Tobacco Documents Online

I r : I , r P.N. Lee Environmental Tobacco SmOke and Mortall"ty A Detailed' Review of Epidemioiogicat Evidence Relating Environmental Tobacco Smoke to the Risk of Cancer, Heart Disease and Other Guses of Death in Adults Who Have Never Smoked 101 tables, 1992 K/~\RGER Easel • Munehen • Paris • l ondon • NeW York • Nrr Delhi • Banskok • Sinpporo -Tokyo • Sydne) 2023512110 ®

Library of Congress Cataloging-in•Pubhcation Data Lec. Pcter N., 1943- Environmental tobacw smoke and morulny; a detailed reviovu or epidemiological evidence rolating environmental tobacco smoke to the nsk of cancer, heart disease- and other wuses of' death in ad'ults who have nevcr smoked'J P.N. Let. Includes bibliograpfiicnl references and index 1I Passive smoking - Health upects., 2. Tobacaosmoke pollution - Health aspects. 3. Environmentally induced diseases-Tozioology. L Tttle. (DNLM: 1. Cause of Dcath, 2~: Hean Diseases - epidcmiolog. 3. Heart Diseases - mortality, f. Neoplasms-epidemiolbgy. S. Neoplasms-mortahq. 6. TobaccoSmoke Pollution -adverx effects: w'A 754 U7g5eji RA12t2 T6L4, 1992 615.9'52379 -dc20 ISBN 3-d055-5529-6 All rights rsserved.. No pan of this publiation may be tnnslated into othenlanguaaesr reproduced or utilized in any form or by any means, eknronic or mechanicaL incJudrng photowpyirrt. eeoording, miesocopying. or by any information storage and retrieval system, without permission ie writing from the pubhshcr, O Copyright 1992 by S. Karger AG. P.O. Bo:.,CH-s009 Base1 (Switzerland) Printed in Switzerland on acid-free paper, by Thur AG OfRctdrudc, Praaela ISBN 3-4055-5529-1>

I I I h I i Contents lndez to Tables ....................................... X . Preface ........................................... XV Acknowledgemenu ..................................... XX 1 I ntroduction 1.1 Background and Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I 1.2 Terms of Reference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.3 Structure of This Book . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.4 Problems of Inference from Epidemiological Studies . . . . . . . . . . . . . . . . . 4 1.5 Dosimetri.c Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.6 Misclassification of Active Smoking Status . . . . . . . . . . . . . . . . . . . . . . . 7 2 The 53 Studies 2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . 9 2.2 Prospective Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . ., . . . . . . . . 9 2?:1 Amerinn Cancer Society Million Person Study - Garfinkel I . . . . . . ., . . . . . 9 2.2.2 !!apanese Study - Hirayama . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. 2:2:3 Wcst ofScotland Study - Hole . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . l5 2.2.4 Alameda County Study - Reynolds . . . . . . . . . . . . . . . . . . . . . . . . . . . 1!8 2.2.5 Rancho Bernardo Study - Garland . . . . . . . . . . ., . . . . . . . . . . . . . . . . 19 ' 2.2.6 Multipte Risk Factor, Intervention Trial - Svendsen ., . . . . . . . . . . . . . . . . 20 2.2.7 Washington County Study - Sandier 11 . ., ., . . . . . . . . ., . . . . . . . . . . . . . 22' 2.2.8 Amsterdam Study- Vandenbroucke . ., . . ., ., ., . ., . . . ., ., . . . . ., . ., . . . . . 25 2.2.9 Georgia Study - Humble II . ., . . . . . . . . . . ., . . . . . . . . . . . . . . . . . . • 25 2.2.10 Californian Seventh-Day Adventists Study - Butler . . . . . . . . . . . . . . . . . . 26 2.3 Gse-Control Studies of Mor¢ than One Disease . . . . . . . . . . . . . . . . . . . ., 27 2.3.1 United' Kingdom Study - Lee . . . . . . . . . . . . . . . . . . . . . . . . . . . . ., . 27 2.3.2 North Carolina Study - Sandler I . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 2.3.3 Erie County Study of Cancer of All Sites - Miller . . . . . . . . . . . . . . . . . . . 32 2:4 Case-Control Studies of Lung Cancer . . . . . . . . . . . . . . . . . . . . . . . . . . 34 2:A:1 Japanese Atomic Bomb Survivors Study- Akiba . . . . . . . . . . . . . . . . . . . 34 2:4.2 Southern Swedish Study - Axelson . . . . . . . . . . . . . . . . . . . . . • . . . . . 35

2.4.3 Colondo Studv of Adenocarcinoma - Brownson . . . . . . . . . . . . . . . . . . . 36 2.4.4 Texas Study - Buffier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 2.4.5 First Hong Kong Study - Chan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 2.4.6 Louisiana Study - Corrca . . . . . . . . . . . . . ., . . . ., . . . . . . . . . . . . . . . U 2.4.7 Shanghai Study- Gao . . . . . ., . ., . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 2.4.8 New JeneylOhio Study - Garfinkel II . . . . . . . . . . . . . . . . . . . . . . . . . . 40 2.4.9 Tianjin Study - Geng . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 2.4.10 New Mexico Study - Humble I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 2.4.11 Kanagawa Study - I'noue . . . . . . . . . . ... . . . . . . . . . . . . . . . . . . . . . 45 2.4.12 First New York Study - Kabat I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 2.4.13, Third New York Studv - Kabat Il l . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 2.4.14 Socond Athens Study - Kalandidi IJ . . .. . . . . . . . . .. . . . . . . . . . . . . . . 49 2:4.1!5 Nara Study - Katada . . . . . . . . .. .. . . . . . . . . . . . . . . . . . . . . . . . . . S I' 2:4'.16 German Studv - Knoth ~ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 2.4:17 Second Hong Kong Study - Koo . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 53 2.4.18 Third Hong Kong Study - Lam 1 . . . . . . . . . . . . . . . . . . . . . . . . . . .. .. 55 2.4.19 Fourth Hong Kong Study - Lam 11 . . . . . . . . . . . . . . . . . . . . . .. . . . . . 56 2.4.20 Shanghai Study, - Li . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 2.4.21 Armadale Study - Uoyd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 2.4.22 Swedish Study - Pershagen . . . . . . . . . . .. . .. .. . . .. . .. . . . . . . . . . . . . . 58 2.4.23 Nagoya Study - Shimizu . ., . . ., . ., . . ., ., ., . . . . . . ., . . . . . . . . . - . . . . 59 2:4.24 Osaka Study - Sobue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 2.4.25 Stockholm Study - Svensson . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 2.4.26 Athens Study - Triehopoulos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 63 2.4.27 Second New York StudN - Varcla . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 2.4.28 Hyrbin Study of Adenocarcinoma - Wang . . . . . . . . . . . . . . . . . .. . . . . . 68 2.4.29 Los Angeles Study - W,w . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 2.4.30 North-East China Study - Wu-Wiiliarns . . . . . . . . .. . . . . . . . . . . . . . . . 69 2.4.31 New Jersey Study - Ziegler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 2.5 Case-Control Studies of Bladder Cancer . . . . . . . . . . . . . . . . . . . . . . . . 73 2.5,1 Canadian Study - Burch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 2.5:2 US Study - Kabat 11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 2.6 Casc-ControliStudies of Cervix Cancer . . . . . . . . . . . . . . . . . . . . . . . . . 75 2.6.1 Gothenburg Study - Hellberg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 2.6.2 Utah Study - Slattery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 2:7 Case-Control Studies of Heart' Disease . . . . . . . . . . . . . . . . . . . . . . . . . 77 2.7.1 Chinese Study - He . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 2.7:2 Utah Study- Martin . . . . . . . . . . . . . . . . ., . . ., ., . . . . . . . . ... ..... . 79 2.7.3 US Study- Palmer . . . . . . . . . . . . . . . . . . . . . . . . . . . . ., . ., . . . . . 79 2.8 Case-Control Study of Chronic Obstructive Pulmonary Disease . . . . . . . . . . . 801 2.8.1 Greek Study - Kalandidi I . . . . . . . . . . . . ., . . . ., . . . . . . . . . . . . . . . 80 , 2.9 Case{ontrol Study of Stroke ............................... , , 82 2.9.1 Australian Study - Donnan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

I I ii Contents 3 Lung Cancer VII 3.1 Seleaion of Studirs for Detailed Consideration .................... $3 3.2 Features of the 29 Studies Included . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 3.2:1 Study Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 . 3.2.2 (i.ocation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 3:2.3 Time of Death/Diagnosis . . . . . . . . . . . . . . . . . . . . . . . . . . ., . . . . . $4 3:2.4 The Cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 3.2.5 The Controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ., ., . . . . . . . 88 3.2.6 Sources or Information . . ., ., . . . . . . . . . . . ., . . . ., . . . . . . . . . . . . . . 93 3.2.7 Non-Response . . . . . . . . . . . . . . . . ., . . . . . . . . . . . . . . . . . . . . . 95 3.2.8' indices of ETS Exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 3.2.9 Smoking Status of the Subjects . . . . . . . . . . . . i . . . . . . . . . . . . . . . . . 98 3.2.10 Confounding Variablt:s . . . . . ., . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 3.3 Summary of Results from the 29 Studies . . . . . . . . . . . . . . . . . . . . . . . . 101 3.3.1 Main Index - Spouse Smoking or Nearest Equivalent . . . . . . . . . .. . .. . .. . . 101 3.3.2 Histological: Type of Lung Cancer . . . . . . . . . . . . . . . . . . . . . . . . . . ., . 109 3.3.3 l.eveUof Exposure ., . . . . ... . . . . . . . . . . . . . . . . . . . I1 l, 3.3.4 Variation in Relative Risk According to Other Factors . . . . . . . . . . . . . . . . 113 3.3.5 Other Indices of ETS Exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 3.3,6 'Attributable Risk' from ETS Exposure . . . . . . . . . . . . . . . . . . . . . . . . . 119 3.3.7 Comparability of Associations with ETS Exposure and Active Smoking ...... 123 3.4 bnter9reution of the Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 3:4.1 The Main Findings . . . . . . . . . . . . . . . . . . . . . . ., . . . . 1,24 3.4.2 Plausibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 3.4.3 Explanations to Be Considered . . . . . . . . . . . . . . . . . . ., . . . . . . . . . . . 127 3.4.4 Misclassification of Diagnosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 3.4.5 Misciassifintion of ETS Ezposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 3.4.6 Lack of Comparability of Cases and Controls . . . . . . . . . . . . . . . . . . . . . 130 3.4.7 Other Study Weaknesses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 3:4: 8 Confounding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 3.4'.8.1' Age . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ., 133 3.4.8.2 Marital Status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ., ., . . . . . . 134 3.4.8.3 Other Confounding Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 3.4.8.4 Conclusions Regarding Confounding . . . . . . . . . . . . . . . . . . . . . . . . . . 142 3.4.9 Misclassification of Active Smoking Status . . . . . . . . . . . . . . . . . . . . . . . 142 3.4.9.1 The Problem . . . . . . . . . . . . . . ... . . .. . . . . . . . . . . . . . . .. . ., . . . . 142 3.4.9.2 Major and Minor Sourtxs of Bias . . . . . . . . . . . . . . . . . ., ., . . . . . . . . . 143 3.4.9.3 What I,eve1 of Misclassification Would Be Needed to Enplain~the Observed Associations of Spouse Smoking with Lung Canoer7' . . . . . . . . . . . . . . . . . 146 3.4.9.4 Evidence on Extent of Misclassification . . . . . . . . . . . . . . . . . . . . . . . ., ., 150 3.4.9.5 Evidence on Extent of Smoking Habit Concordance . . . . . . . . . . .. . .. . . . . 158 3.4.9.6 Conclusions Regarding Misclassi6aation . . . . . . . . . . . . . . . . . . . . . . . . 162 3.4.10 Publintion Bias and Other Issues in Meta-Analysis . . . . . . . . . . . . . . . . . . 164 3.5 Overall Assessment of the Evidence on ETS and Lung Caneer; ........... 166 202351209:114

V1I1 Contents 4 Other Cancers 4.1 Studies Providing Data .................................. . 168 4.2 Results ., . . . . . . . . . . ., . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168 4.2.1 Weaknesses of Some Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168 4.2.2' Ovcrall Risk . . . . . . . . . . . . . . . . . . . . . . . . . . ., . . . . . . . . . . . . . 169 4.2.3 Lip: Oral Cavity and Pharynz . ., . . . . . . . . . . . . ., . . . . . . . . . . . . . . . 172 4.2.4 Digestive System . . . . . . . . ., . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 4.2.5 Respiratory Sites Other Than the Lung . . . . . . . . . . . . . . . . . . . . . . . . . 174 4.2.6 Bone, Connective Tissue and Skin . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174 4.2.7 Breast . . . . . ., . . . . . ., . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . 175 4.2.8 Female Genital System. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176 4.2.9 Male Genital System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178 4.2.10 Unnary Tract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .~ . . . . . . II79 4.211 Eyc, Brain and Nervous System . . . . . . . . . . . . . . . . . . . . . . . . . . . . ., 179 4.2.12 Endocrine Glands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ., . . . 180 4.2.13' Lymphatic and Haematopoictic Tissue . . . . . . . . . . . . . . . . . . . . . . . . . 180 4.2.14: Other Cancer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181 4.3 Discussion and Conclusions . . . ., . . . ., . . . . . . . . . . . . . . . . . . . . . . . 181 5 Heart Disease 5,1 Studies Providing Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184 5.2 Features of the Studies Included . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184 5.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186 5.4 Discussion and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . ., . . 193 6 Other Diseases 6.1 Studies Providing Data . . . . . . . . . . ., . . ., . ., . ., . . . . . . . . . . . . . . . . 197' 6.2 Results . . . . . . . . . . . . ., ., . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197 6.2.1 Tuberculosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ., . . . . . . . . . . 197 6.2.2 Diabetes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ., . . . . . . . . . 197 6.2.3 Cerebrovascular Disease . . . . . . . . . . . . . . . . . . . . ., . . . . . . . . . . . . 198 6.2.4 Respiratory Diseases . . . . . . . . . . . . . . . . . . . . ., ., . . . . . . . . . . . . . 199 6,2.5 Diseases of the Digestive System . . . . . . . . . ., . . . . . . . . . . . . . . . . . . 200 6.2.6 Gastro-Iotestinal1 Ulcer . . ., ., . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200 6.2.7 Cirrbosis of the Liver . . . ., ., . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201 6.2.8 Diseases of the Genito-Urinary System . . . . . . . . . . . . . . . . . . . . . . . . . 201 6.2.9 Accidental Causes of Death . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201 6.2.10 Suicide . . . . ., . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201. 6.2.11 Ot her Diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201 6.3 Discussion and Conclusions . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202

Contents 7 Overall Risk of Mortality Ix 7.1 Studies Providing Data .................................. 203 7.2 Results ........................................... , 203 7.3 Discussion and Conclusions . . . . . . . . . . . . . . . . . . . . • • . • . . • • . . . 201 8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210 Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21'8' h.

-1 ) Heart Disease 5.1 Studies Providing Data i Eleven studies have provide& information on the relationship ofl ETS to risk of cardiovascular disease. Seven of these are prospective studies, 5 eonducted in the USA (Butler, Garland, Humble 11, Sandler 11, Svendsen), I in Japan (Hirayama) and I in Scotland (Hole). Four are case-control studies, 2 in the USA (Martin, Palmer), I in China (He) and I in England (Lee). Three prospective studies which have reported results for lung cancer (Garfinkel I), totat cancer (Reynolds) or all-cause mortality (Vanden- broucke) have not reported results for heart disease. While the numbers of deaths would not be substantial for the last 2 studies. the fact that the first, the American Cancer Society million person study„ has not provided information is a wastage of resource. This study alone would certainly have had data on more deaths/cases than all! the i l published studies combined! 5:2 Eeatures of the Studies Included It should be noted that very little information is available for 3 of the 1 L studies considered: the prospective study of Butler, and the case-control studies of Martin and Palmer for which the only published data consist of abstracts. The number of deaths/cases in some studies is verv small. There are 4 studies with extremely small numbers (Svendsen, 13; Garland, 19; Martin, 23; He, 34)„4 with quite small numbers (Humble II, 76; Butler, 80; Hole, 84; Lee, l18), and 1' (Paltner, 336 in ever smokers and never smokers combined) which, though of moderate size, has not presented findings in a form~to allow proper evaluation. Considering the prevalence of heart dis-

5 Hean Dssease 185 ease among never smokers - very much higher than from lung cancer - it is surprising that by now there are only 2 published' studies for reporting results based on reasonably substantial numbers of deaths/cases (Hiraya- ma, 494; Sandier It„ 1,358). Of the 7 prospective studies, 4 involved study subjects attending for an examination during which blood pressure, cholesterol and body mass index were measured. It is unfortunate that neither of the 2 substantial studies collected information on these classical risk factors for heart dis- ease. The factors used by the authors for adjustment - occupation by Hirayama„and schooling; housing quality, and marital status by Sandler I1 - are not those which firstoccur as being most relevant in a study of heart disease. Some of the studies have problems regarding representativeness of the subjects. Thus„ in both the Garland and Hole studies, about 20% of the population did not attend for examination, with a possibility of bias if failure to attend was associated both with ETS exposure and risk of death from heart disease. ln the large Sandier lt study, only deaths in Washing- ton County were recorded, again imparting a danger of bias if ETS is associated with the chance of migration out of the county. The Svendsen study, was based on the well-known Multiple Risk Factor Intervention Trial, which involved people at very high risk of heart disease base& on their smoking, blood pressure and cholesterol levels. Since the paper con- cerned never smokers, all the subjects involwed most probably exhibited abnormally high blood pressure and'/or cholesterol levels. The Butler study involved Seventh-Day Adventists, an atypical population with regard to many variables. The 2 studies which provided data on by far the largest number of deaths are both open to criticism„as detailed in sections 2.2.2 (Hirayama) and 2.2.7 (Sandler 1~1), and also in section 4.2:1. It is interesting to note that both studies have publishe& inconsistent results for women. In 1981, Hirayama presented results showing no association of heart disease with husband smoking, based on follow-up of his population to 1979. In 1984, he reported results which showed a significant association, based on fol, low-up of his population to 1981. These results implied an implausibly strong relationship of heart disease to smoking by the husband when deaths occurring in 1979-81 were considered (a fact pointed out by Lre in correspondence in the New Zealand MedicalJoarnal [26, 27J). As a result, Hirayama published revised figures for follow-up to 1979, indicating that the data published in 1981 were incorrect.

166 Environmental,TobacooSmoke and Morulitj Helsing et al. [40]I and Sandier et al. [43] have both presented (in Table 4 of each paper) results for risk of arteriosclerotie heart disease in relation to ETS exposure. Based' on ddentical numbers of deaths an&stated adjustment factors, the reported relative risks and confidence limits for men were identical. Inexplicably, however, this was not the case for wom- en, where relative risks and' confidence limits both varied. Given the rela- tively larger contribution of the Hirayama an& Sandler II results to the overall number of heart disease cases studied4 such~diffcrenccs in reported findings are rather disconcerting. 53 Results Table 5.1 summarizes results for exposure to ETS from the spouse or in the household. Before considering the findings, some points are worth not- ing: (i) As far as can be ascertained, alli the relative risks are for never smok- ers. (ii) The index was based on smoking by the spouse in 7 studies (Butler, Garland; He, Hirayama, Humble il, Lee, Svendsen), only married women being considered, except perhaps in the Chinese study where single women were included with the wives of non-smoking husbands. In 2 studies (Tvlartin, Palmer) thc index used is not known. The Hole study compared people living at the same address as a study partici- pant who had ever smoked with people who lived arthe same address as a study participant who had never smokedl with no other studyy participant at that address ever having smoked. The Sandler 11, , stud. used a complex index of exposure, but for the results in Table 5.1 it amounted to a comparison of people living in the household where some adult had ever smoked with people in a household' where no adult had' ever smoked. (iii) When a study has presented different findings at different time points for apparently the same comparison, the later publication has been used, namely, Hirayama [23) and not his 198'I paper [2), and Sandier et al. [43) and not Helsing et al. (40): (iv) Ifi the authors have presented adjusted relative risks these have nor- mally been given in Table 5: f. There are some exceptions. First, the relative risk of 14.9 by Garland et al. [38), adjusted for age, systolic blood pressure„total cholesterol, obesity index and years of marriagc,