Tobacco Documents Online

i l Tobacco ; Envlronmerital I F Smoke Proceedings of th:: lnternw2onal Symposium at 'UwvQrsi~y 1989 Donald J. Eeohi6y,?n McGill Universiry Joseph A Wu New York Medi~~ fbllege 1N.ZX kQr , , L. M , C~,..pt. a„J Cu. L. s. l~ ~ it. ce : L LErc Lvr¢ c J Zo~. w+.d~~ .. ~ ~•o. ~ ~...4 ZQSO~ ~ K, 13Ci Editors and Osganfzers of the Sym,irosium. ~r3-~ds ~T Leacington Books D.C. Heath and Company/Lexington, MassachusettslToronto

Library o%Congrtss Cataloging-in-Publication Data Environmental tobacco smoke : proceedings oPthe international symposium at McGill UniversiryJ Donald). Ecobichon, Joseph M. Wu,, editors. p. cm. ISBN 0-669-24365-3 (alk: paper) 1. Passive smoking-Health aspeas-Congresses. 2. Tobacco smoke pollution-Hcalth aspects-Congresses. 3. Tobacco smoke- Congresses. 1. Ecobichon, Donald J. II. Wu,,Joseph M. lll. International Symposium on Environmental Tobacco Smoke (1989 : McGill Universiry). RA1242.T6E58 1990 616.86'5'071-d'c20 89-49011 CIP Copyright m 1990 by Lexington 9ooks All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical„including photocopy, recording, or any, information storage or retrieval system, without permission in writing from the publisher. Published'simultaneously in Canada Printed in the United'States of America International Standard Book Number: 0-669-24365-5 Library of Congress Catalog Card'Number: The paper used in this publication meets the minimum requirements of~Amerian NationallStandard for Information Scienees-Permanence of Paper for Printed Library Materials, ANSI Z39.48-1984. O,. 8990919287654321

8 Environmental Tobacco Smoke and Cardiovascular Disease: A Critique of the Epidemiological Literature and Recommendations for Future Research 3 ri LawrenceM. Wexier, Ph.D. New York Medical College T his paper evaluates the current epidemiological literature examining the possible relationship between exposure to environmental tobacco: smoke ("ETS") and cardiovascular disease. Based on the available evidence, it is this author's opinion that it has nor been demonstrated that exposure to ETS increases the risk of cardiovascular disease. This paper eval- uates seven studies that examine this issue (table 8-1). Five of the studies are prospective in nature, one is a case-control design (retrospective); and one is am experimental design examining the biological plausibility of a link be- tween ETS and cardiovascular disease. Several key points of epidemiology need to be mentioned here, and' shoul& be kept in mind when reading,the critiques of the seven studies. To prove causality five criteria need to be met. The first relates to the strength of the association. There are three elements to this criterion. First, there must be a statistically significant increase in the incidence of the disease in the exposed population compared with the non-exposed population. Second, for the association to be regarded as meaningful, a relative risk of 2.0 or greater is generally considered necessary. Third~ the association should also be dose dependent, i.e., higher doses are associated with higher incidence of: disease. The second point is that consistency of the association must exist among the relevant studies. This means that similar rates of' disease musr occur at different times and places, under~ comparable study designs. A third point deals with the temporali aspect of the association. This means that exposure to ETS should have occurred at a reasonable time before the onset off disease, given what is known about how long it takes for cardio- vascular disease to develop. A fourth point is specificity of the assodation. With ETS, this means that exposure to ETS must be shown to be associated with cardiovascular disease while controlling for all confounding variables.

140 - Environmental'Tobacco Smoke Tablt 8-1 Env'uonmentaliTobacco Smoke and Cardiovascular Disease Design 1. Hirayama Sixteen year prospective (1981,,1984)' study of nonsmoking Japanese women classified at start of follow-up by the smoking status of eheir husbands. 142;857 women 40 and over (91,540 nonsmoking wives), 2. Garland Prospective-enrolled (1i985) 82% of'adults ages 50- 79 between 1972-1974 in a community in San Diego. Blood pressure and plasma cholesterol measured at entry; interviewed a111cohort of 695 current married non- smoking women free of heart disease; ten year follow-up. 3. Gillis (1989) Two urban communities in Scotland. Ten year follow-up report. 8,128 adults ages 45-64. Findings l.,Relative risk of 1.31 for ischemic heart disease for nonsmoking,women whose husbands smoked'> 19 cigarettes per day compared with nonsmoking women whose husbands did not smoke. 2. Mantel-Haensze) signifiaant at p < .019, 1984. 3. °Passive smoking did not seem to increase the risk of developing... ischemic heart disease ° -Hirayama, 1981. 1. Elevated cardiac disease deaths in non-smoking women, ages 50-79, whose husbands were former or current smokers. 2. 19 deaths from ischemic heart disease after ten years. Non-smokers exposed totigzrette smoke in, their homes had a slightly higher rate of myocardial infarction than those unexposed. Met6'odological Problems 11. Potential biases. 2. Misciassificanon of smokers and non- smokers. 3. Misclassification of dose response (number of'cigarettes smoked'per day):, 4'. Looked arspouse exposure onl), not workplace. 5, No control for indoor air pollution, e.g., cooking witfl, kerosene stoves. 6. Not representative of Japanese population-only agriculture represented. 7. Non-random samp)e of prefectures-only, a convenience sample. 1. Some misgrouping- wives of former smoker were grouped with wives of current smokers. 2. Small sample sizes, valbe may be inappropriate based on Mantel-Haenszel, and may only be an approximation; still pwasonlyp<.10. 3. 15 of 19 deaths occurred in nonsmoking women married to former smokers-puzzling results. 1. SmallIsample size. 1'. Few of the results were statistically significant. Table l 4. Svenc (198' 5. Hcls (19b

Table 8-1 continued ~logical kms liases. 4. Svendsen ~cation of (1987) d non, cation of se cigarettes r d'ay);. spouse nly, not ~ for pollhtion,. ~g with oves. Kntarivr of -only m sample res-onh_• ce ~ouping-- ~rmer re th wives, mokers. Ie sizes, be te based I'l'aenszelJ, lii be an ion; still p < .10. aths 5'. Helsing. (1988) . g women former puzzling ~le size., t results ically Design 1. Multiple Risk Factor Intervention Trial (MRFIT): 2. Randomized primary prevention trial designed to test the effect of a multifactor, intervention program on mortality fromm coronary heart disease in men with previous cardiac episodes. 3: Memwere chosen for participation ifithey, had at least two of three risk factors for heart disease (smoking, high cholesterol levels, , high blood pressure). 1. Twelve year study, executed in Washington County, Maryland: 2. July, 1963 census of 91,909 people. 3. Whites only. 4. Death certificates collected from July, 1963 through July, 1975: S. Non-smokers, ages 25 and'over. 6. 4,162 men and 14,873 women. ETS and Cardiovascular Disease • 141 Findings 1. No difference between smoking wives and nonsmoking wives for non-smoking men for blood pressure or cholesterol. 2. Roughly two-fold increase in risk of CHD mortality and morbidity among, nonsmoking men exposed to ETS of wives. 1. Death rates from arteriosclerotic heart disease were higher among men (Relative risk = 1.31),and women (relative risk = 1.24) who lived with,smokers in 119631 after adjustmentfor age, marital status„years of'schooling, and quality of!housing index. 2. For women; relative risk increased significantly. (p < .OOS) for dose response (increasing levels of exposure).. 3. Men-4itde evidence of a dose response relationship. MetHodolog,cal Problems 1.,Sample size small. 2. Results-not statistically significant. 1. Only smoking data collected on every person was in 1963. 2. No measurement of', changes in smoking habits. 3. No data on household changes from 1963-1975. 4. Very little other risk factor data for heart disease. S.,No diet, exercise, blood pressure,, cholesterol data, or ETS exposure outof home.

Methodological Design Findings Probli•ms 6. Lee (1986)i Case-conerol; ischemic heart disease Case-control cases and controls did methodological issues. not show a statistically significant difference in their exposure to involuntary smoking, based on smoking habits of spouses or on an index accounting for exposure at home, at work, and during travel and leisure. 7. Aronow Experimental design. ETS aggravates angina 1. Endpoinnof angina (1978) pectoris. based on subjcaivc evaluation. 2. Stress not controlled fon se( etr tic hu W rel th, 0.. he sn hi ri` th Ti pj tl-. The fifth point is that there must be biological plausibility. This means that under experimental conditions exposure to the pertinent substance (or similar substances) must be shown to cause biological changes that can lead to the disease in question. All' five conditions must be met for causality to be established. We will return to these points at the end of the paper, when we examine recommen- dations for future research, I. Summary of Epidemiological Literature A. Prospective Studies 1. Huayama. Hirayama (1984) conducted a prospective cohort study in 29 health center districts in six prefectures in Japan between January 1966 and December 1981. In total, 265,118 adults (122,261 men and 142,857 women) aged 40 years and over were followed. Ninety-five percent of the census pop- ulation was interviewed between October and' December 1965. Also, Hirayama established a record' linkage system under which he gathered and analyzed death, certificates, risk factor records, and a residence list obtained by an annual census. Questions on smoking habits were asked independently of husbands and wives at the beginning of the study. There were 91,540 non• smoking married women whose husbands' smoking habits were reported by qµestionnaire. ex (a n, ir. p st K ft sl c~.. h E Ii

ETS and Cardiovascular Disease • 143 PologicaL bbJems ol gical'issues. ~t of angina k~ subjpctive `on. bt controlled peans. Pe (or ` lead t wil1 men- L~n 29 and irnen) Pop- so, and ~ ined ently non- d by In 1981, Hirayama (1981) concluded that "husbands' smoking habitss seemed to have no effect on their non-smoking wives' risk of'developing isch- emic heart disease." Hirayama reported' age/occupation standardized risk ra- tios for ischemic heart disease in non-smoking women by smoking habit of husband. When the husband was a non-smoker, the relative risk was 1.0. When the husban&was an ex-smoker or srnoke& 1-19 dgarettes per day, the relative risk was .97. When the husband smoked 20:or more cigarettes/day„ the relative risk was 1.03, and the reported p value was not significant at 0:393. Hirayama (19g4)~ in a 1984 paper, reported an elevated risk of'ischemic heart disease morbidity based on, further analyses. The relative risk for non- smoking married women for husbands who were non-smokers was 1.0; for husbands who were ex-smokers or smoked 1-19 cigarettes/day the relative risk was 1.10;,and for husbands who smoke&20 or, more cigarettes per day, the relative risk was 1.31, with a 90% confidence interval of 1.06 to 1.63. The reported p value was significant at .019. Hirayama's study has severali major methodological problems. The firstt problem is potential misclassification of smokers and non-smokers. Many of the wives who stated they were non-smokers may in, fact be ex-smokers or even current smokers, and thus likely to have had or continue to have direct (as opposed to indirect), exposure to cigarette smoke. The second problem is that Hirayama's study included a disproportionate number of women of lower socioeconomic status. In Japan, these women live in much closer proximity to: their cooking quarters and may have more ex- posure to charcoal or kerosene stoves than women of' higher socioeconomic status. This exposure has been associated with lung cancer in women iniHong Kong.. Women in Japan, of a higher socioeconomic status live farther away from their kitchens and are more likely to use electric burners. The Hirayama study failed to controlI for these confounding variables, which may be asso- ciated with ischemic heart disease. A third problem is the misclassification of dose response. Ex-smoking husbands were lumped with currenti cigarette smokers of 1-19 cigarettes/day: Because ex-smokers are very different in their cigarette exposure rates and lifestyles than smokers of 1-19 cigarettes/day, this could skew the data. A fourth problem is that Hirayama only examined the exposure of the wife in the context of the husband's cigarette smoking behavior. No attempt was made to quantify any exposure to ETS outside of the home, such as in the workplace. A fifth problem is that the Hirayama study was not representative of Japanese society but only of an~agriculturally based population, which is not typical for Japan. In addition, six prefectures were chosen to participate in the study based on the fact that they appear to have had the best conditions for collecting data. Hence, random sampling was not used'.

A sixth problem is that the Hirayama study did not control for other risk factors associate& with cardiovascular disease, i.e., systolic blood pressure and plasma cholesterol. Although the Hirayama study offers a large prospective cohort to ex- amine the relationship between presumed exposure to environmental tobacco smoke and ischemic heart disease, one can not draw definitive conclusions because of the aforementioned methodological problems. 2. Garland. Garlan&(1985)' conducted a prospective cohort study commen- cing in 1972-1974 in Rancho Bernardo, a white middle-class suburb of San Diego, California. The entire adult population was invited to participate, of which 82% agreed. The authors report that the respondents were represen- tative of the total population with regard'to age and'sex. All respondents were administered a standardized' inventory, including questions about age, cigarette smoking, history of past hospitalizations for heart attack, heart failure or stroke, and number, of' years marriedi Cigarette smoking was assessed as current, former or never. Only current smokers were asked the number of cigarettes they smoked' per day. No data were obtained for duration of'smoking. In addition, blood pressure and plasma cholesterol were obtained. An annual mailing was utilized to determine vital status for the next ten years. Death certificates were obtained for all' decedents. Diagnosis of isch- emic heart disease was validated by interviews with family and physicians, andlor examination of' hospital records, for 85% of the deceased group: Six hundred ninery-five (695) currently married nonsmoking women, ages 50-79, with no previous hi'story of heart disease or stroke were followe& based on their husband's self-reported smoking status in 1972=1974. The results, after adjusting for age, systolic blood pressure, total plasma cholesterol; obesity index and years of'~ marriage gave a relative risk of 14.9' of deaths from ischemic heart disease for women married', to current or for- mer smokers at entry compared with1 women married to never smokers. The p value was not significant, p<.10! Important methodological problems exi'st, with the Garland study. The first is that Garland later reported a corrected relative risk of 2.7 (not 14.9 as reported in the 1985 publication). The p value is still < .10 and not significant. The second problem i's that after ten years of follow-up, only 19 deaths from ischemic heart disease occurred. This small sample size is compounded by the fact that 15 of the 19 deaths occurred in nonsmoking women married to husbands who had'stopped smoking at entry. Without more detailed char- acterization of these women's exposure to ETS, it is difficult to show an as- sociation between ETS and ischemic heart disease. As the study di& not as- certain number of cigarettes smoked per day in former smokers, it is not pos, grol gro exa hor hea: suc lacl ma 3. of anc (~1` res scr tio an, ba pu 1. 2. 3. 4.

ETS and Cardiovascular Disease • 145 possible to measure any sustained' effects of' ETS in this former smoking group. Another methodological problem is that wives of former smokers were grouped with wives of current smokers, and it is difficult to determine the exact effect of ETS for this former smoking group. Although the Garlan&study does make an attempt in a prospective co- hort study to measure the effects of possible exposure to ETS on ischemic heart disease, and does control for important cardiovascular confounders, such as obesity, blood pressure and cholesterol,,the small sample size and the lack of adequate measurement of' ETS in a former cigarette smoking group make the results only suggestive and' certainly not definitive. 14.9 not kiths Jde& ed as- as- not 3. Gillis. The Gillis study (1989):consists of a prospective cohort comprise& of men and women aged 45-641 years who resided in two towns, Renfrew an& Paisley; in the west of' Scotland, between 1972 and 1976. Residents (15;399):of these two towns who met the age and resid'encgcriteria (an 80% response) agreed to participate; 7,997 were subjected to a cardiorespiratory screening examination, a self-administered questionnaire that included ques- tions on smoking behavior. The eventual sample was comprised of 3,960!men an& 4,037 women where it was possible to study varying exposures to to- bacco smoke by cohabitees. Four groups were established for •analysiss purposes: 1. Control-neither the case nor anyone living at the same address ha&ever smoked. 2. Presumed ETS exposure in the home-the case had' never smoked but lived at the same address as a subjea who had smoked. 3. Single smoking-the case was a smoker or an ex-smoker and lived at the same address as a person who had never smoked. 4. Double smoking: the case was a smoker or an: ex-smoker who lived at the same address as a subject who was also a smoker or ex-smoker. Mortality was used as an endpoint and was obtained'from the National Healthl Service. Cardiovascular signs and' symptoms were also noted. Data presented were complete through December 1985, for an average follow-up of1 LS years. The authors present relative risks and 95% confidence intervals adjusted for age, sex, social' class, diastolic blood pressure, serum cholesterol concen- tration an& body mass index. Total mortality for ischemic heart disease was higher among those reportedly exposed to ETS in the home than controls. Women with ETS exposure in the home were broken into two dose re- sponse categories for further analyses. These included: (I) the high exposure

146 • Environmental Tobacco Smoke group, where the woman's cohabitee smoked 15 or more cigarettes daily, and (2) the low exposure group where the women's cohabitee smoked! less than 15 cigarettes daily. Age-adjusted mortality from ischemic heart disease was higher for those in the high exposure category than in the low exposure group. Relative risk was adjusted for age, sex, social class and cardiovascular variables including diastolic blood pressure, serum cholesterol concentrations and body mass index. Compared with controls, the relative risk was 2.01 for ischemic heart disease and was not significant. The Gillis paper has several methodological problems. The first is that it does not have sufficient power to demonstrate an association~ between~ ETS and ischemic heart disease. The sample size is too small. A second'problem is that the relative risk of 2.01 for ischemic heart dis- ease for non-smokers compared with controls is too similar to the relative risk of 2.27 for active smokers compared with controls to make sense. An explanation for this is not clear, but may be due to small sample size as well. Potential biases also exist in the Gillis study. One potential bias is that those exposed to ETS within the home may have had higher exposures to ETS outside of the home compared with controls. A second potential bias is misdassification of women as non-smokers when they may be former smok- ers or current smokers. Although the Gallis study suggests an ~ association between ETS and' car- diovascular mortality in non-smokers, the data lacks any statistical signifi- cance. Also, the study reports some confusing and similar relative risks for active and passive smokers, and' is confounded by several' important meth- odological biases. This study shoul& be replicated in a much: larger study population, with adequate statistical power. 4. Svcndsen. Svendsen (1987)!reports the results of the Multiple Risk Factor Intervention Trial (MRFIT), conducted from 1973-1982. The trial consisted of inen, aged 35-57, recruited from 18 cities in the United States. Males who felliwithin the upper: 110-15% risk score distribution ~ for heart disease, based on an index eomprised~of serum cholesterol concentration, cigarette smoking and diastolic blood pressure, and free of overt coronary heart disease were randomized to one of'two groups:, (1) special intervention or (2) usual care. Participants in both groups were seen annually over six to eight years for risk factor measurement and a medical' examination. A detailed smoking history was obtained at baseline and at all subsequent annual visits. Cause of death was evaluated by a committee of three cardiologists after examination of death certificates and other medical records. Fourteen hundred of 12,866 men reported that they had never smoked at entry into the study. Of these 1400, 1,245 were married. Of the later group, 286 were married to: women who smoked and 959 were marrie& to women who did'not smoke.