Tobacco Documents Online

i I I I I I I 1 I 1 I I I I I I I 1 REGULATORY TOXICOLOGY AND PkARMACOLOGY 21.180-183(1995) Ischemic Heart Disease and Spousal Smoking in the National Mortality Followback Survey' MAXWELL W. LAYARD Layard Associates, Alameda, Cali(ornia 94501 Received June 25, 1994 Data on never-smoking decedents from the 1986 Na- tional Mortality Followback Survey were used to per- form a case-control analysis of ischemic heart disease in relation to spousal cigarette smoking. The case groups consisted of 475 men and 914 women who died from heart disease. Controls consisted of 998 men and 1930 women who died from other causes. In this study there was no association between spousal smoking and ischemic heart disease in either sex (males, odds ratio = 0.97; females, odds ratio = 0.99). The results of this study are in striking disagreement with risk elevations reported in several previous studies of spousal smoking and heart disease. e tssa Aasae~;a P...., t,a. such an inference (Wells, 1988; Glantz and Parmley, 1991; Steenland, 1992; Taylor et al., 1992). The National'.vlortality Followback Survey (NMFS), which was based on a probability sample of adult deaths in the United States in 1986, offered the opportunity to perform a case-control analysis of ischemic heart dis- ease and spousal smoking that includes more never- smoking cases than any previously published study. This paper reports the results of that case-control study. MATERLALS AND METHODS The National Mortality Followback Survey was conducted in 1986 by the U.S. National Center for INTRODUCTION Several epidemiologic studies have reported associ- ations between ischemic heart disease mortality in never-smokers and exposure to environmental to- bacco smoke (ETS) from spousal smoking (Hirayama, 1984; Garland et a1., 1985; Lee et al., 1986; Martin et a1.,.1986; Svendsen et a1., 1987; Butler, 1988; Helsing et al., 1988; Palmer et a1., 1988; He, 1989; Hole et al., 1989; Jackson, 1989; Humble et a1., 1990; Dobson et al., 1991; LaVecchia et a1., 1993; He et al., 1994). De- spite the high rate of heart disease death among never- smokers, most of those studies have included rela- tively few cases; only three studies (Hirayama, 1984; Helsing et al., 1988; Dobson et al., 1991) included sub- stantially more than 100 cases. Although some review panels concluded that there was insufficient evidence to infer that ETS exposure increases the risk of heart disease (National Research Council, 1986; U.S. Surgeon General, 1986), other reviewers have drawn t This work was supported in part by the Center for Indoor Air Re- eearch. The viewe e:pressed represent the personal opinions of the author and are not neceaearily those of the Center for Indoor Air R.e• eearch. Health Statistics (NCHS) (NCHS, 1988; Seeman et al., 1989). The survey was based on a national proba- bility sample of about 1% of all deaths in 1986 of U.S. residents aged 25 years or older. This sample was drawn from the Current Mortality Sample (CMS), a 10% sample of death certificates that NCHS receives from state authorities approximately 3 months after the deaths occur. The NMFS sample included data fnm 49 states and the District of Columbia; deaths from Oregon were not included because of respondent consent requirements. Of the 18,733 death certificates in the NMFS sam- ple, about 2500 were selected with certainty from the CMS for certain causes of death and population sub- groups. These included all deaths due to ischemic heart disease (International Classification of Dis- eases, Revision 9, codesr410-414) for males aged 25-44 years and for females aged 25-54 years. In addition, some population groups were oversampled. For exam- pie, black decedents were oversampled 2.9 times and those under 55 years of age were oversampled 3.1 times. Next of kin of decedents in the NMFS sample were asked to complete a questionnaire that included ques- tions on demographic characteristics, dietary pat- terns, cigarette smoking (personal and spousal), alco- hol consumption, education, income, and history of 02 73-2300/95 36.00 180 Copyright'C 1995 by Academic Press. Inc. All rights of reproduction in any form reeerved I

~ r HEART DISEASE AND SPOtiSAL SMOKING 181 TABLE 1 ... . . RESULTS Distribution of Male and Female [schemic Heart Dis- ease Cases and Controls by Race and Mean Ages at Table 1 presents data on race and age at death for Death (National Mortality Followback 9urvey, 1986) ischemic heart disease cases and controls. For both sexes Cases Controls Test for No. °b No. . °b case-control difference Race White Black Other 356 96 23 Males 74.9 20.2 4.9 681 264 53 68.2 26.5 5.3 = 0.02 Mean age at death 72.64 64.75 P<0.0001 Females Race White Black Other 675 212 27 73.9 23.2 2.9 1,339 516 75 69.4 26.7 3.9 P=0.04 Mean age at death 78.17 71,89 _ P<00001 sal . sal smoking interactions. I I diseases. The questionnaire response rate was 88.6%. The NMFS is described in more detail in Seeman et al. (1989). The decedents included in the present study were re- stricted to those reported by next of kin to be lifetime never-smokers, defined to be those who had never smoked 100 or more cigarettes in their entire lives. Fur- thermore, decedents were excluded from analysis if they had never married or their marital status was unknown or if it was not known whether their spouse had smoked; 549 male and 692 female subjects were excluded for those reasons. After these exclusions, the case group consisted of 1389 (475 male and 914 female) ischemic heart disease deaths. Deaths from causes generally con- sidered to be smoking related were excluded from the control group. Excluded causes of death were cancers of the lung, mouth, pharynx, larynx, esophagus, pancreas, 95% confidence Caaee Controls Odds ratio• interval M.les chronic obstructive lung disease. The control group then No 378 783 1.0 consisted of 998 male decedenta and 1930 female dece- Yes 97 215 0.97 0.73-1.28 the race distribution of heart disease cases differed sig- nificantly from that of controls, and mean ages at death also differed significantly between cases and controls. All analyses of spousal smoking and heart disease risk were adjusted for age and race. No overall spousal smoking-heart disease risk eleva- tion was observed for either males or females, the ad- justed odds ratios being 0.97 and 0.99, respectively (Ta- w 2) , th f ith y s e nor as ere or e er sex an bl i gn ifi cant tren d in risk with increasing numbers of cigarettes smoked per day by the spouse (trend test P values: males, 0.8; fe- males, 0.3). The spousal smoking results were not appreciably affected by adjustment for history of hypertension, his- tory of diabetes, family history of heart attack, relative weight, alcohol consumption, dietary factors, education, and family income. There were no significant age-spou- . . DISCUSSION AND CONCLUSIONS In this study there was no suggestion of an association between exposure to ETS from spousal smoking and ischemic heart disease. Odds ratios were below 1.0 for both men (odds ratio = 0.97) and women (odds rgtio = TABLE 2 Odds Ratios for Ischemic Heart Disease Death among Never-Smokers by Spousal Smoking Categories (Na- tional Mortality Followback Survey, 1986) bladder, kidney, and cervix, cerebrovascular disease, and SPouset amoking dents. Spousal smoking status was based on answers to two questions asked of questionnaire respondents about the decedents: "Did (any) spouse smoke at least 100 ciga- rettes?"; and "How many cigarettes a day did (this) Spotue'scigs/day <15 38 107 0.76 0.51-1.14 15-34 45 92 1.07 0.72-1.59 35+ 6 12 0.92 0.33-2.55 Trend testt P s 0.8 spouse smoke?" (If more than one spouse smoked, the S i k Females pouea sm(M mg answer to the second question was for the most recent No 459 969 1.0 spouse.) Yea 455 961 0.99 0.84-1.16 Multiple logistic regression methods were used to es- Spooae's cigs/day timate spousal smoking odds ratios adjusted for age and <15 139 336 0.85 0.67-1.07 ~ race (Breslow and Day, 1980). Trend tests were calcu- 15-a4 224 405 .1.15 0.94-1.41 35+ 52 1t1 1.06 0.74-1.52 lated by assigning scores to exposure categories, includ- Trend test P= 0.3 ~ e quan it t a- ng h t d i an a d d ing the no-exposure category 4`2 , tive variable so defined to the logistic regression model. Adjusted for age and race . ~ ~ ~ ~

I I I I 2 182 MAXWELL W. LAYARD , . . ... . ._ .. .. _ ..,. .. 0.99), and there was no indication of an increasing trend Whatever the explanation for the inconsistency, in risk with increasing numbers of cigarettes smoked by taken as a whole the ETS-heart disease epidemiol- the spouse. Spousal smoking is of course a crude surro- ogy clearly does not justify an inference that ETS gate for ETS exposure, so some misclassification of exposure increases heart disease risk, and estimates of exposure is present in all studies that use spousal smok- never-smoker heart disease deaths attributable to ing as an exposure index. Also, death certificate data are ETS exposure are without a valid basis. subject to misclassification of cause of death. If there was an association between ETS exposure and the risk of heart disease, random misclassification of exposure, and of cause of death, would tend to bias an observed associ- ation toward the null. However, such biases could not explain the complete absence of association reported here. The results of the present study are in marked con- trast to risk elevations reported in other ETS-heart dis- ease epidemiologic studies (Hirayama, 1984; Garland et al•, 1985; Lee et al., 1986; Martin et al., 1986; Svendsen et al., 1987; Butler, 1988; Helsing et al., 1988; Palmer et al•, 1988; He, 1989; Hole et al., 1989; Jackson, 1989; Humble et al., 1990; Dobson et al., 1991; LaVecchia et al., 1993; He et al-, 1994). Those studies reported sex-specific rela- tive risks ranging from 0.97 to 4.00, with a pooled rela- tive risk of about 1.3 (Glantz and Parmley, 1991). Largely on the basis of published epidemiologic data, Wells 11988), Glantz and Parmley (1991), and Steenland He Y Lam T H et a4 (1994) Passive smokin at work ss a risk (1992) concluded that exposure to ETS increases the factor for eoronary heea dixase in Chinee women who have never risk of heart disease and estimated that some 50,000 smoked. Br. Med. J. 308, 330-384. heart disease deaths per year among U.S. nonsmokers are attributable to ETS exposure. The results of this study are consistent with those of recent analyses of data on spousal smoking and heart disease from two large American Cancer Society co- hort studies, Cancer Prevention Study I and Cancer Prevention Study II (CPS-Iand CPS•II) (LeVois and Layard, 1995). Neither of those analyses suggested an association between spousal smoking and heart dis- ease: sex-specific relative risks ranged from 0.97 to 1.03. The CPS-I data included 14,898 coronary heart disease (CHD) deaths among never-smokers, and the CPS-II data included 3065 CHD deaths among never- smokers- The total number of never-smoker heart disease cases in the two CPS studies and the present NMFS study is over 19,000, a far greater number than the to- tal of about 3000 cases in all other published studies of ETS and heart diseasB: When the six se,t-specific relative risks from the CPS studies and the NMFS study are combined, the pooled relative risk is 1•00 (95% confidence interval, 0.97-1.04), which is highly significantly different from the pooled relative risk of 1.3 from previous studies (LeVois and Layard, 1995). The inconsistency between the results of the three new spousal smoking-heart disease analyses and those of previous published studies suggests that the 30% risk elevation estimated from previous studies can be partly or wholly explained by publication bias. REFERENCES Breslow, N. E.. and Day, N. E. (1980). The Analysis of Caee-Control . Studies. [ARC, Lyon, France. Butler; T. L. I1988). The relationship of passive smoking to various health outcomes among Seventh Day Adventists in California. Ph.D. dissertation, University of California. Los Angeles. . Dobson, A. J., Alexander. H. M.. Heller. R. F.• and Lloyd. D. M. i 19911. Passive smoking and the risk of heart attack or coronary death. . Med. J. Aua[. 154, 793-797. Garland, C., Barrett-Connor, E.• Suarez, L., Criqui,l4., and Wingard, D. (1985). Effects of passive smoking on ischemic heart disease mor- tality of nonsmokers. Am. J. Epidemiol. 121, 645-650. Glantz, S. A.. and Parmley. W. W. (1991). Passive smoking and heert disease: Epidemiology, physiology and biochemistry. Circulation 83, 1-12. He, Y. (1989). Women's passive smoking and coronary heart disease. (English abstract only). Chung-hua Yu Fong IHsueh Taa Chih 23, - 915-922. . Helsing, K., Sandler, D., Comstock, G., and Chee, E. (1988). Heart disease morrality in nonsmokers living with smokers. Am. J. Epide• ,niot 127, 915-922. - Hiraynma, T. (1984). Lung cancer in Japan: Effects of nutrition and passive smoking. In Lung Canrerr Causes and Preuention (Mizell and Correa, ads.). Verlag-Chemie, New York. . I Hole, D. J., GWis, C. R, et oL /1989). Passive smoking and cardlo- respiratory respiratory health in the west of Scotland. Br, Med. J. 299, 423- ~ 427. Humble,.C., CroR. J.. Gerber, A.. Casper, M., Hamee, C. G., and ~ Tyroler, H. A. (1990). Passive smoking and 20-year cardiovascular i disease mortality among nonsmoking wives, Evans county, Georgia. Am. J. Public Health 80, 599-601. ~ Jackspn, R. (1989). Chapter 6: Passive Smoking. From The Auck- fand Heart Suruty. Ph.D. thesis, University of Auckland. New Zealand. LaVecchie, C., D'Avanzo, B., et al. (1993). Passive smoking and the risk of acute myocardial infarction. Lancet 941, 505-506. - Lee; P. N., Chamberlain, J., and Alderson, M. R(1986). Relationship of paseive smoking to risk of lung cancer and other smoking-aseoci- , ated diseases. Br. J. Cancer S4, 97-105. LeVoie, M. E., and Layard, M: W. (1995). Publication bias in the environmental tobacco smoke/coronary heart disease epidemio- j logic literature. Regul. Tosicot. Pharmacol. 21, 181-188. Martin, M.. Hunt. S.. and Williams, R(1986). Increased incidence - of heart attacks in nonsmoking women married to smokers. Paper I presented at the annual meeting of the American Public Health As- societion. ~ National Center for Health Statistics (NCHS) 11988). National Mor ~.q tality Followback Survey, 1988: Public Use Data Tape Documente• ~. re tion. NCHS/DF/MT-90/014a, Hyattsville, MD. s,t l~ National Research Council (1986). Enufronmentaf Tobacco Smoke: .~ ~ ~ i

HEART D15EASE AND SPOUSAL SMOKING 1&3 Mcosu*•+,g E=ccsa*es and Asaesst..g Health Effects. National Acad- Svendsen. K. H., Kuller, L. H., et af. (1987). Effects of passive smoking ,emy Press, Washington, DC. in the multiple risk,lactor intervention tnal.Am.J. Eademioi. 126, Palmer,J..etaL(1988),Passivesmokingandmyocardialinfarctionin"83-795. . women (Abstract), CVD Epidemiol. Neu•slett, 43,29. Taylor. A. E., Johnson. D. C., et aL (1992). Environmental tobacco Seeman, I., Poe. C. S., and McLaughlin, J, K. (1989). Design of smoke and cardiovascular disease. Circrdatioh 88, 699-702. the 1986 National Mortality Followbeck Surveyi Considerations ~. S. Sur on collecting data on decedents. Public Health Rep. 104, 183- geon General ( 1986). The Health Consequcnces of Incofuntary 198. , Smoking.C,S.DepanmentofHealthandHumanServices. - Steenland.K.(1992).Passivesmokingahdtheriskofheartdisease.J.WeI1s,A.J.(1988).Anestimateofadultmon.al ityintheVnitedStaces . Am. Med. Assoc. 287, 94-99. from passive smoking. Enuiron. lnt. 14, 249-265.