g. (a) the evidence of carcinogenicityis focused as narrowly as the available data on is restricted to a single experiment; (b) there areexposure and other aspects permit. unresolved questions regarding the adequacy of the design, conduct or interpretation of the stud-(b) Carcinogenicity in experimental ies; (c) the agent increases the incidence only of animals benign neoplasms or lesions of uncertain neo- plastic potential; or (d) the evidence of carcino- Carcinogenicity in experimental animals can genicity is restricted to studies that demonstratebe evaluated using conventional bioassays, bioas- only promoting activity in a narrow range of tis-says that employ genetically modified animals, sues or organs.and other in-vivo bioassays that focus on one or Inadequate evidence of carcinogenicity:more of the critical stages of carcinogenesis. In The studies cannot be interpreted as showingthe absence of data from conventional long-term either the presence or absence of a carcinogenicbioassays or from assays with neoplasia as the effect because of major qualitative or quantitativeend-point, consistently positive results in several limitations, or no data on cancer in experimentalmodels that address several stages in the multi- animals are available.stage process of carcinogenesis should be con- Evidence suggesting lack of carcinogenicity:sidered in evaluating the degree of evidence of Adequate studies involving at least two speciescarcinogenicity in experimental animals. are available which show that, within the limits The evidence relevant to carcinogenicity in of the tests used, the agent is not carcinogenic.experimental animals is classified into one of the A conclusion of evidence suggesting lack of car-following categories: cinogenicity is inevitably limited to the species, Sufficient evidence of carcinogenicity: The tumour sites, age at exposure, and conditionsWorking Group considers that a causal relation- and levels of exposure studied.ship has been established between the agent andan increased incidence of malignant neoplasms (c) Mechanistic and other relevant dataor of an appropriate combination of benign andmalignant neoplasms in (a) two or more species Mechanistic and other evidence judged toof animals or (b) two or more independent stud- be relevant to an evaluation of carcinogenicityies in one species carried out at different times and of sufficient importance to affect the over-or in different laboratories or under different all evaluation is highlighted. This may includeprotocols. An increased incidence of tumours in data on preneoplastic lesions, tumour pathol-both sexes of a single species in a well conducted ogy, genetic and related effects, structure–activ-study, ideally conducted under Good Laboratory ity relationships, metabolism and toxicokinetics,Practices, can also provide sufficient evidence.26
    • Preamblephysicochemical parameters and analogous bio- have been focused on investigating a favouredlogical agents. mechanism. The strength of the evidence that any carcino- For complex exposures, including occupa-genic effect observed is due to a particular mech- tional and industrial exposures, the chemicalanism is evaluated, using terms such as ‘weak’, composition and the potential contribution of‘moderate’ or ‘strong’. The Working Group then carcinogens known to be present are consideredassesses whether that particular mechanism is by the Working Group in its overall evaluationlikely to be operative in humans. The strongest of human carcinogenicity. The Working Groupindications that a particular mechanism oper- also determines the extent to which the materi-ates in humans derive from data on humans als tested in experimental systems are related toor biological specimens obtained from exposed those to which humans are exposed.humans. The data may be considered to be espe-cially relevant if they show that the agent in ques- (d) Overall evaluationtion has caused changes in exposed humans thatare on the causal pathway to carcinogenesis. Finally, the body of evidence is considered asSuch data may, however, never become available, a whole, to reach an overall evaluation of the car-because it is at least conceivable that certain com- cinogenicity of the agent to humans.pounds may be kept from human use solely on An evaluation may be made for a group ofthe basis of evidence of their toxicity and/or car- agents that have been evaluated by the Workingcinogenicity in experimental systems. Group. In addition, when supporting data indi- The conclusion that a mechanism operates in cate that other related agents, for which there isexperimental animals is strengthened by find- no direct evidence of their capacity to induceings of consistent results in different experimen- cancer in humans or in animals, may also betal systems, by the demonstration of biological carcinogenic, a statement describing the ration-plausibility and by coherence of the overall data- ale for this conclusion is added to the evaluationbase. Strong support can be obtained from stud- narrative; an additional evaluation may be madeies that challenge the hypothesized mechanism for this broader group of agents if the strength ofexperimentally, by demonstrating that the sup- the evidence warrants it.pression of key mechanistic processes leads to The agent is described according to the word-the suppression of tumour development. The ing of one of the following categories, and theWorking Group considers whether multiple designated group is given. The categorization ofmechanisms might contribute to tumour devel- an agent is a matter of scientific judgement thatopment, whether different mechanisms might reflects the strength of the evidence derived fromoperate in different dose ranges, whether sepa- studies in humans and in experimental animalsrate mechanisms might operate in humans and and from mechanistic and other relevant data.experimental animals and whether a unique Group 1: The agent is carcinogenic tomechanism might operate in a susceptible group. humans.The possible contribution of alternative mecha- This category is used when there is suffi-nisms must be considered before concluding cient evidence of carcinogenicity in humans.that tumours observed in experimental animals Exceptionally, an agent may be placed in thisare not relevant to humans. An uneven level of category when evidence of carcinogenicity inexperimental support for different mechanisms humans is less than sufficient but there is suffi-may reflect that disproportionate resources cient evidence of carcinogenicity in experimental 27
    • IARC MONOGRAPHS – 100Danimals and strong evidence in exposed humans Group 2B: The agent is possibly carcinogenicthat the agent acts through a relevant mechanism to humans.of carcinogenicity. This category is used for agents for whichGroup 2. there is limited evidence of carcinogenicity in This category includes agents for which, at humans and less than sufficient evidence of car-one extreme, the degree of evidence of carcino- cinogenicity in experimental animals. It maygenicity in humans is almost sufficient, as well as also be used when there is inadequate evidencethose for which, at the other extreme, there are of carcinogenicity in humans but there is suffi-no human data but for which there is evidence of cient evidence of carcinogenicity in experimentalcarcinogenicity in experimental animals. Agents animals. In some instances, an agent for whichare assigned to either Group 2A (probably car- there is inadequate evidence of carcinogenicity incinogenic to humans) or Group 2B (possibly humans and less than sufficient evidence of car-carcinogenic to humans) on the basis of epide- cinogenicity in experimental animals togethermiological and experimental evidence of carci- with supporting evidence from mechanistic andnogenicity and mechanistic and other relevant other relevant data may be placed in this group.data. The terms probably carcinogenic and possi- An agent may be classified in this category solelybly carcinogenic have no quantitative significance on the basis of strong evidence from mechanisticand are used simply as descriptors of different and other relevant data.levels of evidence of human carcinogenicity, with Group 3: The agent is not classifiable as to itsprobably carcinogenic signifying a higher level of carcinogenicity to humans.evidence than possibly carcinogenic. This category is used most commonly forGroup 2A: The agent is probably agents for which the evidence of carcinogenicity carcinogenic to humans. is inadequate in humans and inadequate or lim- This category is used when there is limited ited in experimental animals.evidence of carcinogenicity in humans and suffi- Exceptionally, agents for which the evidencecient evidence of carcinogenicity in experimental of carcinogenicity is inadequate in humans butanimals. In some cases, an agent may be classi- sufficient in experimental animals may be placedfied in this category when there is inadequate evi- in this category when there is strong evidencedence of carcinogenicity in humans and sufficient that the mechanism of carcinogenicity in experi-evidence of carcinogenicity in experimental ani- mental animals does not operate in humans.mals and strong evidence that the carcinogenesis Agents that do not fall into any other groupis mediated by a mechanism that also operates are also placed in this category.in humans. Exceptionally, an agent may be clas- An evaluation in Group 3 is not a determi-sified in this category solely on the basis of lim- nation of non-carcinogenicity or overall safety.ited evidence of carcinogenicity in humans. An It often means that further research is needed,agent may be assigned to this category if it clearly especially when exposures are widespread orbelongs, based on mechanistic considerations, to the cancer data are consistent with differinga class of agents for which one or more members interpretations.have been classified in Group 1 or Group 2A. Group 4: The agent is probably not carcinogenic to humans. This category is used for agents for which there is evidence suggesting lack of carcinogenicity28
    • Preamblein humans and in experimental animals. In 2001. Workshop report. IARC Sci Publ, 1571–27.some instances, agents for which there is inad- PMID:15055286 Capen CC, Dybing E, Rice JM, Wilbourn JD (1999).equate evidence of carcinogenicity in humans Species Differences in Thyroid, Kidney and Urinarybut evidence suggesting lack of carcinogenicity in Bladder Carcinogenesis.Proceedings of a consensusexperimental animals, consistently and strongly conference. Lyon, France, 3–7 November 1997. IARC Sci Publ, 1471–225.supported by a broad range of mechanistic and Cogliano V, Baan R, Straif K et al. (2005). Transparencyother relevant data, may be classified in this in IARC Monographs. Lancet Oncol, 6: 747 doi:10.1016/group. S1470-2045(05)70380-6 Cogliano VJ, Baan RA, Straif K et  al. (2004). The sci- ence and practice of carcinogen identification and(e) Rationale evaluation. Environ Health Perspect, 112: 1269–1274. doi:10.1289/ehp.6950 PMID:15345338 The reasoning that the Working Group used Dunson DB, Chen Z, Harry J (2003). A Bayesian approachto reach its evaluation is presented and discussed. for joint modeling of cluster size and subunit-specificThis section integrates the major findings from outcomes. Biometrics, 59: 521–530. doi:10.1111/1541- 0420.00062 PMID:14601753studies of cancer in humans, studies of cancer Fung KY, Krewski D, Smythe RT (1996). A comparisonin experimental animals, and mechanistic and of tests for trend with historical controls in carcinogenother relevant data. It includes concise state- bioassay. Can J Stat, 24: 431–454. doi:10.2307/3315326ments of the principal line(s) of argument that Gart JJ, Krewski D, Lee PN et al. (1986). Statistical meth- ods in cancer research. Volume III–The design andemerged, the conclusions of the Working Group analysis of long-term animal experiments. IARC Scion the strength of the evidence for each group of Publ, 791–219. PMID:3301661studies, citations to indicate which studies were Greenland S (1998). Meta-analysis. In: Rothman, K.J. &pivotal to these conclusions, and an explanation Greenland, S., eds, Modern Epidemiology, Philadelphia, Lippincott Williams & Wilkins, pp. 643–673of the reasoning of the Working Group in weigh- Greim H, Gelbke H-P, Reuter U et  al. (2003).ing data and making evaluations. When there Evaluation of historical control data in carcino-are significant differences of scientific interpre- genicity studies. Hum Exp Toxicol, 22: 541–549. doi:10.1191/0960327103ht394oa PMID:14655720tation among Working Group Members, a brief Haseman JK, Huff J, Boorman GA (1984). Use of historicalsummary of the alternative interpretations is control data in carcinogenicity studies in rodents. Toxicolprovided, together with their scientific rationale Pathol, 12: 126–135. doi:10.1177/019262338401200203and an indication of the relative degree of sup- PMID:11478313 Hill AB (1965). The environment and disease: Associationport for each alternative. or causation? Proc R Soc Med, 58: 295–300. PMID:14283879 Hoel DG, Kaplan NL, Anderson MW (1983). Implication of nonlinear kinetics on risk estimation in carcino-References genesis. Science, 219: 1032–1037. doi:10.1126/sci- ence.6823565 PMID:6823565Bieler GS & Williams RL (1993). Ratio estimates, Huff JE, Eustis SL, Haseman JK (1989). Occurrence and the delta method, and quantal response tests for relevance of chemically induced benign neoplasms in increased carcinogenicity. Biometrics, 49: 793–801. long-term carcinogenicity studies. Cancer Metastasis doi:10.2307/2532200 PMID:8241374 Rev, 8: 1–22. doi:10.1007/BF00047055 PMID:2667783Breslow NE & Day NE (1980). Statistical methods in can- IARC (1977). IARC Monographs Programme on the cer research. Volume I - The analysis of case-control Evaluation of the Carcinogenic Risk of Chemicals to studies. IARC Sci Publ, 325–338. PMID:7216345 Humans. Preamble (IARC intern. tech. Rep. No. 77/002)Breslow NE & Day NE (1987). Statistical methods in can- IARC (1978). Chemicals with Sufficient Evidence of cer research. Volume II–The design and analysis of Carcinogenicity in Experimental Animals – IARC cohort studies. IARC Sci Publ, 821–406. PMID:3329634 Monographs Volumes 1–17 (IARC intern. tech. Rep.Buffler P, Rice J, Baan R et  al. (2004). Workshop on No. 78/003) Mechanisms of Carcinogenesis: Contributions of IARC (1979). Criteria to Select Chemicals for IARC Molecular Epidemiology. Lyon, 14–17 November Monographs (IARC intern. tech. Rep. No. 79/003) 29
    • IARC MONOGRAPHS – 100DIARC (1982). Chemicals, Industrial Processes and Tomatis L, Aitio A, Wilbourn J, Shuker L (1989). Human Industries Associated with Cancer in Humans (IARC carcinogens so far identified. Jpn J Cancer Res, 80: 795– Monographs, Volumes 1 to 29). IARC Monogr Eval 807. PMID:2513295 Carcinog Risk Chem Hum Suppl, 4: 1–292. Toniolo P, Boffetta P, Shuker DEG et  al.editors (1997).IARC (1983). Approaches to Classifying Chemical Proceedings of the Workshop on Application of Carcinogens According to Mechanism of Action (IARC Biomarkers to Cancer Epidemiology. Lyon, France, intern. tech. Rep. No. 83/001) 20–23 February 1996. IARC Sci Publ, 1421–318.IARC (1987). Overall evaluations of carcinogenicity: an Vainio H, Magee P, McGregor D, McMichael Aeditors updating of IARC Monographs volumes 1 to 42. IARC (1992). Mechanisms of Carcinogenesis in Risk Monogr Eval Carcinog Risks Hum Suppl, 7: 1–440. Identification. IARC Working Group Meeting. Lyon, PMID:3482203 11–18 June 1991. IARC Sci Publ, 1161–608.IARC (1988). Report of an IARC Working Group to Review Vainio H, Wilbourn JD, Sasco AJ et  al. (1995). the Approaches and Processes Used to Evaluate the [Identification of human carcinogenic risks in IARC Carcinogenicity of Mixtures and Groups of Chemicals monographs] Bull Cancer, 82: 339–348. PMID:7626841 (IARC intern. tech. Rep. No. 88/002) Vineis P, Malats N, Lang M et al.editors (1999). MetabolicIARC (1991). A Consensus Report of an IARC Monographs Polymorphisms and Susceptibility to Cancer. IARC Sci Working Group on the Use of Mechanisms of Publ, 1481–510. PMID:10493243 Carcinogenesis in Risk Identification (IARC intern. Wilbourn J, Haroun L, Heseltine E et al. (1986). Response tech. Rep. No. 91/002) of experimental animals to human carcinogens: anIARC (2005). Report of the Advisory Group to Recommend analysis based upon the IARC Monographs pro- Updates to the Preamble to the IARC Monographs gramme. Carcinogenesis, 7: 1853–1863. doi:10.1093/ (IARC Int. Rep. No. 05/001) carcin/7.11.1853 PMID:3769134IARC (2006). Report of the Advisory Group to Review the Amended Preamble to the IARC Monographs (IARC Int. Rep. No. 06/001)IARC (2004). Some drinking-water disinfectants and contaminants, including arsenic. IARC Monogr Eval Carcinog Risks Hum, 84: 1–477. PMID:15645577McGregor DB, Rice JM, Venitt Seditors (1999). The use of short-and medium-term tests for carcinogens and data on genetic effects in carcinogenic hazard evaluation. Consensus report. IARC Sci Publ, 1461–536.Montesano R, Bartsch H, Vainio H et al.editors (1986). Long- term and Short-term Assays for Carcinogenesis—A Critical Appraisal. IARC Sci Publ, 831–564.OECD (2002) Guidance Notes for Analysis and Evaluation of Chronic Toxicity and Carcinogenicity Studies (Series on Testing and Assessment No. 35), Paris, OECDPeto R, Pike MC, Day NE et al. (1980). Guidelines for sim- ple, sensitive significance tests for carcinogenic effects in long-term animal experiments. IARC Monogr Eval Carcinog Risk Chem Hum Suppl, 2: Suppl311–426. PMID:6935185Portier CJ & Bailer AJ (1989). Testing for increased carcinogenicity using a survival-adjusted quan- tal response test. Fundam Appl Toxicol, 12: 731–737. doi:10.1016/0272-0590(89)90004-3 PMID:2744275Sherman CD, Portier CJ, Kopp-Schneider A (1994). Multistage models of carcinogenesis: an approximation for the size and number distribution of late-stage clones. Risk Anal, 14: 1039–1048. doi:10.1111/j.1539-6924.1994. tb00074.x PMID:7846311Stewart BW, Kleihues P, editors (2003). World Cancer Report, Lyon, IARC30
    • General RemarksPart D of Volume 100 of the IARC Monographs on the Evaluation of Carcinogenic Risks toHumans considers all forms of radiation that were classified as carcinogenic to humans(Group 1) in Volumes 1–99.Volume 100 – General Information About half of the agents classified in Group 1 were last reviewed more than 20 years ago, beforemechanistic studies became prominent in evaluations of carcinogenicity. In addition, more recentepidemiological studies and animal cancer bioassays have demonstrated that many cancer hazardsreported in earlier studies were later observed in other organs or through different exposure sce-narios. Much can be learned by updating the assessments of agents that are known to cause cancerin humans. Accordingly, IARC has selected A Review of Human Carcinogens to be the topic forVolume 100. It is hoped that this volume, by compiling the knowledge accumulated through severaldecades of cancer research, will stimulate cancer prevention activities worldwide, and will be a valuedresource for future research to identify other agents suspected of causing cancer in humans. Volume 100 was developed by six separate Working Groups: Pharmaceuticals Biological agents Arsenic, metals, fibres, and dusts Radiation Personal habits and indoor combustions Chemical agents and related occupations Because the scope of Volume 100 is so broad, its Monographs are focused on key information.Each Monograph presents a description of a carcinogenic agent and how people are exposed, criticaloverviews of the epidemiological studies and animal cancer bioassays, and a concise review of theagent’s toxicokinetics, plausible mechanisms of carcinogenesis, and potentially susceptible popula-tions, and life-stages. Details of the design and results of individual epidemiological studies and ani-mal cancer bioassays are summarized in tables. Short tables that highlight key results are printed inVolume 100, and more extensive tables that include all studies appear on the Monographs programmewebsite (http://monographs.iarc.fr/). For a few well-established associations (for example, tobacco 31
    • IARC MONOGRAPHS – 100Dsmoke and human lung cancer), it was impractical to include all studies, even in the website tables.In those instances, the rationale for inclusion or exclusion of sets of studies is given. Each section of Volume 100 was reviewed by a subgroup of the Working Group with appropriatesubject expertise, then all the sections of a Monograph were discussed together in a plenary sessionof the full Working Group. As a result, the evaluation statements and other conclusions reflect theviews of the Working Group as a whole. Volume 100 compiles information on tumour sites and mechanisms of carcinogenesis. This infor-mation will be used in two scientific publications that may be considered as annexes to this volume.One publication, Tumour Site Concordance between Humans and Experimental Animals, will ana-lyze the correspondence of tumour sites among humans and different animal species. It will dis-cuss the predictive value of different animal tumours for cancer in humans, and perhaps identifyhuman tumour sites for which there are no good animal models. Another publication, MechanismsInvolved in Human Carcinogenesis, will describe mechanisms known to or likely to cause cancer inhumans. Joint consideration of multiple agents that act through similar mechanisms should facilitatethe development of a more comprehensive discussion of these mechanisms. Because susceptibilityoften has its basis in a mechanism, this could also facilitate a more confident and precise descriptionof populations that may be susceptible to agents acting through each mechanism. This publicationwill also suggest biomarkers that could improve the design of future studies. In this way, IARC hopesthat Volume 100 will serve to improve the design of future cancer studies.Specific remarks about the review of radiation in this volume Solar radiation was classified as Group 1 in Volume 55 (IARC, 1992). At that time, some indi-vidual components of solar radiation, ultraviolet radiation A, B, and C, were classified as probablycarcinogenic to humans (Group 2A), along with sunlamps and sunbeds, which act as artificial sourcesof ultraviolet radiation. These agents are also reviewed in this volume to evaluate whether the epi-demiological and mechanistic studies available today provide sufficient evidence to identify specificcomponents of solar radiation as carcinogenic to humans. In Volume 75 (IARC, 2000), X-radiationand γ-radiation were classified as Group 1, along with neutrons. Internalized radionuclides that emitα particles or β particles were classified as Group 1 in Volume 78 (IARC, 2001). That volume also listedindividually in Group 1 specific radionuclides for which there was sufficient evidence in humans. Ofthese, radon-222 and its decay products had been classified earlier as Group 1 in Volume 43 (IARC,1988). One occupation involving radiation exposure, underground haematite mining with exposureto radon, was reviewed in Volume 1 (IARC, 1972) and classified as Group 1 in Supplement 7 (IARC,1987). In reviewing studies on occupational exposures to ultraviolet radiation, the Working Group foundstrong evidence of ocular melanoma in welders. After a literature search for other studies of weldersand a review of this information, the Working Group concluded that these studies provide sufficientevidence of carcinogenicity. Welding fumes had been classified as possibly carcinogenic to humans(Group 2B) in Volume 49 (IARC, 1990) and this was not scheduled for update in this volume. A fullreview of welding was considered to be outside the scope of this meeting, as concern about weldinghas generally focused on exposures to mixtures of metal and chemical fumes (IARC, 1990). Welders32
    • General Remarksand people who work with them may also be exposed to fumes of thorium-232, which is used in tung-sten welding rods (NCRP, 1988; Nuclear Regulatory Commission, 2001). Although it is not possiblewithout a full review to attribute the occurrence of ocular melanoma to ultraviolet radiation specifi-cally, the review of ocular melanoma in this volume was thorough and the findings are expected toremain after a full review of welding in a subsequent Monograph. Accordingly, the Working Groupmade an evaluation that there is sufficient evidence in humans for the carcinogenicity of welding. A summary of the findings of this volume appears in The Lancet Oncology (El Ghissassi et al.,2009).ReferencesEl Ghissassi F, Baan R, Straif K et al.; WHO International Agency for Research on Cancer Monograph Working Group (2009). A review of human carcinogens–part D: radiation. Lancet Oncol, 10:751–752. doi:10.1016/S1470- 2045(09)70213-X PMID:19655431IARC (1972). Some inorganic substances, chlorinated hydrocarbons, aromatic amines, N-nitroso compounds and natu- ral products. IARC Monogr Eval Carcinog Risk Chem Man, 1:1–184.IARC (1987). Overall evaluations of carcinogenicity: an updating of IARC Monographs volumes 1 to 42. IARC Monogr Eval Carcinog Risks Hum Suppl, 7:1–440. PMID:3482203IARC (1988). Man-made mineral fibres and radon. IARC Monogr Eval Carcinog Risks Hum, 43:1–300.IARC (1990). Chromium, nickel and welding. IARC Monogr Eval Carcinog Risks Hum, 49:1–648. PMID:2232124IARC (1992). IARC Monographs on the evaluation of carcinogenic risks to humans. Solar and ultraviolet radiation. IARC Monogr Eval Carcinog Risks Hum, 55:1–316. PMID:1345607IARC (2000). Ionizing radiation, Part 1: X- and gamma- radiation and neutrons. IARC Monogr Eval Carcinog Risks Hum, 75:1–492. PMID:11203346IARC (2001). Ionizing radiation, Part 2: some internally deposited radionuclides. IARC Monogr Eval Carcinog Risks Hum, 78:1–559. PMID:11421248National Council on Radiation Protection and Measurements (NCRP) (1988). Exposure of the population in the United States and Canada from natural background radiation. Bethesda: No. NCRP Report No. 94Nuclear Regulatory Commission (2001). Systematic Radiological Assessment of Exemptions for Source and Byproduct Materials. Washington: No. NUREG-1717 33
    • SOLAR AND ULTRAVIOLET RADIATIONSolar and ultraviolet radiation were considered by a previous IARC Working Group in 1992(IARC, 1992). Since that time, new data have become available, these have been incorpo-rated into the Monograph, and taken into consideration in the present evaluation.1. Exposure Data 1.1 Nomenclature and units For the purpose of this Monograph, the Terrestrial life is dependent on radiant energy photobiological designations of the Commissionfrom the sun. Solar radiation is largely optical Internationale de l’Eclairage (CIE, Internationalradiation [radiant energy within a broad region Commission on Illumination) are the mostof the electromagnetic spectrum that includes relevant, and are used throughout to defineultraviolet (UV), visible (light) and infrared the approximate spectral regions in whichradiation], although both shorter wavelength certain biological absorption properties and(ionizing) and longer wavelength (microwaves biological interaction mechanisms may domi-and radiofrequency) radiation is present. The nate (Commission Internationale de l’Eclairage,wavelength of UV radiation (UVR) lies in the 1987).range of 100–400 nm, and is further subdivided Sources of UVR are characterized in radio-into UVA (315–400  nm), UVB (280–315  nm), metric units. The terms dose (J/m2) and dose rateand UVC (100–280  nm). The UV component (W/m2) pertain to the energy and power, respec-of terrestrial radiation from the midday sun tively, striking a unit surface area of an irradi-comprises about 95% UVA and 5% UVB; UVC ated object (Jagger, 1985). The radiant energyand most of UVB are removed from extraterres- delivered to a given area in a given time is alsotrial radiation by stratospheric ozone. referred to as ‘fluence’, ‘exposure dose’ and ‘dose’ Approximately 5% of solar terrestrial radia- (see IARC, 1992 for further details).tion is UVR, and solar radiation is the major A unit of effective dose [dose weighted insource of human exposure to UVR. Before the accordance with its capacity to bring about abeginning of last century, the sun was essentially particular biological effect] commonly usedthe only source of UVR, but with the advent of in cutaneous photobiology is the ‘minimalartificial sources the opportunity for additional erythemal dose’ (MED). One MED has beenexposure has increased. defined as the lowest radiant exposure to UVR that is sufficient to produce erythema with sharp margins 24 hours after exposure (Morison, 1983). Another end-point often used in cutaneous 35
    • IARC MONOGRAPHS – 100Dphotobiology is a just-perceptible reddening 1.2 Methods for measuring UVRof exposed skin; the dose of UVR necessary toproduce this ‘minimal perceptible erythema’ is UVR can be measured by chemical or physicalsometimes also referred to as a MED. In unac- detectors, often in conjunction with a monochro-climatized, white-skinned populations, there mator or band-pass filter for wavelength selection.is an approximately 4-fold range in the MED Physical detectors include radiometric devices,of exposure to UVB radiation (Diffey & Farr, which respond to the heating effect of the radia-1989). When the term MED is used as a unit tion, and photoelectric devices, in which incidentof ‘exposure dose’, a representative value for photons are detected by a quantum effect such assun-sensitive individuals of 200  J/m2 is usually the production of electrons. Chemical detectorschosen. Since 1997, the reference action spec- include photographic emulsions, actinometrictrum for erythema on human skin (McKinlay solutions and UV-sensitive plastic films.& Diffey, 1987) has become an International The solar UV irradiation of large portionsStandards Organization(ISO)/CIE norm, which, of the Earth is currently measured using multi-by convolution with the emission spectrum frequency imaging detectors on meteorologicalof any UVR source, enables the calculation of satellites.the erythemal yield of the source. A StandardErythema Dose (SED) has been proposed as a 1.3 Sources and exposureunit of erythemally effective UVR dose equiva-lent to 100 J/m2 (Commission Internationale de 1.3.1 Solar UVRl’Eclairage, 1998). Optical radiation from the sun is modified Notwithstanding the difficulties of inter- substantially as it passes through the Earth’spreting accurately the magnitude of such impre- atmosphere, although about two-thirds of thecise units as the MED and the SED, they have the energy from the sun that enters the atmosphereadvantage over radiometric units of being related penetrates to ground level. The annual variationto the biological consequences of the exposure. in extraterrestrial radiation is less than 10%; the The UV index is a tool intended for the variation in the modifying effect of the atmos-communication of the UVR intensity to the phere is far greater (Moseley, 1988).general public. It has been developed jointly On its path through the atmosphere, solarby the World Health Organization, the United UVR is absorbed and scattered by variousNations Environment Program, the International constituents of the atmosphere. It is scattered byCommission on Non-Ionizing Radiation air molecules, particularly oxygen and nitrogen,Protection and was standardized by ISO/CIE. by aerosol and dust particles, and is scatteredIt expresses the erythemal power of the sun as and absorbed by atmospheric pollution. Totalfollows: solar irradiance and the relative contributions ofUV Index = 40 times the erythemally effective different wavelengths vary with altitude. Cloudspower of the sun in W/m2 attenuate solar radiation, although their effect The clear sky UV Index at solar noon is gener- on infrared radiation is greater than on UVR.ally in the range of 0–12 at the Earth’s surface, Reflection of sunlight from certain groundwith values over 11 being considered extreme. surfaces may contribute significantly to the total amount of scattered UVR (Moseley, 1988). The levels of solar UVB radiation reaching the surface of the Earth are largely controlled36
    • Solar and UV radiationby the stratospheric ozone layer, which has been regions but much less nearer the equatorprogressively depleted as a result of accumula- (Diffey, 1991).tion of ozone-destroying chemicals in the Earth’s • Geographic latitude: Annual UVR expo-atmosphere – mostly chlorofluorocarbons sure dose decreases with increasing dis-(CFCs) and hydrochlorofluorocarbons (HCFCs), tance from the equator (Diffey, 1991).whose main use has been in refrigeration and • Altitude: In general, each 300 metreair-conditioning. The accumulation of ozone- increase in altitude increases the sun-depleting chemicals in the atmosphere ceased burning effectiveness of sunlight by aboutlargely as a result of the Montreal Protocol on 4% (Diffey, 1990).“Substances that deplete the ozone layer,” which • Clouds: Clouds influence UV groundwas opened for signature in 1987, and has been irradiance, through reflection, refrac-ratified by 196 states. tion, absorption and scattering, and Global climate change due to the accumula- may increase or, more usually, decreasetion of carbon dioxide (CO2) in the atmosphere UV ground irradiance. Complete lightcan also adversely affect stratospheric ozone. This cloud cover prevents about 50% of UVRwill influence whether, when, and to what extent energy from reaching the surface of theozone levels will return to pre-1980 values. The Earth (Diffey, 1991). Very heavy cloudcurrent best estimate is that global (60°S–60°N) cover absorbs and can virtually eliminateozone levels will return to pre-1980 levels around UVR even in summer. Even with heavythe middle of the 21st century, at or before the cloud cover, however, the scattered UVRtime when stratospheric concentrations of ozone- component of sunlight (as opposed to thatdepleting gases return to pre-1980 levels. Climate coming directly from the sun) is seldomchange will also influence surface UV radiation less than 10% of that under clear sky.through changes induced mainly to clouds and While most clouds block some UV radia-the ability of the Earth’s surface to reflect light. tion, the degree of protection dependsAerosols and air pollutants are also expected to on the type and amount of clouds; somechange in the future. These factors may result clouds can actually increase the UVin either increases or decreases of surface UV intensity on the ground by reflecting,irradiance, through absorption or scattering. As refracting and scattering the sun’s rays.ozone depletion becomes smaller, these factors For example, under some circumstancesare likely to dominate future UV radiation levels (haze, cirrus skies, solar zenith angles(World Meteorological Organization, 2007). ranging from 40–63°), the solar irradi- The amount of solar UVR measured at the ance at Toowoomba, Australia (27.6°S,Earth’s surface depends upon several factors as 151.9°E), was found to be 8% greater thanfollows: that of an equivalent clear sky (Sabburg & • Time of day: In summer, about 20–30% of Wong, 2000; Sabburg et al., 2001). the total daily amount of UVR is received • Surface reflection: The contribution of between 11:00 and 13:00, and 75% between reflected UVR to a person’s total UVR 9:00 and 15:00 (sun time not local time; exposure varies in importance with sev- Diffey, 1991). eral factors. A grass lawn scatters 2–5% • Season: Seasonal variation in terrestrial of incident UVB radiation. Sand reflects UV irradiance, especially UVB, at the about 10–15%, so that sitting under an Earth’s surface is significant in temperate umbrella on the beach can lead to sun- burn both from scattered UVB from the 37
    • IARC MONOGRAPHS – 100D sky and reflected UVB from the sand. radiation in the visible and ultraviolet ranges. Fresh snow may reflect up to 85–90% of NASA’s Total Ozone Mapping Spectrometer incident UVB radiation while water, in (TOMS) device was installed on several space- particular white foam in the sea, may craft, including the Earth Probe spacecraft for reflect up to 30%. Ground reflectance collecting data during 1996–2005. TOMS is no is important, because parts of the body longer available but the continuity of satellite- that are normally shaded are exposed to derived global UV data is maintained via the reflected radiation (Diffey, 1990). new Ozone Monitoring Instrument (OMI), on • Air pollution: Tropospheric ozone and board the Aura satellite (http://aura.gsfc.nasa. other pollutants can decrease UVR. gov/index.html). The presence of aerosols, clouds and snow or ice cover can lead to significant(a) Measurements of terrestrial solar biases, and new algorithms have been developed radiation to improve the satellite-derived measurement of surface UV irradiance using Advanced Very High Because UVR wavelengths between about Resolution Radiometer (AVHRR) and Meteosat295–320  nm (UVB radiation) in the terrestrial images. Currently the European Solar Data Basesolar spectrum are thought to be those mainly (SoDa) is capable to perform on-the-fly fast inter-responsible for adverse health effects, several polation with a non-regular grid and to providestudies have focused on this spectral region. data for any geographic site with a limitation toAccurate measurements of UVR in this spectral a 5-km grid cell. The SoDa contains informationband are difficult to obtain, however, because going back to the year 1985, available at http://the spectral curve of terrestrial solar irradiance www.soda-is.com/eng/services/services_radia-increases by a factor of more than five between tion_free_eng.php.290–320  nm. Nevertheless, extensive measure- Satellite data have been used to draw maps ofments of ambient UVR in this spectral band have UV exposure, and are available for use for epide-been made worldwide. Measurements of terres- miological and other purposes. For example, datatrial solar UVA are less subject to error than sets of UV irradiance derived from TOMS datameasurements of UVB, because the spectrum for the period 1979 to 2000 are available by date,does not vary widely with zenith angle and the latitude and longitude for UVB and UVA. Dataspectral irradiance curve is relatively flat (IARC, from satellites and ground-level measurements1992). show that UV irradiation does not vary steadily The total solar radiation that arrives at the with latitude but that local conditions mayEarth’s surface is termed ‘global radiation’. Global greatly influence actual UV irradiation levels (aradiation is made up of two components, referred good example of this situation may be found into as ‘direct’ and ‘diffuse’. Approximately 70% of the extremely elevated UV levels recorded in thethe UVR at 300 nm is in the diffuse component summer 2003 during the heat wave that killedrather than in the direct rays of the sun. The ratio thousands of people in France and Northernof diffuse to direct radiation increases steadily Italy).from less than 1.0 at 340  nm to at least 2.0 at300  nm. UVR reflected from the ground (the (b) Personal exposuresalbedo) may also be important (IARC, 1992). Solar UV levels reaching the Earth’s surface Individual sun exposure can be estimatedcan now be measured by satellites using hyper- through questionnaires, which are at bestspectral imaging to observe solar backscatter semi-quantitative, and do not give any detailed38
    • Solar and UV radiationinformation on the wavelength of UV exposure. have been reported, these estimates should beIndividual UV dosimeters have been used in considered to be very approximate. They areepidemiological studies, but cannot be used for also subject to differences in cultural and socialthe large-scale monitoring of UV exposure of behaviour, clothing, occupation, and outdoorpopulations. activities. Exposure data for different anatomicalsites is of value in developing biological dose– 1.3.2 Artificial sources of UVRresponse relationships. The exposure of differentanatomical sites to solar UVR depends not only Cumulative annual outdoor exposure may beon ambient UVR and the orientation of sites increased by exposure to artificial sources of UVR.with respect to the sun, but also on cultural Indoor tanning is a widespread practice in mostand social behaviour, type of clothing, and use developed countries, particularly in northernof sunscreen. The most exposed skin surfaces, Europe and the United States of America, andsuch as the nose, tops of the ears and forehead, is gaining popularity even in sunny countrieshave levels of UVB exposure that range up to one like Australia. The prevalence of indoor tanningorder of magnitude relative to that of the lesser varies greatly among different countries, and hasexposed areas, such as underneath the chin. increased during the last decades (IARC, 2006a).Ground reflectance plays a major role in expo- The majority of users are young women, and asure to UVB of all exposed body parts, including recent survey indicated that in the USA, up tothe eye and shaded skin surfaces, particularly 11% of adolescents aged 11–years had ever used anwith highly reflective surfaces such as snow. The indoor tanning device (Cokkinides et al., 2009).solar exposure of the different anatomical sites The median annual exposure dose from artifi-of outdoor workers has recently been calculated cial tanning is probably 20–30 times the MED.(Milon et al., 2007) [Computerised models that Prior to the 1980s, tanning lamps emitted highintegrate direct, diffuse and reflected radiation proportions of UVB and even UVC. Currentlyare currently being developed]. used appliances emit primarily UVA; and in Sunscreens can be applied to control the dose countries where tanning appliances are regu-of UVR to exposed skin. While undoubtedly lated (e.g. Sweden and France), there is a 1.5%useful when sun exposure is unavoidable (IARC, upper limit UVB. However, commercially avail-2001), their use may lead to a longer duration of able “natural” UV-tanning lamps may emit upsun exposure when sun exposure is intentional to 4% UVB. UV emission of a modern tanning(Autier et al., 2007). appliance corresponds to an UV index of 12, i.e. The cumulative annual exposure dose of equivalent to midday tropical sun (IARC, 2006a).solar UVR varies widely among individuals in a Other sources of exposures to UVR includegiven population, depending to a large extent on medical and dental applications. UVR has beenthe occupation and extent of outdoor activities. used for several decades to treat skin diseases,For example, it has been estimated that indoor notably psoriasis. A variety of sources of UVRworkers in mid-latitudes (40–60°N) receive an are used, emitting either broad-band UVA orannual exposure dose of solar UVR to the face narrow-band UVB. A typical dose in a singleof about 40–160 times the MED, depending course of UVB phototherapy can be in the rangeon their level of outdoor activities, whereas the of 200–300 times the MED (IARC, 2006a).annual solar exposure dose for outdoor workers UVR is also used in many different indus-is typically around 250 times the MED. Because tries, yet there is a paucity of data concerningfew actual measurements of personal exposures human exposure from these applications, 39
    • IARC MONOGRAPHS – 100Dprobably because in normal practice, sources are some more recent studies have made objectivewell contained and exposure doses are expected measures of ambient UV and used clinical signsto be low. In some settings, workers may be of cumulative UV damage to the skin such asexposed to radiation by reflection or scattering solar lentigines and actinic keratoses (Table 2.1from adjacent surfaces. Staff in hospitals who available at http://monographs.iarc.fr/ENG/work with unenclosed phototherapy equipment Monographs/vol100D/100D-01-Table2.1.pdf,are at potential risk of overexposure unless Table  2.2 available at http://monographs.protective measures are taken. Indoor tanning iarc.fr/ENG/Monographs/vol10 0D/10 0D-facilities may comprise 20 or more UVA tanning 01-Table2.2.pdf, and Table  2.3 available atappliances, thus potentially exposing operators http://monographs.iarc.fr/ENG/Monographs/to high levels (> 20W/m2) of UVA (IARC, 2006a). vol100D/100D-01-Table2.3.pdf). Acute overexposures to the eyes are common With regard to basal cell carcinoma, allamong electric arc welders. Individuals exposed studies except one (Corona et al., 2001) showedto lighting from fluorescent lamps may typi- significant positive associations with sunburns atcally receive annual exposure doses of UVR in some stage of life or overall. Of the studies thatthe range of 0–30 times the MED, depending on collected information on the presence of actinicilluminance levels and whether or not the lamps keratoses (Green et al., 1996; Corona et al.,are housed behind plastic diffusers. It is also 2001; Walther et al., 2004; Pelucchi et al., 2007),worth noting that tungsten–halogen lamps used all showed this also to be a strong risk factorfor general lighting may emit broad-band UVR (Tables 2.1 and 2.3 on-line). It was proposed that(including UVC) when not housed behind a glass the association of basal cell carcinoma with sunfilter. exposure may vary by histological subtype and anatomical site (Bastiaens et al., 1998). Although a case–control study showed this variation for2. Cancer in Humans recalled sun exposure (Pelucchi et al., 2007), a cohort study did not (Neale et al., 2007).2.1 Natural sunlight For squamous cell carcinoma, while case– control studies tended to demonstrate little asso-2.1.1 Basal cell carcinoma and cutaneous ciation with sunburns (Table 2.2 on-line), cohort squamous cell carcinoma studies uniformly showed significant positive associations (Table  2.3 on-line). The presence In the previous IARC Monograph (IARC, of actinic keratoses, a proportion of which are1992), the evaluation of the causal association squamous cell carcinoma precursors, was theof basal cell carcinoma and squamous cell carci- strongest risk factor identified (Table 2.3 on-line;noma with solar radiation was based on descrip- Green et al., 1996).tive data in Caucasian populations, which showedpositive associations with birth and/or residence 2.1.2 Cutaneous malignant melanomaat low latitudes and rare occurrence at non-sun-exposed anatomical sites. The evaluation was Cutaneous malignant melanoma occurs inalso based on case–control and cohort studies the pigment cells of the skin. Until 10–15 yearswhose main measures were participants’ retro- ago, with the exception of two histologicalspectively recalled sun exposure. The majority subgroups, melanoma was usually regarded as aof analyticalal studies published since have also single entity in analytical studies assessing theused recalled amount of sun exposure, though association with sunlight. The two subgroups,40
    • Solar and UV radiationlentigo maligna melanoma and acral lentiginous risk (RR) estimates of one of the largest meta-melanoma, were usually excluded from studies, analyses, based on 57 studies published up tothe former paradoxically because of its known September 2002 (Gandini et al.,, 2005a, b) were:causal link with cumulative sun exposure, the sunburn (ever/never), 2.0 (95%CI: 1.7–2.4); inter-latter for the opposite reason because it typically mittent sun exposure (high/low), 1.6 (95%CI:occurs on the soles of the feet. 1.3–2.0); chronic sun exposure (high/low), 1.0 In the previous IARC Monograph (IARC, (95%CI: 0.9–1.0); total sun exposure (high/low),1992), the evaluation of the causal association 1.3 (95%CI: 1.0–1.8); actinic tumours (present,between solar radiation and melanoma was past/none), 4.3 (95%CI: 2.8–6.6).based on descriptive data and on data from Case–control studies and the cohort studycase–control studies. The main measures of (Veierød et al., 2003) that have been publishedexposure were participants’ recalled sun expo- since September 2002 have shown results thatsure. ‘Intermittent’ sun exposure, which loosely are generally consistent with the meta-anal-equated with certain sun-intensive activities, ysis, and have not been included in this reviewsuch as sunbathing, outdoor recreations, and (Table  2.5 available at http://monographs.holidays in sunny climates, generally showed iarc.fr/ENG/Monographs/vol100D/100D-01-moderate-to-strong positive associations with Table2.5.pdf and Table  2.6 available at http://melanoma. However, ‘chronic’ or ‘more contin- mono g r aph s . i a rc . f r/ E NG/ Mono g r a ph s /uous’ exposure, which generally equated with vol100D/100D-01-Table2.6.pdf).‘occupational’ exposure, and total sun expo-sure (sum of ‘intermittent’+‘chronic’), generally (a) Anatomical site of melanomashowed weak, null or negative associations. Melanoma–sun-exposure associations These results were collectively interpreted according to the anatomical site of the melanomaunder the ‘intermittent sun exposure’ hypothesis have recently gained greater consideration.(Fears et al., 1977) as showing that melanoma Several studies reported differences in age-occurs as a result of a pattern of intermittent specific incidence rates by site of melanomaintense sun exposure rather than of more contin- (Holman et al., 1980; Houghton et al., 1980;uous sun exposure. Studies that had also assessed Elwood & Gallagher, 1998; Bulliard & Cox, 2000).objective cutaneous signs of skin damage that The numerous analytical studies of risk factorswere generally assumed to be due to accumulated by site of melanoma (Weinstock et al., 1989; Ursosun exposure, e.g. presence or history of actinic et al., 1991; Green, 1992; Krüger et al., 1992; Riegerkeratoses, or signs of other sun-related skin et al., 1995; Whiteman et al., 1998; Carli et al.,damage, showed, almost uniformly, strong posi- 1999; Håkansson et al., 2001; Winnepenninckxtive associations with melanoma. This inconsist- & van den Oord, 2004; Cho et al., 2005; Purdueency of evidence with the apparently negative et al., 2005; Nikolaou et al., 2008) collectivelyassociations of reported ‘chronic’ sun exposure show that melanomas of the head and neck arewith melanoma was noted but not satisfactorily strongly associated with actinic keratoses, andexplained. melanomas on the trunk are strongly associ- Several systematic reviews and meta-anal- ated with naevi. Similar findings have beenyses of analytical studies of the association of reported from recent detailed case–case studiesmelanoma with sun exposure have been published (Whiteman et al., 2003, 2006; Siskind et al., 2005;since (Table  2.4 available at http://monographs. Lee et al., 2006).iarc.fr/ENG/Monographs/vol100D/100D-01-Table2.4.pdf). The summary melanoma relative 41
    • IARC MONOGRAPHS – 100D(b) Skin pigmentation throughout life were generally predictive of Two observations from epidemiological melanomas at all sites in both case–controlstudies may help explain the paradox of the studies and in the pooled analysis (RR, 1.0–2.0).lack of association of melanoma with chronic Those who had objective signs of cumulativesun exposure. First, outdoor workers are not sun damage were at increased risk of melanomaat a substantially increased risk of melanoma at specific sites: the presence of solar lentigines(IARC, 1992; Armstrong & Kricker, 2001); increased the risk of melanoma on the lowersecond, outdoor workers tend to have a higher- limbs (Naldi et al., 2005; RR, 1.5; 95%CI: 1.0–2.1,than-average ability to develop a tan (Green with reference to absence of solar lentigines),et al., 1996; Chang et al., 2009). Outdoor workers while actinic keratoses increased the risk oftend to be constitutionally protected from solar melanoma on the head and neck (Chang et al.,skin damage and at a lower risk of skin cancer 2009; RR, 3.1; 95%CI: 1.4–6.7; based on threethan workers in other occupations because studies from high to low latitudes in which solarof self-selection based on skin pigmentation. keratoses were measured). [The Working GroupIndeed, such self-selection has been observed noted that the omission from many studies ofin a non-Hispanic white study population from the lentigo maligna melanoma subgroup, whichPhiladelphia and San Francisco, USA, whereby is known to be associated with cumulative sunthe average number of hours outdoors in general exposure, potentially results in an underestima-increases with an increasing ability to tan (Fears tion of the association with melanomas on theet al., 2002). The role of baseline sun sensitivity head and limbs.]in influencing sun exposure in the etiology ofmelanoma has long been recognized (Holman 2.1.3 Cancer of the lipet al., 1986; Nelemans et al., 1995). Cancer of the lip has been associated with outdoor occupations in several descriptive(c) Latitude studies (IARC, 1992). Three early case–control The assessment and reporting of sun expo- studies reported increases in risk for cancer ofsure may vary among studies at different lati- the lip with outdoor work, but use of tobaccotudes, due to latitude differences in sun exposure could not be ruled out as an explanation for thisopportunity and behaviour (Elwood & Diffey, association in any study (Keller, 1970; Spitzer1993; Gandini et al., 2005a, b). One approach to et al., 1975; Dardanoni et al., 1984).avoid the problems of quantifying individual sun Two case–control studies have been publishedexposure at different latitudes has been to use since that include information on tobaccoambient UV flux (Fears et al., 2002; Kricker et al., smoking. The first (Pogoda & Preston-Martin,2007) for individuals through life, calculated 1996), which included women only, foundfrom their residential histories, to accurately increased risks of cancer of the lip with averagequantify at least potential solar UV exposure. annual residential UV flux, recalled average Two case–control studies, both done at annual hours spent in outdoor activities, andcomparatively high latitudes (Connecticut, USA; having played high-school or college sports; riskChen et al., 1996) and (Italy; Naldi et al., 2005), estimates were adjusted for complexion, historyand one pooled analysis stratified by latitude of skin cancer and average number of cigarettes(Chang et al., 2009), have presented site-specific smoked per day. Risk was not increased in womenmelanoma risk estimates in relation to latitude whose last occupation was outdoors (odds ratio(see Table  2.5 on-line). Recalls of sunburns (OR)), 1.2; 95%CI: 0.5–2.8). The dose–response42
    • Solar and UV radiationrelationship with recalled average annual hours (ii) Case–control studiesspent in outdoor activities was inconsistent: with Three small case–control studies included<  30 hours as the reference category, the odds only or mainly cases with conjunctival intraepi-ratios were 2.6 (95%CI: 1.0–6.5) for 30–99 hours; thelial neoplasia (Table  2.9 available at http://1.8 (95%CI: 0.7–4.6) for 100–299 hours; and, 4.7 mono g r a ph s . i a rc . f r/ E NG/ Mono g r a ph s /(95%CI: 1.9–12.1) for >  300 hours. The second, vol100D/100D-01-Table2.9.pdf). Napora et al.which included men only (Perea-Milla López (1990) compared 19 patients with biopsy-provenet al., 2003), found no evidence of an increased conjunctival intraepithelial neoplasia (includingrisk for cancer of the lip with estimates of one with invasive squamous cell carcinoma)cumulative sun exposure during leisure time or with 19 age- and sex-matched controls. Theholiday. Risk was increased with cumulative sun odds ratio for “office work” was 0.21 [95%CI:exposure in outdoor work during the summer 0.04–0.99; Fisher Exact 95%CIs calculated frommonths, but without any dose–response (OR, numbers in authors’ table]. Lee et al. (1994)11.7–12.7; with wide confidence intervals). The included 60 [probably prevalent] cases of ocularodds ratios were adjusted for cumulative alcohol surface epithelial dysplasia (13 were conjunctivaland tobacco intake and “leaving the cigarette on squamous cell carcinoma) diagnosed over 19the lip,” among other things. In a meta-analysis years (40% participation), and 60 age- and sex-of cancer in farmers (Acquavella et al., 1998), the matched hospital-based controls. Among others,pooled relative risk for cancer of the lip from 14 positive associations were observed betweenstudies was 1.95 (95%CI: 1.82–2.09) (P for heter- ocular surface epithelial dysplasia and historyogeneity among studies, 0.22). [The Working of solar keratoses [OR, for history at <  50 andGroup noted that given the relative risks for ≥ 50 years of age combined, 9.4 (95%CI: 2.8–31)]oesophageal cancer and lung cancer were 0.77 and duration of residence at ≤ 30° south latitudeand 0.65, respectively, confounding by smoking for 31–49 years (OR, 2.2; 95%CI: 0.6–8.3), andwas unlikely, but confounding with other farm- for 50 years or more (OR, 3.9; 95%CI: 1.0–14.8)related exposures could not be excluded.] relative to ≤  30 years. Cumulative years of life See Table 2.7 available at http://monographs. in which > 50% of daytime was spent outdoorsiarc.fr/ENG/Monographs/vol100D/100D-01- were similarly but more weakly associated withTable2.7.pdf and Table  2.8 available at http:// ocular surface epithelial dysplasia. Tulvatana etmono g r aph s . i a rc . f r/ E NG/ Mono g r aph s / al. (2003) studied 30 cases of conjunctival squa-vol100D/100D-01-Table2.8.pdf. mous cell neoplasia (intraepithelial or invasive) and 30 age- and sex-matched control patients2.1.4 Cancer of the eye having extracapsular cataract extraction from(a) Squamous cell carcinoma of the conjunctiva whom diseased conjunctiva was taken [site of biopsy not specified]. Solar elastosis [repre-(i) Descriptive studies senting pathologically proven solar damage] was Incidence of squamous cell carcinoma of the observed in the conjunctiva of 53% of cases andeye was inversely correlated with latitude across a 3% of controls, resulting in an odds ratio of 16.0wide range of countries (Newton et al., 1996), and (95%CI: 2.49–671). [The Working Group noteddirectly associated with measured ambient UVB that while pathologists were said to be “masked,”irradiance across the original nine Surveillance it was not stated that tissue sections from casesEpidemiology and End Results (SEER) cancer were free of neoplastic tissue.]registry areas of the USA (Sun et al., 1997). 43
    • IARC MONOGRAPHS – 100D In the only case–control study of exclusively choroidal and ciliary body melanomas, mappedconjunctival squamous cell carcinoma, Newton incident melanomas on the retina and observedet al. (2002) studied 60 Ugandan patients with that rates of occurrence were concentrated ina clinical diagnosis of conjunctival squamous the macula area, and decreased progressivelycell carcinoma and 1214 controls diagnosed with with increasing distance from the macula to theother cancers not known to be associated with ciliary body. It was concluded that this patternsolar UV exposure or infection with HIV, HPV was consistent with the dose distribution of lightor Kaposi Sarcoma herpesvirus. The risk for on the retinal sphere as estimated by Schwartz etconjunctival squamous cell carcinoma increased al. (1997).with “time spent cultivating”: with reference to (ii) Case–control and cohort studies0–9  hours a week, the odds ratios were 1.9 for10–19 hours and 2.4 for ≥  20 hours (P  =  0.05), Nine case–control studies and one cohortadjusted for age, sex, region of residence, HIV-1 study reported on associations of sun exposurestatus, and low personal income. Both HIV-1 with ocular melanoma (Gallagher et al., 1985;status and personal income were strong predic- Tucker et al., 1985; Holly et al., 1990; Seddontors of risk. et al., 1990; van Hees et al., 1994; Pane & Hirst, 2000; Håkansson et al., 2001; Vajdic et al., 2002;(b) Ocular melanoma Lutz et al., 2005 (incorporating also data from(i) Descriptive studies Guénel et al., 2001); and Schmidt-Pokrzywniak et al., 2009). In addition, one previously reported No increase in the incidence of ocular case–control study reported new analyses ofmelanoma was recorded by the US SEER occupation and ocular melanoma (Holly et al.,programme during 1974–98, which is in contrast 1996; Tables 2.8 and 2.9 on-line).with the increasing incidence of cutaneous Four studies (Gallagher et al., 1985; Hollymelanoma over the same period (Inskip et al., et al., 1990; Seddon et al., 1990; Tucker et al., 1985)2003). found an increased risk for ocular melanoma in Three studies have reported on the distri- people with light skin, light eye colour or lightbution of choroidal melanomas within the hair colour. Outdoor activities were associatedeye in relation to the presumed distribution of with ocular melanoma in one study (Tuckerchoroidal sun exposures across the choroid. et al., 1985).The first of these (Horn et al., 1994), which Four studies (Tucker et al., 1985; Seddonanalysed 414 choroidal, 20 ciliary body and 18 et al., 1990; Håkansson et al., 2001; Vajdic et al.,iris melanomas, concluded that choroidal and 2001, 2002) reported statistically significant asso-iris melanomas were located most frequently in ciations between a measure of sun exposure and“the areas that are presumably exposed to the ocular melanoma. Tucker et al. (1985) observedmost sunlight.” Specifically, melanomas in the an increased risk of ocular melanoma in peopleposterior choroid were observed to preferentially born in the south of the USA (south of 40°N) rela-involve the central area. The second (Schwartz tive to those born in the north (OR, 2.7; 95%CI:et al., 1997), which analysed 92 choroidal mela- 1.3–5.9), which appeared to be independent ofnomas, concluded that there was no preferential duration of residence in the south. Seddon et al.location for tumours on the choroid, having (1990) reported on two separate series of casesrigorously estimated “the average dose distribu- and controls. In the first series, increased riskstion on the retina received in outdoor daylight.” of uveal melanoma with residence in the south ofA third study (Li et al., 2000), which analysed 420 the USA were observed (OR, 2.4; 95%CI: 1.4–4.344
    • Solar and UV radiationfor up to 5  years; and OR, 2.8; 95%CI: 1.1–6.9 associations with measures of sun exposurefor more than 5 years). In the second series, the (Vajdic et al., 2002).risk increased with increasing years of “intensesun exposure” (OR, 1.5; 95%CI: 1.0–2.2 for 1–40 (c) Meta-analysesyears; and, OR, 2.1; 95%CI: 1.4–3.2 for >  40 Shah et al. (2005) and Weis et al. (2006)years); this association was only weakly present reported the results of meta-analyses of risk ofin the first series; the odds ratio for uveal mela- ocular melanoma in relation to sun sensitivitynoma with birthplace in the south of the USA characteristics and sun exposure, includingwas 0.2 (95%CI: 0.0–0.7), which was statistically both case–control and cohort studies (Table 2.10independent of the positive association between available at http://monographs.iarc.fr/ENG/duration of residence in the south and uveal Monographs/vol100D/100D-01-Table2.10.pdf).melanoma risk. Vajdic et al. (2001, 2002) found A fixed-effects model was used except whenthat the risk of choroid and ciliary body mela- statistically significant heterogeneity was foundnoma was increased in the highest categories between the effects of individual studies andof total sun exposure (OR, 1.6; 95%CI: 1.0–2.6), a random-effects model was used instead. Aweekdays sun exposure (OR, 1.8; 95%CI: 1.1–2.8), summary relative risk was reported only whenand occupational sun exposure (OR, 1.7; 95%CI: four or more studies were included in the anal-1.1–2.8); the underlying trends across quarters ysis. In the analysis by Shah et al. (2005), neitherof exposure were reasonably consistent and latitude of birth nor outside leisure was appre-statistically significant. These associations were ciably associated with ocular melanoma. Therelargely due to stronger associations confined to was weak evidence that occupational exposure tomen. Finally, the one cohort study (Håkansson the sun increased ocular melanoma risk (RR foret al., 2001), based in the Swedish construction highest exposed category, 1.37; 95%CI: 0.96–1.96).industry’s health service, observed an increasing [The Working group noted that this analysis didrisk of ocular melanoma with increasing occupa- not include results of Lutz et al. (2005) or Schmidt-tional sun exposure based on recorded job tasks Pokrzywniak et al. (2009), but included those of(RR, 1.4; 95%CI: 0.7–3.0, for medium sun expo- Guénel et al. (2001), which are a component ofsure; and, RR, 3.4; 95%CI: 1.1–10.5, for high sun Lutz et al. (2005). When the results of Lutz et al.exposure). (2005) are substituted for those of Guénel et al. Five of the case–control studies limited (2001) and those of Schmidt-Pokrzywniak et al.their study to uveal melanoma (melanoma in (2009) added to the fixed effects meta-analysis,the choroid, ciliary body, and iris), and one of the meta-RR is 1.25 (95%CI: 1.02–1.54).]these excluded iris melanoma because of small The meta-analysis of Weis et al. (2006)numbers. Two studies reported results for iris provides strong evidence that having blue or greymelanoma (Tucker et al., 1985; Vajdic et al., eyes, fair skin and/or burning easily rather than2002). One study observed odds ratios of 3–5 for tanning when exposed to the sun are associatediris melanoma with the use of an eye shade when with an increased risk of ocular melanoma. Hairoutdoors occasionally, rarely or never, relative colour was not associated with this cancer.to almost always (Tucker et al., 1985), and theother observed an increased risk of iris mela- 2.1.5 Other sitesnoma in farmers (OR, 3.5; 95%CI: 1.2–8.9; Vajdicet al., 2002). One study also reported results for Prompted at least in part by the hypothesesconjunctival melanoma, but found no positive arising from ecological studies, case–control and cohort studies have been conducted in 45
    • IARC MONOGRAPHS – 100Dwhich measures of personal exposure to solar for each quintile of exposure in each sex variedradiation (loosely referred to here as sun or from 0.9–1.1, and were not significantly increasedsunlight exposure) have been related to cancers (Kampman et al., 2000).in internal tissues (Table  2.11 available athttp://monographs.iarc.fr/ENG/Monographs/ (b) Cancer of the breastvol100D/100D-01-Table2.11.pdf and Table  2.12 Three case–control and two cohort studiesavailable at http://monographs.iarc.fr/ENG/ have examined the association between meas-Monographs/vol100D/100D-01-Table2.12.pdf). ures of sun exposure and breast cancer. InStudies that infer high sun exposure from a past three studies reporting results for sun expo-history of skin cancer (basal cell carcinoma, sure assessed from location of residence, onesquamous cell carcinoma or melanoma) were found slightly higher risks in women residingexcluded (see for example, Tuohimaa et al., 2007). in California (using ‘south’ as a reference; LadenIt has been argued in respect of these studies that et al., 1997); the other two studies found reduced“the incidence of second cancers in individuals is relative risks (0.73 and 0.74) with residence inelevated by several known and unknown mech- areas of high mean daily solar radiation (Johnanisms, including common etiological factors et al., 1999; Freedman et al., 2002), significantlyand predispositions, and influenced by possible so in one of these studies (Freedman et al., 2002).biases in the ascertainment of second cancers Sun-related behaviour was recorded in three[…] The net direction of these influences will studies (John et al., 1999; Freedman et al., 2002;mostly be in the direction of elevated occurrence Knight et al., 2007) and was inversely associatedof second cancers, against which a possible effect with risk for breast cancer for some measures.of sunlight and vitamin D […] could be difficult For example, the relative risks for breast cancerto detect.” (IARC, 2008). Thus, such studies are with frequent recreational and occupationalunlikely to be a reliable source of evidence for sun exposure relative to rare or no exposuredetermining whether sun exposure causes or were 0.66 (95%CI: 0.44–0.99) and 0.64 (95%CI:prevents any other cancers. 0.41, 0.98), respectively, in 5009 women from the NHANES Epidemiologic Follow-up Study(a) Cancer of the colorectum (John et al., 1999). For the highest category of Two case–control studies have related esti- estimated lifetime number of outdoor activitymates of individual sun exposure to risk of cancer episodes at 10–19 years of age, the odds ratioof the colorectum. Based solely on death certifi- was 0.65 (95%CI: 0.50–0.85) in a large Canadiancates, Freedman et al. (2002) observed a some- case–control study (Knight et al., 2007). In eachwhat reduced risk (OR, 0.73; 95%CI: 0.71–0.74) study, these effect measures were adjusted for awith high ambient sunlight in the state of resi- measure of socioeconomic status and some otherdence at the time of death, adjusted for age, sex, variables associated with breast cancer.race, occupational sun exposure (inferred fromusual occupation), physical activity, and socio- (c) Cancer of the ovaryeconomic status. In a large population-based In a case–control study, based on deathstudy in which participants were interviewed, certificates, the relative risk of cancer of the ovaryno appreciable association was found between was reduced in those residing in areas with highcancer of the colon and sun exposure recalled mean daily solar radiation (OR, 0.84; 95%CI:for each season for the 2 years before case diag- 0.81–0.88), but not in those with high occupa-nosis. With the exception of the second quintile tional sun exposure (Freedman et al., 2002).of exposure in women (OR, 1.3), the odds ratios46
    • Solar and UV radiation(d) Cancer of the prostate occupational title and, in one study, industry Four case–control studies (two hospital- (Freedman et al., 1997; Adami et al., 1999). Thebased) and one cohort study (John et al., 2004, results for residential exposure were conflicting:2007) examined the association between meas- one study, in the USA, found a reduced relative riskures of sun exposure and risk for cancer of the with residence at lower latitudes (Freedman et al.,prostate. In one case–control study conducted 1997); and the other, in Sweden, an increased riskin two consecutive periods and with patients (Adami et al., 1999). They concurred, however, inwith benign prostatic hypertrophy as controls, finding reduced relative risks in people with highthe odds ratio for prostate cancer with highest occupational sun exposure with values of 0.88lifetime sun exposure was [0.32 (95%CI: 0.20– (95%CI: 0.81–0.96) in the USA and 0.92 (95%CI:0.51); combined odds ratio calculated from two 0.88–0.97; combined result for men and women) inreported odds ratios]. Odds ratios were similarly Sweden. Subsequent studies focusing specificallylow with indirect measures of sun exposure, such on occupational sun exposure have not observedas regular foreign holidays or childhood sunburn a reduced risk of non-Hodgkin lymphoma with(Luscombe et al., 2001; Bodiwala et al., 2003). higher exposure (van Wijngaarden & Savitz,Two other studies showed weaker evidence of an 2001; Tavani et al., 2006; Karipidis et al., 2007).inverse association of residence in a high solar A study of non-Hodgkin lymphoma in childrenradiation environment with cancer of the pros- reported a reduced risk in those who had spenttate (Freedman et al., 2002; John et al., 2004, 15 or more days annually at seaside resorts, with2007). Outdoor occupation, self-reported recrea- an odds ratio of 0.60 (95%CI: 0.43–0.83; Petridoutional sun exposure, physician-assessed sun et al., 2007).exposure or actinic skin damage had no effect on All other studies (Hughes et al., 2004;prostate cancer risk in these studies. In a case– Smedby et al., 2005; Hartge et al., 2006; Sonicontrol study that included only cases of primary et al., 2007; Weihkopf et al., 2007; Zhang et al.,advanced cancer of the prostate (John et al., 2005), 2007; Boffetta et al., 2008; Kricker et al., 2008)a reduced risk for cancer of the prostate was were included in a pooled analysis of originalreported with high values of sun exposure index data from 8243 cases of non-Hodgkin lymphoma(based on comparison of the measured reflect- and 9697 controls in ten member studies of theance of usually exposed and usually unexposed InterLymph Consortium (Kricker et al., 2008;skin; OR, 0.51; 95%CI: 0.33–0.80), but with little Table 2.13 available at http://monographs.iarc.fr/evidence of similar associations with residential ENG/Monographs/vol100D/100D-01-Table2.13.ambient solar radiation or total or occupational pdf). [The Working Group noted that results onlifetime outdoor hours. sun exposure and non-Hodgkin lymphoma in three of these studies have not yet been published(e) Non-Hodgkin lymphoma and other separately.] In eight studies in which a composite lymphomas measure of total sun exposure (recreational plus non-recreational exposure) could be defined, the While some early, mainly ecological studies, pooled odds ratio fell weakly with increasing sunsuggested that sun exposure might increase risk exposure to 0.87 (95%CI: 0.71–1.05) in the fourthfor non-Hodgkin lymphoma, studies of indi- quarter of exposure. There was a steeper down-vidual sun exposure suggest that recreational trend for recreational exposure to an odds ratiosun exposure may decrease its risk. of 0.76 (95%CI: 0.63–0.91; P for trend, 0.005), and Two earlier studies in individuals assessed no appreciable downtrend for non-recreationalsunlight exposure based on place of residence, exposure. Physical activity and obesity, which 47
    • IARC MONOGRAPHS – 100Dmight be confounding, were not controlled for of exposure used in the different studies. Allin the analysis of any of the pooled studies. studies that examined age at first exposure found Four case–control studies have reported on the an increased risk for melanoma when exposureassociation between sun exposure and Hodgkin started before approximately 30 years of age, withlymphoma (Table  2.11 on-line); there was no a summary relative risk estimate of 1.75 (95%CI:consistent pattern of decreasing or increasing risk 1.35–2.26) (Table 2.14). The second meta-analysiswith different sun exposure measures (Smedby (Hirst et al., 2009) included an additional nestedet al., 2005; Petridou et al., 2007; Weihkopf et al., case–control study of melanoma (Han et al.,2007; Grandin et al., 2008). The same was true for 2006), bringing the total number of melanomamultiple myeloma in two case–control studies cases to 7855, and the summary relative risk for(Boffetta et al., 2008; Grandin et al., 2008). One melanoma in relation to ever versus never use ofstudy found weak evidence of an increased risk sunbeds was reported as 1.22 (95%CI: 1.07–1.39).of mycosis fungoides [a cutaneous lymphoma] Regarding basal cell carcinoma and squa-in people with high occupational sun exposure mous cell carcinoma, a meta-analysis of the three[OR: 1.3 (95%CI: 1.0–1.9; combined result for studies on ever use of indoor tanning facilitiesmen and women] (Morales-Suárez-Varela et al., versus never use showed an increased risk for2006). squamous cell carcinoma of 2.25 (95%CI: 1.08– 4.70) after adjustment for sun exposure or sun sensitivity (IARC, 2006a, 2007a). One study had2.2 Artificial UV radiation information on age at first exposure of indoor2.2.1 Use of artificial tanning devices tanning facilities and suggested that the risk (sunlamps, sunbeds, solaria) increased by 20% (OR, 1.2; 95%CI: 0.9–1.6) with each decade younger at first use. The four studies(a) Cutaneous melanoma, squamous cell on basal cell carcinoma did not support an asso- carcinoma, and basal cell carcinoma ciation with the use of indoor tanning facilities Two meta-analyses of skin cancer in rela- (IARC, 2006a, 2007a).tion to sunbed use have been undertaken overthe past few years (Table 2.14). The first (IARC, (b) Ocular melanoma2006a, 2007a) was based on 19 informative Four case–control studies have reportedpublished studies (18 case–control, of which explicitly on the association of artificial tanningnine population-based, and one cohort, all in devices and ocular melanoma (Tucker et al., 1985;light-skinned populations) that investigated the Seddon et al., 1990; Vajdic et al., 2004; Schmidt-association between indoor tanning and skin Pokrzywniak et al., 2009; Table 2.15). Odds ratioscancers, and included some 7355 melanoma for the highest exposure categories in each were:cases (Table  2.14). The characterization of the 2.1 (95%CI: 0.3–17.9) (Tucker et al., 1985); 3.4exposure was very varied across reports. The (95%CI: 1.1–10.3) and 2.3 (95%CI: 1.2–4.3) formeta-relative risk for ever versus never use of the population-based comparison and case–indoor tanning facilities from the 19 studies was sibling comparison, respectively (Seddon et al.,1.15 (95%CI: 1.00–1.31); results were essentially 1990); 1.9 (95%CI: 0.8–4.3) (Vajdic et al., 2004);unchanged when the analysis was restricted to and 1.3 to 2.1 depending on the control categorythe nine population-based case–control studies (Schmidt-Pokrzywniak et al., 2009). The onlyand the cohort study. A dose–response model study to analyse dose–response found evidencewas not considered because of the heterogeneity of increasing risk with increasing durationamong the categories of duration and frequency of use (P  =  0.04) and, less strongly, estimated48
    • Table 2.14 Meta-analyses of use of artificial tanning devices and skin cancers Reference, Study description Number of Exposure Exposure Relative risk Adjustment Comments study cases and assessment categories (95%CI) for potential location & controls confounders period IARC 18 case–control Cutaneous All studies except Indoor tanning Melanoma All analyses One study presented (2007a) studies (10 pop- melanoma: two presented Never use 1.0 adjusted for results for men and Europe, based) and 1 7355 cases and estimates for ever Ever use 1.15 (1.00–1.31) the maximum women separately; north cohort published 11275 controls versus never of potential Dose–response was Age first use America and from case– confounders not considered because in 1981–2005, with Never 1.0 Australia of the heterogeneity exposure assessment control studies; < 35 yr 1.75 (1.35–2.26) 1971 to 2001 cohort: 106379 among the categories of to indoor tanning members BCC duration and frequency BCC-SCC (No. Never use 1.0 of exposure between of cases not Ever use 1.0 (0.6–1.9) studies. stated) from 5 SCC case–control Never use 1.0 studies Ever use 2.25 (1.1–4.7) Hirst et al. 18 case–control Cutaneous Indoor tanning Melanoma One study presented (2009) studies and 2 melanoma: 7885 Never use 1.0 results for men and Europe, nested-cohort cases and 24209 women separately; No Ever use 1.22 (1.07–1.39) north studies published controls from summary risk estimate America and all studies BCC/SCC for BCC or SCC in 1981–2006, with Australia exposure assessment BCC/SCC: 1812 Never use 1.0 separately 1971 to 2002 to indoor tanning cases and 2493 Ever use controls from 1.34 (1.05–1.70) 6 case–control studies BCC, basal cell carcinoma; SCC, squamous cell carcinoma. Solar and UV radiation49
    • 50 Table 2.15 Case–control studies of exposure to artificial tanning devices and ocular melanoma Reference, study Cases Controls Exposure assessment Exposure Relative risk Adjustment Comments location and categories (95% CI) for potential period cofounders Tucker et al. 439 White patients 419 White Telephone interview with Sunlamp use Age, eye colour (1985), USA, with intraocular patients with detailed information about Never 1.0 and history of 1974–79 melanoma detached medical history, family cataract Rarely 1.3 (0.8–2.3) confirmed retina not due history, employment, histologically or to tumours; exposure to environmental Occasionally 1.3 (0.5–3.6) from highly reliable matched by age, agents, sunlight; details Frequently 2.1 (0.3–17.9) IARC MONOGRAPHS – 100D ancillary studies; sex, race, date from ophthalmologic participation rate, of diagnosis; examination and medical 89% participation history abstracted from rate, 85% medical records; interview with next-of-kin for 17% of cases and 14% of controls, half of them with spouses Seddon et White patients Series 1: selected Telephone interview Case–control Age, eye and *Series 1: al. (1990), with clinically by random digit including constitutional series 1* skin colour, population- Massachusetts, or histologically dialing, matched factors, ocular and medical Sunlamp use moles, ancestry, based, 197 USA, 1984–87 confirmed 2:1 by sex, age, histories, and exposure eye protection, cases and melanoma of city of residence, to environmental factors Never 1.0 outside work, 385 controls; the choroid, 85% response including natural and Rarely 0.7 (0.4–1.4) fluorescent Series 2: not ciliary body or rate artificial sources of UV Occasionally or 3.4 (1.1–10.3) lighting, population- both, identified Series 2: living frequently southern based, 337 at local hospital sibling of cases, residence, cases and Case–control or by mailing to up to 4 siblings yr of intense 800 sibling series 2* ophthalmologists, per case, median, exposure controls. diagnosed within 2; participation Sunlamp use 140 cases were previous yr; age rate, 97% Never 1.0 included in range, 17–88 Rarely 0.9 (0.6–1.4) both series. yr, mean, 57 yr; Occasionally or 2.3 (1.2–4.3) participation rate, frequently 89% (see comments)
    • Table 2.15 (continued) Reference, study Cases Controls Exposure assessment Exposure Relative risk Adjustment Comments location and categories (95% CI) for potential period cofounders Vajdic et 246 White 893 controls Self-administered Sunlamp use* Age, sex, *Sunlamps use al. (2004), Australian matched 3:1 questionnaire, and Never 1.0 place of birth, includes use of Australia, residents, aged by age, sex, telephone interview Ever 1.7 (1.0–2.8) eye colour, sunbeds and 1996–98 18–79 yr, with residence, regarding sun exposure, ability to tan, tanning booths Duration of use histopathologically selected from sun-protective wear and squinting as a or clinically electoral rolls; quantitative exposure to ≤ 1 mo 1.2 (0.5–2.8) child and total diagnosed participation welding equipment and 2 mo to 1 yr 1.8 (0.8–3.9) personal sun melanoma rate, 47% sunlamps > 1 yr 2.3 (0.9–5.6) exposure at 10, originating in the Lifetime hours 20, 30 and 40 yr choroid, ciliary of use of age body; participation 0.1–1.4 1.3 (0.5–3.2) rate, 87% among 1.5–7.8 1.8 (0.8–4.2) those eligible > 7.8 1.9 (0.8–4.3) Period of first use < 1980 1.4 (0.7–2.7) 1980–90 2.0 (0.8–4.7) > 1990 4.3 (0.7–27.9) Age at first use > 20 yr 1.5 (0.8–2.6) ≤ 20 yr 2.4 (1.0–6.1) † Schmidt- 459 cases of Control 1: 827 Self-administered Regular sunlamp Results Pokrzywniak incident primary population- postal questionnaire use presented for et al. (2009), uveal melanoma based, selected and computer-assisted No 1.0 population Germany, diagnosed at 1 from mandatory telephone interview Yes 1.3 (0.9–1.8) controls. Odds 2002–05 clinic, aged 20–74 list of residence, ratios with Age at first use yr matched 2:1 on sibling controls age (5-yr age Never used 1.0 were somewhat groups), sex and > 20 yr 1.3 (0.9–1.9) higher, but region < 20 yr 1.7 (0.8–3.6) with wider Control 2: 187 confidence sibling controls, intervals and matched 1:1 not significant; by (+/− 10 yr) *Sunlamps use and sex when includes use of possible sunbeds and tanning booths yr, year or years Solar and UV radiation51
    • IARC MONOGRAPHS – 100Dcumulative time of exposure (P = 0.06) (Vajdic 2.2.2 Weldinget al., 2004). The two most recent studies (Vajdicet al., 2004; Schmidt-Pokrzywniak et al., 2009) Six separate case–control studies (sevencalculated odds ratios for exposure that started reports) and one meta-analysis have reported onat or before 20 years of age and after this age; associations between welding and risk of ocularin both, the odds ratio was greater for exposure melanoma (Table 2.17). All studies reported anstarting at the younger age. The results of Seddon odds ratio for ocular melanoma above unity inet al. (1990) and Vajdic et al. (2004) were adjusted most categories of exposure to welding. Seddonfor sun sensitivity and personal sun exposure. et al. (1990) reported on two sets of cases and[The Working Group noted that Schmidt- controls and found an increased risk in only onePokrzywniak et al. (2009) found little evidence of of them. Lutz et al. (2005) found an increasedassociations between measures of personal sun risk with a “history of at least 6 months’ employ-exposure and ocular melanoma.] ment in welding or sheet metal work,” but not for “working with welding”; the increase observed(c) Internal cancers was restricted to the French component of the study, which Guénel et al. (2001) had previously Five case–control studies (Table 2.16) have reported. The strongest associations of weldingreported on the association of the use of artifi- with ocular melanoma (although based on smallcial tanning devices and cancer of the breast (one numbers) were reported in those studies thatstudy), non-Hodgkin lymphoma (four studies), restricted the exposure definition to “work as aHodgkin lymphoma (three studies), multiple welder,” i.e. not including being in proximity tomyeloma (two studies), and lymphoprolifera- welding (Tucker et al., 1985; Siemiatycki, 1991;tive syndrome (one study) (Smedby et al., 2005; Guénel et al., 2001; Lutz et al., 2005). SeveralHartge et al., 2006; Knight et al., 2007; Boffetta studies showed evidence of dose–response rela-et al., 2008; Grandin et al., 2008). In all the tionships (Holly et al., 1996; Guénel et al., 2001;studies of non-Hodgkin lymphoma, the risk was Vajdic et al., 2004) with duration of employmentlower in people who had used artificial tanning or of use.devices than in those who had not; in two there The meta-analysis (Shah et al., 2005) estimatedwas also a dose–response relationship across a meta-relative risk of 2.05 (95%CI: 1.20–3.51)exposure categories with a P value for trend of for welding, using a random-effects model. [The≤ 0.01 (Smedby et al., 2005; Boffetta et al., 2008). Working Group noted that this study includedOdds ratios were also below unity for cancer of results from Ajani et al. (1992), which overlapthe breast (Knight et al., 2007) and for Hodgkin with those from case–control Series 1 of Seddonlymphoma (Smedby et al., 2005; Boffetta et al., et al. (1990), and did not include those from the2008), with a significant dose–response relation- case–control Series 2 of Seddon et al. (1990). Itship (P value for trend  =  0.004) in one study also did not include results from Siemiatyckiof Hodgkin lymphoma (Smedby et al., 2005). (1991).]Confounding with exposure to natural sunlightcannot be ruled out as an explanation for theseinverse relationships because none of the studies 2.3 UVA, UVB, and UVCadjusted the results for sun exposure. Epidemiology has little capacity to distinguish between the carcinogenic effects of UVA, UVB, and UVC. UVC is not present in natural sunlight at the surface of the earth and is therefore not52
    • Solar and UV radiationrelevant; in almost all circumstances humans are increasing use of sunscreens blocking UVB andexposed simultaneously to UVB and UVA, and the increasing risk of melanoma.UVB and UVA exposures vary more or less in Moan et al. (1999) plotted the relationships ofparallel (see Section 1). Several epidemiological UVB and UVA irradiances and incidence ratesapproaches have been used in an attempt to of cutaneous basal cell carcinoma, squamousdistinguish the effects of UVA and UVB on skin cell carcinoma and melanoma using data fromcancer risk. Their major focus has been to assess Australia, Canada, the Czech Republic, Denmark,whether solar UVA exposure contributes to the Finland, Iceland, Norway, New Zealand, Sweden,increased risk of cutaneous melanoma, for which Scotland, USA, and the United Kingdom. Asthere is some conflicting evidence in experimental expected, all were inversely related to latitudestudies (see Section 4). These include studies on but the slope of the fitted linear relationship wasexposure to UVA for artificial tanning, effect of numerically smaller for UVA than for UVB, andsunscreens on melanoma risk, and UVB photo- for melanoma than for basal cell carcinoma andtherapy without associated exposure to PUVA squamous cell carcinoma. Estimates of biological(psoralen-UVA photochemotherapy). amplification factors (relative increase in risk per PUVA is the combination of psoralen with unit increase in exposure) based on these slopesUVA radiation, and is used in the treatment of for UVB were, in men and women respectively,psoriasis. PUVA has been reviewed previously by 2.8 and 2.8 for basal cell carcinoma, 3.1 and 2.9two IARC Working Groups and there is sufficient for squamous cell carcinoma, and 1.3 and 1.0 forevidence that PUVA therapy is carcinogenic to melanoma. Those for UVA and melanoma werehumans (Group 1), causing cutaneous squamous 3.8 and 2.9, respectively, suggesting that UVAcell carcinoma (IARC, 1986, 2012), and these may play a significant role in the induction ofstudies will not be reviewed here. melanomas.2.3.1 Descriptive studies 2.3.2 Exposure to artificial UVA for tanning Garland et al. (1993) noted that “rising purposestrends in the incidence of and mortality from Early artificial tanning devices emitted bothmelanoma have continued since the 1970s and UVB and UVA. UVB emissions were subse-1980s, when sunscreens with high sun protec- quently reduced relative to UVA, presumably totion factors became widely used.” They related reduce skin cancer risk, but have been increasedthis observation to the fact that commonly used again recently to mimick the sun and to producechemical sunscreens had blocked UVB but not longer lasting tans (see Section 1). In principleUVA; and the possibility that by preventing these periods of different relative exposures toerythema, sunscreens would permit extended UVA and UVB during artificial tanning couldsun exposure and thus substantially increase be used to evaluate the relative effects of UVAexposure to UVA. However, nearly half of the and UVB on skin cancer risk. Veierød et al.melanoma mortality increase between 1950–54 (2003, 2004) attempted this analysis in a cohortand 1990–94 in the USA in white men and more study of Norwegian and Swedish women whothan half of that in white women had occurred had reported their use of a sunbed or sunlampby 1970–74, with only a minor upward pertur- (solarium) in different age periods on entry tobation in the trend after 1970–74. Thus, there the cohort. They defined three subgroups ofprobably was not a close association between women: those who had used solaria in the period 1963–83 (mainly before they became mainly 53
    • 54 Table 2.16 Associations of use of artificial tanning devices with cancers of internal tissuesa Reference, study Cancer type Exposure Exposure categories Relative risk location and period assessment Knight et al. (2007), Breast cancer Telephone interview Ever sunlamp use Canada, 2003–04 Age 10–19 No 1.0 Yes 0.81 (0.57–1.14) Age 20–29 No 1.0 IARC MONOGRAPHS – 100D Yes 0.88 (0.66–1.18) Age 45–54 No 1.0 Yes 0.84 (0.64–1.11) Hartge et al. (2006), Non-Hodgkin Self-administered Use of sunlamp or tanning booth USA, 1998–2000 lymphoma questionnaire and Never 1.0 computer assisted Ever 0.88 (0.66–1.19) personal interview Only after age 20 0.97 (0.69–1.37) Before age 20 0.72 (0.45–1.14) < 5 times 0.78 (0.46–1.32) 5–9 times 0.90(0.52–1.58) 10+ times 0.90 (0.61–1.30) Smedby et al. (2005), Non-Hodgkin Computer assisted Solaria/sun lamp use Denmark and Sweden, lymphoma telephone interview Never 1.0 1999–2002 < 10 times 1.0 (0.9–1.2) 10–49 times 0.9 (0.8–1.0) 50 times or more 0.8 (0.7–1.0) Hodgkin lymphoma Never 1.0 < 10 times 0.8 (0.6–1.0) 10–49 times 0.7 (0.5–0.9) 50 times or more 0.7 (0.5–0.9)
    • Table 2.16 (continued) Reference, study Cancer type Exposure Exposure categories Relative risk location and period assessment Boffetta et al. (2008), Non-Hodgkin Interviewer Sunlamp use France, Germany, lymphoma administered Never 1.0 Ireland, Italy, and questionnaire 1–24 times 0.79 (0.59–1.04) Spain, 1998–2004 25 times or more 0.69 (0.51–0.93) Hodgkin lymphoma Never 1.0 1–24 times 0.86 (0.53–1.39) 25 times or more 0.93 (0.57–1.50) Multiple myeloma Never 1.0 1–24 times 0.76 (0.41–1.41) 25 times or more 1.10 (0.59–2.05) Grandin et al. (2008), Non-Hodgkin Self and interviewer Aesthetic use of artificial UV radiation France, 2000–04 lymphoma administered No 1.0 questionnaires Yes 1.1 (0.7–1.7) Regularly 0.5 (0.2–1.3) Occasionally 1.4 (0.8–2.3) Hodgkin lymphoma No 1.0 Yes 1.6 (0.7–3.6) Regularly 0.6 (0.1–3.3) Occasionally 2.2 (0.9–5.5) Lymphoproliferative No 1.0 syndrome Yes 1.5 (0.7–3.5) Regularly 0.9 (0.2–4.6) Occasionally 1.9 (0.7–4.7) Multiple myeloma No 1.0 Yes 1.2 (0.4–3.6) Regularly 0.8 (0.1–7.3) Occasionally 1.4 (0.4–4.9) a In none of these studies was potential confounding with exposure to natural sunlight controlled in the analysis yr; year or years Solar and UV radiation55
    • 56 Table 2.17 Case–control studies on welding and ocular melanoma Reference, Cases Controls Exposure assessment Exposure Relative risk Adjustment Comments study location categories for potential and period confounders Tucker et al. 439 White patients 419 White patients Telephone interview with Ever worked as a Age, eye colour (1985), USA, with intraocular with detached detailed information welder and history of 1974–79 melanoma confirmed retina not due to about medical history, No 1.0 cataract histologically or tumours; matched family history, Yes 10.9 (2.1–56.5) from highly reliable by age, sex, race, employment, exposure ancillary studies; date of diagnosis; to environmental IARC MONOGRAPHS – 100D participation rate, participation rate, agents, sunlight; details 89% 85% from ophthalmologic examination and medical history abstracted from medical records; interview with next-of- kin for 17% of cases and 14% of controls, half of them with spouses Seddon et White patients Series 1: selected Telephone interview Case–control Age, eye Series 1: al. (1990), with clinically by random digit including constitutional series 1 and skin population- Massachusetts, or histologically dialing, matched factors, ocular and Exposure to colour, moles, based, 197 cases USA, 1984–87 confirmed melanoma 2:1 by sex, age, city medical histories, welding arc ancestry, use and 385 controls of the choroid, of residence, 85% and exposure to No 1.0 of sunlamps, Series 2: not ciliary body or response rate environmental factors eye protection, population- Yes 1.3 (0.5–3.1) both, identified Series 2: living including natural and outside work, based, 337 cases at local hospital sibling of cases, up artificial sources of UV Case–control fluorescent and 800 sibling or by mailing to to 4 siblings per series 2 lighting, controls. ophthalmologists, case, median, 2; Exposure to southern 140 cases were diagnosed within participation rate, welding arc residence, included in both previous yr; age 97% No 1.0 yr of intense series. range, 17–88 Yes 0.9 (0.6–1.5) exposure Result for case yr, mean, 57 yr; series 1 also participation rate, was reported 89% (see comments) by Ajani et al. (1992) using the same numbers but with fewer covariates in the logistic regression model (see below).
    • Table 2.17 (continued) Reference, Cases Controls Exposure assessment Exposure Relative risk Adjustment Comments study location categories for potential and period confounders Siemiatycki [33] incident male 533 population Personal interview and Occupational Age, family 4 exposed cases (1991), cases of uveal controls; collection of detailed exposure to arc income, Montreal, melanoma, aged 35– participation rate, occupational history welding fumes ethnicity, Canada, 70 yr, histologically 72% No 1.0 respondent 1979–85 confirmed; response Yes 8.3 (2.5–27.1) type, cigarette rate, 69.6% and alcohol indexes Ajani et al. 197 White 385 controls Telephone interview with Exposure to Age, ancestry, Same population (1992), USA, patients with selected by occupational history and welding arc skin colour, as in study by 1984–87 uveal melanoma, random digit exposures related to work No 1.0 moles, use of Seddon et al. histologically dialling, matched occurring 15 yr before Yes 0.99 (0.48 sunlamps, past (1990) in case confirmed, diagnosed 2:1 for age (+/− 8 the interview. –2.05) income level series 1 using the during the previous yr), sex, telephone same numbers yr, residents of 6 exchange; mean but with more New England States; age, 58.3 yr, range covariates in mean age, 59.2 yr, 19–88 yr; response the logistic range 18–88 yr; rate, 85% regression model participation rate, (see above). 92% Holly et al. 221 male White 447 controls Interviewer administered Welding* Age, number * Self welding or (1996), USA, patients with selected by questionnaire with No 1.0 of large nevi, in proximity to 1978–87 histologically random digit demographic and Yes 2.2 (1.3–3.5) eye colour, others for > 3 h a confirmed uveal dialling, matched phenotypic caracteristics, tanning or wk for > 6 mo Years from start melanoma, age 20–74 2:1 by age (5-yr occupational history, burning of occupation yr residing in 11 age group) and exposure to chemicals. response to to diagnosis or States; participation residential area; 30 min. sun interview rate, 93% interview rate, exposure in 77% ≤ 10 1.2 (0.2–6.6) the summer 11–29 1.5 (0.7–3.0) noond sun ≥ 30 2.1 (1.1–4.0) Solar and UV radiation57
    • 58 Table 2.17 (continued) Reference, Cases Controls Exposure assessment Exposure Relative risk Adjustment Comments study location categories for potential and period confounders Guénel et al. 50 cases (29 men and 479 (321 men, Face-to-face interview, Worked for six Age Data also (2001), France 21 women) identified 158 women) or occasionally telephone mo or more as a included in 1995–96 from records of controls selected interview welder or sheet analysis of Lutz local pathology from electoral metal worker et al. (2005). departments for rolls, frequency No 1.0 Results shown surgery, and from matched by age here for men 2 cancer treatment (5-yr interval), sex Yes 7.3 (2.6–20.1) only; only one IARC MONOGRAPHS – 100D centres in France; and study area; Duration of woman in diagnosis confirmed participation rate, employment as a this study had by pathologists or 76% welder worked as a ophthalomogic Less than 20 yr 5.7 (1.6–19.8) welder and she report; participation was a case. rate, 100% 20 yr or more 11.5 (2.4–55.5) Vajdic et 246 White Australian 893 controls Self-administered Own welding Age, sex, al. (2004), residents, aged matched 3:1 by questionnaire, and Never 1.0 place of birth, Australia, 18–79 yr, with age, sex, residence, telephone interview eye colour, Ever 1.2 (0.8–1.7) 1996–98 histopathologically selected from regarding sun exposure, ability to tan, or clinically electoral rolls; sun-protective wear and Duration of use squinting as a diagnosed melanoma participation rate, quantitative exposure to 0.1–4.0 yr 0.8 (0.4–1.4) child and total originating in the 47% welding equipment and 4.1 to 22.0 yr 1.2 (0.7–2.2) personal sun choroid, ciliary body; sunlamps > 22 yr 1.7 (1.0–2.7) exposure at 10, participation rate, 20, 30, and 40 87% among those Lifetime hours yr of age eligible of use 0.1–52.0 1.1 (0.6–1.9) 52.1–858.0 1.4 (0.8–2.3) > 858 1.1 (0.6–1.9) Age at first use > 20 yr 1.2 (0.8–1.9) ≤ 20 yr 1.2 (0.7–1.9)
    • Table 2.17 (continued) Reference, Cases Controls Exposure assessment Exposure Relative risk Adjustment Comments study location categories for potential and period confounders Lutz et al. 292 incident cases 2062 population Questionnaire with Worked for six Data from (2005), of uveal melanoma, controls selected face-to-face or telephone mo or more as a France reported Denmark, identified from from population interview welder or sheet in analysis of Latvia, France, ophthalmologic registers, metal worker Guénel et al. Germany, departments, hospital electoral rolls No 1.0 (2001). Results Italy, Sweden, records or cancer or practitioner, Yes 2.2 (1.2–4.0) shown here for Portugal, registries aged 35–69 frequency men only; only Working with Spain, and yr; participation rate, matched by one woman in welding the United 91% region, sex and this study had Kingdom, 5-yr birth cohorts; No 1.0 worked as a 1995–96 participation Yes 0.9 (0.6–1.5) welder and was rate, 61%; 1094 a case. cancer controls randomly selected from colon cancer patients; participation rate, 86% d, day or days; h, hour or hours; min, minute or minutes; mo, month or months; wk, week or weeks; yr, year or years Solar and UV radiation59
    • IARC MONOGRAPHS – 100DUVA-emitting), the period 1979–91 (mainly <  0.0001 (Gorham et al., 2007). [The Workingafter solaria were designed to emit mainly UVA) Group noted that although these observationsor the period 1975–87 (covering both catego- might be explained by a lack of effectiveness ofries of solarium) when they were 20–29 years early sunscreens against higher wavelengths ofof age. The odds ratios for solarium use in these UVA, there are other possible, and probably moresubgroups were 3.75 (95%CI: 1.73–8.13) for use plausible, explanations. First, there is undoubtedin 1963–83, 3.19 (95%CI: 1.22–8.32) for use in positive confounding between sunscreen use and1979–91, and 1.28 (95%CI: 0.46–3.60) for use sun exposure, and probably also sun sensitivity.in 1975–87. These results show little difference Although this confounding can, in principle, bebetween those exposed in the earlier and later dealt with by adjustment for sun exposure andperiods of solarium use. [The Working Group sun sensitivity in multiple variable models of thenoted that only seven cases of melanoma were association of sunscreen use with melanoma risk,observed in each of these periods, and there was inaccurate measurement of these confounderslittle statistical power to see a difference.] A recent limits the ability of modelling to control theirmeta-analysis of use of artificial tanning devices confounding. Thus, residual confounding couldand skin cancer (IARC, 2007b) reported that the easily explain the lack of protective effect ofrelative risks of melanoma associated with ever sunscreens seen in observational studies ofuse of a sunbed or sunlamp did not vary with melanoma (IARC, 2001). Second, there is clearyear of publication of a study or the first year of evidence of adaptation to the use of sunscreensa study period, where available. [The Working such that people who apply sunscreens beforeGroup noted that the most relevant time metric outdoor recreation may increase their dura-would be year of first reported use of a sunbed or tion of exposure to the sun (Autier et al., 2007)sunlamp, rather than the year of publication or so that their dose of erythemal UV radiationfirst year of study period.] may not change. Thus, observed associations of sunscreens with risk of melanoma (or other skin2.3.3 Use of sunscreens and risk for cancers) in observational studies do not provide melanoma useful information regarding the relative effects of UVB and UVA on cancer risk.] Initially, sunscreens contained only UVBabsorbers; more recently they have covered a 2.3.4 UVB phototherapybroader spectrum with the addition of UVAreflectors or absorbers, although many are still UVB phototherapy is used to treat a varietyless effective against the higher wavelengths of of skin conditions. Lee et al. (2005) reviewedUVA than they are against UVB (see Section the literature and concluded that there was no1). Recent meta-analyses of published observa- evidence of an increased risk of skin cancer intional studies of sunscreen and melanoma, each those who had received UVB phototherapyincluding slightly different subsets of studies, as their only form of UV phototherapy. [Thehave found meta-relative risks close to unity with Working Group noted that only three cases ofhighly significant heterogeneity among studies: melanoma were identified among about 10001.11 (95%CI: 0.37–3.32) with a P value for hetero- who had received this therapy.]geneity < 0.001 (Huncharek & Kupelnick, 2002); Lim & Stern (2005) extended follow-up of1.0 (95%CI: 0.8–1.2) with a P value for hetero- 1380 patients with severe psoriasis who had beengeneity <  0.001 (Dennis et al., 2003); and 1.2 treated with variations of PUVA, methotrexate,(95%CI: 0.9–1.6) with a P value for heterogeneity UVB, topical tar, and ionizing radiation. In60
    • Solar and UV radiationpatients who had less than 100 PUVA treatments, 2.4 Synthesisthe incidence rate ratio for cutaneous squamouscell carcinoma with ≥ 300 UVB treatments was 2.4.1 Solar radiation0.81 (95%CI: 0.34–1.93) for chronically sun- In Caucasian populations, both basal cellexposed sites, and 2.75 (95%CI: 1.11–6.84) for carcinoma and squamous cell carcinoma arerarely to intermittently sun-exposed sites. The strongly associated with solar radiation, as meas-corresponding values for basal cell carcinoma ured by indicators of accumulated solar skinwere 1.38 (95%CI: 0.80–2.39) for chronically sun- damage (e.g. increasing age, especially for squa-exposed sites and 3.00 (95%CI: 1.30–6.91) for mous cell carcinoma; and presence of actinicintermittently sun-exposed sites. [The Working keratoses), and secondarily by recalled episodesGroup noted that the possibility that the observed of acute solar skin damage (multiple sunburns).effect required interaction with PUVA or another The causal association of cutaneous mela-treatment for psoriasis cannot be ruled out in this noma and solar exposure is established, thisstudy.] Hearn et al. (2008) described the results link has become clearer in the last decade orof follow-up of 3867 patients who had received so through the observation of the site-specificnarrow-band UVB phototherapy, a quarter of heterogeneity of melanoma, the lower-than-whom had also received PUVA. In comparison average phenotypic risk for skin carcinogenesiswith data from the Scottish Cancer Registry, among outdoor workers, and the recognitionthere were near 2-fold increases in the risk of first that the different associations of melanoma withsquamous cell carcinoma [two observed cases] sun exposure observed among Caucasian peopleand of first basal cell carcinoma [14 observed at different latitudes around the world correlatecases] for treatment with narrow-band UVB with marked variations in sun exposure oppor-only, but their 95% confidence intervals included tunity and behaviour.unity. For melanoma, the relative risk was just Five case–control studies of cancer of thebelow 1. For those who had more than 100 UVB lip have been published. The three earliesttherapy treatments, the risks, relative to those studies found apparent increases in risk withwho received 25 or less such treatments, were 1.22 outdoor work, but use of tobacco could not be(95%CI: 0.28–4.25) for basal cell carcinoma, 2.04 ruled out as an explanation for these associa-(95%CI: 0.17–17.8) for squamous cell carcinoma, tions. The two later studies both took account ofand 1.02 (95%CI: 0.02–12.7) for melanoma. Two possible confounding of outdoor exposure withprevious small studies of narrow-band UVB, tobacco smoke. One of them, in women, showedof 126 (Weischer et al., 2004) and 484 patients increased risks for cancer of the lip with several(Black & Gavin, 2006), observed only one skin measures of exposure, together with strong andcancer between them, an in-situ melanoma, in moderately consistent dose–response relation-less than 10 years of follow-up. ships. The other, in men, found no increase in Given the few cases of skin cancer so far risk with leisure time or holiday sun exposurereported in patients given UVB phototherapy as but a substantial increase in risk with cumula-their only form of phototherapy, the statistical tive exposure during outdoor work during thepower of currently available studies to detect summer months, without any indication ofother than a large increase in relative risk of any dose–response across four categories. This lacktype of skin cancer with this therapy, and, there- of dose–response suggests bias rather than afore, of UVB specifically is weak. causal effect. 61
    • IARC MONOGRAPHS – 100D Four case–control studies reported at least small number of conjunctival melanomas foundone result each suggesting that sun exposure no such association.is associated with conjunctival intraepithelial The associations of sun exposure with severalneoplasia or squamous cell carcinoma of the internal cancers have been investigated in case–eye. Only one study was exclusively of conjunc- control and cohort studies, generally with thetival squamous cell carcinoma; in this study hypothesis that sun exposure might be protectiveand another, the relevant exposure variables against such cancers. The cancers investigated(office work and cultivating the fields) were only included cancer of the colorectum (two studies),indirect measures of sun exposure. A very large of the breast (five studies), of the ovary (onedifference between cases and controls in preva- study), of the prostate (four studies), and severallence of conjunctival solar elastosis in another cancers of the lymphatic tissue, principally non-study raised concerns about possible bias. The Hodgkin lymphoma and Hodgkin disease (15remaining study reported a strong association studies). Exposure metrics used in these studiesof ocular surface dysplasia with solar keratoses included residential or occupational ambientand increasing risk with increasing duration of solar radiation, recreational or non-recreationalresidence at ≤ 30° south latitude. However, only sun exposure, recent and lifetime sun exposure,22% of its cases had conjunctival squamous cell and sun-related behaviour. The results werecarcinoma. mostly inconsistent. Two out of three studies that examined thedistribution of choroidal melanomas found 2.4.2 Artificial sources of UVthem to be concentrated in the central area orthe macula area of the choroid, which coincides (a) Tanning applianceswith the estimated distribution of light in the Two meta-analyses investigated the associa-retinal sphere. Of ten case–control studies of tion between indoor tanning and skin cancers.ocular melanoma published from 1985 to 2009, The summary relative risk for ever versusfour reported statistically significant associa- never use of indoor tanning facilities wastions of one or more measures of sun exposure significantly increased for melanoma, with nowith ocular melanoma. In two studies, these consistent evidence for a dose–response relation-associations were with the latitude of birth or of ship. All studies that examined age at first expo-residence in early life, with some inconsistency sure found an increased risk for melanoma whenbetween them. In the other two, which were more exposure started before approximately 30 yearsrecent and had better measures of exposure than of age, with a summary relative risk estimate ofmany previous studies, one study related only to 1.75.occupational sun exposure and showed a strong For squamous cell carcinoma, the threeassociation with a dose–response relationship, available studies found some evidence for anand the strongest association seen in the other increased risk, especially when age at first usewas with occupational sun exposure and showed was below 20 years. Studies on basal cell carci-evidence of a dose–response relationship. These noma did not support an association with use ofresults relate principally to choroid and ciliary indoor tanning facilities.body melanomas (the dominant types). Two Four case–control studies reported on asso-studies reported results consistent with a posi- ciations between artificial tanning devices andtive association of small numbers of iris mela- ocular melanoma. Each observed an increase innomas with sun exposure. One study with a risk of ocular melanoma in the highest category of exposure to these devices, and there were62
    • Solar and UV radiationindications of a dose–response relationship in with the relative contributions of UVB and UVAthree of the studies. In two studies, the risk was emitted from the devices. There is little or nohigher in people who began exposure before 20 evidence to suggest that the use of sunscreensyears of age than those who began after this age. that block mainly UVB radiation increased thePossible confounding with natural sun exposure risk for melanoma. Studies of patients exposedwas explicitly addressed in two of the studies. exclusively to UVB phototherapy show weak Five studies reported on the association evidence of an increase in risk of squamous cellof use of indoor tanning devices with internal carcinoma and basal cell carcinoma, based on acancers, specifically breast cancer, non-Hodgkin few cases.lymphoma, Hodgkin lymphoma, and multiplemyeloma. Most studies found little evidence ofan association. Two studies observed inverse 3. Cancer in Experimental Animalsassociations between the use of internal tanningdevices and non-Hodgkin lymphoma, and The previous IARC Monograph on solar andone study with Hodgkin lymphoma. Possible ultraviolet radiation concluded that there wasconfounding with exposure to natural sunlight sufficient evidence for the carcinogenicity of solarcannot be ruled out in any of these studies. radiation, broad-spectrum ultraviolet radiation, ultraviolet A, ultraviolet B and ultraviolet C(b) Welding radiation in experimental animals (IARC, 1992). Six case–control studies reported on the asso- The experimental induction of skin cancers inciation between welding and ocular melanoma. mice following exposure to a mercury-arc lampAll found evidence of a positive association, was first reported by Findlay (1928). Initially,which was strong in three studies, each of which haired albino mice were used, but hairless Skh-1related specifically to working as a welder or sheet (albino) and Skh-2 (pigmented) immunocompe-metal worker (other studies included working in tent mice and eventually immunodeficient nudeproximity to welding in the definition of expo- mice or transgenic mice are now used.sure). In each of three studies in which it was Hundreds of studies have clearly establishedexamined, there was evidence of a dose–response the carcinogenic activity of UVR in mice. Therelationship. action spectrum for ultraviolet-induced skin carcinogenesis in albino hairless mice has been2.4.3 UVA, UVB, UVC determined and shows a peak in the UVB range (280–315 nm) and a steep decrease in the UVA Several sources of evidence were examined range (315–400 nm). However, while the induc-to see if the carcinogenic effects of UVA and tion of non-melanoma skin cancer is regularlyUVB could be distinguished: descriptive studies obtained in mice, the induction of melanomaof skin cancer have shown that the slope of lati- was exceptional.tude variation in incidence of melanoma is less Solar radiation was tested for carcinogenicitythan that in incidence of squamous cell carci- in a series of studies in mice and rats. Largenoma and basal cell carcinoma, suggesting that numbers of animals were studied (600 rats andmelanoma incidence is more influenced by UVA 2000 rats and mice), and incidences of squamous-irradiance than are squamous cell carcinoma and cell carcinoma of the skin and of the conjunctivabasal cell carcinoma. Present data on the risk for were clearly increased in most of the survivingmelanoma associated with the of UV-emitting mice and rats (Roffo, 1934, 1939; IARC, 1992).tanning devices show little evidence that it varies 63
    • IARC MONOGRAPHS – 100D Broad-spectrum UVR (solar-simulated radi- cocarcinogenicity and even tumour inhibition.ation and ultraviolet lamps emitting in the entire Chemical immunosuppressive agents have beenUV wavelength range) was tested for carcino- shown to enhance the probability of developinggenicity in two large studies in mice (Grady et al., UVR-induced tumours in mice (IARC, 1992).1943; Blum, 1959; IARC, 1992), several studies in Studies released since the previous Monographrats, and one study in hamsters and guinea-pigs are summarized below.(Freeman & Knox, 1964; IARC, 1992). Incidencesof squamous-cell carcinoma of the skin and ofthe cornea/conjunctiva were clearly increased in 3.1 Non-melanoma skin cancerrats and mice. Hamsters developed malignant See Table 3.1tumours of the cornea. No eye tumours wereobserved in guinea-pigs. 3.1.1 Mouse In several studies in mice exposed to sourcesemitting mainly UVA radiation, squamous-cell Most of the recent studies were not designedcarcinomas of the skin were clearly induced. to test whether or not the radiation used wasBoth short-wavelength UVA (UVA2, 315–340 carcinogenic per se but to investigate the processnm) and long-wavelength UVA (UVA1, 340–400 of UV carcinogenesis, or to test enhancementnm) were effective (IARC, 1992). or inhibition of photocarcinogenicity by drugs In several studies in mice exposed to sources and chemical agents. Methods for testing photo-emitting mainly UVB radiation, the predomi- carcinogenicity have been standardized to meetnant tumour type was squamous-cell carcinoma the requirements of regulatory agencies (Forbesof the skin. Skin papillomas were observed in one et al., 2003; Sambuco et al., 2003).study in rats and one study in hamsters. Invasive Recent studies have mainly focused on themelanomas were induced in two experiments mechanisms of UV-induced carcinogenesisin platyfish-swordtail hybrid fish. In two out of and have used specific strains of mice. Sencarthree studies in opossums (Monodelphis domes- mice were derived by selective breeding fortica), squamous-cell carcinomas were shown to susceptibility to chemical carcinogens. Theydevelop; in one of these three studies, malig- are more sensitive than other mouse strains tonant tumours of the cornea were observed and a variety of chemical initiators and promotersmelanocytic neoplasms of the skin were reported (e.g. 7,12-dimethylbenz(a)anthracene (DMBA)in another one (IARC, 1992). and 12-o-tetradecanoylphorbol-13-acetate In some studies in mice exposed to sources (TPA)) as well as to UV radiation. Sencar miceemitting mainly UVC radiation, squamous-cell have been widely used to study multistage skincarcinomas of the skin were clearly induced. carcinogenesis. Using these mice, squamous cellIn one study in rats, keratoacanthomas of the carcinomas (SCCs) and malignant spindle cellskin were observed. In none of the experi- tumours (SCTs) appeared within 16-18 weeksments involving UVC was it possible to exclude and 30 weeks of irradiation respectively (Tongcompletely a contribution of UVB, but the size of et al., 1997, 1998). Tong et al. (1997, 1998) havethe effects observed indicate that they cannot be also shown that alterations in the Tp53 gene aredue to UVB alone (IARC, 1992). frequent events in SCCs induced by chronic UV UVR has been studied in protocols involving exposure in Sencar mouse skin, and that over-two-stage chemical carcinogenesis. UVR has expression of H-Ras-p21 in conjunction withbeen reported to exert many effects on the carci- aberrant expression of keratine K13 is a frequentnogenic process, including initiation, promotion, event in UVR-induced SCCs in Sencar mouse64
    • Table 3.1 Non-melanoma skin cancers induced in mice and opossum exposed to ultraviolet radiation Species, strain Animals/group at start Results Comments (sex) Exposure regimen: radiation type, dose, dose rate Target organ Duration Incidence and/or multiplicity of Reference tumours (%) Mice, Sencar (F) 63, 10 controls Among all 73 tumours biopsied: SCCs begin to appear 18 wk after up to 60 wk exposed to UVR from Dermalight 2001 sun lamp (3 ×/wk, 4% papillomas, initiation of UV irradiation. Tong et al. (1997, 8 min each time, for 18 wk). Total UVB dose = 139.2 J/m 2 54% SCCs, Among the 8 mutations, 3/8 (38%) were 1998) UVR-induced skin tumours biopsied when 1.5 × 1.5 mm C → T changes (codons 146 and 158)-a 36% spindle cell tumours (SCT), for histological examination, immunohistochemical typical “UV-signature” mutation-and detection of p53, Hras-p21 and keratin K13 expression, 6% dermal fibromas and BCCs. 3/8 (38%) were C → A changes (codons and DNA isolation. Skin biopsies from untreated control Tp53 mutations (exon 5) in 10/37 (27%) 150 and 193), which is also a frequent mice. of SCCs and 12/24 (50%) of SCTs mutation pattern induced by UVR. Hras-p21 expressed in 24/36 (67%) of SCCs but not in normal skin SCTs or UV-exposed skin. Co-expression with K13 in 47% SCCs. Mice, Tg.AC (F) 10 animals/group Papillomas develop from 4 wk in a UV-induced tumours harbour few Tp53 20 wk 3 exposures (every other d) on shaved back dose dependent manner, that progress mutations. In contrast, UV-exposed skin Trempus et al. 2.6 to 43.6 kJ/m 2 (cumulative). to malignancy in the high UV show Tp53 activation in the basal layer. (1998) FS40T12 sunlamp (60% UVB, 40% UVA, total output 1.6 exposure groups: mW/cm 2) - 21.8 kJ/m2: 6/10 (60%) mice with SCC or SCT at 23–30 wk - 43.6 kJ/m2: 5/9 (55%) mice with SCC or SCT at 18–30 wk Mice, PKCε number/group at start (NR) SCC develop earlier and more PKCε overexpression sensitizes skin transgenic exposed to UVR (2 kJ/m2) from a bank of 6 Kodacell frequently in transgenic mice than in to UVR-induced cutaneous damage FVN/B starins filtered FS40 sunlamps, 3 ×/wk, up to 38 wk. normal littermates. and development of squamous cell 215, 224 Up to 60% mice developed SCC by carcinoma possibly at the promotion sex and duration 38 wk. step of carcinogenesis, and this is (NR) probably accomplished by promoting Wheeler et al. the enhanced induction and release of (2004, 2005) specific cytokines such as TNFα. Solar and UV radiation65
    • 66 Table 3.1 (continued) Species, strain Animals/group at start Results Comments (sex) Exposure regimen: radiation type, dose, dose rate Target organ Duration Incidence and/or multiplicity of Reference tumours (%) Mice, Skh-1 (F) 15 animals/group 2.4-fold increase in yield of tumours Low (non toxic) concentrations of 26 wk 1.7 kJ/m2 solar UVR (mostly UVB) 3 ×/wk, ± 10 mg/L in mice given arsenite compared with arsenite can enhance the onset and Rossman et al. sodium arsenite in drinking-water for 26 wk mice given UVR alone. growth of malignant skin tumours (2002) Tumors (mostly SCCs) appeared only induced by a low (non erythemic) dose of in UVR treated mice, and only on the UVR in mice. exposed area (backs) of the mice. Tumors occurring in mice given UVR IARC MONOGRAPHS – 100D plus arsenite appeared earlier (time to first tumour < 60 d vs > 80 d after UVR exposure alone) and were much larger than in mice given UVR alone. Mice, Skh-1 (F) Number at start (NR) More than 95% of the tumours were This study was designed to establish 182 d Mice were fed sodium arsenite continuously in drinking- SCCs. dose–response relationship for cancer Burns et al. water starting at 21 d of age at concentrations of 0.0, 1.25, Only UVR irradiated mice developed enhancement in a new mouse skin model (2004) 2.5, 5.0, and 10 mg/L. At 42 d of age, solar spectrum UVR locally invasive SCCs. using arsenite in drinking-water in exposures were applied every other d (3 ×/wk) to the Mice exposed only to UVR: 2.4 ± 0.5 combination with chronic topical UVR dorsal skin at 1.0 kJ/m 2 per exposure until the experiment cancers/mouse at 182 d. exposures ended at 182 d. Arsenite enhanced the UVR-induced Arsenite alone and UVR alone induced cancer yield in a linear pattern up to a epidermal hyperplasia, but the combined peak of 11.1 ± 1.0 cancers/mouse at 5.0 exposures have a greater than additive mg/L arsenite (i.e. peak enhancement effect. ratio: 4.63 ± 1.05). A decline occurred 50% cancer incidence occurred at 140 d to 6.8 ± 0.8 cancers/mouse at 10.0 in the UVR only group, whereas in the mg/L arsenite. highest response group (UVR plus 5.0 mg/L), 50% incidence occurred at 109 d.
    • Table 3.1 (continued) Species, strain Animals/group at start Results Comments (sex) Exposure regimen: radiation type, dose, dose rate Target organ Duration Incidence and/or multiplicity of Reference tumours (%) Mice, SK1-hrBD 15-30 animals/group ~95% of the tumours were SCCs. The first tumour appeared in mice (F) Weanling mice were exposed to solar spectrum UVR alone Few papillomas, fibrosarcomas and exposed to UVR + arsenite at 10 wk after 6 mo (1 kJ/m2  3 x/wk) or to UVR + sodium arsenite (5 mg/L in premalignant hyperplasias were also beginning UVR exposure, whereas the Uddin et al. drinking-water) and fed laboratory chow supplemented seen. first tumour in mice exposed to UVR (2005) or not with Vitamin E (α-tocopheryl acetate, 62.5 IU/kg Average tumour/mouse: alone appeared after 12 wk of UVR diet) or organoselenium (1,4-phenylenebis(methylene) UVR–3.60 exposure. selenocyanate (p-XSC), 10 mg/kg diet) for 26 wk. Mice exposed to UVR plus arsenite UVR + Vitamin E–2.53 exhibited an enhanced tumour yield UVR + p-XSC–3.33 (1.94-fold) compared with mice exposed UVR + arsenite–7.0 to UVR alone. UVR + arsenite + Vit E–3.27 Vitamin E and p-XSC reduce tumour yield in mice given UVR + arsenite (2.1 UVR + arsenite + p-XSC–3.40 and 2.0 fold respectively). Vitamin E but not p-XSC reduces tumour yield induced by UVR alone Mice, SK1-hrBR 12-19 animals/group No tumour in untreated mice and Proportion of malignant tumours per (F) Animals were exposed to: mice treated with chromium alone. mouse: 182 d - UVR alone (1.2 kJ/m 2, from 3 FS 20 and 1 F-20T12-BL Dose-dependent increase in the - UV alone: 0.55. Davidson et al. lamps; 85% UVB, 4% UVA), number of skin tumours (SCCs > 2 - UVR + 5 ppm K 2CrO4: 0.73 (2004) - K 2CrO4 alone (2.5 and 5.0 ppm in drinking-water), mm) in mice exposed to K 2CrO4 and - or combination of UVR + K 2CrO4 (0.5, 2.5, and 5.0 ppm). UV compared with mice exposed to Exposure to UV started 1 mo after the initial chromate UV alone: 2.63 and 5.02 tumours/ exposure, 3 ×/wk (every other d) for the first 3 mo, then 2 x mouse for 2.5 and 5.0 ppm K 2CrO4 vs /wk (Monday and Wednesday) for 3 further mo. 0.8 (P < 0.05). Solar and UV radiation67
    • 68 Table 3.1 (continued) Species, strain Animals/group at start Results Comments (sex) Exposure regimen: radiation type, dose, dose rate Target organ Duration Incidence and/or multiplicity of Reference tumours (%) Mice, SK1-hrBD 10 animals/group 96% of the tumours were SCCs and 4% Chromium and nickel significantly (M, F) Weanling mice were exposed to: were papillomas increase the UVR-induced skin cancer 6 mo - UVR (1.0 kJ/m 2, 3 ×/wk) for 26 wk, Cancers/mouse: yield in mice. Uddin et al. - UVR + 2.5 or 5.0 ppm potassium chromate, - Male: Chromate caused a more rapid cancer (2007) - UVR + 20, 100 or 500 ppm nickel chloride in drinking- induction (percentage of mice with UVR: 1.9 ± 0.4 water. cancer) in mice given UVR plus IARC MONOGRAPHS – 100D Vitamin E or selenomethionine was added to the UVR + 2.5 ppm K 2CrO4: 5.9 ± 0.8 chromate: laboratory chow for 29 wk beginning 3 wk before the start UVR + 5 ppm K 2CrO4: 8.6 ± 0.9 at 18 wk of UVR exposure, 50% of mice of UVR exposure. - Female given UVR alone developed at least one UVR: 1.7 ± 0.4 cancer compared to 80% of mice given UVR + 20 ppm NiCl2: 2.8 ± 0.9 UVR + 2.5 ppm chromate and 100% of mice given UVR + 5.0 ppm chromate. UVR + 100 ppm NiCl2: 5.6 ± 0.7 Final cancer incidence: – UVR: 80% UVR + 500 ppm NiCl2: 4.2 ± 1.0 - UVR + chromate (2.5 and 5.0 ppm): 100%. Neither vitamin E nor selenomethionine reduced the cancer yield enhancement by chromium. Mice Skh:HR-1 15 animals/group First tumours appear hairless (F) Animals were: - by d 84 in mice fed the highest 232 d - pre-fed for 4 wk on diets designed to provide 20% polyunsaturated fat, Reeve et al. by weight of fat, comprising 0.5%, l%, 15% or 20% - by d 113 in mice fed the lowest (1996) polyunsaturated sunflower oil (balance: hydrogenated polyunsaturated fat. cottonseed oil), CHS reactions in those groups - exposed to an incremental SSUV radiation regime for 10 supporting the highest tumour loads wk, 5 d per wk, cumulative doses 111 J/m2 UVB and 2 106 (fed 15% or 20% polyunsaturated kJ/m2 UVA. fat), were significantly suppressed in Feeding of the prepared diets continued until d 232 comparison with the mice bearing from commencement of the UV irradiation, when the smaller tumour loads (fed 0.5% or 10% experiment was terminated. polyunsaturated fat).
    • Table 3.1 (continued) Species, strain Animals/group at start Results Comments (sex) Exposure regimen: radiation type, dose, dose rate Target organ Duration Incidence and/or multiplicity of Reference tumours (%) Opossum M. 32-62 animals/group; 31 controls Corneal tumours develop in nearly The South American opossum domestica (M, F) Shaved or unshaved animals exposed to 250–500 J/m 2, 3 x/ 100% animals. Monodelphis domestica possesses 12–24 mo wk, from a bank of FS40 lamps (280 to 400 nm), rate 4 W/ 154 tumours examined histologically: a photolyase enzyme that catalyses Sabourin et al. m2, for ≈1 yr. - 134 fibrosarcomas, the monomerization of UV-induced (1993), Immediately after UV irradiation, half of the animals pyrimidine dimers in DNA. UVR - 18 haemangiosarcomas Kusewitt et al. are exposed to visible light (60 or 90 minutes) to remove effects reduced by photoreactivation (2000) pyrimidine dimers by photoreactivation. - 2 squamous cell carcinomas can be attributed to pyrimidine dimers Controls exposed to photoreactivating light. overlaying sarcomas formation. To prevent photoreactivation, animals are maintained Post-UVR exposure to under red light. photoreactivating light delays the onset of eye tumours and reduces overall tumour incidence BCCs, Basal Cell Carcinomas; CHS, Contact Hypersensitivity; d, day or days; h, hour or hours; min, minute or minutes; mo, month or months; NR, not reported; SCCs, Squamous Cell Carcinomas; SCTs, Spindle Cell Tumours; SSUV, simulated solar UVR; TNFα, tumour necrosis factor α; vs, versus; wk, week or weeks; yr, year or years Solar and UV radiation69
    • IARC MONOGRAPHS – 100Dskin. Using the v-Ha-ras transgenic Tg.AC corneal tumours; post-UVR exposure to photo-mouse line, sensitive to tumour promoters, reactivating light delays the onset of eye tumoursTrempus et al. (1998) have shown that SCCs and and reduces overall tumour incidence (SabourinSCTs developed within 18-30 weeks following et al., 1993, Kusewitt et al., 2000).the initial UVR exposure and that in contrastto other mouse stains used in photocarcinogen-esis studies, few Tp53 mutations were found in 3.2 MelanomaTg.AC UV-induced skin tumours, although all 3.2.1 Transgenic mice exposed to ultravioletTg.AC tumours express the v-Ha-ras transgene. radiationOther strains of transgenic mice, FVN/B strains215 and 224, which overexpress protein kinase C See Table 3.2epsilon (PKCε) and are highly susceptible to the In the mouse, wild-type animals are resistantinduction of skin tumours by chemical carcino- to malignant melanoma (MM) development evengens, also show increased susceptibility to the when exposed to repeated treatments with ultra-induction of skin tumours by UVR. PKCε trans- violet radiation. Chronic UVR treatment regi-genic mice were observed to be highly sensitive mens, however, have increased MM penetranceto the development of papilloma-independent by up to 26% in mice carrying various transgenesmetastatic squamous cell carcinomas elicited by capable of inducing spontaneous MM develop-repeated exposure to UVR (Wheeler et al., 2004, ment, or melanocytic hyperplasia.2005). In studies using Skh-1 mice, exposure to Inbred lines of transgenic Tyr-SV40EUVR induced a statistically significant increase mice, having an integrated recombinant genein the number of malignant skin tumours per comprised of the tyrosinase promoter, expressedmouse, mainly SCCs when compared to controls in pigment cells, and the simian virus 40 early-(Rossman et al., 2002; Burns et al., 2004; Davidson region transforming sequences spontaneouslyet al., 2004; Uddin et al., 2005, 2007). Dietary develop ocular and cutaneous melanomas (Bradlpolyunsaturated fat enhances the development et al., 1991). UVB irradiation of 2–4-day oldof UVR-induced tumours in Skh-1 mice, this Tyr-SV40E transgenic mice of either moderateenhancement being mediated by a modulation of or low susceptibility lines induce skin melanomathe immunosuppression caused by chronic UV (Klein-Szanto et al., 1994; Kelsall & Mintz, 1998).irradiation (Reeve et al., 1996). The pigment-producing cells in TPras trans- genic mice express a mutated human T-24 Ha-ras3.1.2 Opossum (Monodelphis domestica) driven by a 2.5 kb promoter region from the mouse tyrosinase gene. The ras transgenic mice Unlike laboratory rodents, a small marsu- exhibit an altered phenotype, including melano-pial, the South American opossum Monodelphis cytic hyperplasia and a muted agouti coat, indic-domestica possesses the ability to remove ative of hyperproliferative melanocytes. Topicalcyclobutane-pyrimidine dimers by photoreacti- 7,12-dimethylbenz[a]anthracene (DMBA) treat-vation, a light-dependent process of enzymatic ment of TPras mice resulted in a high incidencemonomerization. M. domestica is sensitive to of melanomas. UV light exposures inducedUVR, and, when photoreactivation is prevented, papillomas in TPras-negative littermate anddevelops primary tumours of the skin and eye melanomas in some albino TPras mice (Broomein response to chronic exposure to low doses Powell et al., 1999). When Hacker et al. (2005)of UVR. Virtually all M. domestica chronically treated brown mice (mixed C3H/Sv129 strainexposed to low doses of UVR develop primary background) carrying a melanocyte-specific70
    • Table 3.2 Melanomas induced in transgenic mice exposed to ultraviolet radiation Species, strain (sex) Animals/group at start Results Comments Duration Exposure regimen: radiation type, dose, Target organ Reference dose rate Incidence and/or multiplicity of tumours (%) Mice, C57Bl/6 Tyr- 19, 11 controls Melanocytic lesions resembling macules, Eye melanomas develop before any skin SV40E, moderately Exposed to 328 mJ/cm2 UVB (20 min/d) for nevi, or early melanomas gradually melanomas and are fatal in young mice of susceptible line 9 (M, up to 4 consecutive d. appeared in the irradiated mice (not in the more susceptible lines; less susceptible F) unirradiated transgenic controls of similar mice have much later onset eye tumours and Duration (NR) age). 20 wk after irradiation, skin samples longer lives. Skin melanomas were obtained Klein-Szanto et al. containing 26 selected lesions were grafted in the absence of advanced eye melanomas (1994) to low susceptibility line 12 mice. by grafting skin from high susceptibility - 10/26 selected lesions in 7 of the grafts gave (unirradiated) donors to low susceptibility rise to melanomas hosts, thereby greatly prolonging the life of - all melanomas had ulcerated and two had the donor skin. metastasized. Mice, C57Bl/6 112, 71 controls 14 melanomas in 80 (18%) mice surviving at Among the 80 transgenic survivors, Tyr-SV40E, low Exposed on each of 3–10 d to 0.22–0.42 J/ 4 wk, latency: 37–115 wk, metastases in 5/14 40% of the mice had from one to four susceptibility line 12 cm2 UV radiation from F40 sunlamps (65% (35%) mice keratoacanthomas on the tail. Most arose (M, F) UVB), totaling 1.1–3.7 J/cm2 6–8 mo after UVR; one-fifth of the lesions The most favourable protocol (1.9 J/cm2 Duration (NR) (8 protocols) regressed spontaneously in 8–20 mo after total UVR, at 0.38 J/cm2/d for 5 d starting Kelsall & Mintz (1998) Controls: non transgenic C57BL/6 mice. detection. Keratoacanthomas also arose on at 3 d of age) led to the highest incidence of the tails of 4 of the group of 16 surviving melanoma, 5 of 19 (26%) mice and one of the C57BL/6 nontransgenic controls treated with lowest mortality rates, 2 of 19 (10%). UVR. Mice, TPras (M, F) 10 animals/group, 18 controls (TPras- Melanocytic naevi and melanomas develop The TPras mice that developed melanoma 45 wk negative littermates) in 20% irradiated mice. had an albino coat colour Broome Powell et al. - Irradiated 2 x/wk for 38 wk (1999) from FS40T12 UVB lamps (> 90% UVB), - Initial dose 5.6 kJ/m 2, increased twice by 20%, up to a total final dose of 8.06 kJ/m2. Mice C3H/Sv129, 10-18, 42 controls UVR irradiated mice (n = 14) developed TPras (M, F) Exposed to a single total dose of 8.15 kJ/m 2 in situ cutaneous MM with a penetrance Duration (NR) from FS40 lamps (UVA 320–400 nm, 2.36 of 57% by 12 mo, none of the untreated Hacker et al. (2005) kJ/m2 UVB 280–320 nm, 5.77 kJ/m2, UVC controls (n = 42) developed tumours 250–280 nm, 0.02 kJ/m 2) Solar and UV radiation71
    • 72 Table 3.2 (continued) Species, strain (sex) Animals/group at start Results Comments Duration Exposure regimen: radiation type, dose, Target organ Reference dose rate Incidence and/or multiplicity of tumours (%) Mice C3H/Sv129 2–3 d old Cdk4R24C/R24C mice did not develop Lesions were mainly dermal melanomas, Cdk4/TPras (M, F) Cdk4 R24C/R24C , melanoma, spontaneously or after neonatal often multicentric, usually accompanied Duration (NR) Cdk4 R24C/R24C /TPras, UVR. TPras mice developed neonatal UVR- by epidermal hyperplasia in UVR treated Hacker et al. (2006) Cdk4 R24C/+/TPras mice induced, but not spontaneous, melanomas. animals. Exposed to a single total dose of 8.15 kJ/m 2 The increased melanoma susceptibility in 58% of mice homozygous for the Cdk4-R24C from FS40 lamps mice carrying both activated Cdk4 and Hras IARC MONOGRAPHS – 100D mutation and also carrying the melanocyte- is underlined by their increased propensity specific activated Hras (Cdk4R24C/R24C /TPras) to develop multiple primary melanomas. All developed melanoma spontaneously. melanoma-bearing UVR-treated Cdk4R24C/ UVR treatments increased the penetrance R24C /TPras animals developed more than of tumour development to 83% (and from one primary lesion, significantly more than 0% to 40% in Cdk4R24C/+/TPras mice) and untreated Cdk4R24C/R24C /TPras mice (40%, decreased the age of onset compared with P = 0.012) or UVR-treated TPras mice (16%, untreated animals. P = 0.001). Mice, albino FVB, Number/group at start (NR) Only mice from groups A and C developed The second UV exposure increased the HGF/SF (M, F) UV-irradiated at: – group A, 3.5 d and melanoma. multiplicity of melanocytic lesions as well as 13 mo again at 6 wk; No melanoma in non-transgenic or the incidence of non-melanocytic tumours. Noonan et al. (2001) – group B, 6 wk; untreated transgenic mice (observation: 13 – group C, 3.5 d; – group D, no UV mo). treatment. Melanoma development in HGF/SF Neonatal mice received a single treatment transgenic mice after UV irradiation at both of 9.58 kJ/m2 from Phillips F40 UV lamps 3.5 d and 6 wk (group A) identical to that (UV-A, 320–400 nm, 3.31 kJ/m2; UV-B, seen after only a single exposure at 3.5 d 280–320 nm, 6.24 kJ/m 2; UV-C, 250–280 (group C). UV irradiation (group B) was not nm, 0.03 kJ/m2). tumorigenic. 6-wk-old mice received a single treatment of 19.16 kJ/m2.
    • Table 3.2 (continued) Species, strain (sex) Animals/group at start Results Comments Duration Exposure regimen: radiation type, dose, Target organ Reference dose rate Incidence and/or multiplicity of tumours (%) Mice, albino FVB, Number/group at start (NR) Incidence of MM UVB highly effective at initiating melanoma. HGF/SF (M, F) Neonatal HGF/SF-transgenic mice Xenon UVB filter: 10/18 A further group of animals was irradiated 14 mo irradiated at 3–5 d of age with a source unfiltered F40 lamp: 6/23 with 4.5 kJ/m2 of UVB (7 SED), which was De Fabo et al. (2004) coupling UV interference or cutoff filters to also effective at initiating melanoma (not solar simulator: 5/29 a 2.5 kW xenon arc lamp, to produce UVB shown). UVA radiation did not initiate any or UVA wavebands or solar simulating mylar filtered F40: 1/20 melanomas. radiation (UVB + UVA + visible radiation Xenon UVA filter: 0/23 Removal of UVB from the broadband F40 in proportions approximating sunlight). source prevented the initiation of melanoma. Neonatal transgenic animals also Median time to first melanoma (d): irradiated with F40 sunlamps, (UVB + Xenon UVB filter: 127 UVA radiation and visible light). unfiltered F40 lamp: 169 Total UVvis doses (kJ/m2): solar simulator: 284 Xenon UVB filter: 14.0 unfiltered F40 lamp: 14.7 solar simulator: 322.1 mylar filtered F40: 14.1 Xenon UVA filter: 150 (UVB and solar simulator doses are equivalent to 23 standard erythemal doses) UVB dose 14.0 kJ/m2 UVA dose 150 kJ/m2 Mice XPA (−/−), SCF- Number/group at start (NR) 55% of UV-treated XPA (−/−), SCF-Tg mice Tg Irradiated 3 x/wk for 10 wk, 5 J/cm 2 UVB develop melanoma at 70 wk after UVB 24 mo (total dose: 150 J/cm2), from FL.20SE.30; radiation. Yamazaki et al. (2005) fluorescent lamps (55% radiation within Lentigo maligna melanoma appear 4 mo the UVB range (305 nm), 25% and less than after the termination of UVR exposure. 1%, within UVA and UVC, respectively). At 6 mo, some mice developed nodular melanoma. No melanoma develop in UV-treated XPA- normal, SCF-Tg mice and non-treated XPA -(−/−), SCF-Tg mice. d, day or days; F, female; h, hour or hours; M, male; min, minute or minutes; MM, malignanat melonoma; mo, month or months; NR, not reported; SED, standard erythemal dose; wk, week or weeks; yr, year or years Solar and UV radiation73
    • IARC MONOGRAPHS – 100Dmutant Hras (G12V) transgene (TPras), with a activating gene-1 (RAG-1) knockout micesingle neonatal UVR dose of (8.15 kJ m2), 57% of (Atillasoy et al., 1998). Chronic UVB irradiationthe UV irradiated mice developed in situ cuta- with or without an initiating carcinogen canneous MM by 12 months, whereas none of the induce human melanocytic lesions, includinguntreated controls developed tumours. In another melanoma. It was further shown that overex-study by the same author, UVR treatment greatly pression of basic fibroblast growth factor (bFGF)increased the penetrance and decreased the age via adenoviral gene transfer in human skin xeno-of onset of melanoma development in Cdk4R24C/ grafted to severe combined immunodeficiencyR24C /TPras animals compared with TPras alone mice led to black pigmented macules within 3(Hacker et al., 2006). weeks of treatment, and to melanoma when bFGF However, murine melanocytic tumours are was combined with UVB (Berking et al., 2001).dermal in origin and lack the epidermal compo- In contrast with experiments using neonatalnent that characterizes human melanoma. foreskin, no melanocytic lesions were inducedHowever, the skin of transgenic mice in which a when adult skin was used (Berking et al., 2002).metallothionein-gene promoter forces the over- In normal human skin grafted onto severeexpression of hepatocyte growth factor/scatter combined immunodeficient mice (SCID), anfactor (HGF/SF) has melanocytes in the dermis, increased expression of a combination of threeepidermis and dermal–epidermal junction, and growth factors, bFGF, stem cell factor, andis thus more akin to human skin. Untreated endothelin-3, along with exposure to UVB canHGF/SF-transgenic mice are already genetically transform normal melanocytes into a melanomapredisposed to late-onset melanoma. Using these phenotype within 4 weeks. Invasion of mela-transgenic mice, Noonan et al. (2001) showed that noma lesions was found in skin from newborna single UV irradiation of neonates is sufficient donors, whereas melanomas in adult skin wereto induce early onset melanoma in the majority of a non-invasive in situ type only. This suggestsof animals, while UV irradiation of 6-week-old that susceptibility of skin to exogenous tumourmice is insufficient. Using the same model, it was promoters is dependent on age (Berking et al.,further shown that UVB and not UVA is effec- 2004).tive at initiating melanoma (De Fabo et al., 2004). Xeroderma pigmentosum group A gene-defi- 3.2.3 Opossumscient (XPA–/–), stem cell factor-transgenic (SCF-Tg) mice are defective in the repair of damaged See Table 3.4DNA and do have epidermal melanocytes. Chronic UVB irradiation of suckling youngFollowing chronic UVB irradiation, these mice opossums (M. domestica) induces nevi anddevelop lentigo maligna and nodular melanomas melanoma that progress to metastasis (Robinson(Yamazaki et al., 2005). et al., 1994, 1998) suggesting that in this species, UVB can act as a complete carcinogen, inducing precursor lesions and driving progression to3.2.2 Human melanocytes grafted to metastatic melanoma. immunodeficient mice exposed to ultraviolet radiation 3.2.4 Fish See Table 3.3 See Table 3.5 Atilasoy et al., have developed an experi- Interspecies hybrids and backcrosses of plat-mental model in which full-thickness human yfish (Xiphophorus maculatus) and swordtailsskin is grafted to immunodeficient recombinase74
    • Table 3.3 Melanomas induced in human melanocytes grafted to immunodeficient mice exposed to ultraviolet radiation Species, strain (sex) Animals/group at start Results Comments Duration Exposure regimen: radiation type, dose, dose rate Target organ Reference Incidence and/or multiplicity of tumours (%) Mice, RAG-1 (M, F) Number/group at start (NR) 9/40 (23%) of human foreskin grafts Duration (NR) 8–12 wk old mice grafted with full-thickness human treated with UVB only, and 18/48 (38%) Atillasoy et al. (1998) foreskin. of grafts treated with the combination 4 groups: after 4-6 wk: of DMBA + UVB developed solar no treatment, a single treatment with 7,12- dimethyl(a) lentigines within 5 to 10 mo. benzanthracene (DMBA), 73% of of all UVB-treated xenografts UVB irradiation at 500 J/m 2 alone 3 ×/wk, and a develop melanocytic hyperplasia combination of DMBA and UVB. 1 melanoma (nodular type) out of 48 DMBA+UVB treated xenografted mice. Mice, SCID (M, F) Number/group at start (NR) 1 lentiginous melanoma in an adult Duration (NR) Human skin xenograft injected intradermally with abdominal skin graft after 2 mo (7 Berking et al. (2001) adenoviral vector bFGF/Ad5, exposed 3 ×/wk for 10 min to bFGF/Ad5 injections and 26 UVB 30–50 mJ/cm2 UVB from FS72/T12 UVB lamps throughout irradiations). a period of 2 to 10 mo. Mice, SCID and RAG-1 155 adult human skin specimens grafted onto SCID or Only actinic keratoses and 1 squamous Melanocytes from young (M, F) RAG-1 mice, irradiated 2–3 ×/wk with 40 mJ/cm 2 UVB cell carcinoma. individuals may be Up to 22 mo over a period of up to 10 mo with or without beforepical No melanocytic lesions. more susceptible to the Berking et al. (2002) application of DMBA. transforming effect of genotoxic agents than melanocytes from adults. Mice, SCID (M, F) Human skin xenografts (neonatal foreskin or adult skin) 17/50 invasive melanomas in newborn Lesions regressed upon Duration (NR) injected intradermally with adenoviral vectors bFGF/Ad5, foreskin. withdrawal of the growth Berking et al. (2004) ET-3/Ad5, SCF/Ad5 exposed 3 ×/wk to 30 – 50 mJ/cm 2 UVB in situ melanomas in 45–56% of adult factor stimulation after 4 wk. from FS72/T12 UVB lamps throughout a period of 4 wk. skin grafts exposed to the three growth factors independent from the exposure to UVB. bFGF, basic fibroblast growth factor; d, day or days; ET-3, endothelin-3; F, female; M, male; mo, month or months; NR, not reported; SCF, steam cell factor; wk, week or weeks Solar and UV radiation75
    • 76 Table 3.4 Melanomas induced in South American opossum M. domestica exposed to ultraviolet radiation Species, strain Number/group at start Results Comments Reference Exposure regimen: radiation type, dose, dose rate Target organ Incidence and/or multiplicity of tumours (%) South American - 43 litters of suckling young were irradiated with sunlamps Of 358 sucklings exposed, 217 (60%) opossum with a spectral emission peak at 302 nm (UVB) to induce survived to weaning, and 22 (6%) (M. domestica) melanocytic nevi. possessed a nevus at weaning. Nevi of 8 Duration (NR) Total doses of 0.87–5.0 kJ/m 2 were divided equally among of the 20 chronically-exposed animals Robinson et al. (1994) up to 14 exposures during the 19 d from birth. – 13 litters progressed to malignant melanoma with IARC MONOGRAPHS – 100D received doses of 125 J/ m2 of UVB every other d, for up to metastases to lymph node(s). 19 d after birth, with a maximum total dose of 1.12 kJ/rn 2. - 30 litters received different total doses, up to a total dose of 5.0 kJ/ m2. Affected animals were then exposed 3 times/wk to 125 J/ m2 of UVB for up to 45 wk to promote progression to malignancy. South American 620 suckling young were exposed to ultraviolet radiation The lowest dose (175 J/ m2) administered In the opossum, UVR can opossum (UVR, predominantly UVB: 290–320 nm) to determine three times a wk for three wk led to the act as a complete carcinogen (M. domestica) an optimal protocol for induction and progression of highest incidence of melanotic lesions for progression to widely Duration (NR) melanoma (7 protocols). with melanoma potential (8.1%) among disseminated disease and Robinson et al. (1998) young (5-mo-old) adults. Among exposure of sucklings can 101 much older animals (> 17 mo at lead, in old age, to widespread necropsy), 43% showed metastatic metastatic melanoma in this melanoma to the lymph nodes and model. almost one-third of these had progressed to widespread dissemination. d, day or days; NR, not reported; wk; week or weeks
    • Table 3.5 Melanomas induced in Platyfish-swordtail hybrids exposed to ultraviolet radiation Species, strain (sex) Number/group at start Results Comments Duration Exposure regimen: radiation type, dose, dose rate Target organ Reference Incidence and/or multiplicity of tumours (%) Platyfish (X. maculatus)- A total of 5000 fish. Tumor prevalence: Exposure of the fish to visible swordtail (X. Helleri) Multiple exposures on 5-20 consecutive days beginning 20% to 40% at 4 mo of age, (background light after UV exposure hybrids (M, F) on d 5 after birth (150 to 1700 J/ m2/d) or a single exposure rates: 12% in strain 1, and 2% in strain 2). reduces the prevalence to up to 6 mo of = 200 J/ m2/d of λ > 304 nm from FS-40 sunlamps, background. Setlow et al. (1989) filtered by a thin acetate film (λ > 290 nm), or a thin Mylar film (λ > 304 nm), or a thick plastic sheet (λ > 360 nm). Platyfish (X. maculates- Groups of five 6-d-old fish submitted to a single irradiation Single exposures to filtered sunlamp Only heavily pigmented swordtail (X. Helleri) - from a filtered (λ > 304 nm) sunlamp radiation: animals are susceptible to hybrids (M, F) - or with narrow wavelength bands at 302, 313, 365, 405, - up to 42% melanomas for (850 J/m 2). melanoma induction by 4 mo and 436 nm and scored for melanomas 4 mo later. a single, relatively small Setlow et al. (1993) The action spectrum (sensitivity exposure to UV. per incident photon as a function of wavelength) for melanoma induction shows appreciable sensitivity at 365, 405, and probably 436 nm. d, day or days; F, female; M, male; mo, month or months Solar and UV radiation77
    • IARC MONOGRAPHS – 100D(Xiphophorus helleri) eventually develop geneti- 4. Other Relevant Datacally determined spontaneous melanoma (theGordon-Kosswig melanoma). Setlow et al. (1989)have developed two strains of these fishes that are 4.1 Transmission and absorption insusceptible to invasive melanoma induction by biological tissuesexposure to filtered radiation from sunlamps in UVR may be transmitted, reflected, scatteredthe wavelength ranges λ > 290 nm and λ > 304 nm. or absorbed by chromophores in any layer ofIrradiation of these fishes and of X. maculatus/X. tissue, such as the skin and the eye. Absorptioncouchianus hybrids with narrow wavelength is strongly related to wavelength, as it depends onbands show that the action spectrum for mela- the properties of the responsible chromophore(s)noma induction shows appreciable sensitivity at (IARC, 1992).365, 405, and probably 436 nm, suggesting that UVC (200–280  nm) has the highest energywavelengths not absorbed directly in DNA are and thus is potentially the most damaging toeffective in induction (Setlow et al., 1993). biological tissues. However, because of its absorp- tion by the ozone layer, its impact on human3.3 Synthesis health is largely theoretical except for occasional artificial UV sources. UVB (280–315 nm) makes Recent studies have mainly focused on the up only 5–10% of the UVR that penetrates themechanisms of UV-induced carcinogenesis and ozone layer but because of its ability to directlyhave used specific stains of mice (sencar mice). damage DNA-forming modified bases, under-Several studies conducted examining the tumo- standing molecular and cellular links betweenrigenic effects of solar radiation, broad-spectrum UVB exposure and carcinogenesis has continuedultraviolet radiation, UVA, UVB and UVC in to be a major focus since the previous IARCexperimental animals, since 1992, support and Monograph (IARC, 1992). The role of non-DNAconfirm the conclusions of the previous IARC chromophores in UV carcinogenesis has beenMonograph. extensively studied over the past 15 years in Solar radiations cause squamous-cell carci- particular in relation to UVA (315–400 nm) expo-noma of the skin and of the conjunctiva in mice sure. UVA, in addition to inducing a variety ofand rats. DNA damage, also penetrates the dermis where Broad-spectrum UVR cause squamous-cell it interacts with proteins and lipids resulting incarcinoma of the skin and of the cornea/conjunc- skin ageing (for a review, see Ridley et al., 2009).tiva in mice and rats. UVA cause squamous-cell carcinoma of the 4.1.1 Eyeskin in mice. UVB cause squamous-cell carcinoma of the The eye is a complex multilayered organ.skin in mice and opossum and invasive skin The retina at the back of the eye receives visiblemelanomas in platyfish-swordtail hybrid fish radiation and the intermediate layers attenuateand opossum. UVR to different degrees, thereby protecting UVB cause skin melanomas in transgenic the retina from photodamage. The outermostmice and skin melanomas in genetically engi- cornea absorbs UVC (from artificial sources) andneered immunocompromised mice grafted with a substantial amount of UVB, which is furtherhuman melanocytes. attenuated by the lens and the vitreous humour UVC cause squamous-cell carcinoma of the in front of the retina. UVA is less attenuated byskin in mice. the cornea than by the internal structures, and78
    • Solar and UV radiationdoes not reach the retina (for a review, see Young, occurs indirectly via photosensitizers, which2006). Age-related changes in lens crystallins include endogenous melanins or proteinsaffect their structure and function causing the containing porphyrin, haem or flavin groups.lens to increasingly scatter light on the retina, They can also be exogenous, e.g. antibacterialand causing the lens to become opaque (for a agents such as naladixic acid and fluoroqui-review, see Sharma & Santhoshkumar, 2009) nolones or the immunosuppresive drug azathio- prine (for a review, see Ridley et al., 2009), and4.1.2 Skin 8-methoxypsoralen (methoxsalen) in combina- tion with UVA (PUVA) used for photochemo- The skin comprises two main layers (for a therapy. These exogenous chemicals absorb in thereview, see Young, 2006): UVA range and release reactive oxygen species 1) the outer acellular and cellular (IARC, 2012), and thus mediate UVA-induced epidermis, and DNA damage. The excited sensitizers may react 2) the inner largely extracellular dermis. with DNA directly by one-electron transfer Keratinocytes are the main epidermal cell (Type I mechanism) and/or via the generation oftype, which differentiate to create the outermost, singlet oxygen (1O2) by energy transfer to molec-non-living, terminally differentiated, cornified ular oxygen (major Type II mechanism), givingand protective stratum corneum. The dividing rise to guanine modifications including 8-oxog-cell population is located in the innermost uanine. The excited sensitizer can also transferbasal layer of the epidermis. Dendritic pigment- an electron to oxygen resulting in the formationproducing melanocytes and immunocompe- of superoxide anion radical (O2-) (minor Type IItent dendritic Langerhans cells are also present mechanism). Disproportionation of O2- can givein the epidermis. The dermal connective tissue rise to hydrogen peroxide (H2O2), and reactiveis mostly collagen synthesized by fibroblasts. species formed through the interaction of H2O2The dermis contains the skin’s vascular supply. with metal ions may induce DNA damage (forSignificant differences have been found in the reviews, see Ridley et al., 2009 and Hiraku et al.,UVA and UVB absorption properties of different 2007).skin types (Antoniou et al., 2009). In addition to the generation of reactive oxygen species, reactive nitrogen species such4.2 Genetic and related effects: as nitric acid and peroxynitrite are formed after consequences of UVR exposure UVA exposure. UVA irradiation can also lead to the long-term cellular generation of both reac-4.2.1 Photoproduct formation tive nitrogen species and reactive oxygen species (Didier et al., 1999; Valencia et al., 2006), indi-(a) DNA photoproducts: direct and indirect cating the possibility of a prolonged effect of a formation single UVA exposure (see Section 4.2.3). A multitude of photoproducts, the ratio of Several studies in vitro have shown a predom-which depends markedly on wavelength, are inance of oxidized purines after UVA-inducedformed in cellular DNA by solar UVR (IARC, oxidative damage with relatively few strand breaks1992). The question of which types of DNA or oxidized pyrimidines (Kielbassa et al., 1997;damage are formed by UVA, UVB and UVC Pouget et al., 2000). However, thymidine-derivedhas been extensively studied. Unlike UVB, UVA cyclobutane–pyrimidine dimer lesions haveis weakly absorbed by DNA and the primary been detected after UVA exposure in several cellmethod of DNA-damage induction by UVA models (e.g. Chinese hamster ovary cells) (Douki 79
    • IARC MONOGRAPHS – 100Det al., 2003), and in human skin (Courdavault are of no significance as a source of damage inet al., 2004; Mouret et al., 2006), recently reviewed natural sunlight (IARC, 1992).by Ridley et al. (2009). Moreover, in human skin DeMarini et al. (1995) evaluated the muta-after exposure to UVA, cyclobutane–pyrimidine genicity and mutation spectra of a commercialdimer lesions rather than oxidative lesions were tanning salon bed, white fluorescent light andthe main type of DNA damage induced (Mouret natural sunlight in four DNA-repair backgroundset al., 2006). It has been suggested that UVA may of Salmonella. Approximately 80% of the radia-generate cyclobutane–pyrimidine dimer lesions tion emitted by the tanning bed was within thevia a photosensitized triplet energy transfer in UV range (250–400  nm), whereas only ~10%contrast to formation via direct excitation of of the sunlight and 1% of the fluorescent lightDNA by UVB (Douki et al., 2003; Rochette et al., were in the UV range. The tanning bed emitted2003). similar amounts of UVA (315–400 nm) and UVB (280–315 nm), whereas sunlight and fluorescent(b) Other chromophores light emitted, respectively, 50–60 times and 5–10 In addition to DNA, many other cellular times more UVA relative to UVB. Based on totalcomponents absorb and/or are damaged by solar dose (UV + visible, 400–800 nm), the mutagenicUVR (IARC, 1992). Non-DNA chromophores potencies (revertants × 10−3/J/m2) of the expo-and targets are particularly relevant at longer sures in strain TA100 were 3.5 for sunlight, 24.9wavelengths. For instance trans-urocanic acid, a for fluorescent light, and 100.6 for the tanningdeamination product of histidine, is an impor- bed. Thus, the tanning bed was 29 times moretant chromophore found in high concentrations mutagenic than sunlight. The mutagenic potencyin the stratum corneum. Trans-urocanic acid of the tanning bed was similar to that producedundergoes a photoisomerization to cis-urocanic by pure 254-nm UV (DeMarini et al., 1995).acid in the presence of UVR, which has immu- DNA-sequence analysis of the revertants ofnoregulatory properties (Norval, 2006). strain TA100, which is a base-substitution strain, was performed at the doses that produced 10-fold4.2.2 Mutagenicity increases in the mutant yields (revertants/plate) compared to the control plates for sunlight Numerous reports show that sunlight or solar- and fluorescent light, and a 16-fold increase forsimulated radiation induces mutations in bacteria, the tanning bed. Thus, more than 90% of theplants, mammalian cells, Chinese hamster ovary mutants analysed were induced by the expo-and lung (V79) cells, mouse lymphoma cells, sures as opposed to being spontaneous in origin.and human skin fibroblasts. Studies in bacteria More than 80% of mutations induced by all threeexposed to radiation throughout the solar exposures were G:C→A:T transitions, and 3–5%UV spectrum demonstrate mutagenic activity were presumptive or identified multiple muta-unambiguously. UVA (320–400  nm) is muta- tions. The frequencies of the multiple mutationsgenic to yeast and cultured mammalian cells; were increased 38–82-fold in TA100 by the expo-UVB (290–320  nm) to bacteria and cultured sures, with 83% (19/23) of these multiple muta-mammalian cells; and, UVC (200–290  nm) to tions induced by the tanning bed being CC→TTbacteria, fungi, plants, cultured mammalian tandem mutations. Thus, DeMarini et al. (1995)cells, including Chinese hamster ovary and V79 also showed that a tanning bed produced acells, and human lymphoblasts, lymphocytes mutation spectrum similar to that found in theand fibroblasts. Because wavelengths in the UVC TP53 gene in sunlight-associated skin tumoursrange do not reach the surface of the Earth, they (Dumaz et al., 1994).80
    • Solar and UV radiation4.2.3 Mutation profiles and target genes tandem CC→TT transitions (Daya-Grosjean & Sarasin, 2005). The study of the mutation profiles in skin Based on the reactivity of different wave-tumours and in particular those from individuals lengths of UVR with DNA, these G:C→A:Twith either a defect in the repair processes that transition mutations induced at dipyrimidineremove UV-induced DNA damage (e.g. xero- sites were considered for many years as specifi-derma pigmentosum (XP) patients or other rare cally resulting from UVB-induced cyclobu-syndromes associated with increased skin cancer tane–pyrimidine dimers or pyrimidine (6–4)risk) has allowed the assessment of the relative pyrimidone photoproducts, and termed thecontribution of bipyrimidine photoproducts “UV-signature” or “UV-fingerprint mutations”and oxidative damage to the mutagenic effects (Wikonkal & Brash, 1999), and A:T→C:G trans-of UVR, and has provided invaluable models to versions were considered as UVA “fingerprintdelineate the genes affecting crucial pathways mutations” (Drobetsky et al., 1995; Robert et al.,involved in skin carcinogenesis. 1996). However, the wavelength specificity of Point mutations found in the TP53 gene in skin these mutations has been challenged based ontumours from normal individuals and repair- recent findings in rodent cell models, mousedeficient XP patients are mainly G:C→A:T tran- models, and human skin. The UVA-inducedsitions in skin tumours (74% in non-XP, 87% in mutation profile in exon 2 of adenine phospho-XP), and also to a lesser extent in internal tumours ribosyltransferase (Aprt) gene in rodent cells(47%) where, however, they are mainly located at showed a high proportion of mutations recov-5′CG-3′ dinucleotide (CpG; 63%) sequences— ered opposite thymine–thymine–dipyrimidineprobably due to the deamination of the unstable damage sites supporting the notion that cyclob-5-methylcytosine (Dumaz et al., 1994). In XP skin utane–pyrimidine dimers are a premutagenictumours, 100% of the mutations are targeted at lesion in UVA-induced mutagenesis (Rochettepyrimidine–pyrimidine (py–py) sequences and et al., 2003). C→T transition mutations in the lacZ55% of these are tandem CC→TT transitions. In transgene have been detected in the epidermisskin tumours from normal individuals, 14% of and dermis of UVA-treated mice, correspondingthe TP53 mutations are double mutations and, to the formation of cyclobutane–pyrimidineas in XP skin tumours, all these are CC→TT dimers (Ikehata et al., 2008), in the Tp53 gene oftransitions. In contrast, internal tumours rarely UVA- or UVB-induced skin tumours in hairlesscontain tandem mutations (0.8%) and, of these, mice (van Kranen et al., 1997), in the TP53 geneonly 2/14 were CC→TT transitions. A similar of benign solar keratoses and malignant skinmutation profile of C→T or tandem CC→TT UV squamous cell carcinomas, in humans (Agarsignature transitions, occurring at bipyrimi- et al., 2004), and in UVA-irradiated human skindine sequences, has been found in several other cells under certain experimental conditionsgenes including PTEN (phosphatase and tensin (Courdavault et al., 2004; Rünger & Kappes,homologue deleted on chromosome 10; Ming & 2008).He, 2009; Wang et al., 2009). Ras, Ink4a-Arf as Another characteristic of mutations inwell as alterations of the different partners of the epithelial skin cancers is the preference of theirmitogenic sonic hedgehog signalling pathway occurrence for a CpG sequence, which is the(patched, smoothed, and sonic hedgehog) have consensus target motif for epigenetic DNA meth-also been found in XP tumours and sporadic skin ylation in vertebrates. Mutation hotspots in suchcancers. The majority of mutations are at C→T or a sequence context within the Tp53 gene have been identified, and it has been suggested that 81
    • IARC MONOGRAPHS – 100Dtheir presence could be used as a marker of solar occurrence of secondary genetic or epigeneticUV exposure (Ikehata & Ono, 2007; Rochette alterations that can lead to skin cancer develop-et al., 2009). However, the specificity of dinucle- ment. Patients with Gorlin syndrome (or basalotide mutability in skin cancer is complex. Lewis cell nevus syndrome) suffer with multiple basalet al. (2008) compared the base-substitution cell carcinoma. This syndrome is associated withsignatures obtained in several mutation assay mutations in the Patched (PTCH) gene, an essen-model systems after exposure to UVB, UVC tial component in Hedgehog signalling (Epstein,or simulated sunlight and cancer-specific base 2008). Aberrant activation of sonic hedgehogsubstitutions collated in the IARC TP53 database homologue (SHH) signalling, usually because(IARC, 2006b), for exons 5, 7 and 8 of the TP53 of mutations either in the PTCH or smoothenedgene. The UVB, UVC and skin cancer profiles (SMOH) genes (Reifenberger et al., 2005) orfor exon 5 and 8 all showed relatively high levels because of hyperactivation of this pathway, isof G:C→A:T mutations primarily at TC and CC often found in sporadic basal cell carcinomas.sites, and to a lesser extent at CT sites. However Dysfunctional p53 is likely to affect protec-the exon 7 profiles did not group with the skin tive responses to DNA damage and oncogeniccancer profiles which showed a relatively high signalling. Experiments in both humans andlevel of G:C→A:T mutations at CpG sites. mice have shown that clusters of epidermal cells Based on these findings, the back-extrapola- with mutant p53 occur long before squamous celltion from a mutation to an exposure to a single carcinoma becomes visible (de Gruijl & Rebel,wavelength region of the UVR spectrum is not 2008). Although TP53 mutations cause geneticpossible. instability and facilitate the carcinogenic process, The study of syndromes associated with they are not enough to cause basal cell carcinomaincreased skin cancer risk has been instrumental or squamous cell carcinoma, and the activationin the identification of genes critical for UV of signalling cascades (normally needed forcarcinogenesis. Germline mutations in PTEN cell proliferation and homeostasis) is often alsoresulting in altered PTEN function, detected in involved. Based on the molecular, pathological andpatients with Cowden disease and Bannayan– functional dissection of such signalling cascades,Riley–Ruvalcaba syndrome (Bonneau & Longy, evidence has accumulated linking an activated2000), are associated with an increased risk of receptor tyrosine kinase (RTK)/RAS pathway inbasal cell carcinoma, squamous cell carcinoma, combination with dysfunctional p53 to the devel-and melanoma (Nuss et al., 1978; Camisa et al., opment of squamous cell carcinoma; activated1984; Liaw et al., 1997; Trojan et al., 2001; Ming Hedgehog pathway with possibly dysfunctional& He, 2009). Mice with Pten deletion and muta- p53 to the development of basal cell carcinoma;tion are highly susceptible to tumour induction and in cutaneous melanoma, activated RTK/(Suzuki et al., 1998). Conditional knockout of RAS pathway in combination with inactivationPten in skin leads to neoplasia (Li et al., 2002; of the inhibitor of cycline-dependant kinase 4Suzuki et al., 2003; Backman et al., 2004). & 6 (INK4a) locus (de Gruijl et al., 2001). ThePten deficiency in mice causes increases in cell Notch signalling pathway has also been identi-proliferation, apoptotic resistance, stem-cell fied as a key regulator of epidermal homeostasisrenewal/maintenance, centromeric instability, and implicated in skin carcinogenesis; aberrrantand DNA double-strand breaks (Groszer et al., Notch signalling leads to skin cancer including2001; Kimura et al., 2003; Wang et al., 2006; basal cell carcinoma, squamous cell carcinoma,He et al., 2007; Shen et al., 2007), which can and melanoma (Okuyama et al., 2008).enhance susceptibility to carcinogens and the82
    • Solar and UV radiation4.2.4 Genomic instability, bystander effect, For instance, instability was observed for telomere shortening several generations in the GM10115 human– hamster hybrid cell line after combined treatment Another potential mechanism for inducing of UVA with bromodeoxyuridine and Hoechstgenomic instability in cells not directly hit by 33258 dye (Limoli et al., 1998). Both UVA andradiation is via the bystander effect. Bystander UVB are able to induce delayed mutations ineffects via both gap-junction and extracellular the hypoxanthine-guanine phosphoribosyl-signalling have been observed in cells following transferase (Hprt) gene of V79 Chinese hamsterUVB treatment (Banerjee et al., 2005, Dahle et al., fibroblast cells (Dahle & Kvam, 2003), which2005), and an UVA-induced bystander effect has could be inhibited by reactive oxygen speciesbeen reported that can be attenuated by the use scavengers (Dahle et al., 2005). Mutations in theof a nicotinamide adenine dinucleotide phos- HPRT gene have shown to be increased 7 daysphate (NADPH) oxidase inhibitor, suggesting after UVA irradiation in human keratinocytesa possible role of reactive oxygen species in the HaCaT (Phillipson et al., 2002). In the same cellinduction of this effect (McMillan et al., 2008; model, UVA treatment led to continued reduc-Whiteside & McMillan, 2009). After UVA expo- tions in survival of UVA-treated HaCaT for oversure, such mechanisms have been extensively 21 days following treatment, and an increaseinvestigated partly because of the action spectra in the number of micronuclei per cell over theof UVA’s interaction with DNA. There is an same period. The addition of catalase was shownincreasing body of evidence that suggests that to reverse these effects to near-control levels. AUVA-induced (and to some extent UVB-induced) bystander effect was induced in human keratino-damage cannot only remain but also be gener- cytes HaCaT and fibroblasts MRC5 cells treatedated for prolonged periods in the irradiated cell, with UVA radiation but not UVB radiationits progeny, and also in surrounding cells and (Whiteside & McMillan, 2009). One potentialtissues which were not themselves exposed. The mechanism for the generation of reactive oxygenprogeny of cells which have survived irradiation species under such experimental conditionsshow changes in chromosomal structure and copy involves the UVA-induction of enzyme activity.number, the generation of micronuclei, changes One potential target is a NADPH oxidasein gene expression and cell survival (Little, 2000; (Valencia & Kochevar, 2008). This enzyme hasMorgan, 2003), and are all seen as end-points been shown to cause increased superoxide gener-of genomic instability. Such persistent genomic ation in response to UVA in mouse, monkey, andinstability defined as the persistent induction of human cell lines (Hockberger et al., 1999). TheDNA and cellular damage in irradiated cells and resulting increase in superoxide and its conver-their progeny (Ridley et al., 2009) can lead to a sion to other reactive oxygen species wouldhypermutator phenotype where genetic altera- lead to increased cellular and DNA damage.tions increase generation upon generation in a Prolonged generation of reactive oxygen specieslarge proportion of the progeny of the irradiated by such mechanisms in the initially exposed cellscells, increasing the risk of malignant transfor- and their progeny therefore have the potentialmation. Conversely, another characteristic of to enable persistent genomic instability (Ridleypersistent genomic instability can be increased et al., 2009).cell-kill of the progeny, meaning that the risk of Another mechanism for inducing persistentcancer arising from these cells is reduced rather genomic instability is via the shortening andthan increased (Ridley et al., 2009). loss of telomeres. The shortening of telomeres or the dysfunction of proteins associated with 83
    • IARC MONOGRAPHS – 100Dthe telomeres can lead to large scale transfers of of senescent cells in skin leading to skin aging.sequences between chromosomes, which lead to Once cells have entered into senescence, theythe amplification or deletion of sequences (Bailey undergo a series of morphological and metabolic& Murnane, 2006). It has been demonstrated that changes, and gene-expression profiles are alteredUVA can increase the rate of telomere shortening as has been shown in human skin fibroblasts after(Oikawa et al., 2001; Ridley et al., 2009), therefore exposure to UVB (Chen et al., 2008).suggesting a possible link between UVA irra-diation and increasing instability over severalgenerations. 4.3 Genetic susceptibility: host It has also been shown that irradiation factors modulating the responsewith UVA and UVB is able to trigger increased to UVmicrosatellite instability in radial growth phasemelanoma cells (Hussein et al., 2005). 4.3.1 DNA repair capacity and single nucleotide polymorphisms (SNPs) in4.2.5 Cell killing – apoptosis and senescence DNA repair genes Apoptosis and premature senescence are Many of the directly formed UV photoprod-protective mechanisms against the presence ucts are repaired via the nucleotide excisionof unrepaired DNA lesions in the genome that repair (NER) pathway, and those formed indi-could otherwise induce mutations increasing rectly via the modification of DNA by reactivethe risk of carcinogenesis induced after UV irra- oxygen species and reactive nitrogen speciesdiation. The fact that nucleotide excision repair require components of the base-excision repair(NER)-deficient cells are very sensitive to the pathway.cell-killing effect of UV light is a clear indica- NER operates through two subpathwaystion that unrepaired photoproducts constitute in the early stages of damage recognition,the main apoptosis-triggering signal after UV depending on whether the damage is locatedirradiation (Batista et al., 2009). How these anywhere throughout the genome [globallesions are processed to generate a toxic signal genome (GG) repair] or in an actively tran-is unclear. While some data suggest transcrip- scribed gene [transcription-coupled (TC) repair].tion blockage is the main reason behind this GG repair begins with recognition of the damageapoptosis induction, other data suggest that the by the XPC-RAD23B-centrin2 complex, aided information of DNA double-strand breaks during some cases by the UV damaged DNA-bindingthe replication of cyclobutane-pyrimidine activity (UV-DDB) that includes the subunitsdimers-containing DNA is necessary for the DDB1 and DDB2/XPE. The mechanisms for TCcommitment to cell death (Batista et al., 2009). repair are not completely understood; a currentUV light (mainly UVA and UVB) is also able to model postulates that the pathway is initiateddirectly activate membrane death receptors that by the arrest of RNA polymerase II at a lesiontrigger apoptosis independently of DNA damage. on the transcribed strand of an active gene, inMitogen-activated protein kinases (MAPKs) are a process that requires several factors includingalso directly activated by UV light and whether the Cockayne syndrome A (CSA), CSB, andthis activation is DNA-damage dependent or XPA-binding protein-2 (XAB2) proteins (Sarasinindependent is still unclear. & Stary, 2007; Hanawalt & Spivak, 2008). The The hallmark of cellular senescence is the loss recognition events in GG-NER and TC-NER areof proliferative capacity, with the accumulation followed by a common pathway involving the84
    • Solar and UV radiationunwinding of the damaged DNA, dual incisions In addition, rare cases have been describedin the damaged strand, removal of the damage- showing a complex pathological phenotype withcontaining oligonucleotide, repair synthesis in combined symptoms of XP, Cockayne syndromethe resulting gap, and ligation of the repair patch and/or NER syndrome defects that have beento the contiguous parental DNA strand. These associated with combinations of mutations in XP,steps require the coordinated action of several CS, and other unidentified genes (for instance,factors and complexes, including the repair/tran- Itoh et al., 1994; Lehmann, 2003; Spivak, 2005;scription complex factor TFIIH, and the repair Nardo et al., 2009).factors XPA, XPG, and excision repair cross- The rarity of these syndromes associatedcomplementing rodent deficiency, complementa- with mutations in NER genes and compromisedtion group 1 (ERCC1)-XPF, in addition to those repair excludes a direct major public healthrequired for repair replication and ligation. impact on skin cancer risk, however, suboptimal The mismatch repair enzyme hMSH2 has also NER capacity could also result in increasedbeen linked to the NER pathway. This enzyme is cancer risk. There is increasing evidence thata TP53 target gene and induced by UVB radia- more frequently found genetic variation suchtion, suggesting a role for mismatch repair in skin as SNPs can also impact on protein expres-cancer development (Rass & Reichrath, 2008). sion and function, and thus, potentially cancer Defects in NER are associated with three risk. It is hypothesized that polymorphisms inmajor autosomal recessive disorders, xeroderma genes implicated in the responses to the DNApigmentosum (XP), Cockayne syndrome, and damage and oxidative stress following exposuretrichothiodystrophy. At the clinical level, XP to UV constitute genetic susceptibility factorsis characterized by a highly increased inci- for skin cancers. This has been assessed in manydence of tumours in sun-exposed areas of the molecular epidemiological studies using eitherskin (Stefanini & Kraemer, 2008). In contrast, a candidate gene approach or more recentlyCockayne syndrome and trichothiodystrophy genome-wide association studies (GWAS). SNPsare cancer-free disorders characterized by in NER genes have been extensively investigated.developmental and neurological abnormalities For instance, for melanoma, significant associa-and premature aging, associated in trichothi- tions were found for the NER genes ERCC1 andodystrophy with typical hair abnormalities XPF (which act together in a rate-limiting step in(Lehmann, 2003). The two genes identified the repair pathway) in a study population of 596as responsible for the NER-defective form of Scottish melanoma patients and 441 population-Cockayne syndrome (CSA and CSB) are specifi- based controls, with the strongest associations forcally involved in transcription-coupled repair melanoma cases aged 50 years and under (ERCC1TC-NER. Seven NER-deficient complementation OR, 1.59; 95%CI: 1.11–2.27, P = 0.008; XPF OR,groups have been identified in XP patients (desig- 1.69; 95%CI: 1.18–2.43, P = 0.003)] (Povey et al.,nated XPA to XPG); these XP cases are defective 2007). Significant associations between mela-in one of seven genes called XPA to XPG. An noma and XPD SNPs have also been reportedeighth complementation group, the so-called XP (e.g. Manuguerra, et al., 2006). Variants in genesvariant form (XPV) was latter identified with a involved in the signalling cascades activated indefective gene encoding the DNA polymerase ε. response to UVR have been investigated. ForThis enzyme is required for the replication of the instance the TP53 Arg72Pro polymorphism,UV-damaged DNA pathway, called translesion but not p73 G4C14 → A4T14 and p21 Ser31Arg,DNA synthesis (Stefanini & Kraemer, 2008). contribute to the risk of developing cutaneous melanoma (Li et al., 2008). 85
    • IARC MONOGRAPHS – 100D Over the past few years several groups have of UVA-generated cyclobutane-pyrimidineassessed the DNA repair capacity in different dimers is lower than that of dimers produced bypopulations in an attempt to identify “at-risk” UVB irradiation in human skin using an in-vitrosubpopulations in the general population (Li model system (Mouret et al., 2006). The mecha-et al., 2009). Several DNA-repair phenotypic nistic basis of these differences in repair capacitystudies have been developed using cultured remains unknown.blood lymphocytes including the mutagen The base-excision repair and single-strandsensitivity assay, the host-cell reactivation break repair pathways are the main routes forassay, RT–PCR gene expression, microarray for oxidative DNA damage. Attenuation of theprotein expression, and DNA repair capacity. repair of 8-oxoguanine via downregulationFor instance, lower DNA repair capacity meas- of the base-excision repair pathway results inured in a UV-based host-cell reactivation assay hypersensitivity to UVA in a murine cell modelhas been found in individuals with basal cell (Kim et al., 2002). In humans there is substan-carcinoma and cutaneous melanoma (Li et al., tial inter-individual variation in 8-oxoguanine2009), and increased mutagen sensitivity meas- repair (Paz-Elizur et al., 2007), and the presenceured as in vitro UVB-induced chromatid breaks of the Ser326Cys SNP in the human 8-oxogua-was found in basal cell carcinoma and squamous nine DNA glycosylase (hOGG1) gene has beencell carcinoma patients (Wang et al., 2005). The shown to impact on its constitutive activity, withunderlying molecular basis of this reduced repair the Cys variant protein having a lower enzy-capacity remains to be fully determined. matic activity and a greater sensitivity to oxida- Several studies have reported an age-asso- tive stress (Bravard et al., 2009). UVA irradiationciated decline in NER (Moriwaki & Takahashi, induces relocalization of the OGG1 to nuclear2008), which could result in an accumulation of speckles where apurinic/apyrimidinic endonu-damage, and reduced DNA-repair capacity has clease-1 (APE1) is also found (Campalans et al.,been found to be an independent risk factor for 2007). APE1 is also known as redox factor-1basal cell carcinoma and single or non-aggressive (REF-1), a redox regulator of multiple stress-squamous cell carcinoma but not for multiple inducible transcription factors such as nuclearprimaries, local aggressiveness, or recurrence of factor–kappa B (NF-κB). Haploinsufficiency innon-melanoma skin cancer (Wang et al., 2007). mice of APE1 increases the apoptotic response Differences have also been reported between to oxidative stress (Unnikrishnan et al., 2009).keratinocytes and fibroblasts in terms of thelethal effects of UVB and oxidative stress, which 4.3.2 SNPs in genes other than those involvedcould in part be explained by differences in in DNA repairrepair capacity and the induction of apoptosis.Keratinocytes have a more efficient NER global The hypothesis that polymorphisms in genesgenome repair (GGR) subpathway and are char- implicated in the responses to the DNA damageacterized by a strong anti-oxidant capacity and a and oxidative stress induced following exposurehigher susceptibility to reactive-oxygen-species- to UV constitute genetic susceptibility factors forinduced apoptosis than fibroblasts (D’Errico skin cancers has been assessed in many molec-et al., 2005; D’Errico et al., 2007). ular epidemiological studies using either a candi- Studies following the persistence of DNA date gene approach or more recently GWAS. Forphotoproducts using high-performance liquid instance, using in a GWAS of 930 Icelanderschromatography coupled with tandem mass with basal cell carcinoma and 33117 controls,spectrometry have shown that the rate of removal common variants on 1p36 and 1q42 were found86
    • Solar and UV radiationto be associated with cutaneous basal cell carci- known as pigmentation genes: membrane-noma but not with melanoma or pigmentation associated transporter protein gene (MATP),traits (Stacey et al., 2008). SNPs in immune- interferon regulatory factor 4 (IRF4), tyrosinaseregulating components such as cytokines may (TYR), blue eye oculocutaneous albinism type IIlead to inter-individual differences in immuno- (OCA2), and melanocortin-1 receptor (MC1R).suppression response and susceptibility to mela- These are similar to the hair-colour-related locinoma. For instance, in the interleukin-6 receptor detected in the GWAS of hair colour (Han et al.,gene (IL-6R), four SNPs (rs6684439, rs4845618, 2008).rs4845622, and rs8192284) in linkage disequi-librium were associated with an increased riskof melanoma (Gu et al., 2008). An elevated risk 4.4 Other effectsof melanoma was observed in the heterozygous 4.4.1 Immune response and photoadapationgroups of these SNPs with odds ratios of 1.74(95%CI: 1.07–2.81) for rs6684439; 1.72 (95%CI: The development of skin cancer appears to be1.04–2.84) for rs4845618; 1.69 (95%CI: 1.03–2.75) controlled in part by the immune system. Withinfor rs4845622; and 1.68 (95%CI: 1.04–2.73) for the skin all the necessary cellular requirementsrs8192284. These associations were not observed are present to induce and elicit antitumouralin the homozygous variant group with odds immunity (Schröder et al., 2006). Almost 30 yearsratios ranging from 0.93 to 1.03. ago, Fisher and Kripke were the first to demon- Associations have been found between poly- strate that UVR caused suppression of certainmorphisms in the promoter of the vitamin D aspects of the immune system (Fisher & Kripke,receptor gene and malignant melanoma (Povey 1977). It has been well documented that patientset al., 2007; Barroso et al., 2008; Mocellin & Nitti, with organ transplants that are maintained with2008) and non-melanoma skin cancer (Gandini immunotherapy are very prone to skin canceret al., 2009). (e.g. Bordea et al., 2004). Immunosuppression by There is some evidence for a contribution of solar-simulated UV in men has been observed atpigmentation genetic variants, in addition to the doses three times lower than those required formelanocortin-1 receptor variants, to variation immunosuppression in women (Damian et al.,in human pigmentary phenotypes and possibly 2008).the development of skin cancer (Sturm, 2009). The major steps of UV-induced immuneA first multistage GWAS of tanning response suppression have been determined but it shouldafter exposure to sunlight in over 9000 men and be noted that, in many instances, these detailswomen of European ancestry who live in the were obtained following a single or a few expo-USA was recently reported (Nan et al., 2009). An sures of a rodent model or human subjects toinitial analysis of 528173 SNPs genotyped on 2287 UVR and that the dose chosen was sufficient towomen identified with LOC401937 (rs966321) cause burning. In addition, the source used toon chromosome 1 as a novel locus highly asso- emit UVR frequently contained more than 50%ciated with tanning ability. This association was UVB (wavelength 280–315  nm), considerablyconfirmed in 870 women controls from a skin more than natural sunlight. In experimentalcancer case–control study with a joint P value systems, there are differences between what isof 1.6  ×  10−9. However this association was not termed local and systemic immunosuppression.replicated in two further studies. Several SNPs In the former, the antigen is applied directly toreaching the genome-wide significance level the irradiated body site soon after UV exposure.were located in or adjacent to the loci previously In the latter, following UV exposure of one part 87
    • IARC MONOGRAPHS – 100Dof the body, the antigen is applied to a distant, and by inhibiting the activation of effector andnon-irradiated body site (Applegate et al., 1989). memory T cells. Some of the mechanisms impli- Following UVB exposure, convincing cated in UVA-induced immunosuppression,evidence has been published to indicate that such as increased COX-2 activity, are common tothe chromophores for immunosuppression those observed after exposure to UVB. In addi-include DNA, urocanic acid (UCA), and cell tion, the production of reactive oxygen speciesmembranes. Studies by Kripke’s team were the and reactive nitrogen species by UVA alters thefirst to suggest that DNA (and most likely, the redox equilibrium and targets proteins, lipidspyrimidine dimer) may be the chromophore for and DNA, and modulates the immune cellsUVB-induced immunosuppression (Applegate resulting in aberrant behaviour and migrationet al., 1989), and evidence linking DNA damage of antigen-presenting cells, the inhibition ofwith immune modulation has come from studies T-cell activation, and generates suppressor cellson XP patients (Suzuki et al., 2001). Trans-UCA (Norval, 2006; Norval et al., 2008, Halliday &is a natural component of the strateum corneum, Rana, 2008).and UV induces a photoisomeric isomerization The T helper1 (Th1) cytokine response isof trans-UCA to cis-UCA, which appears to be the main adaptive immune mechanism thatan initiator of the UV-immunosuppression, offers protection from many infectious diseases.although its mechanism of action is still uncer- As UVR suppresses this preferentially, whiletain (Halliday & Rana, 2008; Norval et al., 2008). promoting the Th2 cytokine response, there isUVA immunosuppression is likely to involve the potential for UV exposure to increase thedifferent chromophores than those required severity of infection, to alter viral oncogenicity,for UVB immunosuppression: molecules like to cause reactivation from latency or to decreaseporphyrins have been proposed (Halliday & the resistance to re-infection. Alteration ofRana, 2008). immune responses to microorganisms has been UVB irradiation triggers the production of shown in rodent models following exposure tovarious immunomodulatory mediators in the UVR (Norval, 2006). In humans, infections byskin. These include cyclooxygenase-2 (COX-2), herpes simplex virus (HSV) and human papil-receptor activator of NF-κB ligand (RANKL), loma virus (HPV) are influenced by exposureprostaglandins, platelet activating factor, hista- to sunlight (see IARC, 2007b for details on UVmine, neuropeptides and cytokines such as and HPV). UVR is a recognized stimulus of HSVtumour necrosis factor (TNF) that modulate reactivation (Ichihashi et al., 2004) through thethe reactivity of the immune cells in the skin suppression of the local immune response as a(Beissert & Loser, 2008; Halliday & Rana, 2008; result of the UV exposure or a direct interactionNorval et al., 2008). For instance, TNF induces between the UVR and the virus through modu-Langerhans cell activation and migration out lation of the host transcription factors and theof the skin into draining lymph nodes, thus activation of HSV promoters, and hence reacti-limiting the capacity for antigen processing vation of the virus.and presentation. Therefore, UVB ultimately There is also some evidence that there aresuppresses the immune system by inducing genetic and other differences in the way thatthe production of immunosuppressive media- individuals respond to vaccination depending ontors, by damaging and triggering the premature UVR exposure (Norval, 2006). For instance, themigration of antigen-presenting cells required findings from a meta-analysis of Bacille Calmette–to stimulate antigen-specific immune responses, Guérin clinical trials such as the increase of theby inducing the generation of suppressor cells efficacy of Bacille Calmette–Guérin vaccination88
    • Solar and UV radiationwith the increasing distance from the equator these to UVC and a subset of the UVB responsivesuggested there might be an association between genes also responsive to UVA.reduced vaccine efficacy and UVR (Colditz et al., Analysis of the UVR response genes in human1994). melanocytes identified the tyrosine kinase ephrin Human and rodent skins have the capacity receptor A2 (EPHA2) as an essential mediator ofto adapt as a result of repeated suberythemal UVR-induced apoptosis (Zhang et al., 2008).UV exposures. This photoadaption can attenuate Chronic UVR exposure can also modulatethe quantity of UVR that reaches the basal and gene expression. For instance, chronic UVAsuprabasal cells of the epidermis, and results in radiation of human HaCaT keratinocytes resultsan enhanced ability to repair UV-induced DNA in decreased PTEN expression (He et al., 2006).damage and an induction of protective enzymes MicroRNAs are very small endogenous RNAsuch as superoxide dismutase. Whether photoa- molecules about 22–25 nucleotides in lengthdaption can lead to photoprotection against the capable of post-transcriptional gene regulation.normal downregulation of immunity induced MicroRNAs bind to their target mRNA leading toby a high UV dose remains to be established as cleavage or suppression of translation. MicroRNAthere are considerable gaps in the knowledge and profiles have been examined in melanomas (andthere are many variables involved, including the melanoma cell lines) and Kaposi sarcoma (seeacknowledged genetic diversity in the response Sand et al., 2009 and table therein). For instance,of individuals to UVR. Evidence for the devel- the skin specific microRNA miR-203 thatopment of photoadaption is only apparent for represses p63 expression, an important factor inepidermal DNA damage, no evidence exists epidermal cell proliferation and differentiation, iswhen other parameters were considered such as downregulated in melanoma lines; miR-221 andtotal urocanic acid content or cis isomerization, miR-222 are linked to melanoma progressionLangerhans cells and dendritic cell numbers and through the downregulation of cyclin-dependentfunction, natural killer cell numbers and func- kinase inhibitor 1b (p27Kip1/CDKN1B) and thetion, dermal mast cell numbers or contact and tyrosine kinase c-KIT receptor.delayed hypersensitivity responses (Norval et al.,2008 and references therein). Thus, it is probablethat repeatedly irradiating individuals with UVR 4.5 Synthesisis likely to continue to result in downregulation In addition to what is stated in the summaryof immunity. of Volume 55 of the IARC Monographs, it is now known that following exposure to the individual4.4.2 Modulation of gene expression components of UVR, i.e. UVA, UVB or UVC, there is an overlapping profile of DNA damage Differential gene expression in a variety of detectable, in particular for cyclobutane-pyrimi-cell types has been demonstrated after exposure dine dimers. However, the proportion of differentto different UV wavebands. For example Koch- base-pair changes shows variation depending onPaiz et al. (2004) used cDNA microarrays to the wavelength of radiation and cell type/species.analyse the responses in human cell line MCF-7 The mechanisms leading to their formation maycells following exposure to equitoxic doses of also be different. Recent experimental evidence inUVA, UVB, and UVC radiation. Under these human cells shows that cyclobutane-pyrimidineexperimental conditions, 310 of the 7684 genes dimers at cytosine-containing DNA sequenceson the array were UVB responsive, a subset of is formed following exposure to both UVA and UVB individually in human skin ex vivo. 89
    • IARC MONOGRAPHS – 100D Human cells have DNA-repair pathways primarily on results observed in the choroid andthat repair DNA photoproducts: the absence of the ciliary body of the eye.these enzymes, as seen in XP patients, leads to There is sufficient evidence in humans foran increase risk of developing squamous cell the carcinogenicity of the use of UV-emittingcarcinomas and melanomas lending support tanning devices. UV-emitting tanning devicesto a major role of DNA photoproducts in cause cutaneous malignant melanoma andphotocarcinogenesis. ocular melanoma (observed in the choroid and UVR exposure gives rise to mutations in the ciliary body of the eye). A positive asso-several genes in several human cell model ciation has been observed between the use ofsystems, and mutations have been detected in UV-emitting tanning devices and squamous cellseveral genes in human tumours, for example carcinoma of the skin.the TP53 gene in squamous cell carcinoma and There is sufficient evidence in humans for thesolar keratosis, at DNA bases where known carcinogenicity of welding. Current evidencephotoproducts could have been formed lending establishes a causal association for ocularsupport to a major role of DNA photoproducts in melanoma although it is not possible without aphotocarcinogenesis. full review of welding to attribute the occurrence Mutations can be detected in human cells of ocular melanoma to UV radiation specifically.exposed to UVA, UVB and UVC: the base-pair There is sufficient evidence in experimentalchanges involved in some of these mutations animals for the carcinogenicity of solar radia-overlap. In particular, mutations found involving tion, broad-spectrum UVR, UVA radiation,C→T transitions are found in cells treated with UVB radiation, UVC radiation.either UVA, UVB or UVC. The same situation is Solar radiation is carcinogenic to humansfound when the base-pair changes, for instance (Group 1).in the TP53 gene, are analysed in human squa- Use of UV-emitting tanning devices is carci-mous cell carcinoma and solar keratosis. As C→T nogenic to humans (Group 1).transitions are not a specific “fingerprint” for Ultraviolet radiation (bandwidth 100–400UVA, UVB or UVC, either radiation type could nm, encompassing UVC, UVB and UVA) ishave been at the origin of the exposure initiating carcinogenic to humans (Group 1).the carcinogenic process. Based on the above mechanistic considera-tions, UVA, UVB and UVC are carcinogenic in Referenceshuman cells. Acquavella J, Olsen G, Cole P et al. (1998). Cancer among farmers: a meta-analysis. Ann Epidemiol, 8: 64–74.5. Evaluation doi:10.1016/S1047-2797(97)00120-8 PMID:9465996 Adami J, Gridley G, Nyrén O et al. (1999). Sunlight and There is sufficient evidence in humans for the non-Hodgkin’s lymphoma: a population-based cohortcarcinogenicity of solar radiation. Solar radiation study in Sweden. Int J Cancer, 80: 641–645. doi:10.1002/causes cutaneous malignant melanoma, squa- (S I C I )10 9 7- 0 21 5 (1 9 9 9 0 3 01) 8 0 : 5 < 6 4 1 : : A I D - IJC1>3.0.CO;2-Z PMID:10048959mous cell carcinoma of the skin and basal cell Agar NS, Halliday GM, Barnetson RS et  al. (2004). Thecarcinoma of the skin. A positive association has basal layer in human squamous tumors harbors morebeen observed between exposure to solar radia- UVA than UVB fingerprint mutations: a role for UVA in human skin carcinogenesis. Proc Natl Acad Sci Ution and cancer of the lip, conjunctival squamous S A, 101: 4954–4959. doi:10.1073/pnas.0401141101cell carcinoma and ocular melanoma, based PMID:1504175090
    • Solar and UV radiationAjani UA, Seddon JM, Hsieh CC et  al. (1992). Berking C, Takemoto R, Binder RL et  al. (2002). Occupation and risk of uveal melanoma. An Photocarcinogenesis in human adult skin grafts. exploratory study. Cancer, 70: 2891–2900. Carcinogenesis, 23: 181–187. doi:10.1093/carcin/23.1.181 doi:10.1002/1097-0142(19921215)70:12<2891::AID- PMID:11756239 CNCR2820701228>3.0.CO;2-1 PMID:1451071 Berking C, Takemoto R, Satyamoorthy K et al. (2001). BasicAntoniou C, Lademann J, Schanzer S et  al. (2009). Do fibroblast growth factor and ultraviolet B transform different ethnic groups need different sun protection? melanocytes in human skin. Am J Pathol, 158: 943–953. Skin Res Technol, 15: 323–329. doi:10.1111/j.1600- doi:10.1016/S0002-9440(10)64041-2 PMID:11238042 0846.2009.00366.x PMID:19624429 Berking C, Takemoto R, Satyamoorthy K et  al. (2004).Applegate LA, Ley RD, Alcalay J, Kripke ML (1989). Induction of melanoma phenotypes in human skin Identification of the molecular target for the suppres- by growth factors and ultraviolet B. Cancer Res, sion of contact hypersensitivity by ultraviolet radiation. 64: 807–811. doi:10.1158/0008-5472.CAN-03-3438 J Exp Med, 170: 1117–1131. doi:10.1084/jem.170.4.1117 PMID:14871803 PMID:2529340 Black RJ & Gavin AT (2006). Photocarcinogenic riskArmstrong BK & Kricker A (2001). The epidemiology of UV of narrowband ultraviolet B (TL-01) phototherapy: induced skin cancer. J Photochem Photobiol B, 63: 8–18. early follow-up data. Br J Dermatol, 154: 566–567. doi:10.1016/S1011-1344(01)00198-1 PMID:11684447 doi:10.1111/j.1365-2133.2005.07085.x PMID:16445801Atillasoy ES, Seykora JT, Soballe PW et al. (1998). UVB Blum HF (1959). On the mechanism of cancer induction induces atypical melanocytic lesions and melanoma by ultraviolet radiation. IV. The size of the replicated in human skin. Am J Pathol, 152: 1179–1186. unit. J Natl Cancer Inst, 23: 343–350. PMID:13801686 PMID:9588887 Bodiwala D, Luscombe CJ, Liu S et  al. (2003). ProstateAutier P, Boniol M, Doré JF (2007). Sunscreen use and cancer risk and exposure to ultraviolet radiation: further increased duration of intentional sun exposure: still support for the protective effect of sunlight. Cancer a burning issue. Int J Cancer, 121: 1–5. doi:10.1002/ Lett, 192: 145–149. doi:10.1016/S0304-3835(02)00710-3 ijc.22745 PMID:17415716 PMID:12668278Backman SA, Ghazarian D, So K et al. (2004). Early onset Boffetta P, van der Hel O, Kricker A et  al. (2008). of neoplasia in the prostate and skin of mice with Exposure to ultraviolet radiation and risk of malig- tissue-specific deletion of Pten. Proc Natl Acad Sci nant lymphoma and multiple myeloma–a multicentre U S A, 101: 1725–1730. doi:10.1073/pnas.0308217100 European case-control study. Int J Epidemiol, 37: 1080– PMID:14747659 1094. doi:10.1093/ije/dyn092 PMID:18511490Bailey SM & Murnane JP (2006). Telomeres, chromosome Bonneau D & Longy M (2000). Mutations of the human instability and cancer. Nucleic Acids Res, 34: 2408–2417. PTEN gene. Hum Mutat, 16: 109–122. doi:10.1002/1098- doi:10.1093/nar/gkl303 PMID:16682448 1004(200008)16:2<109::AID-HUMU3>3.0.CO;2-0Banerjee G, Gupta N, Kapoor A, Raman G (2005). PMID:10923032 UV induced bystander signaling leading to apop- Bordea C, Wojnarowska F, Millard PR et al. (2004). Skin tosis. Cancer Lett, 223: 275–284. doi:10.1016/j. cancers in renal-transplant recipients occur more canlet.2004.09.035 PMID:15896462 frequently than previously recognized in a temperateBarroso E, Fernandez LP, Milne RL et al. (2008). Genetic climate. Transplantation, 77: 574–579. doi:10.1097/01. analysis of the vitamin D receptor gene in two epithe- TP.0000108491.62935.DF PMID:15084938 lial cancers: melanoma and breast cancer case-control Bradl M, Klein-Szanto A, Porter S, Mintz B (1991). studies. BMC Cancer, 8: 385–392 doi:10.1186/1471- Malignant melanoma in transgenic mice. Proc Natl 2407-8-385 PMID:19105801 Acad Sci USA, 88: 164–168. doi:10.1073/pnas.88.1.164Bastiaens MT, Hoefnagel JJ, Bruijn JA et  al. (1998). PMID:1846036 Differences in age, site distribution, and sex between Bravard A, Vacher M, Moritz E et  al. (2009). Oxidation nodular and superficial basal cell carcinoma indi- status of human OGG1-S326C polymorphic variant cate different types of tumors. J Invest Dermatol, determines cellular DNA repair capacity. Cancer Res, 110: 880–884. doi:10.1046/j.1523-1747.1998.00217.x 69: 3642–3649. doi:10.1158/0008-5472.CAN-08-3943 PMID:9620293 PMID:19351836Batista LF, Kaina B, Meneghini R, Menck CFM (2009). Broome Powell M, Gause PR, Hyman P et  al. (1999). How DNA lesions are turned into powerful killing Induction of melanoma in TPras transgenic structures: insights from UV-induced apoptosis. Mutat mice. Carcinogenesis, 20: 1747–1753. doi:10.1093/ Res, 681: 197–208. doi:10.1016/j.mrrev.2008.09.001 carcin/20.9.1747 PMID:10469620 PMID:18845270 Bulliard JL & Cox B (2000). Cutaneous malignant mela-Beissert S & Loser K (2008). Molecular and cellular mech- noma in New Zealand: trends by anatomical site, anisms of photocarcinogenesis. Photochem Photobiol, 1969–1993. Int J Epidemiol, 29: 416–423. doi:10.1093/ 84: 29–34. PMID:18173698 ije/29.3.416 PMID:10869312 91
    • IARC MONOGRAPHS – 100DBurns FJ, Uddin AN, Wu F et al. (2004). Arsenic-induced Corona R, Dogliotti E, D’Errico M et  al. (2001). Risk enhancement of ultraviolet radiation carcinogenesis factors for basal cell carcinoma in a Mediterranean in mouse skin: a dose-response study. Environ Health population: role of recreational sun exposure early in Perspect, 112: 599–603. PMID:15064167 life. Arch Dermatol, 137: 1162–1168. PMID:11559211Camisa C, Bikowski JB, McDonald SG (1984). Cowden’s Courdavault S, Baudouin C, Charveron M et  al. (2004). disease. Association with squamous cell carcinoma Larger yield of cyclobutane dimers than 8-oxo-7,8-di- of the tongue and perianal basal cell carcinoma. hydroguanine in the DNA of UVA-irradiated human Arch Dermatol, 120: 677–678. doi:10.1001/arch- skin cells. Mutat Res, 556: 135–142. PMID:15491641 derm.120.5.677 PMID:6721530 D’Errico M, Lemma T, Calcagnile A et  al. (2007). CellCampalans A, Amouroux R, Bravard A et al. (2007). UVA type and DNA damage specific response of human skin irradiation induces relocalisation of the DNA repair cells to environmental agents. Mutat Res, 614: 37–47. protein hOGG1 to nuclear speckles. J Cell Sci, 120: PMID:16879839 23–32. doi:10.1242/jcs.03312 PMID:17148573 D’Errico M, Teson M, Calcagnile A et al. (2005). DifferentialCarli P, Massi D, Santucci M et  al. (1999). Cutaneous role of transcription-coupled repair in UVB-induced melanoma histologically associated with a nevus and response of human fibroblasts and keratinocytes. melanoma de novo have a different profile of risk: Cancer Res, 65: 432–438. PMID:15695384 results from a case-control study. J Am Acad Dermatol, Dahle J & Kvam E (2003). Induction of delayed mutations 40: 549–557. doi:10.1016/S0190-9622(99)70436-6 and chromosomal instability in fibroblasts after UVA-, PMID:10188672 UVB-, and X-radiation. Cancer Res, 63: 1464–1469.Chang YM, Barrett JH, Bishop DT et al. (2009). Sun expo- PMID:12670891 sure and melanoma risk at different latitudes: a pooled Dahle J, Kvam E, Stokke T (2005). Bystander effects in analysis of 5700 cases and 7216 controls. Int J Epidemiol, UV-induced genomic instability: antioxidants inhibit 38: 814–830. doi:10.1093/ije/dyp166 PMID:19359257 delayed mutagenesis induced by ultraviolet A and BChen W, Kang J, Xia J et  al. (2008). p53-related apop- radiation. J Carcinog, 4: 11 doi:10.1186/1477-3163-4-11 tosis resistance and tumor suppression activity in PMID:16091149 UVB-induced premature senescent human skin fibrob- Damian DL, Patterson CR, Stapelberg M et  al. (2008). lasts. Int J Mol Med, 21: 645–653. PMID:18425358 UV radiation-induced immunosuppression is greaterChen YT, Dubrow R, Holford TR et al. (1996). Malignant in men and prevented by topical nicotinamide. J Invest melanoma risk factors by anatomic site: a case-control Dermatol, 128: 447–454. PMID:17882270 study and polychotomous logistic regression analysis. Dardanoni L, Gafà L, Paternò R, Pavone G (1984). A case- Int J Cancer, 67: 636–643. doi:10.1002/(SICI)1097- control study on lip cancer risk factors in Ragusa (Sicily). 0215(19960904)67:5<636::AID-IJC8>3.0.CO;2-V Int J Cancer, 34: 335–337. doi:10.1002/ijc.2910340309 PMID:8782651 PMID:6480154Cho E, Rosner BA, Colditz GA (2005). Risk factors for Davidson T, Kluz T, Burns F et  al. (2004). Exposure to melanoma by body site. Cancer Epidemiol Biomarkers chromium (VI) in the drinking water increases suscep- Prev, 14: 1241–1244. doi:10.1158/1055-9965.EPI-04- tibility to UV-induced skin tumors in hairless mice. 0632 PMID:15894679 Toxicol Appl Pharmacol, 196: 431–437. doi:10.1016/j.Cokkinides V, Weinstock M, Lazovich D et  al. (2009). taap.2004.01.006 PMID:15094314 Indoor tanning use among adolescents in the US, 1998 Daya-Grosjean L & Sarasin A (2005). The role of UV to 2004. Cancer, 115: 190–198. doi:10.1002/cncr.24010 induced lesions in skin carcinogenesis: an overview of PMID:19085965 oncogene and tumor suppressor gene modifications inColditz GA, Brewer TF, Berkey CS et al. (1994). Efficacy xeroderma pigmentosum skin tumors. Mutat Res, 571: of BCG vaccine in the prevention of tuberculosis. 43–56. PMID:15748637 Meta-analysis of the published literature. JAMA, 271: De Fabo EC, Noonan FP, Fears T, Merlino G (2004). 698–702. doi:10.1001/jama.271.9.698 PMID:8309034 Ultraviolet B but not ultraviolet A radiation initi-Commission Internationale de l’Eclairage [International ates melanoma. [P.]Cancer Res, 64: 6372–6376. Commission on Illumination] (1987). Vocabulaire doi:10.1158/0008-5472.CAN-04-1454 PMID:15374941 International de l’Eclairage [International Lighting de Gruijl FR & Rebel H (2008). Early events in UV carcino- Vocabulary] (CIE Publication No. 17.4), 4th genesis–DNA damage, target cells and mutant p53 foci. edition, Geneva, Bureau Central de la Commission Photochem Photobiol, 84: 382–387. doi:10.1111/j.1751- Electrotechnique Internationale. 1097.2007.00275.x PMID:18221455Commission Internationale de l’Eclairage [International de Gruijl FR, van Kranen HJ, Mullenders LH (2001). Commission on Illumination] (1998). CIE Standard. UV-induced DNA damage, repair, mutations and onco- Erythema Reference Action Spectrum and Standard genic pathways in skin cancer. J Photochem Photobiol Erythema Dose (CIE 007/E-1998), Vienna B, 63: 19–27. doi:10.1016/S1011-1344(01)00199-3 PMID:1168444892
    • Solar and UV radiationDeMarini DM, Shelton ML, Stankowski LF Jr (1995). Fisher MS & Kripke ML (1977). Systemic alteration induced Mutation spectra in Salmonella of sunlight, white in mice by ultraviolet light irradiation and its rela- fluorescent light, and light from tanning salon beds: tionship to ultraviolet carcinogenesis. Proc Natl Acad induction of tandem mutations and role of DNA repair. Sci U S A, 74: 1688–1692. doi:10.1073/pnas.74.4.1688 Mutat Res, 327: 131–149. PMID:7870082 PMID:300876Dennis LK, Beane Freeman LE, VanBeek MJ (2003). Forbes PD, Beer JZ, Black HS et al. (2003). Standardized Sunscreen use and the risk for melanoma: a quantitative protocols for photocarcinogenesis safety testing. Front review. Ann Intern Med, 139: 966–978. PMID:14678916 Biosci, 8: d848–d854. doi:10.2741/975 PMID:12700109Didier C, Emonet-Piccardi N, Béani JC et  al. (1999). Freedman DM, Dosemeci M, McGlynn K (2002). Sunlight L-arginine increases UVA cytotoxicity in irradiated and mortality from breast, ovarian, colon, prostate, and human keratinocyte cell line: potential role of nitric non-melanoma skin cancer: a composite death certifi- oxide. FASEB J, 13: 1817–1824. PMID:10506585 cate based case-control study. Occup Environ Med, 59:Diffey BL (1990). Human exposure to ultraviolet radia- 257–262. doi:10.1136/oem.59.4.257 PMID:11934953 tion. Semin Dermatol, 9: 2–10. PMID:2203439 Freedman DM, Zahm SH, Dosemeci M (1997). ResidentialDiffey BL (1991). Solar ultraviolet radiation effects and occupational exposure to sunlight and mortality on biological systems. Phys Med Biol, 36: 299–328. from non-Hodgkin’s lymphoma: composite (threefold) doi:10.1088/0031-9155/36/3/001 PMID:1645473 case-control study. BMJ, 314: 1451–1455. PMID:9167561Diffey BL & Farr PM (1989). The normal range in diag- Freeman RG & Knox JM (1964). ultraviolet-induced corneal nostic phototesting. Br J Dermatol, 120: 517–524. tumors in different species and strains of animals. J doi:10.1111/j.1365-2133.1989.tb01325.x PMID:2730842 Invest Dermatol, 43: 431–436. PMID:14216522Douki T, Reynaud-Angelin A, Cadet J, Sage E (2003). Gallagher RP, Elwood JM, Rootman J et  al. (1985). Bipyrimidine photoproducts rather than oxidative Risk factors for ocular melanoma: Western Canada lesions are the main type of DNA damage involved in the Melanoma Study. J Natl Cancer Inst, 74: 775–778. genotoxic effect of solar UVA radiation. Biochemistry, PMID:3857374 42: 9221–9226. doi:10.1021/bi034593c PMID:12885257 Gandini S, Raimondi S, Gnagnarella P et  al. (2009).Drobetsky EA, Turcotte J, Châteauneuf A (1995). A role Vitamin D and skin cancer: a meta-analysis. Eur J for ultraviolet A in solar mutagenesis. Proc Natl Acad Cancer, 45: 634–641. doi:10.1016/j.ejca.2008.10.003 Sci U S A, 92: 2350–2354. doi:10.1073/pnas.92.6.2350 PMID:19008093 PMID:7892270 Gandini S, Sera F, Cattaruzza MS et  al. (2005a). Meta-Dumaz N, Stary A, Soussi T et al. (1994). Can we predict analysis of risk factors for cutaneous melanoma: II. solar ultraviolet radiation as the causal event in human Sun exposure. Eur J Cancer, 41: 45–60. doi:10.1016/j. tumours by analysing the mutation spectra of the p53 ejca.2004.10.016 PMID:15617990 gene? Mutat Res, 307: 375–386. PMID:7513818 Gandini S, Sera F, Cattaruzza MS et  al. (2005b). Meta-Elwood JM & Diffey BL (1993). A consideration of ambient analysis of risk factors for cutaneous melanoma: solar ultraviolet radiation in the interpretation of III. Family history, actinic damage and phenotypic studies of the aetiology of melanoma. Melanoma Res, factors. Eur J Cancer, 41: 2040–2059. doi:10.1016/j. 3: 113–122. PMID:8518549 ejca.2005.03.034 PMID:16125929Elwood JM & Gallagher RP (1998). Body site distribution Garland CF, Garland FC, Gorham ED (1993). Rising of cutaneous malignant melanoma in relationship to trends in melanoma. An hypothesis concerning patterns of sun exposure. Int J Cancer, 78: 276–280. sunscreen effectiveness. Ann Epidemiol, 3: 103–110. doi:10.1002/(SICI)1097-0215(19981029)78:3<276::AID- PMID:8287144 IJC2>3.0.CO;2-S PMID:9766557 Gorham ED, Mohr SB, Garland CF et  al. (2007). DoEpstein EH (2008). Basal cell carcinomas: attack of the sunscreens increase risk of melanoma in populations hedgehog. Nat Rev Cancer, 8: 743–754. doi:10.1038/ residing at higher latitudes? Ann Epidemiol, 17: 956–963. nrc2503 PMID:18813320 doi:10.1016/j.annepidem.2007.06.008 PMID:18022535Fears TR, Bird CC, Guerry D 4th et  al. (2002). Average Grady H, Blum HF, Kirby-Smith JS (1943). Types of tumor midrange ultraviolet radiation flux and time outdoors induced by ultraviolet radiation and factors influencing predict melanoma risk. Cancer Res, 62: 3992–3996. their relative incidence. J Natl Cancer Inst, 3: 371–378. PMID:12124332 Grandin L, Orsi L, Troussard X et al. (2008). UV radia-Fears TR, Scotto J, Schneiderman MA (1977). Mathematical tion exposure, skin type and lymphoid malignancies: models of age and ultraviolet effects on the incidence results of a French case-control study. Cancer Causes of skin cancer among whites in the United States. Am J Control, 19: 305–315. doi:10.1007/s10552-007-9093-6 Epidemiol, 105: 420–427. PMID:860705 PMID:18040875Findlay GM (1928). Ultra-violet light and skin cancer. Lancet, Green A (1992). A theory of site distribution of melanomas: 212: 1070–1073. doi:10.1016/S0140-6736(00)84845-X Queensland, Australia. Cancer Causes Control, 3: 513–516. doi:10.1007/BF00052747 PMID:1420853 93
    • IARC MONOGRAPHS – 100DGreen A, Battistutta D, Hart V et  al.The Nambour He XC, Yin T, Grindley JC et al. (2007). PTEN-deficient Study Group (1996). Skin cancer in a subtropical intestinal stem cells initiate intestinal polyposis. Nat Australian population: incidence and lack of associa- Genet, 39: 189–198. doi:10.1038/ng1928 PMID:17237784 tion with occupation. Am J Epidemiol, 144: 1034–1040. He YY, Pi J, Huang JL et al. (2006). Chronic UVA irradia- PMID:8942434 tion of human HaCaT keratinocytes induces malignantGroszer M, Erickson R, Scripture-Adams DD et al. (2001). transformation associated with acquired apoptotic Negative regulation of neural stem/progenitor cell resistance. Oncogene, 25: 3680–3688. doi:10.1038/ proliferation by the Pten tumor suppressor gene in vivo. sj.onc.1209384 PMID:16682958 Science, 294: 2186–2189. doi:10.1126/science.1065518 Hearn RM, Kerr AC, Rahim KF et al. (2008). Incidence of PMID:11691952 skin cancers in 3867 patients treated with narrow-bandGu F, Qureshi AA, Niu T et  al. (2008). Interleukin and ultraviolet B phototherapy. Br J Dermatol, 159: 931–935. interleukin receptor gene polymorphisms and suscep- doi:10.1111/j.1365-2133.2008.08776.x PMID:18834483 tibility to melanoma. Melanoma Res, 18: 330–335. Hiraku Y, Ito K, Hirakawa K, Kawanishi S (2007). doi:10.1097/CMR.0b013e32830658b2 PMID:18781131 Photosensitized DNA damage and its protection via aGuénel P, Laforest L, Cyr D et  al. (2001). Occupational novel mechanism. Photochem Photobiol, 83: 205–212. risk factors, ultraviolet radiation, and ocular mela- doi:10.1562/2006-03-09-IR-840 PMID:16965181 noma: a case-control study in France. Cancer Causes Hirst N, Gordon L, Gies P, Green AC (2009). Estimation of Control, 12: 451–459. doi:10.1023/A:1011271420974 avoidable skin cancers and cost-savings to government PMID:11545460 associated with regulation of the solarium industryHacker E, Irwin N, Muller HK et  al. (2005). Neonatal in Australia. Health Policy, 89: 303–311. doi:10.1016/j. ultraviolet radiation exposure is critical for malignant healthpol.2008.07.003 PMID:18760857 melanoma induction in pigmented Tpras transgenic Hockberger PE, Skimina TA, Centonze VE et al. (1999). mice. J Invest Dermatol, 125: 1074–1077. doi:10.1111/ Activation of flavin-containing oxidases underlies j.0022-202X.2005.23917.x PMID:16297212 light-induced production of H2O2 in mammalian cells.Hacker E, Muller HK, Irwin N et al. (2006). Spontaneous Proc Natl Acad Sci U S A, 96: 6255–6260. doi:10.1073/ and UV radiation-induced multiple metastatic mela- pnas.96.11.6255 PMID:10339574 nomas in Cdk4R24C/R24C/TPras mice. Cancer Res, Holly EA, Aston DA, Ahn DK, Smith AH (1996). Intraocular 66: 2946–2952. doi:10.1158/0008-5472.CAN-05-3196 melanoma linked to occupations and chemical expo- PMID:16540642 sures. Epidemiology, 7: 55–61. doi:10.1097/00001648-Håkansson N, Floderus B, Gustavsson P et  al. (2001). 199601000-00010 PMID:8664402 Occupational sunlight exposure and cancer incidence Holly EA, Aston DA, Char DH et  al. (1990). Uveal among Swedish construction workers. Epidemiology, melanoma in relation to ultraviolet light exposure and 12: 552–557. doi:10.1097/00001648-200109000-00015 host factors. Cancer Res, 50: 5773–5777. PMID:2393851 PMID:11505175 Holman CD, Armstrong BK, Heenan PJ (1986).Halliday GM & Rana S (2008). Waveband and dose Relationship of cutaneous malignant melanoma to dependency of sunlight-induced immunomodulation individual sunlight-exposure habits. J Natl Cancer Inst, and cellular changes. Photochem Photobiol, 84: 35–46. 76: 403–414. PMID:3456458 PMID:18173699 Holman CD, Mulroney CD, Armstrong BK (1980).Han J, Colditz GA, Hunter DJ (2006). Risk factors for Epidemiology of pre-invasive and invasive malignant skin cancers: a nested case-control study within the melanoma in Western Australia. Int J Cancer, 25: Nurses’ Health Study. Int J Epidemiol, 35: 1514–1521. 317–323. doi:10.1002/ijc.2910250303 PMID:7390655 doi:10.1093/ije/dyl197 PMID:16943234 Horn EP, Hartge P, Shields JA, Tucker MA (1994). SunlightHan J, Kraft P, Nan H et al. (2008). A genome-wide associ- and risk of uveal melanoma. J Natl Cancer Inst, 86: ation study identifies novel alleles associated with hair 1476–1478. doi:10.1093/jnci/86.19.1476 PMID:8089868 color and skin pigmentation. PLoS Genet, 4: e1000074 Houghton A, Flannery J, Viola MV (1980). Malignant doi:10.1371/journal.pgen.1000074 PMID:18483556 melanoma in Connecticut and Denmark. Int J Cancer,Hanawalt PC & Spivak G (2008). Transcription-coupled 25: 95–104. doi:10.1002/ijc.2910250113 PMID:7399748 DNA repair: two decades of progress and surprises. Hughes AM, Armstrong BK, Vajdic CM et  al. (2004). Nat Rev Mol Cell Biol, 9: 958–970. doi:10.1038/nrm2549 Sun exposure may protect against non-Hodgkin PMID:19023283 lymphoma: a case-control study. Int J Cancer, 112:Hartge P, Lim U, Freedman DM et al. (2006). Ultraviolet 865–871. doi:10.1002/ijc.20470 PMID:15386383 radiation, dietary vitamin D, and risk of non-Hodgkin Huncharek M & Kupelnick B (2002). Use of topical lymphoma (United States). Cancer Causes Control, sunscreens and the risk of malignant melanoma: a 17: 1045–1052. doi:10.1007/s10552-006-0040-8 meta-analysis of 9067 patients from 11 case-control PMID:16933055 studies. Am J Public Health, 92: 1173–1177. doi:10.2105/ AJPH.92.7.1173 PMID:1208470494
    • Solar and UV radiationHussein MR, Haemel AK, Sudilovsky O, Wood GS (2005). Biol, 89-90: 549–552. doi:10.1016/j.jsbmb.2004.03.067 Genomic instability in radial growth phase melanoma PMID:15225836 cell lines after ultraviolet irradiation. J Clin Pathol, 58: John EM, Koo J, Schwartz GG (2007). Sun exposure and 389–396. doi:10.1136/jcp.2004.021519 PMID:15790703 prostate cancer risk: evidence for a protective effect ofIARC (1986). Some naturally occurring and synthetic early-life exposure. Cancer Epidemiol Biomarkers Prev, food components, furocoumarins and ultraviolet radi- 16: 1283–1286. doi:10.1158/1055-9965.EPI-06-1053 ation. IARC Monogr Eval Carcinog Risk Chem Hum, 40: PMID:17548698 1–415. PMID:3472998 John EM, Schwartz GG, Dreon DM, Koo J (1999). VitaminIARC (1992). IARC Monographs on the evaluation of D and breast cancer risk: the NHANES I Epidemiologic carcinogenic risks to humans. Solar and ultraviolet follow-up study, 1971–1975 to 1992. National Health radiation. IARC Monogr Eval Carcinog Risks Hum, 55: and Nutrition Examination Survey. Cancer Epidemiol 1–316. PMID:1345607 Biomarkers Prev, 8: 399–406. PMID:10350434IARC (2001). IARC Handbooks of Cancer Prevention – John EM, Schwartz GG, Koo J et al. (2005). Sun exposure, Sunscreens. Lyon, France: IARC. vitamin D receptor gene polymorphisms, and risk ofIARC (2006a). IARC Working Group Reports – Exposure advanced prostate cancer. Cancer Res, 65: 5470–5479. to artificial UV radiation and skin cancer. Lyon, France: doi:10.1158/0008-5472.CAN-04-3134 PMID:15958597 IARC. Kampman E, Slattery ML, Caan B, Potter JD (2000).IARC (2006b). TP53 mutation database. Available at Calcium, vitamin D, sunshine exposure, dairy products http://www-p53.iarc.fr. and colon cancer risk (United States). Cancer CausesIARC (2007b). Human papillomaviruses. IARC Monogr Control, 11: 459–466. doi:10.1023/A:1008914108739 Eval Carcinog Risks Hum, 90: 1–636. PMID:18354839 PMID:10877339IARC (2007a). The association of use of sunbeds with cuta- Karipidis KK, Benke G, Sim MR et al. (2007). Occupational neous malignant melanoma and other skin cancers: exposure to ionizing and non-ionizing radiation A systematic review. Int J Cancer, 120: 1116–1122. and risk of non-Hodgkin lymphoma. Int Arch Occup doi:10.1002/ijc.22453 PMID:17131335 Environ Health, 80: 663–670. doi:10.1007/s00420-007-IARC (2008). IARC Working Group Reports – Vitamin D 0177-0 PMID:17334774 and Cancer. Lyon, France: IARC. Keller AZ (1970). Cellular types, survival, race, nativity,IARC (2012). Pharmaceuticals IARC Monogr Eval occupations, habits and associated diseases in the Carcinog Risk Chem Hum, 100A: 1–437. pathogenesis of lip cancers. Am J Epidemiol, 91:Ichihashi M, Nagai H, Matsunaga K (2004). Sunlight is an 486–499. PMID:5438996 important causative factor of recurrent herpes simplex. Kelsall SR & Mintz B (1998). Metastatic cutaneous Cutis, 74: Suppl14–18. PMID:15603217 melanoma promoted by ultraviolet radiation in miceIkehata H, Kawai K, Komura J et al. (2008). UVA1 geno- with transgene-initiated low melanoma susceptibility. toxicity is mediated not by oxidative damage but by Cancer Res, 58: 4061–4065. PMID:9751610 cyclobutane pyrimidine dimers in normal mouse Kielbassa C, Roza L, Epe B (1997). Wavelength depend- skin. J Invest Dermatol, 128: 2289–2296. doi:10.1038/ ence of oxidative DNA damage induced by UV and jid.2008.61 PMID:18356809 visible light. Carcinogenesis, 18: 811–816. doi:10.1093/Ikehata H & Ono T (2007). Significance of CpG methyla- carcin/18.4.811 PMID:9111219 tion for solar UV-induced mutagenesis and carcino- Kim KJ, Chakrabarty I, Li GZ et al. (2002). Modulation of genesis in skin. Photochem Photobiol, 83: 196–204. base excision repair alters cellular sensitivity to UVA1 PMID:16620158 but not to UVB1. Photochem Photobiol, 75: 507–512.Inskip PD, Devesa SS, Fraumeni JF Jr (2003). Trends doi:10.1562/0031-8655(2002)075<0507:MOBERA>2. in the incidence of ocular melanoma in the United 0.CO;2 PMID:12017477 States, 1974–1998. Cancer Causes Control, 14: 251–257. Kimura T, Suzuki A, Fujita Y et  al. (2003). Conditional doi:10.1023/A:1023684502638 PMID:12814204 loss of PTEN leads to testicular teratoma and enhancesItoh T, Ono T, Yamaizumi M (1994). A new UV-sensitive embryonic germ cell production. Development, 130: syndrome not belonging to any complementation 1691–1700. doi:10.1242/dev.00392 PMID:12620992 groups of xeroderma pigmentosum or Cockayne Klein-Szanto AJ, Silvers WK, Mintz B (1994). Ultraviolet syndrome: siblings showing biochemical characteristics radiation-induced malignant skin melanoma in of Cockayne syndrome without typical clinical mani- melanoma-susceptible transgenic mice. Cancer Res, festations. Mutat Res, 314: 233–248. PMID:7513056 54: 4569–4572. PMID:8062242Jagger J (1985). Solar-UV Actions on Living Cel/s. New Knight JA, Lesosky M, Barnett H et al. (2007). Vitamin D York: Praeger and reduced risk of breast cancer: a population-basedJohn EM, Dreon DM, Koo J, Schwartz GG (2004). case-control study. Cancer Epidemiol Biomarkers Residential sunlight exposure is associated with a Prev, 16: 422–429. doi:10.1158/1055-9965.EPI-06-0865 decreased risk of prostate cancer. J Steroid Biochem Mol PMID:17372236 95
    • IARC MONOGRAPHS – 100DKoch-Paiz CA, Amundson SA, Bittner ML et al. (2004). Li C, Wang LE, Wei Q (2009). DNA repair phenotype and Functional genomics of UV radiation responses in cancer susceptibility–a mini review. Int J Cancer, 124: human cells. Mutat Res, 549: 65–78. PMID:15120963 999–1007. doi:10.1002/ijc.24126 PMID:19065660Kricker A, Armstrong BK, Goumas C et al.for the GEM Li G, Robinson GW, Lesche R et al. (2002). Conditional Study Group (2007). Ambient UV, personal sun expo- loss of PTEN leads to precocious development and sure and risk of multiple primary melanomas. Cancer neoplasia in the mammary gland. Development, 129: Causes Control, 18: 295–304. doi:10.1007/s10552-006- 4159–4170. PMID:12163417 0091-x PMID:17206532 Li W, Judge H, Gragoudas ES et  al. (2000). Patterns ofKricker A, Armstrong BK, Hughes AM et al.Interlymph tumor initiation in choroidal melanoma. Cancer Res, Consortium (2008). Personal sun exposure and risk of 60: 3757–3760. PMID:10919647 non Hodgkin lymphoma: a pooled analysis from the Liaw D, Marsh DJ, Li J et al. (1997). Germline mutations of Interlymph Consortium. Int J Cancer, 122: 144–154. the PTEN gene in Cowden disease, an inherited breast doi:10.1002/ijc.23003 PMID:17708556 and thyroid cancer syndrome. Nat Genet, 16: 64–67.Krüger S, Garbe C, Büttner P et al. (1992). Epidemiologic doi:10.1038/ng0597-64 PMID:9140396 evidence for the role of melanocytic nevi as risk Lim JL & Stern RS (2005). High levels of ultraviolet B markers and direct precursors of cutaneous malignant exposure increase the risk of non-melanoma skin melanoma. Results of a case control study in mela- cancer in psoralen and ultraviolet A-treated patients. noma patients and nonmelanoma control subjects. J Invest Dermatol, 124: 505–513. doi:10.1111/j.0022- J Am Acad Dermatol, 26: 920–926. doi:10.1016/0190- 202X.2005.23618.x PMID:15737190 9622(92)70133-Z PMID:1607409 Limoli CL, Day JP, Ward JF, Morgan WF (1998).Kusewitt DF, Hubbard GB, Warbritton AR et al. (2000). Induction of chromosome aberrations and delayed Cellular origins of ultraviolet radiation-induced genomic instability by photochemical processes. corneal tumours in the grey, short-tailed South Photochem Photobiol, 67: 233–238. doi:10.1562/0031- American opossum (Monodelphis domestica). J 8 6 5 5 (1 9 9 8) 0 6 7< 0 2 3 3 : I O C A A D >2 . 3 . C O ; 2 Comp Pathol, 123: 88–95. doi:10.1053/jcpa.2000.0390 PMID:9487801 PMID:11032660 Little JB (2000). Radiation carcinogenesis. Carcinogenesis,Laden F, Spiegelman D, Neas LM et al. (1997). Geographic 21: 397–404. doi:10.1093/carcin/21.3.397 variation in breast cancer incidence rates in a cohort PMID:10688860 of U.S. women. J Natl Cancer Inst, 89: 1373–1378. Luscombe CJ, Fryer AA, French ME et al. (2001). Exposure doi:10.1093/jnci/89.18.1373 PMID:9308708 to ultraviolet radiation: association with susceptibilityLee E, Koo J, Berger T (2005). UVB phototherapy and skin and age at presentation with prostate cancer. Lancet, cancer risk: a review of the literature. Int J Dermatol, 358: 641–642. doi:10.1016/S0140-6736(01)05788-9 44: 355–360. doi:10.1111/j.1365-4632.2004.02186.x PMID:11530156 PMID:15869531 Lutz JM, Cree I, Sabroe S et  al. (2005). OccupationalLee EY, Williamson R, Watt P et  al. (2006). Sun expo- risks for uveal melanoma results from a case-control sure and host phenotype as predictors of cutaneous study in nine European countries. Cancer Causes melanoma associated with neval remnants or dermal Control, 16: 437–447. doi:10.1007/s10552-004-5029-6 elastosis. Int J Cancer, 119: 636–642. doi:10.1002/ PMID:15953986 ijc.21907 PMID:16572428 Manuguerra M, Saletta F, Karagas MR et al. (2006). XRCC3Lee GA, Williams G, Hirst LW, Green AC (1994). Risk and XPD/ERCC2 single nucleotide polymorphisms and factors in the development of ocular surface epithelial the risk of cancer: a HuGE review. Am J Epidemiol, 164: dysplasia. Ophthalmology, 101: 360–364. PMID:8115157 297–302. doi:10.1093/aje/kwj189 PMID:16707649Lehmann AR (2003). DNA repair-deficient diseases, McKinlay, AE & Diffey, B.L. (1987). A reference action xeroderma pigmentosum, Cockayne syndrome spectrum for ultraviolet induced eryhema in human and trichothiodystrophy. Biochimie, 85: 1101–1111. skin. CIE (Commission Internationale de l’Eclairage) doi:10.1016/j.biochi.2003.09.010 PMID:14726016 1, 6, 17–22Lewis PD, Manshian B, Routledge MN et  al. (2008). McMillan TJ, Leatherman E, Ridley A et  al. (2008). Comparison of induced and cancer-associated Cellular effects of long wavelength UV light (UVA) in mutational spectra using multivariate data analysis. mammalian cells. J Pharm Pharmacol, 60: 969–976. Carcinogenesis, 29: 772–778. doi:10.1093/carcin/ doi:10.1211/jpp.60.8.0004 PMID:18644190 bgn053 PMID:18296683 Milon A, Sottas PE, Bulliard JL, Vernez D (2007). EffectiveLi C, Chen K, Liu Z et al. (2008). Polymorphisms of TP53 exposure to solar UV in building workers: influence Arg72Pro, but not p73 G4C14>A4TA4 and p21 Ser31Arg, of local and individual factors. J Expo Sci Environ contribute to risk of cutaneous melanoma. J Invest Epidemiol, 17: 58–68. doi:10.1038/sj.jes.7500521 Dermatol, 128: 1585–1588. doi:10.1038/sj.jid.5701186 PMID:16926862 PMID:1804945096
    • Solar and UV radiationMing M & He YY (2009). PTEN: new insights into DNA damage. Proc Natl Acad Sci U S A, 106: 6209–6214. its regulation and function in skin cancer. J Invest doi:10.1073/pnas.0902113106 PMID:19329487 Dermatol, 129: 2109–2112. doi:10.1038/jid.2009.79 Neale RE, Davis M, Pandeya N et  al. (2007). Basal cell PMID:19340009 carcinoma on the trunk is associated with excessiveMoan J, Dahlback A, Setlow RB (1999). Epidemiological sun exposure. J Am Acad Dermatol, 56: 380–386. support for an hypothesis for melanoma induc- doi:10.1016/j.jaad.2006.08.039 PMID:17097387 tion indicating a role for UVA radiation. Photochem Nelemans PJ, Rampen FH, Ruiter DJ, Verbeek AL (1995). Photobiol, 70: 243–247. doi:10.1111/j.1751-1097.1999. An addition to the controversy on sunlight exposure tb07995.x PMID:10461463 and melanoma risk: a meta-analytical approach. JMocellin S & Nitti D (2008). Vitamin D receptor polymor- Clin Epidemiol, 48: 1331–1342. doi:10.1016/0895- phisms and the risk of cutaneous melanoma: a system- 4356(95)00032-1 PMID:7490596 atic review and meta-analysis. Cancer, 113: 2398–2407. Newton R, Ferlay J, Reeves G et al. (1996). Effect of ambient doi:10.1002/cncr.23867 PMID:18816636 solar ultraviolet radiation on incidence of squamous-Morales-Suárez-Varela MM, Olsen J, Johansen P et  al. cell carcinoma of the eye. Lancet, 347: 1450–1451. (2006). Occupational sun exposure and mycosis doi:10.1016/S0140-6736(96)91685-2 PMID:8676629 fungoides: a European multicenter case-control study. Newton R, Ziegler J, Ateenyi-Agaba C et  al.Uganda J Occup Environ Med, 48: 390–393. doi:10.1097/01. Kaposi’s Sarcoma Study Group (2002). The epide- jom.0000194160.95468.20 PMID:16607193 miology of conjunctival squamous cell carcinomaMorgan WF (2003). Non-targeted and delayed effects in Uganda. Br J Cancer, 87: 301–308. doi:10.1038/ of exposure to ionizing radiation: I. Radiation- sj.bjc.6600451 PMID:12177799 induced genomic instability and bystander effects Nikolaou VA, Sypsa V, Stefanaki I et al. (2008). Risk asso- in vitro. Radiat Res, 159: 567–580. doi:10.1667/0033- ciations of melanoma in a Southern European popu- 75 8 7(2 0 0 3)159 [0 5 6 7: NA DE O E]2 . 0 .C O ; 2 lation: results of a case/control study. Cancer Causes PMID:12710868 Control, 19: 671–679. doi:10.1007/s10552-008-9130-0Morison WL (1983). Phototherapy and Photochemotherapy PMID:18307049 of Skin Disease. New York: Praeger Noonan FP, Recio JA, Takayama H et al. (2001). NeonatalMoriwaki S & Takahashi Y (2008). Photoaging and DNA sunburn and melanoma in mice. Nature, 413: 271–272. repair. J Dermatol Sci, 50: 169–176. doi:10.1016/j. doi:10.1038/35095108 PMID:11565020 jdermsci.2007.08.011 PMID:17920816 Norval M (2006). The mechanisms and consequencesMoseley H (1988). Non-ionising Radiation: Microwaves, of ultraviolet-induced immunosuppression. Prog Ultraviolet and Laer Radiation, Medical Physics Biophys Mol Biol, 92: 108–118. doi:10.1016/j.pbio- Handbook Volume 18. Bristol: Adam Hilger, pp. 110–154 molbio.2006.02.009 PMID:16564073Mouret S, Baudouin C, Charveron M et  al. (2006). Norval M, McLoone P, Lesiak A, Narbutt J (2008). The Cyclobutane pyrimidine dimers are predominant effect of chronic ultraviolet radiation on the human DNA lesions in whole human skin exposed to UVA immune system. Photochem Photobiol, 84: 19–28. radiation. Proc Natl Acad Sci U S A, 103: 13765–13770. doi:10.1111/j.1751-1097.2007.00239.x PMID:18173697 doi:10.1073/pnas.0604213103 PMID:16954188 Nuss DD, Aeling JL, Clemons DE, Weber WN (1978).Naldi L, Altieri A, Imberti GL et al.Oncology Study Group Multiple hamartoma syndrome (Cowden’s disease). of the Italian Group for Epidemiologic Research in Arch Dermatol, 114: 743–746. doi:10.1001/arch- Dermatology (2005). Sun exposure, phenotypic char- derm.114.5.743 PMID:646396 acteristics, and cutaneous malignant melanoma. An Oikawa S, Tada-Oikawa S, Kawanishi S (2001). Site-specific analysis according to different clinico-pathological DNA damage at the GGG sequence by UVA involves variants and anatomic locations (Italy). Cancer Causes acceleration of telomere shortening. Biochemistry, 40: Control, 16: 893–899. doi:10.1007/s10552-005-2300-4 4763–4768. doi:10.1021/bi002721g PMID:11294644 PMID:16132799 Okuyama R, Tagami H, Aiba S (2008). Notch signaling:Nan H, Kraft P, Qureshi AA et  al. (2009). Genome- its role in epidermal homeostasis and in the patho- Wide Association Study of Tanning Phenotype in a genesis of skin diseases. J Dermatol Sci, 49: 187–194. Population of European Ancestry J Invest Dermatol, doi:10.1016/j.jdermsci.2007.05.017 PMID:17624739 129: 2250–2257. doi: 10.1038 PMID:19340012. Pane AR & Hirst LW (2000). Ultraviolet light exposureNapora C, Cohen EJ, Genvert GI et  al. (1990). Factors as a risk factor for ocular melanoma in Queensland, associated with conjunctival intraepithelial neoplasia: Australia. Ophthalmic Epidemiol, 7: 159–167. a case control study. Ophthalmic Surg, 21: 27–30. PMID:11035552 PMID:2325992 Paz-Elizur T, Elinger D, Leitner-Dagan Y et  al. (2007).Nardo T, Oneda R, Spivak G et al. (2009). A UV-sensitive Development of an enzymatic DNA repair assay syndrome patient with a specific CSA mutation reveals for molecular epidemiology studies: distribution of separable roles for CSA in response to UV and oxidative OGG activity in healthy individuals. DNA Repair 97
    • IARC MONOGRAPHS – 100D (Amst), 6: 45–60. doi:10.1016/j.dnarep.2006.08.003 152: 43–51. doi:10.1111/j.1365-2133.2005.06353.x PMID:16982217 PMID:15656799Pelucchi C, Di Landro A, Naldi L, La Vecchia COncology Ridley AJ, Whiteside JR, McMillan TJ, Allinson SL (2009). Study Group of the Italian Group for Epidemiologic Cellular and sub-cellular responses to UVA in rela- Research in Dermatology (GISED) (2007). Risk factors tion to carcinogenesis. Int J Radiat Biol, 85: 177–195. for histological types and anatomic sites of cuta- doi:10.1080/09553000902740150 PMID:19296341 neous basal-cell carcinoma: an italian case-control Rieger E, Soyer HP, Garbe C et  al. (1995). Overall and study. J Invest Dermatol, 127: 935–944. doi:10.1038/ site-specific risk of malignant melanoma associated sj.jid.5700598 PMID:17068478 with nevus counts at different body sites: a multicenterPerea-Milla López E, Miñarro-Del Moral RM, Martínez- case-control study of the German Central Malignant- García C et  al. (2003). Lifestyles, environmental and Melanoma Registry. Int J Cancer, 62: 393–397. phenotypic factors associated with lip cancer: a case- doi:10.1002/ijc.2910620406 PMID:7635564 control study in southern Spain. Br J Cancer, 88: 1702– Robert C, Muel B, Benoit A et  al. (1996). Cell survival 1707. doi:10.1038/sj.bjc.6600975 PMID:12771984 and shuttle vector mutagenesis induced by ultravioletPetridou ET, Dikalioti SK, Skalkidou A et  al.Childhood A and ultraviolet B radiation in a human cell line. J Hematology-Oncology Group (2007). Sun exposure, Invest Dermatol, 106: 721–728. doi:10.1111/1523-1747. birth weight, and childhood lymphomas: a case control ep12345616 PMID:8618011 study in Greece. Cancer Causes Control, 18: 1031–1037. Robinson ES, Hubbard GB, Colon G, Vandeberg JL (1998). doi:10.1007/s10552-007-9044-2 PMID:17653828 Low-dose ultraviolet exposure early in developmentPhillipson RP, Tobi SE, Morris JA, McMillan TJ (2002). can lead to widespread melanoma in the opossum UV-A induces persistent genomic instability in human model. Int J Exp Pathol, 79: 235–244. PMID:9797719 keratinocytes through an oxidative stress mechanism. Robinson ES, VandeBerg JL, Hubbard GB, Dooley TP Free Radic Biol Med, 32: 474–480. doi:10.1016/S0891- (1994). Malignant melanoma in ultraviolet irradiated 5849(01)00829-2 PMID:11864787 laboratory opossums: initiation in suckling young,Pogoda JM & Preston-Martin S (1996). Solar radiation, metastasis in adults, and xenograft behavior in nude lip protection, and lip cancer risk in Los Angeles mice. Cancer Res, 54: 5986–5991. PMID:7954432 County women (California, United States). Cancer Rochette PJ, Lacoste S, Therrien JP et al. (2009). Influence Causes Control, 7: 458–463. doi:10.1007/BF00052672 of cytosine methylation on ultraviolet-induced cyclob- PMID:8813434 utane pyrimidine dimer formation in genomic DNA.Pouget JP, Douki T, Richard MJ, Cadet J (2000). DNA Mutat Res, 665: 7–13. PMID:19427505 damage induced in cells by gamma and UVA radia- Rochette PJ, Therrien JP, Drouin R et  al. (2003). tion as measured by HPLC/GC-MS and HPLC-EC UVA-induced cyclobutane pyrimidine dimers form and Comet assay. Chem Res Toxicol, 13: 541–549. predominantly at thymine-thymine dipyrimidines doi:10.1021/tx000020e PMID:10898585 and correlate with the mutation spectrum in rodentPovey JE, Darakhshan F, Robertson K et al. (2007). DNA cells. Nucleic Acids Res, 31: 2786–2794. doi:10.1093/nar/ repair gene polymorphisms and genetic predisposition gkg402 PMID:12771205 to cutaneous melanoma. Carcinogenesis, 28: 1087–1093. Roffo AH (1934). Cancer and the sun: carcinomas and doi:10.1093/carcin/bgl257 PMID:17210993 sarcomas caused by the action of the sun in toto (Fr.).Purdue MP, From L, Armstrong BK et al.for the Genes, Bull Assoc Fr Etud Cancer, 23: 590–616. Environment, and Melanoma Study Group (2005). Roffo AH (1939). Physico-chemical etiology of cancer Etiologic and other factors predicting nevus-associated (with special emphasis on the association with solar cutaneous malignant melanoma. Cancer Epidemiol radiation) (Ger.). Strahlentherapie, 66: 328–350. Biomarkers Prev, 14: 2015–2022. doi:10.1158/1055- Rossman TG, Uddin AN, Burns FJ, Bosland MC (2002). 9965.EPI-05-0097 PMID:16103454 Arsenite cocarcinogenesis: an animal model derivedRass K & Reichrath J (2008). UV damage and DNA repair from genetic toxicology studies. Environ Health in malignant melanoma and nonmelanoma skin cancer. Perspect, 110: Suppl 5749–752. PMID:12426125 Adv Exp Med Biol, 624: 162–178. doi:10.1007/978-0- Rünger TM & Kappes UP (2008). Mechanisms of muta- 387-77574-6_13 PMID:18348455 tion formation with long-wave ultraviolet light (UVA).Reeve VE, Bosnic M, Boehm-Wilcox C (1996). Dependence Photodermatol Photoimmunol Photomed, 24: 2–10. of photocarcinogenesis and photoimmunosuppres- doi:10.1111/j.1600-0781.2008.00319.x PMID:18201350 sion in the hairless mouse on dietary polyunsaturated Sabburg J, Parisi AV, Wong J (2001). Effect of cloud on UVA fat. Cancer Lett, 108: 271–279. doi:10.1016/S0304- and exposure to humans. Photochem Photobiol, 74: 3835(96)04460-6 PMID:8973605 412–416. doi:10.1562/0031-8655(2001)074<0412:EOCOReifenberger J, Wolter M, Knobbe CB et al. (2005). Somatic UA>2.0.CO;2 PMID:11594054 mutations in the PTCH, SMOH, SUFUH and TP53 Sabburg J & Wong J (2000). The effect of clouds on genes in sporadic basal cell carcinomas. Br J Dermatol, enhancing UVB irradiance at the earth’s surface:98
    • Solar and UV radiation a one year study. Geophys Res Lett, 27: 3337–3340. Cell, 128: 157–170. doi:10.1016/j.cell.2006.11.042 doi:10.1029/2000GL011683 PMID:17218262Sabourin CL, Kusewitt DF, Fry RJ, Ley RD (1993). Siemiatycki J (1991). Risk factors for cancer in workplace. Ultraviolet radiation-induced corneal tumours in Boca Raton, FL: CRC Press the South American opossum, Monodelphis domes- Siskind V, Whiteman DC, Aitken JF et al. (2005). An anal- tica. J Comp Pathol, 108: 343–359. doi:10.1016/S0021- ysis of risk factors for cutaneous melanoma by anatom- 9975(08)80206-X PMID:8366202 ical site (Australia). Cancer Causes Control, 16: 193–199.Sambuco CP, Forbes PD, Davies RE et  al. (2003). doi:10.1007/s10552-004-4325-5 PMID:15947871 Photocarcinogenesis: measuring the reproducibility Smedby KE, Hjalgrim H, Melbye M et al. (2005). Ultraviolet of a biologic response to ultraviolet radiation expo- radiation exposure and risk of malignant lymphomas. sure in mice. Front Biosci, 8: a26–a33. doi:10.2741/933 J Natl Cancer Inst, 97: 199–209. doi:10.1093/jnci/dji022 PMID:12456327 PMID:15687363Sand M, Gambichler T, Sand D et al. (2009). MicroRNAs Soni LK, Hou L, Gapstur SM et al. (2007). Sun exposure and the skin: tiny players in the body’s largest and non-Hodgkin lymphoma: a population-based, organ. J Dermatol Sci, 53: 169–175. doi:10.1016/j. case-control study. Eur J Cancer, 43: 2388–2395. jdermsci.2008.10.004 PMID:19058951 doi:10.1016/j.ejca.2007.06.018 PMID:17686627Sarasin A & Stary A (2007). New insights for understanding Spitzer WO, Hill GB, Chambers LW et  al. (1975). The the transcription-coupled repair pathway. DNA Repair occupation of fishing as a risk factor in cancer of the (Amst), 6: 265–269. doi:10.1016/j.dnarep.2006.12.001 lip. N Engl J Med, 293: 419–424. PMID:1152953 PMID:17194629 Spivak G (2005). UV-sensitive syndrome. Mutat Res, 577:Schmidt-Pokrzywniak A, Jöckel KH, Bornfeld N et  al. 162–169. PMID:15916784 (2009). Positive interaction between light iris color Stacey SN, Gudbjartsson DF, Sulem P et  al. (2008). and ultraviolet radiation in relation to the risk of Common variants on 1p36 and 1q42 are associated uveal melanoma: a case-control study. Ophthalmology, with cutaneous basal cell carcinoma but not with mela- 116: 340–348. doi:10.1016/j.ophtha.2008.09.040 noma or pigmentation traits. Nat Genet, 40: 1313–1318. PMID:19091418 doi:10.1038/ng.234 PMID:18849993Schröder JM, Reich K, Kabashima K et al. (2006). Who is Stefanini M, Kraemer KHK (2008). Xeroderma pigmen- really in control of skin immunity under physiological tosum. In: Neurocutaneous Disorders: Phakomatoses circumstances - lymphocytes, dendritic cells or kerati- and Hamartoneoplastic Syndromes. Ruggieri M, nocytes? Exp Dermatol, 15: 913–929. PMID:17002689 Pascual-Castroviejo I, Di Rocco C, editors. New York:Schwartz LH, Ferrand R, Boelle PY et al. (1997). Lack of Springer, pp. 771–792. correlation between the location of choroidal melanoma Sturm RA (2009). Molecular genetics of human pigmen- and ultraviolet-radiation dose distribution. Radiat Res, tation diversity. Hum Mol Genet, 18: R1R9–R17. 147: 451–456. doi:10.2307/3579502 PMID:9092925 doi:10.1093/hmg/ddp003 PMID:19297406Seddon JM, Gragoudas ES, Glynn RJ et  al. (1990). Host Sun EC, Fears TR, Goedert JJ (1997). Epidemiology of factors, UV radiation, and risk of uveal melanoma. A squamous cell conjunctival cancer. Cancer Epidemiol case-control study. Arch Ophthalmol, 108: 1274–1280. Biomarkers Prev, 6: 73–77. PMID:9037556 PMID:2400347 Suzuki A, de la Pompa JL, Stambolic V et al. (1998). HighSetlow RB, Grist E, Thompson K, Woodhead AD (1993). cancer susceptibility and embryonic lethality associ- Wavelengths effective in induction of malignant ated with mutation of the PTEN tumor suppressor gene melanoma. Proc Natl Acad Sci USA, 90: 6666–6670. in mice. Curr Biol, 8: 1169–1178. doi:10.1016/S0960- doi:10.1073/pnas.90.14.6666 PMID:8341684 9822(07)00488-5 PMID:9799734Setlow RB, Woodhead AD, Grist E (1989). Animal model Suzuki A, Itami S, Ohishi M et al. (2003). Keratinocyte- for ultraviolet radiation-induced melanoma: platyfish- specific Pten deficiency results in epidermal hyperplasia, swordtail hybrid. Proc Natl Acad Sci USA, 86: 8922– accelerated hair follicle morphogenesis and tumor 8926. doi:10.1073/pnas.86.22.8922 PMID:2813430 formation. Cancer Res, 63: 674–681. PMID:12566313Shah CP, Weis E, Lajous M et al. (2005). Intermittent and Suzuki H, Kalair W, Shivji GM et  al. (2001). Impaired chronic ultraviolet light exposure and uveal melanoma: ultraviolet-B-induced cytokine induction in xero- a meta-analysis. Ophthalmology, 112: 1599–1607. derma pigmentosum fibroblasts. J Invest Dermatol, doi:10.1016/j.ophtha.2005.04.020 PMID:16051363 117: 1151–1155. doi:10.1046/j.0022-202x.2001.01525.xSharma KK & Santhoshkumar P (2009). Lens aging: effects PMID:11710926 of crystallins. Biochim Biophys Acta, 1790: 1095–1108. Tavani A, Bosetti C, Franceschi S et  al. (2006). PMID:19463898 Occupational exposure to ultraviolet radiation andShen WH, Balajee AS, Wang J et al. (2007). Essential role for risk of non-Hodgkin lymphoma. Eur J Cancer Prev, nuclear PTEN in maintaining chromosomal integrity. 15: 453–457. doi:10.1097/00008469-200610000-00011 PMID:16912575 99
    • IARC MONOGRAPHS – 100DTong Y, Smith MA, Tucker SB (1997). Chronic ultravi- Vajdic CM, Kricker A, Giblin M et al. (2002). Sun expo- olet exposure-induced p53 gene alterations in Sencar sure predicts risk of ocular melanoma in Australia. mouse skin carcinogenesis model. J Toxicol Environ Int J Cancer, 101: 175–182. doi:10.1002/ijc.10579 Health, 51: 219–234. doi:10.1080/00984109708984023 PMID:12209995 PMID:9183379 Vajdic CM, Kricker A, Giblin M et  al. (2004). ArtificialTong Y, Tucker SB, Smith MA (1998). Expression of Hras- ultraviolet radiation and ocular melanoma in Australia. p21 and keratin K13 in UVR-induced skin tumors in Int J Cancer, 112: 896–900. doi:10.1002/ijc.20476 Sencar mice. J Toxicol Environ Health A, 53: 439–453. PMID:15386378 doi:10.1080/009841098159178 PMID:9537281 Valencia A & Kochevar IE (2008). Nox1-based NADPHTrempus CS, Mahler JF, Ananthaswamy HN et  al. oxidase is the major source of UVA-induced reac- (1998). Photocarcinogenesis and susceptibility to UV tive oxygen species in human keratinocytes. J Invest radiation in the v-Ha-ras transgenic Tg.AC mouse. Dermatol, 128: 214–222. doi:10.1038/sj.jid.5700960 J Invest Dermatol, 111: 445–451. doi:10.1046/j.1523- PMID:17611574 1747.1998.00237.x PMID:9740239 Valencia A, Rajadurai A, Carle AB, Kochevar IE (2006).Trojan J, Plotz G, Brieger A et  al. (2001). Activation of 7-Dehydrocholesterol enhances ultraviolet A-induced a cryptic splice site of PTEN and loss of heterozy- oxidative stress in keratinocytes: roles of NADPH gosity in benign skin lesions in Cowden disease. J oxidase, mitochondria, and lipid rafts. Free Radic Invest Dermatol, 117: 1650–1653. doi:10.1046/j.0022- Biol Med, 41: 1704–1718. doi:10.1016/j.freeradbi- 202x.2001.01954.x PMID:11886535 omed.2006.09.006 PMID:17145559Tucker MA, Shields JA, Hartge P et  al. (1985). Sunlight van Hees CL, de Boer A, Jager MJ et al. (1994). Are atypical exposure as risk factor for intraocular malignant nevi a risk factor for uveal melanoma? A case-control melanoma. N Engl J Med, 313: 789–792. PMID:4033707 study. J Invest Dermatol, 103: 202–205. doi:10.1111/1523-Tulvatana W, Bhattarakosol P, Sansopha L et  al. (2003). 1747.ep12392754 PMID:8040610 Risk factors for conjunctival squamous cell neoplasia: van Kranen HJ, de Laat A, van de Ven J et al. (1997). Low a matched case-control study. Br J Ophthalmol, 87: incidence of p53 mutations in UVA (365-nm)-induced 396–398. doi:10.1136/bjo.87.4.396 PMID:12642297 skin tumors in hairless mice. Cancer Res, 57: 1238–Tuohimaa P, Pukkala E, Scélo G et al. (2007). Does solar 1240. PMID:9102205 exposure, as indicated by the non-melanoma skin van Wijngaarden E & Savitz DA (2001). Occupational cancers, protect from solid cancers: vitamin D as a sunlight exposure and mortality from non-Hodgkin possible explanation. Eur J Cancer, 43: 1701–1712. lymphoma among electric utility workers. J Occup doi:10.1016/j.ejca.2007.04.018 PMID:17540555 Environ Med, 43: 548–553. doi:10.1097/00043764-Uddin AN, Burns FJ, Rossman TG (2005). Vitamin 200106000-00008 PMID:11411327 E and organoselenium prevent the cocarcinogenic Veierød MB, Weiderpass E, Lund E et  al. (2004). Re: A activity of arsenite with solar UVR in mouse skin. prospective study of pigmentation, sun exposure, and Carcinogenesis, 26: 2179–2186. doi:10.1093/carcin/ risk of cutaneous malignant melanoma in women. J bgi180 PMID:16014701 Natl Cancer Inst, 96: 335–338.Uddin AN, Burns FJ, Rossman TG et al. (2007). Dietary Veierød MB, Weiderpass E, Thörn M et  al. (2003). A chromium and nickel enhance UV-carcinogenesis prospective study of pigmentation, sun exposure, and in skin of hairless mice. Toxicol Appl Pharmacol, risk of cutaneous malignant melanoma in women. J 221: 329–338. doi:10.1016/j.taap.2007.03.030 Natl Cancer Inst, 95: 1530–1538. PMID:14559875 PMID:17499830 Walther U, Kron M, Sander S et  al. (2004). Risk andUnnikrishnan A, Raffoul JJ, Patel HV et  al. (2009). protective factors for sporadic basal cell carcinoma: Oxidative stress alters base excision repair pathway and results of a two-centre case-control study in southern increases apoptotic response in apurinic/apyrimidinic Germany. Clinical actinic elastosis may be a protective endonuclease 1/redox factor-1 haploinsufficient mice. factor. Br J Dermatol, 151: 170–178. doi:10.1111/j.1365- Free Radic Biol Med, 46: 1488–1499. doi:10.1016/j. 2133.2004.06030.x PMID:15270887 freeradbiomed.2009.02.021 PMID:19268524 Wang LE, Li C, Strom SS et al. (2007). Repair capacity forUrso C, Giannotti V, Reali UM et al. (1991). Spatial associa- UV light induced DNA damage associated with risk tion of melanocytic naevus and melanoma. Melanoma of nonmelanoma skin cancer and tumor progression. Res, 1: 245–250. doi:10.1097/00008390-199111000- Clin Cancer Res, 13: 6532–6539. doi:10.1158/1078-0432. 00004 PMID:1823633 CCR-07-0969 PMID:17975167Vajdic CM, Kricker A, Giblin M et  al. (2001). Eye color Wang LE, Xiong P, Strom SS et al. (2005). In vitro sensi- and cutaneous nevi predict risk of ocular melanoma tivity to ultraviolet B light and skin cancer risk: a case- in Australia. Int J Cancer, 92: 906–912. doi:10.1002/ control analysis. J Natl Cancer Inst, 97: 1822–1831. ijc.1281 PMID:11351315 doi:10.1093/jnci/dji429 PMID:16368944100
    • Solar and UV radiationWang S, Garcia AJ, Wu M et al. (2006). Pten deletion leads Wikonkal NM & Brash DE (1999). Ultraviolet radiation to the expansion of a prostatic stem/progenitor cell induced signature mutations in photocarcinogenesis. subpopulation and tumor initiation. Proc Natl Acad J Investig Dermatol Symp Proc, 4: 6–10. doi:10.1038/ Sci U S A, 103: 1480–1485. doi:10.1073/pnas.0510652103 sj.jidsp.5640173 PMID:10537000 PMID:16432235 Winnepenninckx V & van den Oord JJ (2004).Wang Y, Digiovanna JJ, Stern JB et al. (2009). Evidence of p16INK4A expression in malignant melanomas with ultraviolet type mutations in xeroderma pigmentosum or without a contiguous naevus remnant: a clue to melanomas. Proc Natl Acad Sci U S A, 106: 6279–6284. their divergent pathogenesis? Melanoma Res, 14: doi:10.1073/pnas.0812401106 PMID:19329485 321–322. doi:10.1097/01.cmr.0000134855.12474.f3Weihkopf T, Becker N, Nieters A et al. (2007). Sun expo- PMID:15305164 sure and malignant lymphoma: a population-based World Meteorological Organization (WMO) (2007). case-control study in Germany. Int J Cancer, 120: Scientific Assessment of Ozone Depletion: 2006. Global 2445–2451. doi:10.1002/ijc.22492 PMID:17311289 Ozone Research and Monitoring Project- Report n°50.Weinstock MA, Colditz GA, Willett WC et  al. (1989). Geneva. pp. 1–572. Moles and site-specific risk of nonfamilial cutaneous Yamazaki F, Okamoto H, Matsumura Y et  al. (2005). malignant melanoma in women. J Natl Cancer Inst, 81: Development of a new mouse model (xeroderma 948–952. doi:10.1093/jnci/81.12.948 PMID:2733040 pigmentosum a-deficient, stem cell factor-transgenic)Weis E, Shah CP, Lajous M et  al. (2006). The asso- of ultraviolet B-induced melanoma. J Invest Dermatol, ciation between host susceptibility factors and uveal 125: 521–525. doi:10.1111/j.0022-202X.2005.23753.x melanoma: a meta-analysis. Arch Ophthalmol, 124: PMID:16117793 54–60. doi:10.1001/archopht.124.1.54 PMID:16401785 Young AR (2006). Acute effects of UVR on human eyes andWeischer M, Blum A, Eberhard F et  al. (2004). No skin. Prog Biophys Mol Biol, 92: 80–85. doi:10.1016/j. evidence for increased skin cancer risk in psoriasis pbiomolbio.2006.02.005 PMID:16600340 patients treated with broadband or narrowband UVB Zhang G, Njauw CN, Park JM et al. (2008). EphA2 is an phototherapy: a first retrospective study. Acta Derm essential mediator of UV radiation-induced apoptosis. Venereol, 84: 370–374. doi:10.1080/00015550410026948 Cancer Res, 68: 1691–1696. doi:10.1158/0008-5472. PMID:15370703 CAN-07-2372 PMID:18339848Wheeler DL, Li Y, Verma AK (2005). Protein kinase C Zhang Y, Holford TR, Leaderer B et  al. (2007). epsilon signals ultraviolet light-induced cutaneous Ultraviolet radiation exposure and risk of non-Hodg- damage and development of squamous cell carcinoma kin’s lymphoma. Am J Epidemiol, 165: 1255–1264. possibly through Induction of specific cytokines in a doi:10.1093/aje/kwm020 PMID:17327216 paracrine mechanism. Photochem Photobiol, 81: 9–18. doi:10.1562/2004-08-12-RA-271.1 PMID:15458367Wheeler DL, Martin KE, Ness KJ et  al. (2004). Protein kinase C epsilon is an endogenous photosensitizer that enhances ultraviolet radiation-induced cutaneous damage and development of squamous cell carci- nomas. Cancer Res, 64: 7756–7765. doi:10.1158/0008- 5472.CAN-04-1881 PMID:15520180Whiteman DC, Parsons PG, Green AC (1998). p53 expression and risk factors for cutaneous melanoma: a case-control study. Int J Cancer, 77: 843–848. doi:10.1002/(SICI)1097-0215(19980911)77:6<843::AID- IJC8>3.0.CO;2-U PMID:9714052Whiteman DC, Stickley M, Watt P et  al. (2006). Anatomic site, sun exposure, and risk of cutaneous melanoma. J Clin Oncol, 24: 3172–3177. doi:10.1200/ JCO.2006.06.1325 PMID:16809740Whiteman DC, Watt P, Purdie DM et  al. (2003). Melanocytic nevi, solar keratoses, and divergent path- ways to cutaneous melanoma. J Natl Cancer Inst, 95: 806–812. doi:10.1093/jnci/95.11.806 PMID:12783935Whiteside JR & McMillan TJ (2009). A bystander effect is induced in human cells treated with UVA radia- tion but not UVB radiation. Radiat Res, 171: 204–211. doi:10.1667/RR1508.1 PMID:19267546 101
    • X- AND γ-RADIATIONX-and γ-radiation were considered by a previous IARC Working Group in 1999 (IARC, 2000).Since that time, new data have become available, these have been incorporated into theMonograph, and taken into consideration in the present evaluation.1. Exposure Data 1.1.1 X- and γ-rays X- and γ-rays are both electromagnetic1.1 Physical properties radiations distinguished mainly by their origin. X-rays are photons emitted from the electron Radiation sources can be either external to shells surrounding the atomic nucleus or duringthe body, such as medical X-rays, or through the slowing down of electrons or other chargeddeposition on the Earth’s surface, or internal. particles. The term γ-rays is usually applied toInternal exposure can result from the ingestion of radiation originating from the atomic nucleus,contaminated foods, inhalation, dermal absorp- and from particle annihilation. The energytion, or injection of radionuclides. The effects of ranges of X- and γ-rays overlap considerably withradiation are directly related to the dose that an X-rays having energies upwards from a few tensorgan receives, and any differences between the of eV (the shortest ultraviolet wavelengths), andeffects of external and internal sources is in large γ-ray energies extending up to a few tens of MeV.part related to the distribution of dose withinand among body organs (IARC, 2001). (a) X-rays The activity of a radionuclide is defined as the Characteristic X-rays are emitted duringnumber of nuclear transformations occurring per transitions of electrons in excited atomic shellsunit of time. The standard unit is the becquerel to lower energy states: they have line spectra(Bq), which is 1 disintegration per second. characteristic of the corresponding element. AHistorically, the curie (Ci) (1 Ci = 3.7 × 1010 Bq) continuous X-ray spectrum is produced whenwas also used. The energy of radiation emitted charged particles, normally electrons, are decel-during the nuclear transformation is normally erated or deflected (in an electric or magnetic fieldmeasured in units of electron-volts (eV), as this such as that close to a nucleus). This is known asis a small unit, it is commonly represented as kilo ‘bremsstrahlung’ from the German for ‘brakingeV (keV) (1000 eV) or mega eV (MeV) (106 eV). radiation’. For example, X-ray tubes generate bremsst- rahlung and characteristic X-rays (see Fig.  1.1). X-rays for medical exposures are classified, 103
    • IARC MONOGRAPHS – 100D Fig. 1.1 Bremsstrahlung X-ray spectrum from a tungsten target at 90 kVp with 1 mm aluminium filtration. The peaks between 57 and 70 keV are due to characteristic X-rays of tungsten 100 80Fluence (arbitrary units) 60 40 20 0 0 20 40 60 80 100 Photon Energy (keV) Adapted from IPEM (1997) according to their kVp (the peak applied voltage a limited area of the body, with restricted irradia- for an exposure) from ultrasoft (5–20  kVp), to tion of adjacent tissue (IARC, 2000). very hard (>  250  kVp). Extremely hard X-rays are generated with betatrons, synchrotrons, and (b) γ-Rays linear accelerators in the MeV range. γ-Ray photons are usually emitted during X-rays are used in many medical and tech- transformations in atomic nuclei. They have nical applications. The most common are diag- widely different energies in the range of 0.01–17.6 nostic X-ray examinations of the human body, MeV. Such radiation can also be produced by the and the analysis of materials. In X-ray therapy, decay of elementary particles, the annihilation the biological effect of X-rays is used to destroy of electron–positron pairs, and the acceleration malignant tissue. It is applied mainly to treat and deceleration of high-energy electrons in cancer patients, when high doses are delivered to 104
    • X- and γ-radiationcosmic magnetic fields or in elementary particle 1.1.4 β-Particlesaccelerators. β-Particles are emitted from the nucleus of a radionuclide and consist of electrons or1.1.2 Neutrons anti-electrons, these electrons have a mass of Neutrons are uncharged particles which, approximately 0.00055 of an atomic mass unit. β-along with protons, form the nuclei of atoms. (negatron) radiation is the result of the conversionWhereas X- and γ-rays interact primarily with of a neutron into a proton, a negatively chargedorbital electrons, neutrons interact with the electron being emitted as a result. β+ (positron)nucleus of atoms. Neutrons are emitted from radiation is a result of the opposite conversion,nuclei in several ways, in the interaction of a proton is converted to a neutron and an anti-high-energy cosmic radiation with the Earth’s electron, the positively charged equivalent ofatmosphere, and in the fission or fusion of nuclei. an electron, known as a positron, is emitted. βFission neutrons have energies up to several MeV, radiation also results in the production of a thirdand fusion neutrons approximately 10  MeV. body, the first two bodies being the daughterNeutrons can also be produced by the collision nuclide and the electron/positron. The thirdof energetic charged particles (e.g. α-particles, body is an anti-neutrino in the case of β- emis-ions from an accelerator) with a suitable target sion and a neutrino in the case of β+ emission.material. The neutrons emitted are used for radi- Because the energy from β radiation is sharedography and radiotherapy. between the emitted particle and the third body, the energy of β-particles varies, even when the1.1.3 α-particles parent radionuclide is the same (i.e. their energy is not characteristic). The continuum of energies α-particles are emitted from the nucleus of a for a β-particle goes from a lower energy limitradionuclide and consist of two protons, giving of zero to an upper limit set by the maximumthem a +2 charge, and two neutrons bound available energy from the transmutation of thetogether, resulting in an atomic mass of 4, so parent into the daughter (the reaction energy ‘Q’,they are, in effect, high energy helium-4 (4He) typically around 1 MeV). Many β emitters alsoatome The energy of α-particles typically varies emit γ-rays, those that do not are known as ‘pure’between 4 and 8 MeV, the energy increasing with β emitters. High-energy β-particles can producethe mass of the parent nucleus which emitted it. bremsstrahlung. Emitted β- particles quickly (in aConsequently, emissions from any particular few tens of picoseconds) lose their excess energy,radionuclide are mono-energetic and have a char- and are then indistinguishable from other elec-acteristic energy. Because the energy and mass of trons in the environment. As positrons are anti-an α-particle are significant on an atomic scale, electrons, they are normally rapidly annihilatedthe emission of an α-particle causes the parent/ after they are emitted as a result of collisionsdaughter nucleus to recoil. This α-recoil effect with electrons in the surrounding environment,represents a small, but not negligible, percentage which are also annihilated. The released energy(~2%) of the overall energy released during α manifests itself as two characteristic 0.511 MeVdecay. α-particles rapidly lose energy and acquire γ-rays.electrons from the surrounding environment to For the sake of clarity, β- particles will hence-become inert Helium-4 (their typical lifetime is forth be referred to as β-particles and β+ particlesa few picoseconds). as positrons. 105
    • IARC MONOGRAPHS – 100DFig. 1.2 (a) Depth of penetration of α and β-particles in tissue, for selected energy values; (b) depthof penetration of X- and γ-rays in tissue in which 50% of the radiation energy is lost(a) α-particles 2700 keV 5000 keV 7700 keV β-particles 200 keV 1000 keV 5000 keV(b) X- and y-rays 10 keV 100 keV 1000 keV 10000 keV 0.01 0.1 1 10 100 1000 Range (mm)IARC (2000)1.2 Interactions with matter For epidemiological purposes, the basic physics quantity of the gray (Gy, i.e. joule per kilogram) Different radiation types penetrate matter to a should be used where possible. For X- and γ-rays,different extent and in different ways (Fig. 1.2). X- the radiation-weighting factor has always been 1,and γ-rays, especially those with high energy, can and values for individual organs could thereforepenetrate matter easily, while α- and β-particles equally well be expressed in terms of absorbedare much less penetrating. dose in grays or equivalent dose in sieverts. Ultimately, virtually all the radiation energy Doses may be expressed in terms of effec-from ionizing radiation is transferred to elec- tive (whole-body equivalent) dose (ICRP, 2007).trons, which lose their energy by ionizing the Effective doses should only be used for radiationirradiated medium. protection and regulatory purposes, and with For radiation protection purposes, the caution for general comparisons.International Committee for Radiation Protection(ICRP, 2007) introduced radiation-weighting 1.2.1 X- and γ-raysfactors to take into account the fact the variousradiation types have different relative biological The interaction of X- and γ-rays with mattereffects (RBE). The primary dosimetric quantity is described by the photoelectric effect, Comptonunit of dose taking radiation-weighting factors scattering, and pair production. Photoelectricinto account is the sievert (Sv), which should be absorption dominates at low energies followedused with caution (note that values of radiation- by Compton scattering, and then pair produc-weighting factors have changed over the years). tion as the energy increases. Absorption of very106
    • X- and γ-radiationhigh energy photons results in nuclear disinte- Ca, producing many lower energy particles suchgration. The intensity of X- and γ-rays generally as α-particles, protons, and other neutrons withdecreases with depth. The ability to penetrate a broad distribution of LET. Exposure to high-matter increases with increasing energy and energy neutrons is thus quite distinct from expo-decreases with increasing atomic number of the sure to low-energy neutrons. Neutrons as theyabsorbing material. interact with matter generate γ-rays. The above processes (apart from photodis-integration) all result in the production of elec- 1.2.3 α and β Radiationtrons (or their anti-matter equivalent, positrons)and lower energy X-rays, which undergo further Charged particle radiation, such as α and βabsorption and scattering. The energy of the radiation, is not very penetrating, the maximuminitial photon is thus transferred to electrons that range of an α-particle in tissue is less than 100create ionization leading to significant chemical microns and for β-particles only about a centi-and biological effects such as degradation of metre. This means that, for external exposures,DNA. these types of radiation are often a much lower or, in the case of α-particles, insignificant radiolog- ical hazard when compared to highly penetrating1.2.2 Neutrons radiation such as X- and γ-rays. However, when Neutrons are captured or scattered by matter. α- and β-particle emitters become internallyThe likelihood of interactions occurring between deposited within living tissues, their radiationsneutrons and atoms of a material (i.e. the neutron deposit most, if not all, of their energy withincross-section) is unique for each nuclide, and the that tissue. α-particles in particular are rela-nature of these interactions are complex. Thermal tively massive, doubly charged, and very denselyneutron-capture cross-sections are generally ionizing. Consequently, they have a substantiallymuch greater than those at higher energies: in enhanced effect on living tissues per unit energy,nuclear power reactors, neutron energies must be compared to X- and γ-rays, and β-particles. Therereduced by collisions with a moderating medium is also some evidence, from radionuclides such(usually water or graphite) to thermal energies as tritium, that β-particle radiation may have awhere the cross-sections allow a chain reaction slightly greater radiological effect per unit energyto proceed. than X- and γ-rays. The mean free path of neutrons in tissues Neutrinos and anti-neutrinos interact veryvaries with their energy from a fraction of, to weakly with matter, therefore present no radio-several tens of centimetres. In tissue, neutrons logical hazard, and will not be considered further.interact with hydrogen nuclei. The recoilingnuclei (low-energy proton) form densely ionizing 1.2.4 Otherstracks, with a high linear energy transfer (LET)which are efficient in producing biological injury. Other types of ionizing radiation that interactThe ICRP (2007) has therefore defined radiation- with matter include cosmic rays, protons, muons,weighting factors for estimating the risks associ- and heavy ions. As for the other forms of radia-ated with exposure to neutrons, which are larger tion described above, these will all ultimatelythan those for X- or γ-rays for the same tissue produce ionizing electrons.dose. In tissue, neutrons with energy >  50 MeVinteract mainly with nuclei such as C, N, O, and 107
    • IARC MONOGRAPHS – 100D1.2.5 Energy loss process and Physics’ (Lide, 2005–2006), World Nuclear Association Reference Documents ‘Radiological As described above, the indirectly ionizing and Chemical Fact Sheets to Support Health Riskradiations all interact to produce ionizing parti- Analyses for Contaminated Areas’ (Argonnecles; electrons, protons, α–particles, and heavy National Laboratory, 2007) and ICRP (1983,ions. 2008). The information provided is not intended All ionizing particles interact with the atomic to be definitive or comprehensive.electrons of the medium through which theypass to produce secondary electrons with a range (a) Tritiumof energies. In turn, these electrons create more Tritium (3H) is an isotope of the hydrogenelectrons (mainly low energy) until all electrons atom. 3H is naturally produced by interactionsare completely slowed down in the medium. At between cosmic radiation and nitrogen andthe end of their tracks, electrons of less than oxygen in the atmosphere at a rate of approxi-about 500 eV form clusters of ionization. An mately 0.4  kg/year. However, environmentalanalysis of low-energy electron track structure concentrations of naturally occurring 3H arein liquid water is given by Wilson et al. (2004). low (the total steady-state global inventory from this route of production is ~7 kg) due to global1.2.6 Radionuclides, internally deposited dispersal, and because they are constantly being For the purposes of this IARC Monograph, depleted by radioactive decay as a result of itsinternally deposited radionuclides are defined comparatively short half-life. 3H gas will tend toas radionuclides that have been taken into the bond with any available moisture to form tritiatedbody (encapsulated radionuclides entering the water, which, from a biochemical perspective,body, as in brachytherapy, are not discussed in behaves like any other water in the environment.this Monograph because they are considered as 3 H is a pure, low energy, β emitter that hasexternal exposure). These radionuclides may a half-life of 12.35 years, it decays to helium-3,emit any form of radiation, but in practice it is which is stable.those that emit charged particles, α (α) and β (β-)/ Although 3H is not a particularly abundant(β+) radiation, that tend to be the most radiologi- fission product (uranium-235 fission yield iscally significant. 0.01%) and the atmospheric testing of nuclear In theory, any radionuclide could become weapons has largely ceased, the quantity of 3Hinternally deposited but only a subset of radio- in the environment from previous tests stillnuclides which are relatively available from exceeds that from natural cosmogenic produc-nuclear weapons tests, the Chernobyl accident, tion. However, once again due to global dispersalor from radiotherapy and radiodiagnosis, and of this material, concentrations involved are low.known to have the potential to affect cancer 3 H is a strategic material in the production ofrisks are considered here. To understand the nuclear weapons; and because of the nature of thisoccurrence of radionuclides within the environ- application, specific information on the amountsment and their potential to result in significant of 3H generated and used for this purpose areindividual exposures, it is necessary to have difficult to obtain. Production of 3H for weaponsome knowledge of their physical and chemical purposes involves neutron bombardment ofproperties as well as their abundance—this lithium-6 in nuclear reactors. The 6Li atom,information has been collated from various with three protons and three neutrons and thesources: The CRC ‘Handbook of Chemistry captured neutron combine to form a lithium-7 atom, with three protons and four neutrons,108
    • X- and γ-radiationwhich instantaneously splits to an atom of 3H (c) Strontium-90(one proton and two neutrons) and one atom of Strontium is a relatively abundant, chemically4 He (two protons and two neutrons). The United reactive metal, which oxidizes readily. NaturallyStates of America is thought to have produced occurring strontium has four stable isotopes 84Sr,over 200 kg of 3H for military purposes but much 86 Sr, 87Sr, and 88Sr. The chemistry of strontium hasof this has now decayed to 3He, and only ~75 kg similarities to that of calcium.remains (Argonne National Laboratory, 2007). 90 Sr is a man-made isotope that is a pure β Heavy water (2H2O) moderated reactors, such particle emitter with a half-life of 29.12 years. Itas the CANada Deuterium Uranium (CANDU) decays to Yttrium-90, which is a short-lived highdesigns, produce substantial amounts of 3H energy β particle emitter, which greatly increasesas a by-product, due to neutron capture in the the radiological effect of 90Sr exposures. 90Sr ismoderator. 3H is routinely removed from the mostly produced as a result of nuclear fission,heavy water used in CANDU reactors in Canada, either in nuclear weapons or batteries/reactors,and approximately 1–2 kg are recovered per year. and is one of the most commonly occurring 3 H can be produced in a particle accelerator fission products (235U fission yield is ~6%). Its rela-by bombarding 3He with neutrons. In addition, tively long half-life results in it being persistent in3 H is used in the manufacture of radionuclide- the environment if it is released. Levels of 90Srlabelled materials for application in medicine, in surface soil due to fallout from atmosphericresearch and industry (and can be released from nuclear weapons tests are around 3.7  Bq/kg onsuch manufacturing plants), and in the use and average.disposal of these materials. 3H has also been usedin luminous paint used in some wristwatches (d) Iodine-131and compasses, and in emergency exit signs, and Iodine is a halogen, it is both volatile andgun-sights (HPA, 2007). reactive, and is not found in its elemental form(b) Phosphorus-32 in nature but rather, most commonly, as iodide ions. Only one isotope of iodine is stable, 127I. Phosphorus is an abundant, naturally occur- Iodine is an essential element and the humanring, reactive non-metal, and is never found in its body contains about 20 mg mainly in the thyroidelemental form in the environment. Compounds gland.containing phosphorus are essential to life and 131 I is a man-made isotope that is a β andare involved in many metabolic processes. Only γ emitter with a short half-life of 8.04 days. Itone phosphorus isotope is not radioactive, 31P, decays to xenon-131, a small percentage to itsand this is the only isotope found in nature. metastable state, which is a γ emitter, but mostly 32 P is a man-made isotope, generally used (~99%) to its ground state, which is stable.for medical purposes. It is produced by neutron As it is a common fission product (235U fissionbombardment of sulfur-32 (32S, this involves a yield is ~3%), 131I is produced by nuclear weapons‘n,p’ reaction, where a neutron is captured and and in nuclear batteries/reactors. Because ita proton is ejected), is a pure β particle emitter is volatile, 131I can more readily escape fromwith a half-life of 14.29 days, and decays back to containment than other fission products, but its32 S, which is stable. Because of its short radioac- relatively short half-life means it does not persisttive half-life 32P must be used relatively quickly in the environment for long periods.after it is produced, and it cannot be stockpiled. 131 I is also produced via neutron bombard- ment of tellurium-130 for medical diagnostic and 109
    • IARC MONOGRAPHS – 100Dtreatment purposes. Because of its short half-life, 222 Rn is an α-particle emitter with a shortit cannot be stockpiled for this purpose. Global half-life of 3.82 days, it decays to polonium-218,demand for 131I for medical purposes is approxi- which is also an α-emitter, and has in turn furthermately 600 tera (T)Bq (600 × 1012 Bq). short-lived radioactive daughter products (see Fig. 1.3). The presence of this decay chain greatly(e) Caesium-137 increases the overall radiological significance of Caesium is a rare naturally occurring, highly this isotope. Although 222Rn is a gas, its short-reactive alkali metal with only one stable isotope lived progeny are electrically charged particles133 Cs. The chemistry of caesium has some simi- that can become attached to environmental dustlarities to that of potassium. particles in the air, the existence and extent of 137 Cs is a man-made isotope that is a β and γ this ‘attached’ fraction has a considerable impactemitter with a half life of 30 years. It decays to on dose to the upper airways of the lung.barium-137, mostly (~95%) to its metastable state, Like its parent radiosotopes (see Fig. 1.3), 222Rnwhich is a short-lived energetic gamma emitter, is omnipresent in nature but levels vary becausebut also to its ground state, which is stable. 137Cs certain types of rocks and soils (e.g. granite,is mostly produced as a result of nuclear fission, phosphate rocks, and alum shales) contain moreeither in nuclear weapons or batteries/reac- of its parents than others (Appleton, 2007). 222Rntors, and is one of the most commonly occur- rapidly disperses into the troposphere when itring fission products (235U fission yield is ~6%). escapes into the free atmosphere, i.e. outside ofIts relatively long half-life results in it being enclosed spaces. Consequently, concentrations ofpersistent in the environment if released. Levels 222 Rn in breathing air in open spaces is relativelyof 137Cs in surface soil due to fallout from atmos- low, typically around 10 Bq/m3.pheric nuclear weapons tests are around 15 Bq/ 222 Rn can also be found in building materialskg on average. albeit at low concentrations (de Jong et al., 2006). Building materials such as concrete, wallboard,(f) Radon brick and tile usually have concentrations similar Radon is a noble (chemically inert) gas to those of major rock types used for their manu-mostly produced through the radioactive decay facture, and levels also vary according to the typeof environmental uranium/thorium and their of rock used for construction (Mustonen, 1984;radioactive daughters. All of the isotopes of Ackers et al., 1985). Although building materialsradon are radioactive: 222Rn is the isotope with generally contribute only a very small percentagethe longest radioactive half-life, and its naturally of the indoor air 222Rn concentrations, in a fewabundant parent is 226Ra, itself a daughter of 238U areas, concrete, blocks, or wallboard incorpo-(see Fig. 1.3), 222Rn is the most prevalent in the rating radioactive shale or waste products fromenvironment. 220Rn (also known as thoron) is the uranium mining can make an important contri-only other isotope of radon that is found in any bution to the indoor 222Rn levels (Man & Yeung,significant quantity in nature. That isotope and 1998; Åkerblom et al., 2005).its radioactive daughters typically contribute less (g) Radiumthan 20% of the total dose from radon, and itscontribution is often not included in radon expo- Radium is a naturally occurring rare earthsure assessments. Henceforth, the term radon metal. Ubiquitous in the environment, in smallshould be taken as referring to Radon-222 unless quantities, it is found in soils, uranium/thoriumotherwise indicated. ores (e.g. pitchblende), minerals, ground water, and seawater, because the common radium110
    • X- and γ-radiationFig. 1.3 Uranium-238 decay chain isotopes are products of the main uranium/ thorium decay chains. All the isotopes of radium Radionuclide Half-life are radioactive, 226Ra has the longest half-life, Uranium-238 4,468,000,000 years and therefore is the predominant isotope found -particle in nature. Thorium-234 24.1 days 226 Ra is an α-particle emitter with a half-life -particle of 1600 years, and decays to 222Rn, which is also Protactinium-234m 1.17 minutes an α-particle emitter. -particle 228 Ra is a β and gamma emitter with a half-life Uranium-234 2,444,500 years of 5.75 years, and decays to actinium-228, which -particle is a β-particle and gamma emitter. Thorium-230 75,400 years 226 Ra concentrations in soil vary consid- -particle erably, typically between 10–50  Bq/kg, with Radium-226 1,600 years approximately 25 Bq/kg considered to be average -particle (UNSCEAR, 1982), concentration in seawater is Radon-222 3.82 days 4–5 orders of magnitude lower than this. -particle 223 Ra and 224Ra are both α-particle emitters Polonium-218 3.11 minutes with a half-life of 11.43 days and 3.6 days, respec- -particle tively. 224Ra can be found in ground water. Lead-214 26.8 minutes -particle (h) Thorium-232 Bismuth-214 19.9 minutes Thorium is a naturally occurring dense metal -particle Polonium-214 0.000163 seconds that is usually found in minerals such as mona- -particle zite, thorite, and thorianite. Thorium is thought Lead-210 22.3 years to be about three times more abundant than -particle uranium in the environment. All of the isotopes Bismuth-210 5.01 days of thorium are radioactive, therefore the isotope -particle with the longest radioactive half-life, 232Th, is by Polonium-210 138 days far the most prevalent in nature. -particle 232 Th is an α-particle emitter with a half-life Lead-206 Stable of 1.41 × 1010 years, and decays to 228Ra, which is a β-particle emitter. 230 Th is present in soil and ores with 232Th. 230Th is a decay product of 234U. 230Th is an α-particle emitter with a half-life of 7.54 × 104 years. (i) Uranium Uranium is a naturally occurring very dense metal, which is widespread in the environ- ment, including seawater, at low concentrations. All of the isotopes of uranium are radioactive, therefore the isotopes with the longest radioac- tive half-lives are the most prevalent in nature. Environmental uranium is made up of three 111
    • IARC MONOGRAPHS – 100Disotopes: 234U, 235U, and 238U. 238U is predominant purposes annually by facilities in the USA,by mass at 99.284%; 235U, accounting for 0.711%; Canada, France, the Russian Federation, and theand, 234U only 0.005% (it should be noted that United Kingdom.natural isotopic composition can vary slightly). A total of over 2000 tonnes of highly enriched 234 U is an α-particle emitter with a half-life of uranium are though to have been produced for2.445 × 105 years, and decays to 230Th, which is military purposes (World Nuclear Association,also an α-particle emitter. 2009). 235 U is an α-particle and gamma emitter witha half-life of 7.03 × 108 years, and decays to 231Th, (j) Plutoniumwhich is a β-particle and gamma emitter. Plutonium is a man-made (predominantly), 238 U is an α-particle emitter with a half-life of very dense, rare earth metal, which has a complex4.468 × 109 years, and decays to 234Th, which is a chemistry. All the isotopes of plutonium areβ-particle and gamma emitter. radioactive, the most commonly occurring Of the three naturally occurring uranium isotopes are the α-particle emitters 239Pu, 240Puisotopes, only 235U has the capacity to support and, increasingly, the β-particle emitter, 241Pu.sustained nuclear fission through a chain reac- Shortly after its discovery, 239Pu was identified astion. Hence, uranium is commonly classified a strategic material for nuclear weapons produc-into types depending on the percentage of 235U it tion, because it has the capacity (greater thancontains, as compared to that in naturally occur- that of 235U) to support sustained nuclear fission.ring uranium ores (0.711% by mass). Natural Most of the plutonium now in existence hasuranium, as its name would suggest, has the same been man-made as a result of nuclear weaponspercentage of 235U as uranium ores. Depleted and power production programmes. However,uranium, which is a common by-product of the small quantities of plutonium have also beennuclear fuel cycle, has a lower percentage of 235U found at the site of the so-called ‘natural reactor’than natural uranium. Enriched uranium typi- at Oklo in Gabon West Africa. 239Pu is producedcally contains about 2.5–3.5% by mass of 235U, and through neutron capture by 238U, within nuclearis widely produced on an industrial scale for use batteries/reactors. This yields 239U which decaysin the manufacture of power reactor fuel assem- to 239Np by β-particle emission, which decaysblies. Highly enriched uranium is almost all further to 239Pu, also by β-particle emission. The235 U, greater than 80% by mass, and is produced longer that nuclear fuel is used (‘burned’) in ain much more limited quantities than normal reactor, the greater the number of plutoniumenriched uranium for use in nuclear propulsion isotopes that appear in increasing quantities, e.g.reactor systems, and for nuclear weapons. neutron capture by 239Pu yields 240Pu, which can, Approximately 50000 tonnes of natural in turn, capture neutrons to produce 241Pu. 238Puuranium are mined annually, about more than is also increasingly produced from 235U throughhalf of this amount is produced by mines in neutron-capture reactions and radioactive decay.Kazakhstan, Canada, and Australia with the 238 Pu is a high-energy α-particle emitter withremainder coming from mines in many coun- a half-life of 88 years, and decays to 234U, whichtries throughout the world. is also an α-particle emitter. World stockpiles of depleted uranium are 239 Pu is an α-particle emitter with a half-lifecurrently more than 1 million tonnes, with over of 24065 years, and decays to 235U, which is also50000 tonnes being added per year. an α-particle emitter. Approximately 60000 tonnes of enricheduranium is produced for nuclear fuel production112
    • X- and γ-radiation 240 Pu is an α-particle emitter with a half-life because good personal neutron dosimetry isof 6500 years, and decays to 236U, which is also an difficult to achieve over all energy ranges (ener-α-particle emitter. gies of importance cover a range > 109 eV), and 241 Pu is primarily a β-particle emitter with a detection thresholds are often high, particularlyhalf-life of 14 years, and decays to 241Am, which in the early days of monitoring.is a radiologically significant α-particle emitter. The only plutonium isotope required for (a) Accidentsnuclear weapons purposes is 239Pu, and the pres- The production and transport of nuclearence of other isotopes of plutonium, such as weapons have resulted in several accidents. The240 Pu, can also be a hindrance to this application. two most serious accidents in nuclear weaponsTherefore, plutonium is classified into different production were at the Mayak complex neargrades depending on its 239Pu and 240Pu content. Kyshtym in the Russian Federation (formerlyThe primary distinction is between weapons- the Soviet Union), and at the Windscale plantgrade material, which is more than 93% 239Pu, at Sellafield in the United Kingdom. A majorand other grades, for example reactor grade, accident in a nuclear power plant occurred inwhich contain lower percentages of 239Pu. Chernobyl, Ukraine. Because of the secrecy surrounding nuclear (i) Southern uralsweapons, precise figures on weapons-grade pluto-nium production are difficult to obtain, however, Mayak, the former Soviet Union’s maintotal worldwide production is thought to have production facility for weapons-grade pluto-been of the order of several hundred tonnes. nium was built near the town of Ozersk in theGlobal stockpiles of weapons-grade plutonium southern urals, the Russian Federation, in thehave diminished as a result of strategic arms limi- 1940s. Operations at this facility resulted intation agreements, and are currently believed to several major, and persistent minor, uncontrolledbe about 250 tonnes. releases of activity into the surrounding environ- Approximately 70 tonnes of reactor-grade ment, particularly the Techa river.plutonium are produced by power-generating In 1957, a Mayak waste storage facility locatednuclear reactors every year, this adds to an near Kyshtym exploded as a result of a chem-existing inventory of about 1300 tonnes globally, ical reaction, this incident is referred to as themuch of this is still contained in spent fuel (World Kyshtym accident. The region contaminated byNuclear Association, 2009). this accident had a population of approximately 238 Pu is used as a heat source in radiothermal 273000 people and around 11000 of these had to begenerators to produce electricity for a variety of relocated, including 1500 people who had previ-purposes (Argonne National Laboratory, 2007). ously been resettled from the Techa River area. In Mayak, the total collective effective dose to an exposed population of 273000 was 2500 man.Sv1.3 Exposure (UNSCEAR, 2000a). This and other discharges from the plant (routine and accidental) resulted1.3.1 X-rays, γ-rays and neutrons in substantial doses to workers (see Vasilenko Detailed information on the different et al., 2007) and to the local population (Degtevamethods of measurement (present and histor- et al., 2006).ical) of all types of external radiation and their As a result of a drought in 1967, Karachayassociated uncertainty can be found in NCRP Lake, which had been used as an open depot(2007). Estimates of neutron dose are uncertain for liquid radioactive waste from Mayak, dried 113
    • IARC MONOGRAPHS – 100Dup and the winds associated with a subsequent (b) External exposurestorm picked up radionuclide-loaded sediments (i) Natural sourcesfrom the lake and distributed them over a widearea. Exposure to external radiation accounts for about 40% of the average worldwide natural(ii) Windscale fire radiation dose, the rest being due to internal In October 1957, at the Windscale Works, now exposure, mainly from 222Rn (Table 1.1).part of the Sellafield site, in the United Kingdom, Most of the natural exposure to X- and γ-raysa nuclear reactor used to produce plutonium for is from terrestrial sources, and depends on theweapons caught fire. Before the fire could be concentration of (natural) radioactive materialsextinguished, damage occurred to the irradiated in the soil and building materials. Cosmic raysfuel contained in the reactor, and radionuclides contribute substantially to the effective dose andwere released in the environment. Because of are practically the only natural source of neutronthe design of this reactor, which incorporated exposure. Cosmic ray dose at sea level is mainlythe filtration of exhausted coolant air, mainly from muons, electrons, and photons with aboutgaseous and volatile radioisotopes escaped. In 8% of the effective dose from neutron interac-the Windscale accident, doses were mainly due tions. The neutron fraction increases to a peakto internal ingestion, and are reported in Section of about 40% at a height of around 4000 m. The1.3. cosmic ray dose increases with altitude and also is greater at higher latitudes (UNSCEAR, 2000a).(iii) Three Mile Island UNSCEAR (2000a) gives detailed data for Failure to maintain coolant fluid in a commer- exposure in various regions of the world. Averagecial light-water reactor at Three Mile Island in outdoor external dose rates for different coun-the USA resulted in the reactor core becoming tries cover the range 18–93 nGy/h. The popula-exposed to the air, and led to a partial meltdown tion-weighted average is 59 nGy/h (0.52 mSv perof the fuel load. year). Areas of very high dose rates above ~10000(iv) Chernobyl nGy/h have been reported from various sites In the accident at Chernobyl in the Ukraine in throughout the world.April 1986, a Russian reactor Bolshoy Moschnosti The population-weighted average effectiveKanalniy (RMBK) became uncontrollable dose of neutrons was estimated to be 100 μSv percreating a steam explosion and a subsequent fire, year by UNSCEAR (2000a).which resulted in a loss of containment and ulti- (ii) Medical usesmately to the complete destruction of the reactor. The medical uses of radiation include diag-In the Chernobyl accident, the main contributor nostic examinations and therapy. Radiotherapyto the dose from external irradiation was 137Cs. is intended to deliver high doses to target organsThe doses to individuals throughout the northern of the order of tens of Gy (UNSCEAR, 2000a).hemisphere varied widely, some staff and rescue Assessing the risk to non-target organs may beworkers on duty during the accident receiving important in some cases.fatal doses >  4 Sv (Savkin et al., 1996). Yearly The dose per medical diagnostic examinationaveraged doses to operation recovery workers of is generally of the order of 0.1–20  mGy. WhileBelarus, the Russian Federation, and Ukraine lower than doses from radiotherapy, diagnosticwere in the range of 20–185 mGy during 1986–89 examinations are the main source of radiation(UNSCEAR, 2008a). from medical use. The use of X- and γ-rays for114
    • X- and γ-radiationmedical purposes is distributed very unevenlythroughout the world (Table  1.2). UNSCEAR Table 1.1 Average radiation dose from natural sources(2000a) reported an increase in the overallfrequency of diagnostic X-ray examinations but Source Worldwide average Typical rangethe frequency was static or had shown decreases annual effective (mSv) dose (mSv)in some countries (Fig.  1.4). The majority ofthe world population receives no exposure in a External exposure Cosmic rays 0.4 0.3–1.0agiven year from X- and γ-irradiation in medical Terrestrial γ-rays 0.5 0.3–0.6bdiagnosis, while the effective dose may be up to Internal exposure100 mSv for a small number of people. Doses due Inhalation (mainly 1.2 0.2–10cto diagnostic X-rays are changing rapidly with radon)time as technologies develop (NCRP, 2009). Ingestion 0.3 0.2–0.8d The average levels of radiation exposure due to Total 2.4 1–10the medical uses of radiation has been increasing a Range from sea level to high-ground elevation b Depending on radionuclide composition of soil and building(Fig.  1.5; UNSCEAR, 2000a), in particular due materialsto increasing use of computed tomography (CT), c Depending on indoor accumulation of radon gas d Depending on radionuclide composition of foods and drinking-angiography, and interventional procedures in waterdeveloped countries (Fig.  1.6). The estimated From UNSCEAR (2000a)global annual effective dose from all diagnosticuses of radiation was estimated to be 1.2  mSv 31% on women 30–39 years old (Klug et al., 2005).per person in 1991–96, compared to 1.0 mSv in In USA, 60% of women had their first mammog-1985–90. In 2006, US citizens received a collec- raphy exams by the age of their 40th year, andtive effective dose from medical procedures 7.3 in France 45.8% during the age of 45–50 yearstimes greater than was the case in the early 1980s (Spyckerelle et al., 2002; Colbert et al., 2004).(NCRP, 2009). Computed tomography scanning has become For the same examination, doses may vary by widely available in many developed countries.an order of magnitude, and reducing the highest The effective dose per examination is consider-doses can reduce collective dose without a reduc- ably higher than that from most conventionaltion in diagnostic information (Watson et al., radiographic procedures, and its use is increasing2005). (Brenner & Hall, 2007). Doses per procedure Conventional radiographs form the majority are in the range of 1.5  mSv to over 25  mSvof radiographic examinations with doses from (UNSCEAR, 2000a).<  0.01 up to ~10  mSv per procedure (Watson Fluoroscopy results in much higher doseset al., 2005). The use of digital imaging tech- than radiography. The doses may vary widely:niques to replace film-screen combinations has modern equipment with image amplifiersbecome widespread in some countries (see e.g. results in lower doses than older equipment withHart et al. (2005) for a detailed review of prac- fluorescent screens, but high doses may still betices in the United Kingdom). received. Advances in technology have facili- Doses to the breast from mammography tated the development of increasingly complexexaminations are of the order of 1.5  mGy with radiological procedures for angiography andlarge variations depending on breast character- interventional radiology, and effective dosesistics (Young & Burch, 2000; Schubauer-Berigan per procedure from under 10 to over 80 mSv,et al., 2002). In Germany, 18% of first mammog- depending on the complexity of the procedure,raphies were on women less than 30 years old and have been reported (UNSCEAR, 2000a). 115
    • IARC MONOGRAPHS – 100D fraction due to neutrons (~1%) and some internalTable 1.2 Radiation exposures from diagnostic exposure. For the survivors, the latest estimatesmedical examinations of the doses using dosimetry system DS02Population per Annual Average annual (Young & Kerr, 2005) are available for 113251physician number of effective dose to persons in the Life Span Study of whom 93741 examinations per population (mSv) 1000 population were within 10 km of the hypocentres. Of these, 44464 had doses < 0.5 mGy and 35393 had doses< 1000 920 1.21000–3000 150 0.14 >  10  mGy (Cullings et al., 2006). The mean of3000–10000 20 0.02 known doses for survivors at about 1600 m was> 10000 < 20 < 0.02 roughly 170 mGy (Preston et al., 2004).Worldwide 330 0.4 Atmospheric nuclear explosions were carriedaverage out, mostly in the northern hemisphere, betweenFrom UNSCEAR (2000a) 1945 and 1980. The most intense period of testing was between 1952 and 1962. In all, approximately The medical use of neutrons and protons in 543 atmospheric tests have been carried out, withradiotherapy is limited at present. a total yield of 440 Mt (megatonne) explosive(iii) General population power (UNSCEAR, 2000a). Since 1963, nuclear Estimates of the average doses received by tests have been conducted mainly underground,the general population are reviewed regularly and the principal source of worldwide exposureby the United Nations Scientific Committee on due to weapons testing is the earlier atmos-the Effects of Atomic Radiation (UNSCEAR, pheric tests. The global average committed effec-2000a), and by many national bodies, such as the tive dose (which includes the sum of all dosesBundesministerium für Umwelt, Naturschutz that will be received over a period of 50 yearsund Reaktorsicherheit in Germany (BMU, 2007), from internal irradiation) is 3.5  mSv, of which,the National Council on Radiation Protection 0.5 mSv is from external irradiation (UNSCEAR,and Measurements in the USA (NCRP, 2009), 2000a). Annual average total radiation doses areand the Radiation Protection Division of the currently ~8 μSv per year, of which, < 3 μSv perHealth Protection Agency in the United Kingdom year is from external irradiation.(Watson et al., 2005). People living near the sites where nuclear Fig. 1.7 shows in a) the distribution of average weapons were tested received doses varyingexposures to ionizing radiation in the United considerably in magnitude. Those near to theKingdom (Watson et al., 2005) and in b) and c) Nevada test site in the USA received an estimatedhow the distribution in the USA has changed average dose of about 3  mSv (Anspaugh et al.,between the early 1980s and 2006 (NCRP, 2009). 1990). After a US test in 1954 at Bikini atoll in theThe distribution of some of the components Marshall Islands, the residents of Rongelap andin different countries may vary by an order of Utirik atolls (230 persons) received high externalmagnitude. exposures (1900  mSv), mainly from short-lived radionuclides, with 67 persons receiving doses(iv) Nuclear explosions and production of of 1750 mSv on Rongelap (Conard et al., 1980). nuclear weapons At Semipalatinsk in the former Soviet Union, The atomic bombings of Hiroshima and atmospheric tests between 1949 and 1963,Nagasaki, Japan, in 1945 exposed hundreds exposed 10000 people in settlements borderingof thousands of people to substantial doses of the test site with doses ranging up to several Gyexternal radiation from γ-rays with a small (Tsyb et al., 1990).116
    • X- and γ-radiationFig. 1.4 Temporal trends in global practice with medical X-ray examinations: average frequenciesand doses for 1991–96 relative to previous estimates for 1985–90 1.5 1.4 RELATIVE INCREASE 1.3 1.2 1.1 1 per caput Annual total Effective dose Annual effective dose Annual collective dose Annual frequency per 1,000 population per examination Global population of examinationsAdapted from UNSCEAR (2000a) 117
    • IARC MONOGRAPHS – 100DFig. 1.5 Comparison of distribution of collective dose values (S) or effective dose (Eus) for thecategories of exposure as reported for the early 1980s and for 2006 2006 Early 1980s Background (50%) Background (83%) Occupational / Occupational / industrial (0.1%) industrial (0.3%) Consumer (2%) Consumer (2%) Medical (15%) Medical (48%) E arly 1980s 2006 S: Collective effective dose 835,000 1,870,000 (person-Sv) E : Effective dose per US 3.6 6.2 individual in the U S population (mSv)Adapted from NCRP (1987, 2009) γ-ray exposures to the local population entire fuel cycle (mining and milling, enrich-resulting from the production of weapons mate- ment and fuel fabrication, reactor operation, fuelrial and chemical separation can be consider- processing, waste disposal, transport of radioac-able. For example, the release of nuclear wastes tive materials) has been estimated by UNSCEARfrom the Mayak complex into the Techa River (2000a). If the present annual generation of 250from a military plant of the former Soviet gigawatt continues for 100 years, the internalUnion, resulted in organ doses up to 5.2  Gy at plus external dose to an individual of the generalbone surfaces (median 0.37  Gy), mainly from population would be less than 0.2 μSv per year.internal radionuclides, with half of the much Evrard et al. reported an estimated dose oflower external doses lying between 0.0017 and 0.17  μSv per year due to gaseous discharge in0.0062 Gy (Degteva et al., 2006). 2107 “communes” located in the vicinity of 23 French nuclear facilities, including all power(v) Nuclear power production plants (Evrard et al., 2006). Most of the exposure Assuming that the generation of electrical is due to internal irradiation.energy by nuclear power reactors lasts for 100years, the maximum collective dose for the118
    • X- and γ-radiationFig. 1.6 Comparison of collective dose values for CT, conventional radiography and fluoroscopy,interventional fluoroscopy, and nuclear medicine (as % of total collective dose) as reported for theearly 1980s (123700 person-Sv) (left: NCRP, 1989), and for 2006 (899000 person-Sv) (right: NCRP,2009). For EUS, the same percentages apply. Collective dose quantities are S for 2006 and collectiveeffective dose equivalent H E for NCRP (1989). Medical exposure of patients (early 1980s) Collective HE Medical exposure of patients(2006)SConventional radiography & fluoroscopy (68%) Computed tomography (49%) Computed tomography (3%) Conventional radiography & fluoroscopy (11%) Nuclear medicine (26%) Nuclear medicine (26%) Interventional fluoroscopy Interventional (14%) fluoroscopy (3%) E arly 1980s 2006 S: C ollective effective dose 123,700 899,000 (person-Sv) E U S : E ffective dose per 0.53 3.00 individual in the U .S. population (mSv)Adapted from NCRP (2009)(vi) Accidents doses are not large. The steady increase in the For the Mayak, Windscale, and Chernobyl use of sources of ionizing radiation has led toaccidents, see Section 1.3.1 above. an increase in the number of fatalities, despite Sealed sources used for industrial and medical progress in radiation protection.purposes have occasionally been lost, stolen or (vii) Occupational exposuresdamaged, resulting in exposure of members of Occupational exposure to radiation occursthe public to these materials. Examples include during nuclear power production and fuelthe sale of a Cobalt-60 (60Co) source as scrap recycling, military activities, industrial opera-metal in the city of Juarez, Mexico, in 1983 tions, flying and medical procedures (see above(Marshall, 1984); the theft and breaking up of a for details). Average annual effective doses are137 Cs source in Goiânia, Brazil, in 1987 (IAEA, in Table  1.3 (UNSCEAR, 2000a). The average1988); and the retrieval of a lost 60Co source in annual effective dose for occupational workersShanxi Province, the People’s Republic of China, has reduced from 1.9 mSv in 1975–79 to 0.6 mSvin 1992 (UNSCEAR, 1993). IAEA publications in 1990–94.contain information on accidental irradiation Mean doses to medical radiation technolo-during medical procedures, in particular Safety gists in the US have reduced from 100 mSv perReport Series No. 17 (IAEA, 2000). While these year before 1940 to 2.3  mSv per year duringincidents result in significant individual doses to 1977–84 (Simon et al., 2006). In recent years,a small number of people, the collective effective worldwide annual doses have also been reduced 119
    • Fig. 1.7 Percentage contribution of each natural and man-made radiation source120 Natural (a) (84%) Internal (9.5%) Radon (50%) Gamma (13%) IARC MONOGRAPHS – 100D Cosmic Artificial (12%) (16%) Medical (15%) Products (<0.1%) Occupational (0.2%) Fallout Disposals (0.2%) (<0.1%) Other background (28%) (b) (c) Radon & thoron (55%) Radon & thoron (37%) Other background (13%) Consumer / Consumer / occupational / industrial occupational / industrial (2%) (2%) Conventional Nuclear medicine radiography / fluoroscopy (4%) (5%) Computed Interventional fluoroscopy tomography (7%) Diagnostic (24%) (11%) Nuclear medicine (12%) (a) UK population, average annual dose to the population = 2.7 mSv (b) Distribution of S or EUS for the major sources of exposure for the early 1980s (NCRP, 1987). The percent values have been rounded to the nearest 1 %. The total for S is 835,000 person- Sv and the total for EUS is 3.6 mSv, using a U.S. population of 230 million for 1980. (c) Distribution of S or EUS for the major sources of exposure for 2006. The percent values have been rounded to the nearest 1 %. The total for S is 1,870,000 person-Sv and the total for EUS is 6.2 mSv, using a U.S. population of 300 million for 2006. The other background category consists of the external (space and terrestrial) and internal subcategories. Adapted from Watson et al. (2005), NCRP (1987), NCRP (2009)
    • X- and γ-radiationfrom 0.6 mSv in 1980–84 to 0.33 mSv in 1990–94(UNSCEAR, 1993, 2000a). Table 1.3 Occupational radiation exposures Occupational exposure to neutrons consti- Source/practice Number of Average annualtutes a small fraction of the total effective dose monitored effective doseand occurs mainly in the nuclear industry. In a workers (mSv) (thousands)United Kingdom compilation of dose to nuclear Man-made sourcesworkers (Carpenter et al., 1994), the upper limit Nuclear fuel cycle 800 1.8of the neutron component was estimated to be (incl. uranium3% of the total exposure. In the USA, more than mining)10000 nuclear workers per year receive measur- Industrial uses of 700 0.5 radiationable neutron doses (NCRP, 1987). Defence activities 420 0.2 Neutron sources are used to chart progress in Medical uses of 2320 0.3the search for gas and oil resources. For oil-well radiationloggers, doses of 1–2 mSv per year were reported Education/veterinary 360 0.1in one study (Fujimoto et al., 1985), and in another Enhanced natural sources(Inskip et al., 1991), only seven of 1344 workers Air travel (crew) 250 3.0received above-threshold (0.02 mGy) doses. Mining (other than 760 2.7 The exposure of commercial aircraft crews coal)to neutrons depends on the flight route and Coal mining 3910 0.7on the number of flight hours with secondary Mineral processing 300 1.0neutrons from galactic cosmic rays contributing Above-ground 1250 4.8 workplaces (radon)about 10–15% of the dose at an altitude of 10 km. Adapted from UNSCEAR (2000a)Watson et al. (2005) reviewed United Kingdomdata by summarizing findings of Warner Jones of moderation undergone by the neutrons. Inet al. (2003) and Irvine & Flower (2005) and esti- most occupational settings, the neutron spec-mated overall average annual doses for all aircrew trum will be a degraded fission spectrum. Foras 2 mSv from natural radiation and 19 μSv from example, for workers occupationally exposedthe transport of radioactive material. to low neutron doses from nuclear reactors or Staff involved in radiotherapy with neutrons similar settings, the important neutron energyare exposed mainly to γ- and β-rays due to acti- range in terms of dose deposition is, on average,vation of the room and equipment. The dose from about 10–100  keV (Worgul et al., 1996).rates are well below 1 μGy/h and are not detect- Doses from radiotherapy-related photoneutronsable by personal dosimetry (Smathers et al., 1978; are dominated by somewhat higher neutronFinch & Bonnett, 1992; Howard & Yanch, 1995). energies (100–1000 keV) (Ongaro et al., 2000),Individuals are exposed to neutrons largely while the dose from secondary neutrons fromthrough the use of high-energy photon beams galactic cosmic rays (De Angelis et al., 2003), or(> 15 MeV), which produce photo-neutrons (Hall from proton radiotherapy (Zheng et al., 2007),et al., 1995; Ongaro et al., 2000), and also through will generally be dominated by higher-energythe use of high-energy proton-therapy beams, neutrons.which produce secondary neutrons (Brenner & Astronauts are exposed to high doses of spaceHall, 2008). radiation, which consists of protons, heavy ions The neutron energy spectrum to which and secondary neutrons produced by galacticindividuals may be exposed varies widely, cosmic ray interactions, particularly if they godepending on the neutron source and the degree 121
    • IARC MONOGRAPHS – 100Dbeyond low earth orbits. Based on data from For example, about 100000 self-illuminatinga human phantom torso, the organ dose rates exit signs were produced per year in the USAoutside the International Space Station have been during 1983–2002 (PSI, 2003), containing a totalderived by Reitz et al. (2009) and are in the range of approximately 100  PBq (petabecquerel, 1015of ~0.2–1.0 mGy/day. Data for average personnel becquerel) of 3H.badge doses for previous space missions givesimilar figures (Cucinotta et al., 2008). The esti- (b) Phosphorus-32mated dose (Cucinotta & Durante, 2006) for a (i) Medical uselunar mission of 180 days is 60 mGy, and for a 32 P, in the form 32PO4, has been used in theMars exploration of 1000 days it is 420 mGy. The treatment of polycythaemia vera since 1939.relative biological effectiveness for these heavy This has been the primary medical use for thisions may be as high as 40. radionuclide, representing ~5% of all therapeutic use of radionuclides in a survey of 17 European1.3.2 α- and β-emitting radionuclides, countries (Hoefnagel et al., 1999) but only ~1% internally deposited worldwide (UNSCEAR, 2000a). Individual treat-(a) Tritium ments typically involve the use of 150–170 MBq of 32PO4 (UNSCEAR, 2000a) administered orally(i) Nuclear weapons producion or intravenously. Because the amount of 3H needed for nuclear 32 P has also been used as a radioactive tracer,weapons purposes is relatively small, the facilities for purposes such as identifying tumours as anused to produce it tend to be much smaller than aid in surgical removal. Historically, 32P was alsothose used to produce plutonium, consequently used in the treatment of leukaemia (both chronicthe number of workers exposed to 3H also tends myelocytic leukaemia and chronic lymphocyticto be small. The secrecy often associated with leukaemia).military 3H production means that there arealso relatively few 3H worker cohorts identified (c) Strontium-90from this activity. Relaxation of secrecy associ- As exposure to 90Sr is mostly in conjunctionated with military 3H production in the United with other fission products, further informationKingdom in the last 10 years has meant that on exposures is given in the mixed fission prod-several hundred workers are now known to have ucts section below.potentially been exposed to 3H at the Capenhurstand Chapelcross sites (HPA, 2007). (d) Iodine-131(ii) Nuclear power production As exposure to 131I can often be in conjunc- tion with other fission products, further infor- With heavy-water moderated reactors, such mation on exposures is also given in the mixedas the CANDU design, 3H exposures normally fission products section below.account for the majority of the workers’ dose. (i) Medical use(iii) Occupational exposure Radioiodine has been used in the treatment 3 H has also been used to produce self-illu- of hyperthyrodism and cancer of the thyroid forminating devices (the β-particle emissions are more than 50 years, and is by far the most commonused to stimulate light production in a suitable internal emitter used for therapeutic purposes. Itphosphorescent material) used in various appli- should also be noted that radioiodine treatmentcations including watches, gun sights, and signs. can be a source of external exposure to other122
    • X- and γ-radiationpeople, and it is the main source of exposure to (f) Radonthe public and relatives from patients who have (i) Natural sourcesreceived unsealed radionuclides (ICRP, 2004). Internal exposures from Naturally Occurring(ii) Accidents Radioactive Materials (NORM) are generally dominated by the isotopes in the 232Th and 238UWindscale fire decay chains, particularly 222Rn and its progeny. The Windscale fire in 1957, in the United 222 Rn makes by far the largest contribution toKingdom, resulted in the release of a total of average individual internal exposures to the1.5  ×  1015 Bq of radioisotopes. Because of the public from natural sources (see Table  1.1).design of this reactor, which incorporated 222 Rn concentration in buildings varies greatly,the filtration of exhausted coolant air, mainly typically from less than 10 Bq/m3 to more thangaseous and volatile radioisotopes escaped (133Xe 100 Bq/m3 (UNSCEAR, 2006), depending on(14 × 1015 Bq), 210Po (0.09 × 1015 Bq)) including factors such as local geology and air movement1.4 × 1015 Bq of 131I. Prompt action to limit expo- (restricted ventilation in places such as caves cansure to 131I resulted in low doses being released lead to much greater 222Rn concentrations).to the general public; however, workers at the Residential 222Rn concentrations can varyplant involved in efforts to extinguish the appreciably in different parts of the home, withfire did receive larger than normal exposures basement 222Rn concentrations typically 50%(UNSCEAR, 1993; IARC, 2001). higher than on the ground floor (Field et al., 2000,Three Mile Island 2006). 222Rn concentrations within homes in the Initially, the activity released during the same neighbourhood can also vary appreciablyThree Mile Island reactor accident in the USA due to subtle aspects of building construction,was largely contained within the primary such as cracks and fissures in the foundation,containment building but gaseous and volatile and ventilation of the home (Radford, 1985).radionuclides including 133Xe (370 × 1015 Bq) and Residential 222Rn concentrations also exhibit131 I (550 × 109 Bq) were subsequently released into seasonal variation, both within and betweenthe environment (UNSCEAR 1993; IARC, 2001). years (Pinel et al., 1995; Krewski et al., 2005). One other source of 222Rn can be from domesticChernobyl water supplies. Following the Chernobyl accident, reported (ii) Occupational exposureindividual thyroid doses ranged up to severaltens of Gy, while average doses range from a few Because 222Rn is formed from the radioactivetens of mGy to several Gy (UNSCEAR, 2000b; decay of 238U which is ubiquitous in the Earth’sCardis et al., 2006a, b). crust, high levels of 222Rn gas have historically been found in underground mines (Committee(e) Caesium-137 on Health Risks of Exposure to Radon (BEIR As exposure to 137Cs is mostly in conjunction VI, 1999)). Since the discovery of lung disease inwith other fission products, further information underground miners exposed to high levels ofon exposures is given in the mixed fission prod- 222 Rn in the 19th century, subsequently confirmeducts section below. to be lung cancer in the 20th century, 222Rn concentrations in mines were greatly reduced in the interest of industrial hygiene. Currently, 222 Rn concentrations in underground mines are 123
    • IARC MONOGRAPHS – 100Dgenerally well below the current occupational as many as 2.5 million individuals may have beenexposure guideline of 2 working-level month/ exposed to it, before it was replaced by otheryear (WLM/yr) in ventilated mines (1 WLM is contrast media in the 1950s (IARC, 2001).exposure for 1 month (170 h) at 1 WL (working-level) corresponding to 130000 MeV of potential (i) Uraniumα energy released by the short-lived progeny in (i) Natural sourceequilibrium with 100  pCi of 222Rn in one litre Uranium is naturally present in smallof air (3.7 kBq/m3)). Assuming a breathing rate amounts almost everywhere in soil, rockof 1.2 m3/h, the cumulative intake of 1  WLM including well water, and groundwater. Higheris 0.755 MBq. Although historical exposures in levels are present in natural uranium ores.underground mines have exceeded residentialexposures by a factor up to a 1000-fold or more, (ii) Occupational exposurethis difference has been much reduced by a factor As it is the raw material for most nuclear powerof 20–30-fold in recent years. generation, uranium is ubiquitous in the nuclear fuel cycle: from mining and initial processing to(g) Radium enrichment and/or fuel manufacturing, power(i) Occupational exposure production, and reprocessing. The practice of painting clock dials with Exposure can involve natural, depleted and/radium-based paint to make them luminous was or enriched uranium, in a wide variety of chem-introduced just before the First World War. The ical forms (IARC, 2001).production and application of luminous paint (j) Plutoniumsoon became an industry, particularly in the USA.Because of the precision required in applying (i) Nuclear weapons production and testingthese radium-based paints, ‘Dial painters’ or The USA was the first nation to pursue‘Luminisers’ (as they were commonly known) plutonium production as a means to construct afrequently ‘tipped’ their brushes (i.e. brought the nuclear weapon, but the populations of exposedbristles to a point) using their mouths, and as individuals tend to be compartmentalised and/a result would ingest some of the paint and the or widely dispersed. The two largest continuousradium it contained. The use of radium-based populations of workers exposed to plutonium arepaints has also occurred in Germany, the United those at the Mayak Production Association inKingdom, and many other countries throughout the southern urals, the Russian Federation, andthe world (IARC, 2001). those at the Sellafield (formerly Windscale) plant in the United Kingdom. Both of these facilities(h) Thorium-232 have plutonium worker cohorts of over 10000(i) Medical use individuals, with exposures starting in the late Thorium dioxide (ThO2) was first used as 1940s (Mayak) and early 1950s (Sellafield).an X-ray contrast medium for splenography in Political pressure to develop nuclear weaponsthe 1920s, and from 1931, a commercial prepa- as rapidly as possible both during and in theration containing it, under the trade name decade after the Second World War resulted‘Thorotrast’, was marketed as a general vascular in considerable internal exposure, primarily tocontrast medium. Thorotrast was administered plutonium. Unfortunately this tends to be theby instillation or injection and was widely used period in which monitoring data is most lacking,throughout the world. It has been estimated that particularly for Mayak, where exposures were124
    • X- and γ-radiationthe largest, with many individuals having no 273000 people and around 11000 of these hadmonitoring information at all. to be relocated, including 1500 that had previ- ously been resettled from the Techa River area(ii) Occupational exposure (UNSCEAR, 2000a). The reprocessing of irradiated nuclear fuel,and to a lesser extent the production of mixed (iv) Karachay lakeoxide ‘MOX’ fuel assemblies, can result in expo- The Karachay lake accident released 0.022 PBqsure to plutonium. of radionuclides into the environment and distributed them over a wide area (UNSCEAR,(k) Mixed fission products 2000a). Information on some major individual fission (v) Chernobylproducts (90Sr, 131I, 137Cs) is given above. However,because of the stochastic nature of fission- The Chernobyl accident released substantialproduct production, fission products are always amounts of radionuclides into the environmentproduced in mixtures; and consequently, expo- including 131I (1760  PBq) and 137Cs (85  PBq), and these radionuclides were dispersed over ansures are often to mixtures of fission products. enormous area. The two main groups exposedAssessment of doses from mixed fission products were individuals working on recovery opera-that have been released into the environment are tions (so called liquidators) at the reactor site andfrequently dependent on environmental trans- members of the general population living in theport models. vicinity of the site. A total of 116000 members(i) Southern urals of the public were evacuated from a 30-km area As stated previously, Mayak, the former around the Chernobyl site following the acci-Soviet Union’s main production facility for dent, and 226000 recovery operators worked atweapons-grade plutonium was built near the the site or within this evacuated zone during thetown of Ozersk in the southern urals, the Russian following year.Federation, in the 1940s. Operations at thisfacility resulted in several major, and persistentminor, releases of activity into the surrounding 2. Cancer in Humansenvironment, particularly the Techa river andthe surrounding area (IARC, 2001). X-radiation and γ-radiation were previously(ii) Techa river classified as Group 1 carcinogens by a previous IARC Monograph (IARC, 2000). This classifi- During 1949–56, 100 PBq (100 × 1015 Bq) of cation was based on increased risk of severalactivity were released into the Techa–Isset–Tobol cancers associated with X- and γ-rays, includingriver system. Of the approximately 28000 people leukaemia (excluding chronic lymphocyticliving in settlements near the Techa river during leukaemia), breast cancer in women exposedthis period, around 7500 were relocated during before the menopause, cancer of the thyroid1953–60 because of their exposure to radionu- gland among people exposed during childhood,clides (UNSCEAR, 2000a). non-melanoma skin cancer, and cancer of the(iii) Kyshtym accident stomach, colon, and lung. The Kyshtym accident released 74  PBq of Epidemiological information on the carci-radionuclides. The region contaminated by this nogenic effects of X- and γ-rays comes fromaccident had a population of approximately studies of people exposed to radiation from the 125
    • IARC MONOGRAPHS – 100Ddetonations of atomic weapons, from medical known as the Life Span Study (LSS), became theprocedures, and in occupational or environ- foundation for much of the ongoing researchmental settings. The epidemiological findings on mortality and cancer incidence among thethat have been reported since the previous IARC Japanese A-bomb survivors (Shimizu et al.,Monograph (IARC, 2000) have been reviewed, 1990; Preston, et al., 1994). The experiences ofwith an emphasis on large, well designed studies the Japanese A-bomb survivors have shown thatwith adequate assessment of radiation doses. the effect of exposure to detonation of atomicMajor reviews of the literature and risk esti- weapons persists for decades, and has an impactmates provided by UNSCEAR (UNSCEAR, on the development of a wide range of malignant2008b) and the US National Academy of Sciences diseases.Council Committee to Assess Health Risks from The LSS provides an extremely importantExposure to Low Levels of Ionizing Radiation source of information about radiation health(National Research Council, 2006) on radiation effects. The study cohort encompasses a largerisks by cancer sites were also reviewed. The recent number of people, including men and women,evidence is summarized by sources of exposure exposed to a wide range of doses at all ages. Anfirst, and then both earlier and more recent important development since of the previousevidence is reviewed by cancer site. Cohort and IARC Monograph has been the introduction ofcase-control studies of cancer following X-ray revised radiation dose estimates for the A-bombexposure are summarized in Table 2.1 available survivors: the Reassessment of the Atomic Bombat http://monographs.iarc.fr/ENG/Monographs/ Radiation Dosimetry for Hiroshima and Nagaskivol100D/100D-02-Table2.1.pdf and Table 2.2 Dosimetry System 2002 (DS02) (Young & Kerr,available at http://monographs.iarc.fr/ENG/ 2005). Individual dose estimates for survivorsMonographs/vol100D/100D-02-Table2.2.pdf, within 2 km of the bombings are based on esti-and following γ-ray exposure in Table 2.3 available mates of penetrating radiation emitted by theat http://monographs.iarc.fr/ENG/Monographs/ bombs and the location and shielding of survi-vol100D/100D-02-Table2.3.pdf and Table 2.4 vors derived from interviews conducted in theavailable at http://monographs.iarc.fr/ENG/ late 1950s and early 1960s. Dose estimates forMonographs/vol100D/100D-02-Table2.4.pdf. other survivors are based on less detailed infor- mation on shielding provided during interviews. The LSS study does not provide information on2.1 Detonation of atomic bombs the impact of radiation on cancer risk during the The study of Japanese atomic bomb (A-bomb) years immediately after the bombings. Follow-upsurvivors holds an important place in the litera- for mortality started in 1950, and follow-up forture on radiation epidemiology. Atomic bombs cancer incidence in 1958. Furthermore, inclu-were dropped on Hiroshima and Nagasaki in sion in the LSS cohort required people to haveAugust 1945. Survivors’ external radiation doses survived for at least 5 years after the bombings.were primarily from exposure to γ-radiation, Questions have been raised about potential biasesalthough there was also a neutron contribution. associated with the impacts of early mortality onSeveral years after the bombings, the Atomic subsequent radiation risks, and about potentialBomb Casualty Commission initiated a large differences between survivors as a function ofpopulation-based study of mortality and disease age at the time of the bombings and distancerisk in relation to survivors’ distance from the from the hypocentres (Cologne & Preston, 2000;hypocentres of the atomic bombings (Francis Pierce et al., 2007). Due to potential “healthyet al., 1955; Ishida & Beebe, 1959). That study, survivor effect,” selection bias might be expected126
    • X- and γ-radiationto attenuate risk estimates or obscure evidence of was clear evidence of excess risk of leukaemiaassociations rather than to induce spurious posi- among the A-bomb survivors, which increasedtive associations in the LSS; values for the magni- with increasing magnitude of estimated dose,tude of dose-related selective survival assumed as illustrated by the ratio of the fitted excessin a recent study suggested a modest potential to the expected background number of casesfor bias in dose–response estimates (Pierce et al., by category of dose (Table 2.5). The largest2007). The DS02 system focuses on the prompt γ excess risks were observed for those exposed atand neutron doses from the bomb detonations, younger ages, the excess tended to diminish inbut survivors could have also received doses magnitude with time since exposure, and thefrom fallout and neutron activation of soil and exposure–response relationship appeared toother materials (Imanaka et al., 2008; Tanaka be linear-quadratic. UNSCEAR (UNSCEAR,et al., 2008), which are not accounted for in 2008b) and the US National Academy of Sciencescurrent epidemiological analyses of the LSS data. Council Committee to Assess Health Risks fromAssumptions about the relative biological effec- Exposure to Low Levels of Ionizing Radiationtiveness of the neutron component of survivors’ (National Research Council, 2006) have alsodoses may have a substantial impact on quantita- reported analyses of leukaemia mortality in thetive estimates of γ-radiation dose effects (Walsh LSS using the DS02 dose estimates and mortalityet al., 2004). data through 2000; both have shown the asso- Since the previous IARC Monograph, reports ciation between leukaemia mortality and theon the associations between the DS02-estimated exposure.dose and mortality due to leukaemia and solid Analyses of mortality by type of leukaemiacancers (Preston et al., 2004) and solid cancer among the Japanese A-bomb survivors duringincidence (Preston et al., 2007) have been 1950–2000 have found that the ERR/Gy forpublished. The extension of follow-up of these acute myeloid leukaemia was best described bycohorts, and the resultant increase in the number a quadratic dose–response function that peakedof cancer cases ascertained, has increased the approximately 10 years after exposure. Mortalityability to conduct site-specific analyses of cancer associated with acute lymphocytic leukaemia orrisks as well as permitted analyses that can char- chronic myeloid leukaemia was best describedacterize the risk of cancer in this population by a linear dose–response function that did notmore than five decades after the bombings. Some vary with time since exposure, while adult T-cellrecent analyses suggest a U-shaped pattern of leukaemia was not associated with estimatedassociation of the excess relative risk per Sievert bone-marrow dose (Richardson et al., 2009).(ERR/Sv) with age at exposure for solid cancers No updates of analyses of leukaemia inci-(Preston et al., 2007; Little, 2009). Results of these dence in the LSS have been reported since theanalyses are discussed below along with results previous IARC Monograph.from a recent analysis examining cancer risksfollowing in-utero exposure to radiation from 2.1.2 Solid cancersthe atomic bombings. Preston et al. (2004) reported an analysis of all solid cancer mortality using DS02 dose esti-2.1.1 Leukaemia mates and mortality follow-up information for Preston et al. (2004) analysed the association the period of 1950–2000. The ratio of the fittedbetween leukaemia mortality during 1950–2000 excess to the expected background number ofand DS02 estimates of bone-marrow dose. There cases increased with dose (Table 2.6). The excess 127
    • IARC MONOGRAPHS – 100DTable 2.5 Association between leukaemia mortality during the period 1950–2000 and DS02estimates of bone-marrow dose among A-bomb survivors in JapanWeighted marrow Subjects Person–years Leukaemia death Expected Fitted excessdose category (Sv) background    < 0.005     37407     1376521     92     84.9     0.1    0.005–0.1     30387     1125891     69     72.1     4.0    0.1–0.2     5841     208445     14     14.5     4.7    0.2–0.5     6304     231149     27     15.6     10.4    0.5–1     3963     144276     30     9.5     18.9    1–2     1972     71485     39     4.9     27.7    2+     737     26589     25     1.6     28.2    Total     86955     3184256     296     203.0     93.0risk of solid cancer appeared to be linear in dose, as a group. Elevated risks were seen for the fivewith modifying effects of gender, age at expo- broadly classified histological groups considered,sure, and attained age. including squamous cell carcinoma, adenocarci- Unlike recent analyses of mortality in the noma, other epithelial cancers, sarcomas, andLSS, which included 86611 people, recent anal- other non-epithelial cancers. While the ERR/Gyyses of cancer incidence in the LSS also include was modelled with a linear term, the fit suggestedthe Hiroshima or Nagasaki residents who were departures at older ages, driven in part by thetemporarily not in either Hiroshima or Nagasaki lung cancer risk.or were more than 10 km from the hypocentre in Although the previous IARC Monographeither city at the time of the bombings. Preston noted that there was no association between radi-et al. (2007) reported analyses of incidence ation dose and thyroid cancer incidence amongdata during 1958–98 from 105427 people who those over the age of 14 years when exposed,had DS02 dose estimates and who were alive, more recent analyses have shown positive asso-and had not been diagnosed with cancer as of ciations between radiation dose and thyroid1958. The data for solid cancer incidence were cancer incidence among adult female A-bombconsistent with a linear dose–response over a survivors (ERR/Gy  =  0.70; 95%CI: 0.20–1.46)range of 0–2 Gy. Approximately 850 (about 11%) (Richardson, 2009a). The ERR/Gy among menof the cases among cohort members with doses was −0.25 (90%CI: < 0–0.35). In that study, theto the colon in excess of 0.005 Gy were estimated number of thyroid cancer cases among womento be associated with A-bomb radiation expo- (n = 241) was nearly 5-fold the number of casessure. Significant radiation-associated increases among men (n = 55).in incidence were reported for cancer of the oral Results for site-specific solid cancers in thecavity, oesophagus, stomach, colon, liver, lung, LLS are discussed later in this section.non-melanoma skin, breast, ovary, bladder,nervous system, and thyroid. Although there wasno indication of a statistically significant dose–response for cancer of the pancreas, prostate, andkidney, the excess relative risks for these siteswere also consistent with that for all solid cancers128
    • X- and γ-radiationTable 2.6 Association between mortality from solid cancers during the period 1950–2000 andDS02 estimates of bone-marrow dose among A-bomb survivors in JapanWeighted marrow dose Subjects Person–years Solid cancer Expected Fitted excesscategory (Sv) death background    < 0.005     38507     1415830     4270     4282     2    0.005–0.1     29960     1105215     3387     3313     44    0.1–0.2     5949     218670     732     691     41    0.2–0.5     6380     232407     815     736     99    0.5–1     3426     125243     483     378     116    1–2     1764     64689     326     191     113    2+     625     22302     114     56     64    Total     86611     3184356     10127     9647     4792.1.3 Cancers after irradiation in utero, and mortality (Izumi et al., 2003b) with regard to pre-conception exposure pre-conception exposure in the F1 cohort of the Japanese A-bomb survivors. The study partici- Preston et al. (2008) reported on cancer pants were conceived between 1  month and 38incidence during the period 1958–2000 among years after the atomic bombings, and one or bothA-bomb survivors exposed to radiation in utero. parents were in either the cities of Hiroshima orWhile prior work had focused on the excess Nagasaki at the time of the bombing and forrisk of cancer in the first years of life following childbirth. During the 40-year period of follow-in-utero irradiation, Preston et al. found evidence up, 575 solid cancer cases and 68 haemopoieticof an association between in-utero irradia- neoplasms were recorded, and no associationstion and excess solid cancer risk in the period were found with either paternal or maternal pre-starting approximately 13 years after the atomic conception dose (P  >  0.1) (Izumi et al., 2003b).bombings in Japan. The optimal model indicated During the 1946–99 period of follow-up, 314 solidrelationships between radiation dose in both cancer deaths were recorded, and no associationsin-utero and childhood exposures and risk of were found with either paternal or maternal pre-solid cancers, with modifications by a (negative) conception dose (P > 0.1) (Izumi et al., 2003a).power of attained age. The ERR/Sv at age 50 yearswas 1.0 (95%CI: 0.2–2.3) for the in-utero cohort,slightly lower but not significantly different from 2.2 Fallout from nuclear weaponsthe ERR in the early childhood-exposed cohort testingat this age (ERR/Sv, 1.7; 95%CI: 1.1–2.5). Excessabsolute rates (EAR) at age 50 years increased 2.2.1 Semipalatinskmarkedly with attained age among those exposed Several hundred nuclear weapons tests,in early childhood (EAR/104 person–year Sv, 56; including above-ground tests, occurred at95%CI: 36–79) but exhibited little change in the Semipalatinsk, Kazakhstan, then part of thein utero group (EAR/104 person–year Sv, 6.8; former Soviet Union. Nearby residents were95%CI: < 0–49) (Preston et al., 2008). exposed to external doses of γ-radiation and There have been updated analyses of cancer internal doses due to the inhalation and inges-incidence (Izumi et al., 2003a) and cancer tion of radioactive fallout from these nuclear 129
    • IARC MONOGRAPHS – 100Dweapons tests (including 131I, 137Cs, and 90Sr). several challenges: (i) the doses may be so highEstimating dose for these residents has shown that cell-killing (the objective of the treatment)to be difficult, and there are conflicting esti- overwhelms cancer initiation, (ii) radiotherapymates of the magnitude of the doses received is often coupled with chemotherapy and theirby individuals living in villages in the vicinity separate impacts may be difficult to distinguish,of Semipalatinsk (Simon et al., 2003). The study and (iii) patients with existing cancers may differwas comprised of two groups: 9850 permanent from the general population (raising questionsinhabitants of rural areas of the Semipalatinsk about making generalisations of radiation riskregion and 9604 permanent inhabitants of estimates derived from studies of cancer survi-villages located several hundred kilometres east vors). In this Monograph, the risk of the secondof the test site. For the first group, individual cancer following radiation therapy reported byinternal and external doses were available, and a recent X-ray studies is reviewed.collective estimate of 20 mSv due to fallout frommultiple atmospheric nuclear testing was used 2.3.1 Cancer of the lungfor the second group. Risk estimates were foundto differ depending on whether they were based The only major X-ray study with good qualityon the total cohort (including the comparison radiation dosimetry and follow-up is an inter-villages) or on the exposed villages only. The national Hodgkin disease study (Gilbert et al.,estimate of the ERR/Sv for all solid tumours was 2003). This resulted in an ERR/Gy of 0.15 (95%CI:1.77 (95%CI: 1.35–2.27) based on the data for 0.06–0.39) after adjusting for chemotherapy andthe total cohort. A significant trend with dose smoking. As with all studies considered, a poten-was observed for cancer of the stomach (ERR/ tial problem with this study is ascertainmentSv, 1.68; 95%CI: 0.83–2.99), lung (ERR/Sv, 2.60; and adjustment for cigarette smoking. Although95%CI: 1.38–4.63), and of the female breast (ERR/ the methods used in this study are thorough,Sv, 1.28; 95%CI: 0.27–3.28). However, selection they are based on data abstracted from medicalbias regarding the comparison group could not records, in which assessment of smoking beforebe ruled out. Based on the data for the exposed the primary cancer was mainly retrospective,group only, the estimate of the ERR/Sv for all so recall bias cannot be excluded. This studysolid tumours was 0.81 (95%CI: 0.46–1.33); for demonstrated that the interaction of radiationcancer of the stomach, 0.95 (95%CI: 0.17–3.49); and chemotherapy risk was consistent with anlung, 1.76 (95%CI: 0.48–8.83), and of the female additive relationship on the logistic scale, andbreast, 1.09 (95%CI:−0.05–15.8) (Bauer et al., a multiplicative relationship could be rejected2005). (P = 0.017). Conversely, the interaction of radia- tion and smoking was consistent with a multi- plicative relationship, but not with an additive2.3 Medical exposures relationship (P < 0.001). There was little indica- tion of modification of ERR by age at exposure, The previous IARC Monograph (IARC, 2000) years since exposure (after a 5-year minimumreviewed several studies of second cancer risk latent period) or attained age (Gilbert et al., 2003).following X- or γ-radiation therapy for a firstcancer. Since then, several reports have beenpublished on second cancers following radio-therapy; in these studies patients were treatedprimarily, or solely, with X-rays. However, studiesof cancers following radiation therapy pose130
    • X- and γ-radiation2.3.2 Cancer of the female breast in DNA-repair genes may modify the effects of low-dose radiation exposure from medical The major X-ray studies with good quality sources. They reported a stronger trend of breastradiation dosimetry and follow-up are nested cancer risk with the number of diagnosticcase–control studies in an international X-rays among women with 2–4 variant codonsHodgkin disease study (Travis et al., 2003) in XRCC3, NBS1, XRCC2, BRCA2 genes than inand the Netherlands Hodgkin disease study women with only 0 or 1 variant codons in those(van Leeuwen et al., 2003), as well as the French- genes. [The Working Group noted, however, thatUnited Kingdom childhood cancer (Guibout the results were inconclusive, being based onlyet al., 2005) and the US scoliosis (Ronckers et al., on self-reported exposure to ionizing radiation2008) cohorts. The excess risk in the first three of from medical sources, which may therefore bethese studies are reasonably consistent, at least for subject to recall bias. The particular genes used,those women not treated with chemotherapy: the and the gene “dose” cut-off points (≥ 2 versus < 2ERR/Gy was 0.15 (95%CI: 0.04–0.73) in Travis et codons), both presumably chosen a posteriori,al. (2003), 0.06 (95%CI: 0.01–0.13) in van Leeuwen may imply uncertainties regarding the statisticalet al. (2003), and 0.13 (95%CI: <  0.0–0.75) in significance in this study].Guibout et al. (2005). A higher point estimateof risk (ERR/Gy, 2.86; 95%CI: −0.07–8.62) wasobserved in the US scoliosis study (Ronckers 2.3.3 Cancer of the brain/central nervouset al., 2008), but in view of the wide confidence systeminterval this can be considered as consistent with The major X-ray studies with good qualitythe other three studies. A complication in some radiation dosimetry and follow-up are theof these radiotherapy studies is radiation dose Israeli tinea capitis study and the Internationalto the ovaries; the analyses of van Leeuwen et Childhood Cancer Study. The ERR/Gy in the firstal. (2003) and Travis et al. (2003) suggested that of these, a cohort study of survivors of tinea capitiswomen receiving large ovarian doses (>  5  Gy) (a fungal infection of the scalp) treated with radi-were at lower risk of radiation-induced breast ation in childhood, was 4.63 (95%CI: 2.43–9.12)cancer, presumably because of ovarian ablation for benign meningioma and 1.98 (95%CI: 0.73–and induced menopause. 4.69) for malignant brain tumour (Sadetzki et al., Ronckers et al. (2008) reported a significantly 2005). In the second study (Neglia et al., 2006), thegreater dose–response (P = 0.03) for women who ERR/Gy was 0.33 (95%CI: 0.07–1.71) for gliomas,reported a family history of breast cancer in first- 1.06 (95%CI: 0.21–8.15) for meningiomas, andor second-degree relatives (ERR/Gy, 8.37; 95%CI: 0.69 (95%CI: 0.25–2.23) for all central nervous1.50–28.16) compared with those without affected system tumours. Therefore, in both studies, thererelatives (ERR/Gy, −0.16; 95%CI: <  0–4.41). is a pattern of increased relative risk per unit doseSusceptibility alleles of single genes that confer a for benign brain tumours compared with malig-high risk of breast cancer are rare in the general nant brain tumours, a pattern also observed inpopulation, but some studies have shown modi- some other earlier studies (Little et al., 1998).fication of breast cancer risk by family history(Easton, 1999). Recent genome-wide associa- 2.3.4 Leukaemiation studies (GWAS) have established severalnew breast cancer susceptibility loci (Pharoah Modern classifications of leukaemia and otheret al., 2008). The study of Millikan et al. (2005) lymphatic and haematopoietic malignancies (e.g.suggests that other common polymorphisms Swerdlow et al., 2008) are based on cytogenetic and 131
    • IARC MONOGRAPHS – 100Dmolecular principles that do not always coincide and were employed in the industry for at leastwith the International Classification of Diseases. 1 year (Cardis et al., 2007). Workers with poten-There are generally considered to be three tial for substantial doses from other radiationmain radiogenic subtypes: acute lymphocytic types and workers with potential for high-dose-leukaemia, which is a leukaemia of precursor rate exposure were excluded from the maincells of either B-cell or T-cell origin; acute myeloid study population. [The Working Group notedleukaemia, whose lineage and subtype are gener- that strengths of the study include a commonally defined according to the French-American- core study protocol and quantitative radiationBritish (FAB) system (Bennett et al., 1982; Harris dose estimates based upon personal dosimetry.et al., 1999); and chronic myeloid leukaemia, Although it was a large study, the 15-countrywhose predominant haematological feature is an study’s statistical power was limited by smallelevated white-cell count in the peripheral blood, numbers of workers with higher doses. As isand which is characterized cytogenetically by the common in occupational cohort mortalityPhiladelphia chromosome (Linet & Cartwright, studies, there was limited information avail-1996). able on confounders, such as cigarette smoking.] The major X-ray studies with good quality Concerns about confounding by smoking wereradiation dosimetry and follow-up are an inter- addressed indirectly by the examination ofnational nested case–control study on testic- associations between radiation dose and non-ular cancer survivors and the New York tinea malignant respiratory disease. Smoking-relatedcapitis cohort. The ERR at 10  Gy in the first and non-smoking-related solid cancers were alsoof these (Travis et al., 2000) was 3.27 (95%CI: analysed separately. No statistically significant1.2–13). In the New York tinea capitis study association was seen between radiation dose and(Shore et al., 2003), the standardized incidence any of the groups of non-malignant respiratoryratio (SIR) for leukaemia (following an average diseases examined. Risk estimates for mortalitydose of about 4  Gy to cranial marrow) was 3.2 from all non-malignant respiratory disease and(95%CI: 1.5–6.1). No dose–response analysis was for chronic bronchitis and emphysema combinedreported [possibly as a consequence of the small were positive but not significantly different fromnumber of cases (eight leukaemias, of which six zero, and risk estimates for chronic pulmonarywere non-chronic lymphocytic leukaemia in the disease not otherwise specified and for emphy-exposed group versus one chronic lymphocytic sema were negative, but not significantly differentleukaemia in the control group)]. from zero. For the risk of leukaemia associated with Among the cancer categories examined, aprenatal exposures, see Section 2.1.3 and Section significant positive dose–response association2.6.19. was reported for lung cancer mortality; no other specific cancer category exhibited a statisti- cally significant dose–response trend. The ERR/2.4 Occupational studies Sv was 1.86 (90%CI: 0.49–3.63) for cancer of2.4.1 IARC 15-country study the lung, 1.93 (90%CI: <  0–7.14) for leukaemia (excluding chronic lymphocytic leukaemia), IARC conducted a collaborative study 0.97 (90%CI: 0.27–1.80) for all cancers excludingof cancer risk among workers in the nuclear leukaemia, 0.59 (90%CI:−0.16–1.51) for allindustry. Analyses include 407391 nuclear cancers excluding leukaemia, lung and pleura,industry workers who were individually moni- and 0.87 (90%CI: 0.16–1.71) for all solid cancerstored for external irradiation (primarily γ-rays), (Cardis et al., 2007). Risk estimates for all132
    • X- and γ-radiationcancers excluding leukaemia and for all cancers 2.4.2 United Kingdom radiation workersexcluding leukaemia, lung and pleural cancerswere very similar and above 200  mSv. [The Although many workers included in theWorking Group noted that, therefore, although United Kingdom National Registry for Radiationconfounding by smoking might be present, it is Workers (NRRW) were included in the IARCunlikely to explain all of the increased risk for 15-country study, Muirhead et al. (2009) reportedall cancers excluding leukaemia in that study.] on an updated and expanded study of mortalityResults by country show that, for all cancers and cancer incidence through December 2001excluding leukaemia, the ERR/Sv estimate for among 174541 people occupationally exposed toCanadian workers (6.65; 90%CI: 2.56–13.0) was ionizing radiation, based on the NRRW. Doseslarger than for workers from most other coun- from the internal deposition of radionuclidestries with sizable numbers of deaths, statistically were not generally available and were not usedsignificant, and exerted a substantial influence in the analysis, nor was individual informa-on the overall pooled analysis. tion available on smoking history. The analyses The ERR/Sv was greater for those exposed focused on doses from penetrating radiation atat ages over 50 years than for those exposed the surface of the body, estimated using personalat younger ages. With regard to all cancers dosimeters. Mortality and cancer incidenceexcluding leukaemia, ERR/Sv by age at exposure were studied in relation to dose after adjusting –was 1.74 (90%CI: 0.24–3.58) for age > 50 years, through stratification – for age, gender, calendar1.32 (90%CI: 0.12–2.71) for age 35–50 years, and period, industrial classification (industrial/non-−1.07 (90%CI: < 0–1.24) for age < 35 years. The industrial/unknown), and first employer. Withinrespective values were 3.87 (90%CI: 0.92–7.93), each stratum, the number of deaths or cases1.52 (90%CI:−0.71–4.36) and 2.51 (90%CI:−1.96– expected in each category for cumulative external8.89) for cancer of the lung, and 5.01 (90%CI: dose (0–, 10–, 20–, 50–, 100–, 200–, 400+ mSv)<  0–14.7), −1.59 (90%CI: <  0–3.02) and 1.51 was calculated, conditional on the total overall(90%CI: <  0–11.6) for leukaemia excluding dose categories, and presuming no effect of dose.chronic lymphocytic leukaemia. There was a highly significant negative associa- An analysis examined the association tion observed between mortality from bronchitis,between radiation dose and chronic lymphocytic emphysema and chronic obstructive diseaseleukaemia mortality among 295963 workers in and dose (ERR/Sv, −1.04; 90%CI: −1.35, −0.59)the seven countries with chronic lymphocytic [The Working Group noted that this would beleukaemia deaths; there were 65 chronic consistent with lower smoking prevalence amonglymphocytic leukaemia deaths in this cohort workers who accrued higher radiation doses(Vrijheid et al., 2008). The relative risk (RR) at an and suggests potential negative confoundingoccupational dose of 100 mSv compared to 0 mSv in analyses of radiation dose–response associa-was 0.84 (95%CI: 0.39–1.48) under the assump- tions for smoking-related cancers]. There was ation of a 10-year exposure lag. [The Working positive association between radiation dose andGroup noted that this study had little power due mortality due to leukaemia excluding chronicto low doses (average cumulative bone marrow lymphocytic leukaemia (ERR/Sv, 1.71; 90%CI:dose, 15  mSv), short follow-up periods, and 0.06–4.29), and also between radiation doseuncertainties in chronic lymphocytic leukaemia and mortality due to all malignant neoplasmsascertainment from death certificates.] excluding leukaemia (ERR/Sv, 0.28; 90%CI: 0.02–0.56). In analyses of cancer incidence, posi- tive associations were also seen with leukaemia 133
    • IARC MONOGRAPHS – 100Dexcluding chronic lymphocytic leukaemia with lung cancer, which was substantially atten-(ERR/Sv, 1.78; 90%CI: 0.17–4.36), and all malig- uated after adjusting for medical X-ray expo-nant neoplasms excluding leukaemia (ERR/Sv, sures (Yiin et al., 2007). Matanoski et al. (2008)0.27; 90%CI: 0.04–0.51). Among the leukaemia reported the results of analyses of leukaemia,subtypes, the strongest evidence of association, lymphohaematopoietic cancers, lung cancer,from both analyses of mortality and incidence and mesothelioma among workers from ship-data, was for chronic myeloid leukaemia; there yards involved in nuclear powered ship over-was no evidence of an association between hauls. The study included 28000 workers withchronic lymphocytic leukaemia (mortality or cumulative doses of 5  mSv or more, 10462incidence) and radiation. workers with cumulative doses less than 5 mSv, and 33353 non-nuclear workers. Exposures were2.4.3 US radiation workers almost exclusively due to γ-radiation. There was evidence of dose-related increases in leukaemia, The results of several epidemiological studies lung cancer, and lymphohaematopoietic cancers.of US radiation workers have been reported, In an internal comparison of workers withproviding results that extend those encompassed 50.0  mSv exposures to workers with exposuresby the US workers included in the 15-country of 5.0–9.9 mSv, the relative risk was 2.41 (95%CI:study. An analysis of leukaemia mortality among 0.5–23.8) for leukaemia, 1.26 (95%CI: 0.9–1.9)workers employed at the Savannah River site, a for lung cancer, and 2.94 (95%CI: 1.0–12.0) forlarge cohort of US nuclear weapons workers that lymphohaematopoietic cancers.is independent of the 15-country study, reporteda positive association between leukaemia 2.4.4 Mayakmortality and radiation dose under a 3-year lagassumption (ERR/Sv, 4; 90%CI: −0–12). The asso- Since the previous IARC Monograph (IARC,ciation was of larger magnitude for leukaemia 2000), updated reports have been published onexcluding chronic lymphocytic leukaemia (ERR/ cancer risk among workers at the Mayak nuclearSv, 8; 90%CI: 1–20) and for myeloid leukaemia complex in the Russian Federation, another(ERR/Sv, 12; 90%CI: 2–35), and these associa- large cohort of nuclear workers not included intions tended to diminish in magnitude with time the IARC study. Exposures at Mayak includedsince exposure to radiation (Richardson & Wing, external γ-radiation exposure as well as internal2007). A positive association was also observed α-particle exposure. A large number of workers,between lymphoma mortality and radiation particularly those employed in the radiochemicaldose under a 5- and 10-year lag (ERR/Sv, 6.99; and plutonium production facilities, had signifi-90%CI: 0.96–18.39 and ERR/Sv, 8.18; 90%CI: cant potential for plutonium exposures. Gilbert et1.44–21.16, respectively; Richardson et al., 2009). al. (2004) investigated lung cancer mortality overA nested case–control study of leukaemia among the period 1955–2000 in a cohort of 21790 Mayakworkers at four US nuclear weapons facilities and workers. The average cumulative external radia-the Portsmouth naval shipyard reported a posi- tion dose among those monitored for radiationtive [but highly imprecise] association between was 0.8 Gy. For external doses, the ERR/Gy wasleukaemia mortality and radiation dose (ERR/ 0.17 (95%CI: 0.052–0.32) among men and 0.32Sv, 1.44; 90%CI: <−1.03–7.59; Schubauer-Berigan (95%CI: < 0–1.3) among women. [The Workinget al., 2007). A case–control study of lung cancer Group noted that uncertainties in plutoniumamong workers at Portsmouth Naval shipyard exposure assessment could lead to inadequatereported some evidence of a positive association adjustment for the effects of internal exposures.]134
    • X- and γ-radiationAnalyses restricted to Mayak workers who were chronic lymphocytic leukaemia (ERR/Gy, 4.7;monitored for plutonium or worked only in the 90%CI: –∞, 76.1) was similar to the estimate forreactor or auxiliary plants led to smaller esti- all leukaemia combined (ERR/Gy, 4.8; 90%CI:mates of ERR/Gy of external dose (ERR/Gy, –∞, 33.1).0.065; 95%CI: <  0–0.25) than obtained via the Romanenko et al. (2008) reported resultsanalysis of the full cohort (ERR/Gy, 0.10; 95%CI: from a nested case–control study of leukaemia in< 0–0.29). The potential confounding by smoking a cohort of clean-up workers identified from thewas investigated in a subset of the cohort, and in Chernobyl State Registry of Ukraine. The studythat subcohort there was sparse data with which included 71 cases of leukaemia diagnosed duringto evaluate the effects of external dose but the 1986–2000, and 501 age- and residence-matchedERR/Gy was smaller when adjusted for smoking controls; bone-marrow doses were estimated bystatus (ERR/Gy, 0.027; 95%CI: < 0–0.18; Gilbert the RADRUE reconstruction method. The ERR/et al., 2004). Shilnikova et al. (2003) reported that Gy of total leukaemia was 3.44 (95%CI: 0.47–solid cancer and leukaemia death rates increased 9.78). Overall, the dose–response relationship forsignificantly with increasing γ-ray dose. For both chronic (ERR/Gy, 4.09; 95%CI: < 0–14.41)external doses, the ERR/Sv (adjusted for pluto- and non-chronic lymphocytic leukaemia (ERR/nium exposure) was 0.15 (90%CI: 0.09–0.20) Gy, 2.73; 95%CI: < 0–13.50) was comparable.for solid tumours and 0.99 (90%CI: 0.45–2.12) While leukaemia and lymphoma incidencefor leukaemia excluding chronic lymphocytic among Chernobyl liquidators from the Russianleukaemia. Federation were examined in the study by Kesminiene et al. (2008), analyses of mortality2.4.5 Chernobyl clean-up workers and cancer incidence among Russian liquidators were also reported by Ivanov (2007). In 1991–98, Kesminiene et al. (2008) reported the results the ERR/Gy of death from malignant neoplasmof a case–control study of leukaemia and was 2.11 (95%CI: 1.31–2.92). In 1991–2001, thelymphoma incidence among Chernobyl liquida- ERR estimation for incident solid cancers wastors from Belarus, the Russian Federation, and positive [but imprecise] (ERR/Gy, 0.34; 95%CI:Baltic countries. The main analyses included 70 −0.39–1.22; Ivanov, 2007).cases (40 leukaemia, 20 non-Hodgkin lymphoma,and ten other types) and 287 age-matchedcontrols. Bone-marrow doses were estimated 2.5 Environmental studiesby the “RADRUE” (realistic analytical dosereconstruction with uncertainty estimation) 2.5.1 Techa Riverindividual reconstruction methods (Kryuchkov Studies of environmental exposures toet al., 2009). The overall ERR/Gy was 6.0 (90%CI: γ-radiation also provide insights into the carci-−0.2, 23.5; Kesminiene et al., 2008). The dose– nogenic effects of protracted exposures. A notableresponse relationship was of larger magnitude investigation of the effects of environmentalfor non-Hodgkin lymphoma (ERR/Gy, 28.1; exposures to γ-radiation concerns releases of90%CI: 0.9–243.0) than for leukaemia (ERR/Gy, radioactive materials into the Techa River in4.8; 90%CI: <  0, 33.1), although the confidence the southern urals, the Russian Federation, asintervals were wide for both outcomes. The ERR/ a result of operations at the Mayak productionGy for leukaemia excluding chronic lymphocytic facility. External exposures were primarily dueleukaemia was 5.0 (90%CI: −0.38, 5.7) based on to γ-radiation from contamination of the river19 cases and 83 controls; the risk estimate for shoreline and floodplains; in addition, internal 135
    • IARC MONOGRAPHS – 100Dexposures resulted from the consumption of food reinforcing steel used to build their apartments.and drink contaminated with radionuclides. The study included 117 cancer cases diagnosedFission products were the largest component of during 1983–2005 among 6242 people with anthe internal dose, and residents thus received average excess cumulative exposure estimateinternal γ-and β-radiation exposures. The ratio of about 48  mGy. There was a significant asso-of external/internal radiation varied according ciation between the estimated radiation doseto the site. and leukaemia excluding chronic lymphocytic Since the previous IARC Monograph, several leukaemia (hazard ratio (HR)/100 mGy, 1.19;reports have been published on associations 90%CI: 1.01–1.31); the HR/100 mGy estimatedbetween radiation exposure and cancer among for breast cancer was 1.12 (90%CI: 0.99–1.21).residents of villages along the Techa river. Nair et al. (2009) reported results on cancerKrestinina et al. (2007) reported results on solid incidence in Kerala, India, in an area knowncancer incidence in a cohort of 17433 people who for high-background radiation from thorium-resided in villages along the Techa river, with containing monazite sand. Cancer incidence infollow-up from 1956–2002, in relation to the esti- a cohort of 69958 residents aged 30–84 yearsmated cumulative stomach dose (approximately was ascertained through to 2005 (average dura-half from internal dose). There was a highly tion of follow-up, 10.5  years); the cumulativesignificant linear dose–response relationship radiation dose for each individual was estimatedbetween cumulative stomach dose and incidence based on outdoor and indoor dosimetry of eachof solid tumours (P = 0.004). Ostroumova et al. household. The median outdoor radiation levels(2008) reported results on breast cancer incidence were approximately 4  mGy per year; medianin a cohort of 9908 women with follow-up from indoor radiation levels were somewhat lower.1956–2004. A significant dose–response rela- The analysis, which included 1379 cancer casestionship (P = 0.01) was reported between cumu- and 30 leukaemia cases, found no cancer sitelative stomach dose and breast cancer incidence, was significantly related to cumulative radiationwith an estimated ERR/Gy of 5.00 (95%CI: 0.80– dose. The estimated ERR/Gy of cancer excluding12.76). Ostroumova et al. (2006) reported results leukaemia was −0.13 (95%CI: −0.58–0.46).from a nested case–control study of leukaemiaamong residents near the Techa river. The studyincluded 83 cases ascertained over a 47-year 2.6 Synthesisperiod of follow-up and 415 controls; in analyses The previous IARC Monograph (IARC, 2000)of leukaemia excluding chronic lymphocytic states there is strong evidence for causal asso-leukaemia, the odds ratio at 1 Gy, estimated via ciations between X- and γ-radiation and severala log-linear model, was 4.6 (95%CI: 1.7–12.3), 7.2 cancer sites, including those listed in Table 2.7.(95%CI: 1.7–30.0), and 5.4 (95%CI: 1.1–27.2) for In this current Monograph, the Working Grouptotal, external and internal red bone-marrow re-evaluated the evidence (the earlier evidencedoses, respectively. and that published after the previous IARC Monograph) for those cancer sites, and simi-2.5.2 High-background radiation areas larly, found strong evidence of causation. The major publications on which the above conclu- Hwang et al. (2008) reported results on cancer sion is based are also listed in Table 2.7. Therisks in a cohort of Chinese residents in Taiwan, United States National Research Council (2006)China, who received protracted low-dose-rate and UNSCEAR (2008b) have also made similarγ-radiation exposures from 60Co-contaminated conclusions for the cancer sites listed in Table 2.7.136
    • X- and γ-radiationTable 2.7 Cancer sites and tumours judged to have sufficient evidence for a causal associationwith X-ray and γ-ray exposureOrgan site Selected key studiesStomach Boice et al. (1988), Mattsson et al. (1997), Carr et al. (2002), Preston et al. (2003, 2007)Colon Darby et al. (1994), Preston et al. (2003, 2007)Lung Weiss et al. (1994), Carr et al. (2002), Gilbert et al. (2003), Preston et al. (2003, 2007)Basal cell skin carcinoma Schneider et al. (1985), Ron et al. (1991, 1998), Little et al. (1997), Shore et al. (2002), Preston et al. (2007)Female breast Howe & McLaughlin (1996), Preston et al. (2002, 2003, 2007)Thyroid Lundell et al. (1994), Lindberg et al. (1995), Ron et al. (1995), Preston et al. (2007)Leukaemia excluding chronic Little et al. (1999), Travis et al. (2000), Preston et al. (2003, 2004), Muirhead et al. (2009)lymphocytic The evidence for the other individual cancer significant positive dose–response relationshipsites is shown in Table  2.8. The focus has been in the Japanese A-bomb survivor incidence dataon relatively large studies of good design, where (Land et al., 1996), and in the study of patientsgood quality dosimetry has been carried out, and who received radiation therapy during child-where the magnitude of the doses is generally hood for benign conditions in the head and necksubstantial. Wherever possible, risk estimates area (Schneider et al., 1998). The estimated ERR/from several studies were provided including the Sv for the incidence data of the Japanese A-bomblatest LSS incidence analysis (Preston et al., 2007) survivors was 4.47 (90%CI: 2.45–8.46) for malig-and in some cases the latest LSS mortality data nant tumours, based on 31 cases, and for benign(Preston et al., 2003), the International Radiation tumours the risk estimate was 1.71 (90%CI: 1.13–Study of Cervical Cancer Patients (IRSCCP; Boice 2.71), based on 64 cases (Land et al., 1996). Theet al. 1988), the United Kingdom ankylosing ERR/Gy in the Schneider et al. (1998) study wasspondylitis data (Weiss et al., 1994), the United −0.06 (95%CI: –∞–4.0) for malignant tumours,Kingdom metropathia haemorrhagica study based on 22 cases, and 19.6 (95%CI: 0.16–∞) for(Darby et al., 1994), the NRRW (Muirhead et al., benign tumours, based on 66 cases. Although2009), and the IARC 15-country study (Cardis data on dose–response are lacking, there areet al., 2007). For certain cancer sites, some of also indications of significant excess risk in thethese studies are largely uninformative (e.g. only Israeli tinea capitis study (Modan et al., 1998),standardized mortality ratios (SMRs) are given and in the Rochester thymus irradiation studyfor various cancer sites in the metropathia haem- (Hildreth et al., 1985; Table  2.8). In the Israeliorrhagica study), which were therefore omitted study as in the LSS, risks for malignant tumoursfrom Table 2.8. (RR, 4.49; 95%CI: 1.45–13.9) were greater than benign tumours (RR, 2.62; 95%CI: 1.10–6.25), in2.6.1 Cancer of the salivary gland contrast to the pattern in the study of Schneider et al. (1998). In the Rochester study, there were This is a rare cancer site and has not been eight benign tumours (RR, 4.4; 95%CI: 1.2–16.7),much studied in most of the major radiation- but no malignant tumour in the irradiatedexposed cohorts (e.g. Boice et al., 1988; Weiss group. A non-significant excess risk (RR, 1.8;et al., 1994; Cardis et al., 2007, Muirhead 95%CI: 0.4–8.9) for salivary gland tumours (twoet al., 2009). Nevertheless, there is a statistically 137
    • IARC MONOGRAPHS – 100Dmalignant and four benign) was reported in the statistically significant excess risk reported in theNew York tinea capitis study (Shore et al., 2003). United Kingdom ankylosing spondylitis study Preston et al. (2007) did not analyse this (Weiss et al., 1994); the ERR/Gy was 0.17 (95%CI:tumour in the most recent analysis of cancer 0.09–0.25), based on 74 deaths.incidence among the Japanese A-bomb survi- In summary, there are strong and highlyvors. [The Working Group analysed the publicly statistically significant trends in the LSS inci-available data set using a linear relative risk dence and mortality data (Preston et al., 2003,model in which the expected number of cases 2007), as is the case in the United Kingdomin stratum i and dose group d is assumed to ankylosing spondylitis data (Weiss et al., 1994).be given by PYidλd [1 + αDid] fitted by Poisson There are (statistically non-significant) indica-maximum likelihood, and profile-likelihood- tions of excess in several other studies (e.g. Boicebounds derived (McCullagh & Nelder, 1989) et al. 1985; Muirhead et al. 2009; Table 2.8).using EPICURE (Preston et al., 1998). Here, PYid is the number of (migration- 2.6.3 Cancer of the small intestine, includingadjusted) person–years of follow-up, λd is the the duodenum(semi-parametric) background hazard rate (esti-mated separately for each stratum), and Did is the This is a rare cancer site and has not beenDS02 organ dose in Sv (brain dose is used as a much studied in most of the major radiation-surrogate), using the neutron quality factor of 10 exposed cohorts (e.g. Weiss et al., 1994; Cardisrecommended by the ICRP (1991). The estimate et al., 2007; Muirhead et al., 2009). There was noof the ERR coefficient α is given in Table 2.8, and significant excess risk and no evidence of a posi-is seen to be statistically significant (2.42 per Sv; tive dose–response in the IRSCCP (Boice et al.,95%CI: 0.48–6.70).] 1988): the odds ratio was 1.0 (90%CI: 0.3–2.9), In summary, although this is a rare cancer based on 22 cases, despite the very high dosessite, there are strong and highly statistically received (estimated to be several hundred Gy onsignificant trends in the LSS data (Land et al., average). Preston et al. (2007) did not analyse1996; Preston et al., 2007), and trends of similar this tumour among A-bomb survivors. [Themagnitude in the study of Schneider et al. (1998). Working Group analysed the publicly availableThere are indications of excess risk in several LSS incidence data set using a linear relative riskother radiotherapeutically exposed groups. model (Formula 1) and obtained an ERR, given in Table 2.8, which is not statistically significant2.6.2 Cancer of the oesophagus (ERR/Sv, 0.65; 95%CI: −0.32–4.89), based on 16 cases.] Cancer incidence data from the latest LSS In summary, for this rare cancer, there aredata show a significant excess risk of oesopha- essentially only two informative studies, thegeal cancer (Preston et al., 2007), as do the LSS incidence data (Preston et al., 2007) and thelatest site-specific mortality data (Preston et al., IRSCCP (Boice et al., 1988), but neither of which2003), as reported in Table 2.8. The estimate of reports a statistically significant excess risk.the ERR/Sv coefficient for the incidence data is0.52 (90%CI: 0.15–1.0), based on 352 cases. For 2.6.4 Cancer of the rectumthe LSS mortality data the ERR/Sv was broadlysimilar with 0.61 (90%CI: 0.15–1.2) for men, Among the survivors of the atomic bomb-based on 224 deaths; and, 1.7 (90%CI: 0.46–3.8) ings, mortality from cancer of the rectumfor women, based on 67 deaths. There was also a was not clearly associated with radiation dose138
    • Table 2.8 Summary of evidence for organ sites initially deemed to be potentially having limited evidence of carcinogenicity or inadequate evidence of carcinogenicity Organ site Study Target Dose range Reference ERR/EOR per Gy/ Mortality/ Cases/ Other (mean)(Gy) Sv unless otherwise incidence deaths comments stated (95%CI) (A-bomb, IARC 15-country and NRRW are Sv-1, all others Gy-1) Salivary A-bomb Uniform whole 0–4 (0.1) Land et al. All malignant: Incidence 31 gland body, mostly high- (1996) 4.47 (2.45–8.46)a energy (2–5 MeV) All benign: Incidence 64 gamma + small 1.71 (1.13–2.71)a amount of high- energy neutrons A-bomb Uniform whole 0–4 (0.1) Preston et al. All malignant: Incidence 34 Stratified linear body, mostly high- (2007) 2.42 (0.48–6.70) RR model fitted energy (2–5 MeV) to publicly gamma + small available data, amount of high- using brain dose energy neutrons Benign head 200 KeV X-rays to 0.01–15.8 (4.2) Schneider et All malignant: −0.06 Incidence 22 & neck RT in head and neck al. (1998) (–∞–4.0) childhood All benign: 19.6 Incidence 66 (0.16–+∞) Thymic Thymus 250 kVp Breast dose Hildreth et al. All malignant: RR, Incidence 11 Women only enlargement X-rays 0.01–19.51 (1985) 0.0 (0.0–34.6) (0.69) All benign: RR, 1 4.4 (1.2–16.7) New York tinea X-rays to scalp (0.39 per Shore et al. RR, 1.8 (0.4–13) Incidence 8 6 exposed, 2 capitis treatment) (2003), Harley unexposed cases et al. (1976) Israeli tinea X-rays to scalp 0.63–2.86 Modan et al. Malignant: RR, Incidence 16 12 exposed, 4 capitis: (0.78) per (1998) 4.49 (1.45–13.9) controls malignant treatment Benign: RR, 22 14 exposed, 8 2.62 (1.10–6.25) controls139 X- and γ-radiation
    • 140 Table 2.8 (continued) Organ site Study Target Dose range Reference ERR/EOR per Gy/ Mortality/ Cases/ Other (mean)(Gy) Sv unless otherwise incidence deaths comments stated (95%CI) (A-bomb, IARC 15-country and NRRW are Sv-1, all others Gy-1) Oesophagus A-bomb Uniform whole 0–4 (0.1) Preston et al. 0.52 (0.15–1.0)a Incidence 352 80% of cases body, mostly high- (2007) confirmed energy (2–5 MeV) histologically IARC MONOGRAPHS – 100D gamma + small amount of high- energy neutrons A-bomb Uniform whole 0–4 (0.1) Preston et al. Men: 0.61 (0.15–1.2)a Mortality 224 body, mostly high- (2003) Women: 67 energy (2–5 MeV) 1.7 (0.46–3.8)a gamma + small amount of high- energy neutrons Ankylosing X-rays to spinal 90% range Weiss et al. 0.17 (0.09–0.25) Mortality 74 spondylitis 0.48–10.16 (1994) (5.55) Metropathia X-rays to ovaries 90% range Darby et al. SMR, 0.97 (0.44–1.84) Mortality 9 haemorrhagica 0.02–0.11 (1994) (0.05) International Mostly 200– (0.35) Boice et al. 0.26 (−1.1–1.3)b Incidence 12 10-year Radiation Study 400 kVp X-ray (1985) survivors of Cervical +radium +gamma following the Cancer Patients to cervix primary cancer IARC Uniform whole 0–> 0.5 Sv Cardis et al. −1.6 (−4.3–1.5)a d Mortality 144 15-country body (0.0194) (2007) nuclear workers United Kingdom Uniform whole 0–> 0.1 Sv Muirhead et 0.15 (−0.84–1.72) Mortality 341 NRRW body (0.0249) al. (2009) 0.15 (−0.91–2.06) Incidence 300
    • Table 2.8 (continued) Organ site Study Target Dose range Reference ERR/EOR per Gy/ Mortality/ Cases/ Other (mean)(Gy) Sv unless otherwise incidence deaths comments stated (95%CI) (A-bomb, IARC 15-country and NRRW are Sv-1, all others Gy-1) Small A-bomb Uniform whole 0–4 (0.1) Preston et al. 0.65 (−0.32–4.89) Incidence 16 Stratified linear intestine body, mostly high- (2007) RR model fitted energy (2–5 MeV) to publicly gamma + small available data, amount of high- using colon dose energy neutrons Cervical Mostly 200– 10–20 Boice et al. OR, 1.0 (0.3–2.9)a Incidence 22 RR trend not cancer 400 kVp X-ray (1988) computed +radium +gamma because of small to cervix number of non- exposed cases Rectum A-bomb Uniform whole 0–4 Sv (0.1) Preston et al. 0.19 (−0.04–0.47)a Incidence 838 90% of cases body, mostly high- (2007) confirmed energy (2–5 MeV) histologically gamma + small amount of high- energy neutrons A-bomb Uniform whole 0–4 (0.1) Preston et al. Men: −0.25 Mortality 172 body, mostly high- (2003) (< −0.3–0.15)a energy (2–5 MeV) Women: 198 gamma + small 0.75 (0.16–1.6)a amount of high- energy neutrons Ankylosing X-rays to spine 90% range Weiss et al. 0.03 (−0.03–0.10)c Mortality 62 spondylitis 0.53–10.20 (1994) (4.12)c Metropathia X-rays to ovaries 90% range Darby et al. 0.04 (−0.09–0.16) Mortality 14 haemorrhagica 3.4–6.3 (4.9) (1994) International Mostly 200– 30–60 Boice et al. 0.02 (0.00–0.04)a Incidence 488 Radiation Study 400 kVp X-ray (1988) of Cervical +radium +gamma Cancer Patients to cervix141 X- and γ-radiation
    • 142 Table 2.8 (continued) Organ site Study Target Dose range Reference ERR/EOR per Gy/ Mortality/ Cases/ Other (mean)(Gy) Sv unless otherwise incidence deaths comments stated (95%CI) (A-bomb, IARC 15-country and NRRW are Sv-1, all others Gy-1) Rectum IARC Uniform whole 0–> 0.5 Sv Cardis et al. 1.27 (< 0–7.62)a Mortality 185 (contd.) 15-country body (0.0194) (2007) nuclear workers IARC MONOGRAPHS – 100D United Kingdom Uniform whole 0–> 0.1 Sv Muirhead et 1.69 (−0.02–4.73) Mortality 303 NRRW body (0.0249) al. (2009) 1.31 (0.04–3.2) Incidence 586 Liver A-bomb Uniform whole 0–4 (0.1) Preston et al. 0.30 (0.11–0.55)a, e Incidence 1494 41% of cases body, mostly high- (2007) histologically energy (2–5 MeV) confirmed gamma + small amount of high- energy neutrons A-bomb Uniform whole 0–4 (0.1) Preston et al. Men: 0.39 (0.11–0.68)a Mortality 722 body, mostly high- (2003) energy (2–5 MeV) Females: 0.35 (0.07, Mortality 514 gamma + small 0.72)a amount of high- energy neutrons Ankylosing X-rays to spine 90% range Weiss et al. RR, 0.81 (0.40–1.44) Mortality 11 Dose–response spondylitis 0.31–3.83 (1994) not calculated (2.13) Metropathia X-rays to ovaries 90% range Darby et al. SMR, 0.33 (0.04, 1.21) Mortality 2 Dose–response haemorrhagica 0.12–0.55 (1994) not calculated (0.27) IARC Uniform whole 0–> 0.5 Sv Cardis et al. 6.47 (< 0–27.0)a Mortality 62 15-country body (0.0194) (2007) nuclear workers United Kingdom Uniform whole 0–> 0.1 Sv Muirhead et −1.50 (< −1.93–8.56) Mortality 40 NRRW primary body (0.0249) al. (2009) −0.65 (< −1.93–7.73) Incidence 56 liver
    • Table 2.8 (continued) Organ site Study Target Dose range Reference ERR/EOR per Gy/ Mortality/ Cases/ Other (mean)(Gy) Sv unless otherwise incidence deaths comments stated (95%CI) (A-bomb, IARC 15-country and NRRW are Sv-1, all others Gy-1) Pancreas A-bomb Uniform whole 0–4 (0.1) Preston et al. 0.26 (< −0.07–0.68)a Incidence 512 52% of cases body, mostly high- (2007) confirmed energy (2–5 MeV) histologically gamma + small amount of high- energy neutrons A-bomb Uniform whole 0–4 (0.1) Preston et al. Men: −0.11 Mortality 163 body, mostly high- (2003) (< −0.3–0.44)a energy (2–5 MeV) Females: −0.01 (−0.28- 244 gamma + small 0.45)a amount of high- energy neutrons Peptic ulcer 250 kVp X-ray 0.9–> 16 (13.5) Carr et al. Irradiated + not: Mortality 59 (2002) 0.04 (0.00–0.08) Irradiated only: −0.03 (−0.10–0.05) Skin Radium-226 < 0.01–> 1.0 Lundell & 25.1 (5.5–57.7) Incidence 9 haemangioma applicators (0.09) Holm (1995) Ankylosing X-rays to spine 90% range Weiss et al. 0.12 (0.05–0.20) Mortality 84 spondylitis 0.53–8.24 (1994) (4.52) Metropathia X-rays to ovaries 90% range Darby et al. SMR, 0.66 (0.30–1.26) Mortality 9 Dose–response haemorrhagica 0.12–0.61 (1994) not calculated (0.29) International Cervix 0–> 3 (1.9) Boice et al. 0.00 (−0.28–0.62) a Incidence 221 Radiation Study (1988) of Cervical Cancer Patients IARC Uniform whole 0–> 0.5 Sv Cardis et al. 2.10 (−0.59–6.77)a Mortality 272 15-country body (0.0194) (2007) nuclear workers United Kingdom Uniform whole 0–> 0.1 Sv Muirhead et −0.05 (−1.11–2.07) Mortality 330 NRRW body (0.0249) al. (2009) 0.08 (−1.07–2.51) Incidence 320143 X- and γ-radiation
    • 144 Table 2.8 (continued) Organ site Study Target Dose range Reference ERR/EOR per Gy/ Mortality/ Cases/ Other (mean)(Gy) Sv unless otherwise incidence deaths comments stated (95%CI) (A-bomb, IARC 15-country and NRRW are Sv-1, all others Gy-1) Bone & A-bomb Uniform whole 0–4 (0.1) Preston et al. Bone: 1.01 (< 0–4.38) Incidence 18 Stratified linear connective body, mostly high- (2007) Connective tissue: 23 RR model fitted tissue energy (2–5 MeV) 1.76 (< 0–6.41) to publicly IARC MONOGRAPHS – 100D gamma + small Bone+connective 41 available data, amount of high- tissue: 1.34 (0.14–3.74) using skeletal energy neutrons dose Retinoblastoma (0.0) Wong et al. 0.19 (0.14–0.32)b Incidence 81 patients (1997) Childhood (27) Tucker et al. 0.06 (0.01–0.2)a b Incidence 54 radiotherapy (1987) (international) United Kingdom 0–> 50 (10) Hawkins et al. 0.16 (0.07–0.37)b Incidence 49 childhood (1996) cancer: bone Ankylosing X-rays to spine 90% range Weiss et al. Bone: RR, Mortality 9 Dose–response spondylitis 1.42–7.82 (1994) 3.29 (1.58–5.92) not calculated (4.54) Connective tissue: RR, 10 2.83 (1.41–4.95) Metropathia X-rays to ovaries 90% range Darby et al. SMR, 0.00 (0.00–4.01) Mortality 0 Dose–response haemorrhagica: 1.0–1.6 (1.3)g (1994) not calculated bone International Cervix 0–> 30 (22.0) Boice et al. RR, 1.34 (0.3–5.6)a Incidence 15 RR trend not Radiation Study (1988) computed of Cervical because of small Cancer Patients: number of non- bone exposed cases
    • Table 2.8 (continued) Organ site Study Target Dose range Reference ERR/EOR per Gy/ Mortality/ Cases/ Other (mean)(Gy) Sv unless otherwise incidence deaths comments stated (95%CI) (A-bomb, IARC 15-country and NRRW are Sv-1, all others Gy-1) Bone & International Cervix 0–> 20 (7.0) Boice et al. −0.05 (−0.11–0.13)a Incidence 46 connective Radiation Study (1988) tissue of Cervical (contd.) Cancer Patients: connective tissue IARC Uniform whole 0–> 0.5 Sv Cardis et al. Bone: −8.4 Mortality 16 15-country body (0.0194) (2007) (−10.0–17.2)a d nuclear workers Connective tissue: 39 0.32 (< 0–11.5)a United Kingdom Uniform whole 0–> 0.1 Sv Muirhead et Bone: < −1.93 Mortality 8 NRRW body (0.0249) al. (2009) (< −1.93–28.51) Bone: 1.18 Incidence 17 (< −1.93–52.16) Connective tissue: Mortality 31 < −1.93 (< −1.93–7.49) Connective tissue: Incidence 58 < −1.93 (< −1.93–1.42) Skin cancers other than basal cell skin cancer Squamous A-bomb Uniform whole 0–4 (0.1) Ron et al. < −0.1 (< −0.1–0.10)a Incidence 69 cell body, mostly high- (1998) carcinoma energy (2–5 MeV) gamma + small amount of high- energy neutrons New York tinea X-rays to scalp 3.3–6 scalp Shore et al. Irradiated 7 cases vs Incidence 7 capitis dose (4.75) (2002) unirradiated 0 cases Israeli tinea X-rays to scalp 5.5–24.4 scalp Ron et al. Irradiated 0 cases vs Incidence 2 capitis dose (6.8) (1991) unirradiated 2 cases145 X- and γ-radiation
    • 146 Table 2.8 (continued) Organ site Study Target Dose range Reference ERR/EOR per Gy/ Mortality/ Cases/ Other (mean)(Gy) Sv unless otherwise incidence deaths comments stated (95%CI) (A-bomb, IARC 15-country and NRRW are Sv-1, all others Gy-1) Melanoma A-bomb Uniform whole 0–4 (0.1) Thompson et 0.22 (< 0–4.14) Incidence 13 Stratified linear body, mostly high- al. (1994) RR model fitted energy (2–5 MeV) to publicly IARC MONOGRAPHS – 100D gamma + small available data, amount of high- using skeletal energy neutrons dose France-United Treatment at 0–51 (3.1) Guérin et al. 0.07 (0.00–0.15) Incidence 16 Kingdom various sites (2003) childhood cancer IARC Uniform whole 0–> 0.5 Sv Cardis et al. 0.15 (< 0–5.44)a Mortality 87 15-country body (0.0194) (2007) nuclear workers (bone) United Kingdom Uniform whole 0–> 0.1 Sv Muirhead et 1.39 (−0.65–5.6) Incidence 261 NRRW body (0.0249) al. (2009) Uterus A-bomb: uterine Uniform whole 0–4 (0.1) Preston et al. Uterine corpus: 0.29 Incidence 184 97% of cases corpus body, mostly high- (2007) (−0.14–0.95)a confirmed energy (2–5 MeV) Uterine cervix + NOS: 978 histologically gamma + small 0.06 (−0.14–0.31)a Cervix: 97% of amount of high- Uterine corpus, 1162 cases confirmed energy neutrons uterine NOS+cervix: histologically 0.10 (−0.09–0.33)a Uterine NOS: 55% of cases confirmed histologically
    • Table 2.8 (continued) Organ site Study Target Dose range Reference ERR/EOR per Gy/ Mortality/ Cases/ Other (mean)(Gy) Sv unless otherwise incidence deaths comments stated (95%CI) (A-bomb, IARC 15-country and NRRW are Sv-1, all others Gy-1) Uterus A-bomb Uniform whole 0–4 (0.1) Preston et al. 0.17 (−0.10–0.52)a Mortality 518 (contd.) body, mostly high- (2003) energy (2–5 MeV) gamma + small amount of high- energy neutrons Ankylosing X-rays to spine 90% range Weiss et al. Uterus including Mortality 10 spondylitis 0.14–10.35 (1994) cervix: ERR/Gy 0.09 (4.94) (−0.02–0.19) Uterus apart from cervix: RR, 1.91 (0.92–3.51) Cervix: RR, 3 0.36 (0.07–1.04) Metropathia X-rays to ovaries 90% range Darby et al. 0.09 (−0.02–0.19) Mortality 25 haemorrhagica: 4.3–6.4 (5.2) (1994) uterine corpus +cervix International Mostly 200– (165) Boice et al. OR, 1.34 (0.8–2.3)a h Incidence 313 RR trend not Radiation Study 400 kVp X-ray (1988) computed of Cervical +radium +gamma because of small Cancer Patients: to cervix number of non- uterine corpus exposed cases IARC Uniform whole 0–> 0.5 Sv Cardis et al. Uterus apart from Mortality 13 15-country body (0.0194) (2007) cervix 0.16 (< 0–94.1)a i nuclear workers: uterus apart from cervix: Cervix −0.11 (< 0, 14 131)a United Kingdom Uniform whole 0–> 0.1 Sv Muirhead et 17.81 (< −1.93–91.96) Mortality 19 NRRW: uterine body (0.0249) al. (2009) 10.52 (−0.50–48.02) Incidence 58 corpus +cervix147 X- and γ-radiation
    • 148 Table 2.8 (continued) Organ site Study Target Dose range Reference ERR/EOR per Gy/ Mortality/ Cases/ Other (mean)(Gy) Sv unless otherwise incidence deaths comments stated (95%CI) (A-bomb, IARC 15-country and NRRW are Sv-1, all others Gy-1) Ovary A-bomb Uniform whole 0–4 (0.1) Preston et al. 0.61 (0.00–1.5)a Incidence 245 88% of cases body, mostly high- (2007) confirmed energy (2–5 MeV) histologically IARC MONOGRAPHS – 100D gamma + small amount of high- energy neutrons A-bomb Uniform whole 0–4 (0.1) Preston et al. 0.94 (0.07–2.0)a Mortality 136 body, mostly high- (2003) energy (2–5 MeV) gamma + small amount of high- energy neutrons Ankylosing X-rays to spine 90% range Weiss et al. RR, 0.97 (0.52–1.67) Mortality 13 spondylitis 0.12–12.28 (1994) (5.53) Metropathia X-rays to ovaries < 4.8–> 6.0 Darby et al. 0.02 (−0.08–0.12) Mortality 18 haemorrhagica (5.3) (1994) International Mostly 200– 0–> 50 (32.1) Boice et al. 0.01 (−0.02–0.14)a Incidence 309 Radiation Study 400 kVp X-ray (1988) of Cervical +radium +gamma Cancer Patients to cervix IARC Uniform whole 0–> 0.5 Sv Cardis et al. −9.1 (−10.0–15.8)a d Mortality 35 15-country body (0.0194) (2007) nuclear workers United Kingdom Uniform whole 0–> 0.1 Sv Muirhead et < −1.93 Mortality 18 NRRW body (0.0249) al. (2009) (< −1.93–121.76) < −1.93 Incidence 15 (< −1.93–88.75)
    • Table 2.8 (continued) Organ site Study Target Dose range Reference ERR/EOR per Gy/ Mortality/ Cases/ Other (mean)(Gy) Sv unless otherwise incidence deaths comments stated (95%CI) (A-bomb, IARC 15-country and NRRW are Sv-1, all others Gy-1) Prostate A-bomb Uniform whole 0–4 (0.1) Preston et al. 0.11 (−0.10–0.54)a Incidence 387 88% of cases body, mostly high- (2007) confirmed energy (2–5 MeV) histologically gamma + small amount of high- energy neutrons A-bomb Uniform whole 0–4 (0.1) Preston et al. 0.21 (< −0.3–0.96)a Mortality 104 body, mostly high- (2003) energy (2–5 MeV) gamma + small amount of high- energy neutrons Ankylosing X-rays to spine 90% range Weiss et al. 0.14 (0.02–0.28) Mortality 88 spondylitis 0.18–0.71 (1994) (0.36)j IARC Uniform whole 0–> 0.5 Sv Cardis et al. 0.77 (< 0–4.58)a Mortality 301 15-country body (0.0194) (2007) nuclear workers United Kingdom Uniform whole 0–> 0.1 Sv Muirhead et 0.42 (−0.42–1.64) Mortality 702 NRRW body (0.0249) al. (2009) −0.18 (−0.73–0.57) Incidence 1516 Bladder A-bomb Uniform whole 0–4 (0.1) Preston et al. 1.23 (0.59–2.1)a e Incidence 469 88% of cases body, mostly high- (2007) confirmed energy (2–5 MeV) histologically gamma + small amount of high- energy neutrons A-bomb Uniform whole 0–4 (0.1) Preston et al. Men: 1.1 (0.2–2.5)a Mortality 83 body, mostly high- (2003) energy (2–5 MeV) Women: 1.2 (0.10–3.1)a 67 gamma + small amount of high- energy neutrons149 X- and γ-radiation
    • 150 Table 2.8 (continued) Organ site Study Target Dose range Reference ERR/EOR per Gy/ Mortality/ Cases/ Other (mean)(Gy) Sv unless otherwise incidence deaths comments stated (95%CI) (A-bomb, IARC 15-country and NRRW are Sv-1, all others Gy-1) Bladder Ankylosing X-rays to spine 90% range Weiss et al. 0.24 (0.09–0.41) Mortality 71 (contd.) spondylitis 0.20–4.85 (1994) (2.18) IARC MONOGRAPHS – 100D Metropathia X-rays to ovaries 90% range Darby et al. 0.40 (0.15–0.66) Mortality 20 haemorrhagica 4.3–6.4 (5.2) (1994) SMR, 3.01 (1.84–4.64) International Mostly 200– 30–60 Gy Boice et al. 0.07 (0.02–0.17)a Incidence 273 Radiation Study 400 kVp X-ray (1988) of Cervical +radium +gamma Cancer Patients to cervix IARC Uniform whole 0–> 0.5 Sv Cardis et al. −2.2 (−5.0–1.0)a d Mortality 145 15-country body (0.0194) (2007) nuclear workers United Kingdom Uniform whole 0–> 0.1 Sv Muirhead et 0.40 (−0.78–2.48) Mortality 301 NRRW body (0.0249) al. (2009) 0.65 (−0.28–1.96) Incidence 748 Kidney A-bomb Uniform whole 0–4 (0.1) Preston et al. 0.13 (−0.25–0.75)a Incidence 167 82% of cases body, mostly high- (2007) EAR, 0.25 × 10−4/PY/ confirmed energy (2–5 MeV) Sv (0.07–0.53)a histologically gamma + small amount of high- energy neutrons A-bomb Uniform whole 0–4 (0.1) Preston et al. Men: −0.02 Mortality 36 body, mostly high- (2003) (< −0.3–1.1)a energy (2–5 MeV) Women: 0.97 31 gamma + small (< −0.3–3.8)a amount of high- energy neutrons Ankylosing X-rays to spine 90% range Weiss et al. 0.10 (0.02–0.20) Mortality 35 spondylitis 0.71–11.74 (1994) (6.08)
    • Table 2.8 (continued) Organ site Study Target Dose range Reference ERR/EOR per Gy/ Mortality/ Cases/ Other (mean)(Gy) Sv unless otherwise incidence deaths comments stated (95%CI) (A-bomb, IARC 15-country and NRRW are Sv-1, all others Gy-1) Kidney Metropathia X-rays to ovaries 90% range Darby et al. SMR, 1.19 (0.39–2.78) Mortality 5 Dose–response (contd.) haemorrhagica 0.17–0.79 (1994) not calculated (0.40) International Mostly 200– 0–> 3 (2.0) Boice et al. 0.71 (0.03–2.24)a Incidence 148 Radiation Study 400 kVp X-ray (1988) of Cervical +radium +gamma Cancer Patients to cervix IARC Uniform whole 0–> 0.5 Sv Cardis et al. 2.26 (< 0–14.9)a Mortality 127 15-country body (0.0194) (2007) nuclear workers United Kingdom Uniform whole 0–> 0.1 Sv Muirhead et −1.03 (−1.57–0.39) Mortality 187 NRRW body (0.0249) al. (2009) −0.41 (−1.32–1.48) Incidence 296 Brain & A-bomb Uniform whole 0–4 (0.1) Preston et al. All brain & CNS: Incidence 281 81% of cases CNS body, mostly high- (2007) 0.62 (0.21–1.2)a confirmed energy (2–5 MeV) Glioma: 0.56 56 histologically gamma + small (−0.2–2.0) amount of high- Meningioma: 0.64 110 energy neutrons (−0.01–1.8) Schwannoma: 64 4.50 (1.9–9.2) A-bomb Uniform whole 0–4 (0.1) Preston et al. Men: 5.3 (1.4–16)a Mortality 14 body, mostly high- (2003) energy (2–5 MeV) Women: 0.51 17 gamma + small (< −0.3–3.9)a amount of high- energy neutrons New York tinea Scalp irradiation 0.75–1.7 (1.4) Shore et al. 1.1 (0.1–2.8) Incidence 7 SIR for brain capitis (2003) RR (treated:control), cancer, +∞ (1.2–+ ∞) 3.0 (1.3–5.9). No brain cancers were observed in the control group151 X- and γ-radiation
    • 152 Table 2.8 (continued) Organ site Study Target Dose range Reference ERR/EOR per Gy/ Mortality/ Cases/ Other (mean)(Gy) Sv unless otherwise incidence deaths comments stated (95%CI) (A-bomb, IARC 15-country and NRRW are Sv-1, all others Gy-1) Brain & Israeli tinea X-rays to scalp 1.0–6.0 (1.5) Sadetzki et al. All malignant: Incidence 44 CNS capitis (2005) 1.98 (0.73–4.69) (contd.) All benign: 81 IARC MONOGRAPHS – 100D 4.63 (2.43–9.12) France-United Exposure of 0–82.7 (6.2) Little et al. All malignant: 0.07 Incidence 12 Kingdom various sites (1998) (< 0–0.62) childhood All benign: > 1000 10 cancer (0.25–> 1 000) Ankylosing X-rays to spine Brain 90% Weiss et al. Spinal cord death Mortality 1 spondylitis range (1994) 3.33 (0.08–18.6) (spinal cord) 0.03–0.40 (0.20) Metropathia X-rays to ovaries 90% range Darby et al. SMR, 1.84 (0.84–3.49) Mortality 9 Dose–response haemorrhagica 0.001–0.004 (1994) not calculated (0.002) IARC Uniform whole 0–> 0.5 Sv Cardis et al. −1.8 (−4.7–1.7)a d Mortality 235 15-country body (0.0194) (2007) nuclear workers United Kingdom Uniform whole 0–> 0.1 Sv Muirhead et −1.36 (−1.85–0.55) Mortality 278 NRRW body (0.0249) al. (2009) −0.88 (−1.56–0.69) Incidence 337 Non- A-bomb Uniform whole 0–4 (0.1) Richardson et 1.12 (0.26–2.51)a Mortality 84 Men, aged 15–64 Hodgkin body, mostly high- al. (2009) yr at exposure lymphoma energy (2–5 MeV) gamma + small amount of high- energy neutrons A-bomb Uniform whole 0–4 Sv (0.1) Preston et al. Combined ERR/Sv, Incidence 170 Stratified linear body, mostly high- (1994) 0.05 (< 0–0.70) RR model fitted energy (2–5 MeV) Men EAR, 0.56 × 10−4/ to publicly gamma + small PY/Sv (0.08–1.39) available data, amount of high- Women EAR, 0 × using bone- energy neutrons 10−4/PY/Sv (< 0–0.28) marrow dose
    • Table 2.8 (continued) Organ site Study Target Dose range Reference ERR/EOR per Gy/ Mortality/ Cases/ Other (mean)(Gy) Sv unless otherwise incidence deaths comments stated (95%CI) (A-bomb, IARC 15-country and NRRW are Sv-1, all others Gy-1) Non- Ankylosing X-rays to spine 90% range Weiss et al. RR, 1.74 (1.23–2.36) Mortality 37 Dose–response Hodgkin spondylitis 1.65–8.41 (1994) not calculated lymphoma (5.10)k because there (contd.) was no clear appropriate organ dose Metropathia X-rays to ovaries 90% range Darby et al. SMR, 0.75 (0.20–1.93) Mortality 4 Dose–response haemorrhagica 1.0–1.6 (1.3) g (1994) not calculated Benign Exposure of pelvic (1.19) Inskip et al. RR (exposed:not), Mortality 53 Dose–response gynaecological area (1993) 0.9 (0.6–1.6)a not calculated disease International Mostly 200– 0–> 12 (7.1) Boice et al. OR, 2.51 (0.8–7.6)a Incidence 94 RR trend not Radiation Study 400 kVp X-ray (1988) computed of Cervical +radium +gamma because of small Cancer Patients to cervix number of non- exposed cases Savannah river Uniform whole 0–> 0.3 Sv Richardson et 7.62 (0.93–20.77)a Mortality 51 site workers body al. (2009) Chernobyl Work in aftermath 0–> 0.5 (0.013) Kesminiene et 28.1 (0.9–243)a Incidence 20 liquidator study of Chernobyl al. (2008) accident IARC Uniform whole 0–> 0.5 Sv Cardis et al. 0.44 (< 0–4.78)a Mortality 248 15-country body (0.0194) (2007) nuclear workers United Kingdom Uniform whole 0–> 0.1 Sv Muirhead et 0.78 (−0.66–3.4) Mortality 237 NRRW body (0.0249) al. (2009) 1.28 (−0.38–4.06) Incidence 305153 X- and γ-radiation
    • 154 Table 2.8 (continued) Organ site Study Target Dose range Reference ERR/EOR per Gy/ Mortality/ Cases/ Other (mean)(Gy) Sv unless otherwise incidence deaths comments stated (95%CI) (A-bomb, IARC 15-country and NRRW are Sv-1, all others Gy-1) Hodgkin A-bomb Uniform whole 0–4 (0.1) Preston et al. 0.48 (< 0–3.96) Incidence 21 Stratified linear disease body, mostly high- (1994) RR model fitted energy (2–5 MeV) to publicly IARC MONOGRAPHS – 100D gamma + small available data, amount of high- using bone- energy neutrons marrow dose Ankylosing X-rays to spine 90% range Weiss et al. RR, 1.65 (0.88–2.81) Mortality 13 Dose–response spondylitis 1.65–8.41 (1994) not calculated (5.10)k Metropathia X-rays to ovaries 90% range Darby et al. SMR, 3.30 (0.90–8.46) Mortality 4 Dose–response haemorrhagica 1.0–1.6 (1.3)g (1994) not calculated Benign Exposure of pelvic (1.19) Inskip et al. RR (exposed:not), Mortality 13 Dose–response gynaecological area (1993) 0.9 (0.3–3.2)a not calculated disease ) International Mostly 200– 0–> 12 (8.2) Boice et al. OR, 0.63 (0.2–2.6)a Incidence 14 RR trend not Radiation Study 400 kVp X-ray (1988) computed of Cervical +radium +gamma because of small Cancer Patients to cervix number of non- exposed cases IARC Uniform whole 0–> 0.5 Sv Cardis et al. −0.18 (< −0.18–7.25)a Mortality 44 15-country body (0.0194) (2007) nuclear workers United Kingdom Uniform whole 0–> 0.1 Sv Muirhead et < −1.93 Mortality 33 NRRW body (0.0249) al. (2009) (< −1.93–32.73) Incidence 67 < −1.93 (< −1.93–12.55)
    • Table 2.8 (continued) Organ site Study Target Dose range Reference ERR/EOR per Gy/ Mortality/ Cases/ Other (mean)(Gy) Sv unless otherwise incidence deaths comments stated (95%CI) (A-bomb, IARC 15-country and NRRW are Sv-1, all others Gy-1) Multiple A-bomb Uniform whole 0–4 (0.1) Preston et al. EAR, 0.08 × 10−4/PY/ Incidence 59 myeloma body, mostly high- (1994) Sv (< 0–0.3) energy (2–5 MeV) gamma + small amount of high- energy neutrons Ankylosing X-rays to spine 90% range Weiss et al. RR, 1.62 (1.07–2.46) Mortality 22 Dose–response spondylitis 1.65–8.41 (1994) not calculated (5.10)k because there was no clear appropriate organ dose Metropathia X-rays to ovaries 90% range Darby et al. SMR, 2.59 (1.19–4.92) Mortality 9 Dose–response haemorrhagica 1.0–1.6 (1.3)g (1994) not calculated Benign Exposure of pelvic (1.19) Inskip et al. RR (exposed:not), Mortality 21 Dose–response gynaecological area (1993) 0.6 (0.3–1.4)a not calculated disease International Mostly 200– 0–> 12 (7.1) Boice et al. RR, 0.26 (0.0–2.6)a Incidence 49 RR trend not Radiation Study 400 kVp X-ray (1988) computed of Cervical +radium +gamma because of small Cancer Patients to cervix number of non- exposed cases IARC Uniform whole 0–> 0.5 Sv Cardis et al. 6.15 (< 0–20.6)a Mortality 83 15-country body (0.0194) (2007) nuclear workers United Kingdom Uniform whole 0–> 0.1 Sv Muirhead et 1.20 (−1.08–7.31) Mortality 113 NRRW body (0.0249) al. (2009) 3.60 (0.43–10.37) Incidence 149155 X- and γ-radiation
    • 156 Table 2.8 (continued) Organ site Study Target Dose range Reference ERR/EOR per Gy/ Mortality/ Cases/ Other (mean)(Gy) Sv unless otherwise incidence deaths comments stated (95%CI) (A-bomb, IARC 15-country and NRRW are Sv-1, all others Gy-1) Chronic Ankylosing X-rays to spine 0–> 7.00 (4.38) Weiss et al. RR, 1.44 (0.62–2.79) Mortality 7 Dose–response lymphocytic spondylitis (1995) not calculated leukaemia IARC MONOGRAPHS – 100D Benign X-rays to spine < 0.2–> 0.5 Damber et al. SIR, 1.07 (0.80–1.41) Mortality 50 Dose–response locomotor and joints (1995) not calculated lesions Benign Exposure of pelvic (1.19) Inskip et al. RR (exposed:not), Mortality 21l Dose–response gynaecological area (1993) 1.1 (0.5–3.0)a l not calculated disease Breast cancer Radiation to chest, (5.3) Curtis et al. RR (exposed:not), Incidence 10 Dose–response supraclavicular (1989) 1.84 (0.5–6.7)a not calculated nodes, axilla, etc. Uterine corpus Radiation to Brachytherapy Curtis et al. RR (exposed:not), Incidence 54 Dose–response cancer vagina, pelvis and 90% range (1994) 0.90 (0.4–1.9) not calculated regional lymph 0.7–2.7 (mean nodes 1.7) External beam 90% range 6.4–14.0 (mean 9.7) (overall mean 5.22) International Mostly 200– 0–> 12 (7.1) Boice et al. OR, 1.03 (0.3–3.9)a Incidence 52 Dose–response Radiation Study 400 kVp X-ray (1988) not calculated of Cervical +radium +gamma Cancer Patients to cervix
    • Table 2.8 (continued) Organ site Study Target Dose range Reference ERR/EOR per Gy/ Mortality/ Cases/ Other (mean)(Gy) Sv unless otherwise incidence deaths comments stated (95%CI) (A-bomb, IARC 15-country and NRRW are Sv-1, all others Gy-1) Chronic Chernobyl Work in aftermath 0–3.22 Romanenko et 4.09 (< 0–14.41) Incidence 39 lymphocytic liquidator study of Chernobyl (0.0764) al. (2008) leukaemia accident (contd.) Chernobyl Work in aftermath 0–> 0.5 (0.013) Kesminiene et 4.7 (–∞–76.1)a f Incidence 21 liquidator study of Chernobyl al. (2008) accident IARC Uniform whole 0–> 0.5 Sv Cardis et al. −1.0 (−5.0–3.7)a d Mortality 47 15-country body (0.0194) (2007) nuclear workers United Kingdom Uniform whole 0–> 0.1 Sv Muirhead et < −1.92 (< −1.92–1.23)a Mortality 69 NRRW body (0.0249) al. (2009) −0.12 (−1.42–2.71)a Incidence 128 a 90%CI b Taken fromUNSCEAR (2008b) c Based on descending & sigmoid colon dose d Computed using a log-linear model (central estimate and confidence bounds given as 10*(RR-1) (RR estimated at 0.1 Sv)) e Sex averaged f Lower confidence bound not determined g Based on total active red bone-marrow dose, using weights to 17 compartments defined byChristy (1981) h Patients receiving less than 100 Gy to uterus were designated as controls I Upper CI computed using a log-linear model j Based on dose on testes k Based on red bone-marrow dose l Chronic lymphocytic leukaemia and lymphocytic leukaemia not otherwise specified (NOS) CNS, central nervous system; SMR, standardized mortality ratio157 X- and γ-radiation
    • IARC MONOGRAPHS – 100D(Preston et al., 2003). For men, there were 2.6.5 Cancer of the liver172 deaths yielding an ERR/Sv of −0.25 (90%CI:<  −0.3–0.15), and for women, there were 198 Among the survivors of the atomic bombings,deaths yielding an ERR/Sv of 0.75 (90%CI: 0.16– liver cancer mortality was clearly associated with1.6). In the analysis of incidence data, a border- radiation dose among men (Preston et al., 2003).line statistically significant dose–response was For men, 722 deaths were reported yieldingreported with an ERR/Sv of 0.19 (90%CI: −0.04– an ERR/Sv of 0.39 (90%CI: 0.11–0.68); and for0.47), based on 838 cases of cancer of the rectum women, 514 deaths yielding an ERR/Sv of 0.35arising evenly between the genders (Preston (90%CI: 0.07–0.72). In the analysis of cancer inci-et al., 2007). There was a highly significant excess dence in the LSS, there were 1494 cases yieldingof cancer of the rectum in the IRSCCP (P = 0.002 a (sex-averaged) ERR/Sv of 0.30 (90%CI: 0.11–for 10-year survivors), yielding an ERR/Gy of 0.55; Preston et al., 2007). [The Working Group0.02 (90%CI: 0.00–0.04) (Boice et al., 1988). There noted that histological confirmation rate ofwas no statistically significant excess risk in the these cancers was low (41%), so it is possible thatUnited Kingdom ankylosing spondylitis data a substantial number were secondary tumours,(Weiss et al., 1994), nor in the IARC 15-country and this might also explain the scatter observedstudy (Cardis et al., 2007). In the latest NRRW in the dose–response.] The dose–response inanalysis (Muirhead et al., 2009), there were the incidence data implies an increase in riskborderline statistically significant elevations of at lower dose, but a reduction above about 2 Sv,ERR in the mortality data (ERR/Sv, 1.69; 95%CI: with a reasonable amount of scatter around the−0.02–4.73), based on 303 deaths, and in the inci- trend line (Preston et al., 2007; Fig.  2.1). Theredence data (ERR/Sv, 1.31; 95%CI: 0.04–3.2), based was little or no evidence of excess in most radio-on 586 cases. Although the confidence intervals therapy studies, e.g. the United Kingdom anky-in the LSS, NRRW and IRSCCP overlap (as they losing spondylitis study of Weiss et al. (1994), thealso do with the other studies), the rather lower metropathia haemorrhagica study (Darby et al.,risks indicated in the LSS compared with the 1994), nor in any occupational studies, e.g. theNRRW, and the even lower risks in the IRSCCP, IARC 15-country study (Cardis et al., 2007) ormight be explained by cell-sterilization effects. the NRRW (Muirhead et al., 2009). However, In summary, there are borderline statisti- the numbers of cases or deaths in all these othercally significant indications of excess risk for this studies is generally small.cancer site in the LSS incidence data (Preston In summary, there is strong and a statisticallyet al., 2007), and for women in the LSS mortality significant excess risk for this cancer site in thedata (Preston et al., 2003). There is a significant LSS incidence and mortality data (Preston et al.,excess risk in the IRSCCP (Boice et al., 1988), but 2003, 2007). However, the shape of the dose–not in other medically exposed groups (Darby response is unusual, and there appears to be a lotet al., 1994; Weiss et al., 1994). There are border- of noise in those data. Possibly the comparativelyline statistically significant indications of excess low percentage of cases that were histologicallyin the NRRW (Muirhead et al., 2009), but not in confirmed in the incidence data might explainthe IARC 15-country study (Cardis et al., 2007). this, and is a cause for concern. There was noWith only a single statistically significant posi- significant excess risk in any other studies (Boicetive study, chance cannot be entirely ruled out as et al., 1988; Darby et al., 1994; Weiss et al., 1994;an explanation for these results. Cardis et al., 2007; Muirhead et al., 2009), but the numbers of cases or deaths is small. With only a single statistically significant positive study,158
    • X- and γ-radiation Fig. 2.1 Liver cancer dose–response in the LSS incidence data (from Preston et al., 2007) Liver Cancer Dose Response 1.5Excess relative risk 1 .5 0 0 1 2 3 4 Weighted liver dose (Gy) The thick solid line is the fitted linear gender-averaged excess relative risk (ERR) dose–response at age 70 after exposure at age 30 based on data in the 0–2-Gy dose range. The points are non-parametric estimates of the ERR in dose categories. The thick dashed line is a non-parametric smooth of the category-specific estimates, and the thin dashed lines are one standard error above and below this smooth. the LSS, chance cannot be entirely ruled out – (Preston et al., 2007). The histological confirma- it is also possible that there is contamination of tion rate of this cancer was low (52%). A statisti- the data for cancer of the liver by that for other cally significant excess risk was reported (ERR/ cancer sites in the LSS. Gy, 0.12; 95%CI: 0.05–0.20, based on 84 cases) in the United Kingdom ankylosing spondylitis data 2.6.6 Cancer of the pancreas (Weiss et al., 1994). There was an indication of excess risk in the Stockholm skin haemangioma Among the survivors of the atomic bombings, study, with nine cases yielding an ERR/Gy of 25.1 pancreatic cancer mortality was not clearly asso- (95%CI: 5.5–57.7; Lundell & Holm, 1995). The ciated with radiation dose (Preston et al., 2003). very large risk predicted by this study is statisti- The ERR/Sv was −0.11 (90%CI: <  −0.3–0.44) cally inconsistent with all the other studies, apart for men, based on 163 deaths, and −0.01 (90%: perhaps from the IARC 15-country study (Cardis −0.28–0.45) for women, based on 244 deaths. et al., 2007), with an ERR/Gy of 2.10 (95%CI: The ERR/Sv for cancer incidence in the LSS was −0.59–6.77), based on 272 cases. In the US peptic 0.26 (90%CI: < −0.07–0.68), based on 512 cases ulcer study of Carr et al. (2002), no excess risk was 159
    • IARC MONOGRAPHS – 100Dreported (ERR/Gy, −0.03; 95%CI: −0.10–0.05, 95%CI: 0.14–0.32), based on 81 cases (Wong et al.,based on 59 deaths). There was also no evidence 1997; risk estimate from UNSCEAR, 2008b); inof excess in the IRSCCP (Boice et al., 1988) and two childhood cancer cohorts of Tucker et al.in the metropathia haemorrhagica study (Darby (1987) (ERR/Gy, 0.06; 95%CI: 0.01–0.2; risk esti-et al., 1994), nor in any occupational study, e.g. mate from UNSCEAR, 2008b), based on 54 cases;the IARC 15-country study (Cardis et al., 2007) in Hawkins et al. (1996) (ERR/Gy, 0.16; 95%CI:or the NRRW (Muirhead et al., 2009). 0.07–0.37; risk estimate from UNSCEAR, 2008b), In summary, there is evidence of an excess risk based on 49 cases; and in the United Kingdomin the United Kingdom ankylosing spondylitis ankylosing spondylitis cohort (RR, 3.29; 95%CI:study (Weiss et al., 1994) and in the Stockholm 1.58–5.92), based on nine deaths (Weiss et al.,haemangioma study (Lundell & Holm, 1995); the 1994). There was no significant excess risk in thelatter was very substantial but based on a small IRSCCP (Boice et al., 1988), nor in various occu-number of cases. However, there is no signifi- pationally exposed groups (Cardis et al., 2007;cant excess risk for this cancer in the LSS inci- Muirhead et al., 2009). In these cohorts, wheredence and mortality data (Preston et al., 2003, data were available (Boice et al., 1988; Weiss2007), nor in the other (radiotherapeutically or et al., 1994; Cardis et al., 2007; Muirhead et al.,occupationally) exposed groups. With only two 2009), the risks for bone and connective tissuestatistically significant positive studies, and one tumours were not markedly different, similar toof these based on a small number of cases that is the findings from the cohort of Japanese A-bombalso inconsistent with most other studies, chance survivors.cannot be entirely ruled out, and coherence is In summary, there is evidence of an excess riskalso not well established. in the LSS incidence data (Preston et al., 2007) and in three other medical radiation cohorts (Tucker2.6.7 Cancers of the bone and connective et al., 1987; Hawkins et al., 1996; Wong et al., tissue 1997). The risks in all cohorts (those with statis- tically significant excess or not) are also reason- This is a rare cancer site. In most studies, ably consistent. There is no evidence that risks forcancers of the bone and connective tissues are bone and connective tissues are dissimilar.analysed together. In most of the cohorts that wereconsidered, bone tumours were outnumbered by 2.6.8 Skin cancers other than basal skinconnective tissue tumours. For example, in the carcinomaUnited Kingdom NRRW, there were 17 bonecancers against 58 connective tissue cancers (a) Squamous cell carcinoma of the skin(Muirhead et al., 2009). [The Working Group Ron et al. (1998) analysed LSS incidence dataanalysed the publicly available LSS incidence data and observed an ERR/Sv of −0.1 (90%CI: < −0.1–set (Preston et al., 2007) using a linear relative 0.10), based on 69 cases (Table 2.8). Updated inci-risk model,and obtained for bone and connective dence data from LSS did not show any significanttissues a statistically significant ERR/Sv of 1.34 association (Preston et al., 2007). Ron et al. (1991)(95%CI: 0.14–3.74), based on 41 cases. The ERR/ observed no cases of squamous cell carcinoma inSv was 1.01 (95%CI: < 0–4.38) for bone tumours, the irradiated Israeli tinea capitis group, and twobased on 18 cases, and 1.76 (95%CI: < 0–6.41) for in the control group. Shore et al. (2002) observedconnective tissues, based on 23 cases (Table 2.8).] seven cases of squamous cell carcinoma in theSignificant excess risks were also reported in a irradiated New York tinea capitis group, andgroup treated for retinoblastoma (ERR/Gy, 0.19; none in the control group.160
    • X- and γ-radiation In summary, for this rarely studied cancer, 2.6.9 Cancer of the uterusthere is essentially only a single quantitativelyinformative study, the LSS incidence data (Ron In the most recent analysis of cancer inci-et al., 1998), which does not indicate an excess dence in the LSS (Preston et al., 2007), 1162risk. Neither of the tinea capitis cohorts (Ron cases were reported yielding an ERR/Sv of 0.10et al., 1991; Shore et al., 2002) are quantitatively (90%CI: −0.09– 0.33). There was a similar (non-informative. significant) risk in the LSS mortality data (ERR/ Sv, 0.17; 90%CI: −0.10–0.52), based on 518 deaths(b) Melanoma (Preston et al., 2003). There are indications in the incidence data that the risks for uterine corpus This is a rare cancer site. In the latest analyses cancer (ERR/Sv, 0.29; 90%CI: −0.14–0.95) isof A-bomb survivors’ data, Preston et al. (2007) greater than for uterine cervix cancer (ERR/Sv,did not analyse this tumour, and the publicly 0.06; 90%CI: −0.14–0.31) [although the uncer-available data were not provided. The much lower tainties are consistent with risks being equalrates of this cancer in the Japanese population for these two cancer sites]. There was little or nothan observed in the western European popula- evidence of an excess in risk of uterine cancer intion (Parkin et al., 2002) imply that even quite most radiotherapy studies, e.g. the metropathialarge ERRs would fail to be statistically signifi- haemorrhagica (Darby et al., 1994), the IRSCCPcant. [The Working Group analysed the older (Boice et al., 1988) or the United Kingdom anky-publicly available LSS data set (with follow-up losing spondylitis study (Weiss et al., 1994),to the end of 1987 rather than the end of 1998) nor in any occupational studies, e.g. the IARCof Thompson et al. (1994). Using a linear relative 15-country study (Cardis et al., 2007) or therisk model, the ERR is not statistically signifi- NRRW (Muirhead et al., 2009). [The occupa-cant (ERR/Sv, 0.22; 95%CI: < 0–4.14), based on tional studies (Cardis et al., 2007; Muirhead13 cases (Table 2.8).] There are few indications of et al., 2009) are particularly uninformative,excess risk in other groups, although a France– for obvious reasons: there were few women inUnited Kingdom childhood cancer study yielded these cohorts, and women tended to have lowera statistically borderline association (excess odds cumulative doses.] In the studies with subtyperatio/Gy, 0.07; 95%CI: 0.00–0.14; Guérin et al., information, the indications, as with the LSS, are2003). There was no significant excess risk in the that ERRs for uterine corpus cancer are greaterNRRW incidence data (ERR/Sv, 1.39; 95%CI: than for uterine cervix cancer (Weiss et al., 1994;−0.65–5.6), based on 261 cases (Muirhead et al., Cardis et al., 2007).2009), nor in the IARC 15-country study (ERR/ In summary, for no cohort are there signifi-Sv, 0.15; 90%CI: <  0–5.44), based on 87 deaths cant excess risks of uterine cancer. In three(Cardis et al., 2007). cohorts with subtype information (Weiss et al., In summary, for this rarely studied cancer, 1994; Cardis et al., 2007; Preston et al., 2007),there are essentially only four quantitatively there were common patterns in risk acrossinformative studies, in none of which are there studies, with greater ERRs for uterine corpusstatistically significant excess risks. The lack of cancer than for uterine cervix cancer. The lackexcess in the LSS is not surprising given the very of excess risks in the two occupational cohortslow rates of this cancer in the Japanese popula- (Cardis et al., 2007; Muirhead et al., 2009) is nottion, even quite large ERRs would fail to be statis- informative, as there were few women in thosetically significant. That said, chance cannot be cohorts, and women tended to have lower cumu-excluded as an explanation of what is reported. lative doses. 161
    • IARC MONOGRAPHS – 100D2.6.10 Cancer of the ovary 2003), were reported in the LLS. In the United Kingdom ankylosing spondylitis data, 88 deaths A borderline significant excess in the inci- were reported yielding a significant ERR/Gy ofdence of cancer of the ovary (ERR/Sv, 0.61; 90%CI: 0.14 (95%CI: 0.02–0.28; Weiss et al., 1994). There0.00–1.5), based on 245 cases (Preston et al., was a non-significant excess of mortality from2007), and a similar excess of mortality (ERR/ cancer of the prostate in occupational studies, e.g.Sv, 0.94; 90%CI: 0.07–2.0), based on 136 deaths the IARC 15-country study (Cardis et al., 2007)(Preston et al., 2003), were reported in the LSS. or the NRRW (Muirhead et al., 2009; Table 2.8).There was little or no evidence of excess in most In summary, the only cohort with significantradiotherapy studies, e.g. the IRSCCP (Boice et al., excess risks of cancer of the prostate is the anky-1988) or the United Kingdom ankylosing spond- losing spondylitis cohort. The risks in the otherylitis study (Weiss et al., 1994), nor in any occu- studies, although not statistically significant, arepational studies, e.g. the IARC 15-country study not incompatible with those in this cohort.(Cardis et al., 2007) or the NRRW (Muirheadet al., 2009; Table 2.8). [The occupational studies(Cardis et al., 2007; Muirhead et al., 2009) are 2.6.12 Cancer of the urinary bladderparticularly uninformative, because there were Significant excess risk for cancer of thefew women in those cohorts, and women tended urinary bladder in the LSS has been reportedto have lower cumulative doses. The lack of in the most recent analysis of cancer incidenceexcess risk in the IRSCCP (Boice et al., 1988) and (ERR/Sv, 1.23; 90%CI: 0.59–2.1; Preston et al.,metropathia haemorrhagica (Darby et al., 1994) 2007) and of mortality with an ERR/Sv of 1.1studies may partly be explained by very large (90%CI: 0.2–2.5) for men and 1.2 (90%CI: 0.10–doses to the ovaries, well into the range at which 3.1) for women (Preston et al., 2003). Significantcell sterilization might occur.] excess risks were also reported from the United In summary, the only cohort with significant Kingdom ankylosing spondylitis data (ERR/excess risks of ovarian cancer is the LSS. The lack Gy, 0.24; 95%CI: 0.09–0.41), based on 71 deathsof excess risks in the other studies, in particular (Weiss et al., 1994), and the IRSCCP study (ERR/the two occupational cohorts (Cardis et al., 2007; Gy, 0.07; 90%CI: 0.02–0.17), based on 273 casesMuirhead et al., 2009), and the IRSCCP (Boice (Boice et al., 1988). [The Working Group notedet al., 1988) and metropathia haemorrhagica that although the risk estimated in the last two(Darby et al., 1994) studies may not be informa- cohorts are lower than those in the LSS, celltive, because of the low number of women, who sterilization resulting from the somewhat higherusually had low cumulative doses, in occupa- average doses might explain this difference.]tional cohorts and potential cell sterilization in The metropathia haemorrhagica study (Darbymedical radiation cohorts. et al., 1994) suggests quite high risks (SMR, 3.01; 95%CI: 1.84–4.64) based on 20 deaths (average2.6.11 Cancer of the prostate dose, 5.2 Gy), and the ERR/Gy was 0.40 (95%CI: 0.15–0.66). There was no significant excess in any A non-significant excess of incidence of occupational study, e.g. the IARC 15-countrycancer of the prostate (ERR/Sv, 0.11; 90%CI: study (Cardis et al., 2007) or the NRRW−0.10–0.54), based on 387 cases (Preston et al., (Muirhead et al., 2009; Table 2.8).2007), and a similar excess (also lacking statistical In summary, there is strong evidence of excesssignificance) of mortality (ERR/Sv, 0.21; 90%CI: risk in the LSS incidence and mortality data< −0.3–0.96), based on 104 deaths (Preston et al., (Preston et al., 2003, 2007), and in three other162
    • X- and γ-radiationmedical radiation cohorts (Boice et al., 1988; 2.6.14 Cancer of the brain and centralDarby et al., 1994; Weiss et al., 1994). The risks nervous systemin all cohorts (those with statistically significantexcess or not) are all reasonably consistent. In the most recent analysis of cancer inci- dence in the LSS (Preston et al., 2007), there were 281 cases resulting in a significant ERR/Sv2.6.13 Cancer of the kidney of 0.62 (90%CI: 0.21–1.2). In the LSS mortality Preston et al. (2007) analysed renal cell carci- analysis, there were very large and significantnomas (comprising 68% of the kidney cancers) in excess risks for men (ERR/Sv, 5.3; 90%CI: 1.4–16)the LSS incidence data set, and obtained a non- based on 14 deaths (Preston et al., 2003). Forsignificant ERR/Sv of 0.13 (90%CI: −0.25–0.75), women, there were 17 deaths yielding a morebased on 167 cases (Table  2.8). However, there modest ERR/Sv of 0.51 (90%CI: < −0.3–3.9). Inwere indications that ERR significantly decreased the New York tinea capitis study, there was alsowith either increasing age at exposure (P = 0.005) a significant association (ERR/Gy, 1.1; 95%CI:or with increasing attained age (P < 0.001). For 0.1–2.8), based on seven cases (Shore et al., 2003).this reason Preston et al. (2007) also fitted an In the Israeli tinea capitis study, there were alsoabsolute risk model, yielding a statistically signif- significantly raised risks of both malignant brainicant dose–response EAR of 0.25×10-4 person– tumours (ERR/Gy, 1.98; 95%CI: 0.73–4.69; basedyear Sv (90%CI: 0.07–0.53). There were similar, on 44 cases) and benign meningiomas (ERR/although non-significant, excess risks in the most Gy, 4.63; 95%CI: 2.43–9.12; based on 81 cases),recent LSS analysis of mortality (Preston et al., with a stronger increase in risk for benign brain2003)—for men, there were 36 deaths resulting tumours (Sadetzki et al., 2005). A similar patternin an ERR/Sv of −0.02 (90%CI: < −0.3–1.1), and of risks was seen in the France–United Kingdomfor women, there were 31 deaths and an ERR/ childhood cancer study; the ERR/Gy was 0.07Sv of 0.97 (90%CI: <  −0.3–3.8). In the United (95%CI: < 0–0.62) based on 12 cases for malig-Kingdom ankylosing spondylitis data, there were nant lesions, and > 1000 (95%CI: 0.25– > 1000)35 deaths yielding a significant ERR/Gy of 0.10 based on ten cases for benign lesions; (Little(95%CI: 0.02–0.20) (Weiss et al., 1994). There is et al., 1998). In the United Kingdom ankylosingalso a significant excess in the IRSCCP (Boice spondylitis data, there was one spinal cord deathet al., 1988); 148 cases resulting in a significant resulting in a significant ERR/Gy of 3.33 (95%CI:ERR/Gy of 0.71 (90%CI: 0.03–2.24). There was no 0.08–18.6; Weiss et al., 1994). There is no signifi-significant excess in any occupational study, e.g. cant excess in any occupational study, e.g. thethe IARC 15-country study (Cardis et al., 2007) IARC 15-country study (Cardis et al., 2007) oror the NRRW (Muirhead et al., 2009; Table 2.8). the NRRW (Muirhead et al., 2009; Table 2.8). In summary, there is evidence of excess risk In summary, there is evidence of significantin the LSS incidence data (Preston et al., 2007) excess brain and central nervous system tumourand in two other medical radiation cohorts risk in the LSS incidence data (Preston et al.,(Boice et al., 1988; Weiss et al., 1994). The risks 2007), in two tinea capitis cohorts (Shore et al.,in all cohorts (those with statistically significant 2003; Sadetzki et al., 2005), in an ankylosingexcess or not) are all reasonably consistent. spondylitis cohort (Weiss et al., 1994) and in the France–United Kingdom childhood cancer study (Weiss et al., 1994). A similar pattern of excess risk being higher for benign tumours than for malignant is in the Israeli tinea capitis and 163
    • IARC MONOGRAPHS – 100DFrance–United Kingdom cohorts. The risks in all cohort of Chernobyl liquidators (Kesminienecohorts (those with statistically significant excess et al., 2008) and in the Savannah River Siteor not) are all reasonably consistent. workers (Richardson et al., 2009).2.6.15 Non-Hodgkin lymphoma 2.6.16 Hodgkin disease In the analysis of haematological malig- Preston et al. (1994) in the LSS did not analysenancy incidence in the LSS cohort (Preston this tumour. [The Working Group analysed theet al., 1994), there was a borderline significant publicly available data set using a linear relativeEAR of 0.56×10-4 /person–years /Sv (90%CI: risk model, and obtained a non-significant ERR/0.08–1.39) for men, but this was not true for Sv of 0.48 (95%CI: < 0–3.96), based on 21 caseswomen (EARx10-4/person–year /Sv, 0; 90%CI: (Table 2.8).] In the United Kingdom ankylosing<  0–0.28). [Fitting a simple linear relative risk spondylitis data, there were 13 deaths yieldingmodel, overall there was no significant excess risk a non-significant relative risk of 1.65 (95%CI:(ERR/Sv, 0.05; 90%CI: < 0–0.70).] These incident 0.88–2.81; Weiss et al., 1994); no dose–responsefindings are consistent with the analysis of male analysis was reported. There was no significantadult LSS mortality data, with a reported ERR/ excess in the IRSCCP (Boice et al., 1988), in theSv of 1.12 (90%CI: 0.26–2.51) based on 84 cases metropathia haemorrhagica cohort (Darby et al.,(Richardson et al., 2009). In the United Kingdom 1994), in a group treated for benign gynaeco-ankylosing spondylitis cohort, there were 37 logical disease (Inskip et al., 1993), in the IARCdeaths yielding a significant relative risk of 1.74 15-country study (Cardis et al., 2007), or in the(95%CI: 1.23–2.36; Weiss et al., 1994); there was NRRW (Muirhead et al., 2009; Table  2.8). [Theno dose–response analysis in this cohort. There Working Group noted that a common featurewas no significant excess risk in the IRSCCP of all the cohorts is the small number of cases,(Boice et al., 1988), in the metropathia haemor- so that large ERRs would be required to detect arhagica cohort (Darby et al., 1994), or in a group significant excess in these groups.]treated for benign gynaecological disease (Inskip In summary, there are no cohorts with signif-et al., 1993; Table  2.8). Among occupational icant excess risks for Hodgkin disease. However,studies, there was a very large excess risk in a the small number of cases in all groups meancohort of Chernobyl liquidators (ERR/Gy, 28.1; that a large ERR would be required to detect90%CI: 0.9–243) based on 20 cases (Kesminiene significant excess risks.et al., 2008), and in the cohort of Savannah RiverSite workers (ERR/Gy, 7.62; 90%CI: 0.93–20.77) 2.6.17 Multiple myelomabased on 51 cases (Richardson et al., 2009).However, there was no significant excess risk In the most recent analysis of haematologicalin the IARC 15-country study (ERR/Sv, 0.44; malignancy incidence in the LSS, Preston et al.90%CI: <  0–4.78) based on 248 deaths (Cardis (1994) used an absolute risk model and obtainedet al., 2007), or in the NRRW cohort (ERR/Sv, a non-significant EAR (EAR/104 person–year1.28; 95%CI: −0.38–4.06) based on 305 cases Sv, 0.08; 95%CI: <  0–0.3), based on 59 cases(Muirhead et al., 2009). (Table 2.8). In the United Kingdom ankylosing In summary, there is evidence of a signifi- spondylitis data, there were 22 deaths yielding acant excess risk of non-Hodgkin lymphoma in borderline significant relative risk of 1.62 (95%CI:men (but not women) in the LSS mortality and 1.07–2.46; Weiss et al., 1994); there was no dose–incidence data (Preston et al., 2003, 2007), in a response analysis in this cohort due to a lack of164
    • X- and γ-radiationappropriate organ dose. There was a significant et al., 2007). In the two Chernobyl liquidatorexcess risk in the metropathia haemorrhagica studies (Kesminiene et al., 2008; Romanenkocohort with an SMR of 2.59 (95%CI: 1.19–4.92), et al., 2008), the risks are both large and positive,based on nine deaths (Darby et al., 1994). There although in neither case conventionally statisti-was no significant excess risk in the IRSCCP cally significant. For example, the ERR/Gy in the(Boice et al., 1988), and in a group treated for study by Kesminiene et al. (2008) was 4.7 (90%CI:benign gynaecological disease (Inskip et al., –∞–76.1), based on 21 cases (Table 2.8). In medi-1993). There was also no excess risk in the IARC cally exposed groups, there was no indication of15-country study (Cardis et al., 2007). There was excess risk in the benign gynaecological diseasea highly significant excess in the incidence of cohort of Inskip et al. (1993) (RR, 1.1; 90%CI:multiple myeloma in the NRRW (ERR/Sv, 3.60; 0.5–3.0; based on 21 deaths), in a group irradi-95%CI: 0.43–10.37), based on 149 cases; and ated for benign locomotor lesions (SIR, 1.07,there was an excess of much smaller size (which 90%CI: 0.80–1.41; based on 50 deaths; Damberwas non-significant) for mortality in that cohort et al., 1995), in the IRSCCP (OR, 1.03, 90%CI:(ERR/Sv, 1.20; 95%CI: −1.08–7.31), based on 113 0.3–3.9; based on 52 cases; Boice et al., 1988),deaths (Muirhead et al., 2009; Table 2.8). and in many other medically irradiated groups In summary, there is no evidence of an excess (Curtis et al., 1989, 1994, Weiss et al., 1994).risk of multiple myeloma in the LSS incidence In summary, there is remarkably littledata (Preston et al., 1994), although an excess evidence of a significant excess risk of chronicrisk has been reported from the NRRW study lymphocytic leukaemia in a large number of(only incidence and not mortality; Muirhead studies.et al., 2009), and also from the ankylosing spond-ylitis study (Weiss et al., 1994) and from the 2.6.19 Exposure in uterometropathia haemorrhagica study (though basedonly on analysis of SMR) (Darby et al.,1994). Preston et al. (2008) reported statistically significant dose-related increases in incidence rates of solid cancers among A-bomb survivors2.6.18 Chronic lymphocytic leukaemia exposed to radiation in utero (see Section 2.1.3). Most of the information on this tumour Excess cancer risk associated with diagnosticcomes from occupationally and medically X-ray exposure was reported in the Oxford Surveyexposed groups. There are very few chronic of Childhood Cancers (Bithell & Stewart, 1975),lymphocytic leukaemias in the LSS cohort – and in various other groups exposed in uteroonly four were documented in the latest reported (Stewart et al., 1958; Monson & MacMahon, 1984;analysis of haematological malignancy incidence Harvey et al., 1985). However, the interpretation(Preston et al., 1994). In general, this is a much of these in-utero studies remains controversialless common tumour in the Japanese population (Boice & Miller, 1999; ICRP, 2003), in partic-than in the European population. In all occupa- ular because the risk for most childhood solidtional cohorts, there were no significant excess. tumour types is increased, at about 40%, by theFor example, the ERR/Sv for the incidence of same magnitude as that for childhood leukaemiachronic lymphocytic leukaemia in the NRRW (see Table  2.9 available at http://monographs.was −0.12 (90%CI: −1.42–2.71), based on 128 cases iarc.fr/ENG/Monographs/vol100D/100D-02-(Muirhead et al., 2009; Table 2.8). In the IARC Table2.9.pdf), implying a possible bias. However,15-country study, there was an ERR/Sv of −1.0 eight cancers among those exposed in childhood(95%CI: −5.0–3.7), based on 47 deaths (Cardis and in utero in the Japanese A-bomb survivors 165
    • IARC MONOGRAPHS – 100Ddeveloped in adolescence (ages 14–19 years), and Survey of Childhood Cancers. Although therewere of various types (Preston et al., 2008). The are no leukaemia cases in the Japanese in-uteroseven of this group that developed after child- cohort in childhood (there were two cases athood exposure included tumours of the stomach, a later age, Yoshimoto et al., 1994), the lack ofbone, soft tissue, skin, thyroid and two tumours excess is nevertheless consistent with the excessof the central nervous system. The single tumour risk observed in the Oxford Survey of Childhoodin this age group that developed after in-utero Cancers, and in other in-utero medically irradi-exposure was a Wilms tumour diagnosed at the ated groups (Wakeford & Little, 2003). The lackage of 14 years (Preston et al., 2008). This spec- of cases among the Japanese and possible incon-trum of tumours after early childhood exposure sistency with some of other groups may alsosuggests that the lack of specificity in the spec- be plausibly accounted for by cell-sterilizationtrum of tumours in the in-utero medical expo- effect (Little, 2008). The fact that risk reducessure cohorts is not necessarily remarkable. with calendar time, almost exactly paralleling It has been suggested that the general eleva- the reduction in in-utero dose (see Fig. 2.2 andtion in risk of most cancer types in the in-utero 2.3) substantially increases the plausibility of themedically exposed groups is related to recall bias observed association in the medical groups (Dollor confounding, possibly by some factors oper- & Wakeford, 1997; Wakeford & Little, 2003), asating in pregnancy that had given rise to the does the dose–response relationship observedneed for radiograp