Tobacco Documents Online

PASSIVE SMOKING AND LUNG CANCER 337 TABt.E 14'- ASSOC]At lON nETW ISEN R1SK OF ADENOCARCINOMA AND OF SQUAMOUS-CELL CANCER OF TnE I UNG AND PASSIb'li SMOKING Ade.nnearelnnn,a Squamous-cdl lung cancer Smoking Tabfl Convols C ves ox' 95°mCI ConcroLs C:axe.e OR' 959oCI n-35R n=1n6 n=35R n-42 Husband's smoking no 195 46 1.0 195 16 1.0 yes 163 60 152 Q96-2.39 163 26 1.94 0.99-3.81 Smoking of other member of a household no 285 84 1.0 285 37 1.0 yes 73 22 0.98 0.57-1.70 73 5 0.48 0.18-1.29 Colleagues' smoking no 291 84 1.0 291 32 1.0 yes 67 22 0.99 0.56-1.73 67 10 1.20 0.54-2.63 Father's smoking no 180 54 1.0 180 19 1.0 yes 178 52 Q9l 0.58-1.42 178 23 1.14 0.59-2.19 Husband's smoking, type of cigarettes no 195 46 1.0 195 16 1.0 filter 72 13 0.74 037-1.48 72 6 1.08 0.40.2.91 no filter 32 15 2.00 0.96-4.16 32 4 1.50 045-5.01 "papirosy" 59 32 2.34 L32-4.14 59 16 3.13 1.44-6.81 t 'OR adjusted for age and education. expected that some association may exist with passive smoking as well. Therefore, we conducted logistic-regression analysis using 2 sub-sets of the control group (i) patients with breast and endome- trial cancer, and (ii) other controls. Using as controls breast-cancer and endometrial-cancer patients only did not materially alter the results, although the estimated OR were somewhat higher. Husbands' smoking was associated with a statistically significant increase of lung cancer in non-smoking wonten (OR, 1.82; 95% Cl, 1 J 8-2$0). Using other controls in the logistic-regression model showed sonlewhat weaker association with the exposure to ETS from husband (OR, 1.22; 95% CI, 0.79-1.88). DISCUSSION The main importance of our study is that it is conducted on a population with a specific snloking pattern from which no informa- tion is available on the health effects of ETS. Another strong aspect of the study is its size, which made it possible to separate sizeable groups of cases and controls with high exposure to ETS, and to conduct analyses for different histological types of lung cancer. The United States Department of Health and Human Services (DHHS) Environmental Protection Agency (EPA) has classified ETS as aa known human lung carcinogen, on the basis of the numerous carcinogens found both in ETS and in mainstream tobacco smoke (IARC, 1985), and of the higher lung-cancer risk seen in ncver-smokers exposed to ETS in epidemiological studies (US DHHS, 1993). Most of these studies concentrated on ETS exposure from spouse's smoking. However, some of them addressed other sources of ETS exposure, such as parents during childhood, household members other than the husband, work environment (US Public Health Service, 1986: National Research Council, 1986; Persha- gen, 1994; Tredaniel et al., 1994). Typically elevated risks of lung cancer were observed in non-smoking women living with smoking men: relative risk (RR) ranged from 1.2 (Brownson et al., 1992; Fontham et a(., 1994) to 3.2 (Pershagen et al., 1987). The meta-analyses performed esti- mated that the RR of lung cancer for non-smoking women living with smoking men was around 1.3 (US Public Health Service. 1986; National Research Council, 1986; Blot and Fraumeni, 1986; Wald et ai., 1986; Saracci and Riboli. 1989). Six studies on risk of lung cancer from spouse's smoke are available from Europe: 2 from Greece (Trichopoulos et al., 1983; Kalandidi et al., 1990) showed a 2-fold increased risk. Of the other studies, only the last study from Sweden (Pershagen et al., 1987) found an increased risk of lung cancer above 45 pack-years of exposure. Two meta-analyses of this data are available. Pershagen (1994) combined 6 studies and estimated an overall RR of 1.47 (95% Cl, 1.12-1.92), while the US DHHS (1993) excluded the Greek studies and calculated a combined RR of 1.17 (95% CI, 0.84-1.62). Our summary OR of 1.53 (95% CI 1.06-2.21) is compatible with cstimates from the meta-analysis of Pershagen (1994), although it suggests that the effect of ETS from spouses in Russia is somehow stronger than in Western Europe. The evidence from the available studies of an association between P.TS exposure during childhood and lung-cancer risk is inconsistent (Pershagen, 1994). Two large studies from the United States reported no increased risk from childhood exposure (Brown- son et aL, 1992; Fontham et al., 1994). In our study we failed to find any association between ETS exposure in childhood and risk of lung cancer in non-smoking women. Our data indicating no effect of workplace exposure to ETS on the risk of lung cancer are in concordance with studies from Europe (Pershagen, 1994). However, in some other studies (Fontham et aL, 1994) RR associated with workplace exposure was higher than that associated with exposure to husband's smoking. Increased relative risk of lung cancer was observed with increasing pack-years of exposure to ETS of the husband (Fnntham et al., 1994) and number of cigarettes per day smoked by him (Cardenas et aL, 1997). We have also studied the association between number of cigarettes smoked and duration of smoking and risk of lung cancer. However, the increase in risk was not uniform. Our study suggests that the association between exposure to husbands' smoking and risk of lung cancer in non-smoking women is somewhat stronger for squamous-ecll carcinoma (OR, 1.94; 95% CI, 0.99-3.81) than for adenocarcinonla (OR, 152; 95% Cf, 0.9(-2 39). Similar results were obtained in other studies on ETS conducted in Europe (Trichopoulos et al., 1983; Pershagen el al., 1987; Kalandidi et al, 1990) and in the United States (Fontham et al., 1994). This result parallels the stronger association found between active smoking and squamous-cell lung carcinoma as compared with adenocarcinoma. We report an association between type of cigarettes smoked by the husband and lungcancer risk. The risk is higher For women whose husbands smoke "papirosy' than for women whose hus- bands smoke cigarettes without filter. This association is stronger for women with squamous-cell carcinoma than for those with adenocarcinoma- In fact, we found no increase in risk of lung

336 RESULTS ZARIDZE EIAL Table I shows the distribution of demographic characteristics. Cases were somewhat older than controls at disease occurrence or interview. Crude occupational classification of cases and controls was similar. Cases were less educated than controls (p = 0.001). There was also a significant difference in marital status between cases and controls (p = 0.03). Among cases, there were more widows than among controls. Among lung-cancer cases, the majority (56.1%) were adenocar- cinoma, while 22.2% of cases were squamous-cell cancer, and 5.3% small-cell carcinoma (Table ID. Diseases among the control women included breast cancer (39.7%), melanoma (20.4%), cancer of the endometrium (10.3%), ovary (7.8%), stomach (4.5%) and uterine cervix (45%), also Hodgkin's disease (5.3%), and other cancers (7.0%) (Table II). Table III presents OR of lung cancer in non-smoking women associated with exposure to ETS. An elevated risk of lung cancer was identified in women whose husbands smoked. The OR adjusted by age and education for husbands smoking was 1.53 (95% CI, 1.06-2.21). Smoking by other members of the family, by colleague or by father during childhood of index patient did not affect the risk of lung cancer. Additional analyses nsing qualitative and quantitative variables for husband's smoking practices showed, that the OR is affected by the type of cigarettes: OR for smoking cigarettes without filter was 1.74 (95% CI, 0.94-3.20). Smoking "papirosy" was associated with a statistically significant increase in risk (OR, 2.12; 95% CI, TABLE I-DESCRIPTIVE D ATA FOR FEMALE LUNG-CANCER CASES AND CONTROLS' Y]nbe Controls, k (N=358) Cases, Ho (N=189) p vNuc Age at interview <55 25.7 23.8 0.14 55-65 39.4 32.8 >65 34.9 43.4 Marital status married 50.8 47.6 0-03 divorced 16.2 11.1 never married widowed 7.0 26.0 4.2 37.1 Educational level sohool(notcompleted) 25.4 27.0 0.001 secondary school 36.0 49.7 some university 38.6 23.3 Occupational classification manual worker 38.3 42.3 0.36 office worker 61.7 57.7 IAll subjects were life-long non-smokers and had lived in Moscow for 20 years or more prior to interview. TABLE II-DIS'1'RIBUTION OF LUNG-CANCER CASES BY HISTOLOOICALTYPE AND CONTROLS BY SITE OF PRIMARY TUMOR Cases Controts FG.ttMogiu,I type Numbrs `,5 Site Number °h Adenocarcinoma 106 56.1 Breast 142 39.7 Squatnous-cell i 42 22.2 Melanoma 73 20.4 carc noma Small-cell card- 10 5.3 Endometrium 37 10.3 noma Large-eell carci- 2 1.1 Ovary 28 7.8 noma Mixed 7 3.7 Stomach 18 5.0 Others(carcinoid) 5 2.6 Cervix ulari 16 4.5 Unknown 17 9.0 Hodkin's 19 5.3 disease Others 25 7.0 Total 189 100.0 Total 358 100.0 TABLE III-ASSOCIATION BETWEEN RISK OF LUNG CANCER AND PASSIVE SMOKING Smnkingbabit Contmis n=358 Orses n=189 OR' 959oCI Trend Husband's smoking no 195 80 1.0 yes 163 109 1.53 1.06-2.21 Smoking of other member of a household no 285 151 1.0 yes 73 38 0.91 0.58-1.42 Colleagues' smoking no 291 153 1.0 yes 67 36 0.88 0.55-1.41 Father's smoking no 180 97 1.0 yes 178 92 0.92 0.64-1.32 Husband's smoking, dura- tion (years) no 195 80 1.0 0.07 1-15 39 31 1.86 1.07-3.22 >15 124, 78 1.42 0.95-2.12 Husband's smoking, quantity of cigarettes no 195 80 1.0 0.10 1-10 90 66 1.66 1.09-2.52 >10 73 43 1.35 0.84-2.18 Husband's smoking, type of cigarettes no 195 80 1.0 filter 72 28 0.95 0.56-1.61 no filter 32 25 1.74 0.94-3.20 "papirosy 59 56 2.12 L32-3.40 Husband's smoking in wife's presence 195 80 1.0 no 47 31 1.48 056-2.53 yes 116 78 1.55 1.04-2.31 'OR adjusted for age and education. 1.32-3.40). Husband's smoking in the wife's presence showed an OR similar to any smoking (OR, 1.55, 95% CI, 1.04-2.31). When husband's smoking was analyzed according to duration, lung-cancer risk was elevated both for short- and for long-term consumption, but there was no uniform increase in risk. Similarly, an increase in risk was present among wives of smokers of 1 to 10 and more than 10 cigarettes per day, but the risk was higher in the first than in the second category of consumption. Risk of lung adenocarcinoma was also found to be associated with husbands smoking (OR, 1.52; 95% CI, 0.96-2.39). No effect was present for ETS from family members other than the husband, workmates or the father. Risk was affected by the type of cigarettes smoked by the husband. OR was 2.0 (95% CI, 0.96-4.16) for husbands smoking cigarettes without filter and 2.34 (95% CI, 1.32-4.14) for husbands smoking "papirosy", ["Papirosy" are a special type of Russian cigarettes with a long mouth piece; they usually have very high levels of tar (higher than 30 mg/cig) and nicotine (higher than 1.8 mg/cig) (Zaridze et at., 1986).) (Table IV). The association between husbands' smoking and risk of squa- mous-cell carcinoma in non-smoking women is somewhat stronger than for all lung cancer combined or for adenocarcinoma OR for husbands smoking is 1.94 (95% CI, 0,99-3.81) (Table IV). There is a strong association between type of cigarettes smoked and risk of squamous-cell cancer. OR for husband's smoking of cigarettes without filter is 1.50 (95% CI, 0.45-5.01), for smoking "papirosy" is 3.13 (95% CI, 1.44-6.81). Also for squamous-cell carcinoma, there was no effect of ETS from sources other than the husband. The other histological types were not numerous enough to allow separate analyses. - In this study, cancer patients were used as controls. The risk for some of the cancer sites, such as cancer of the stomach and uterine cervix, has been associated with active smoking and it may be ! 99035513

Int. J. Cancer: 75, 335-338 (1998) © 1998 Wiley-Liss, Inc. Publlcatlon of the International Union Against Cancer Publication de Union In[ernaVOnale Contre Ie Cancer EXPOSURE TO ENVIRONMENTAL TOBACCO SMOKE AND RISK OF LUNG CANCER IN NON-SMOKING WOMEN FROM MOSCOW, RUSSIA David Za.rzmzE'*, Dimitri MnxLMovtTcta1, Galina ZEMC,YArrAYA', Zailil N. AtTAxov' and Paolo BoFFgrrAZ i tDeparhnentof'EpidemiologyandPreventlon,InstituteofCarcinogenests,RAMSAtN.BloklsinCancerResearchCent re, _ Moscow, Russia 2Unit ofEnvironmental Epidemiology, International Agency for Research on Caacer, Lyon, France kq l: The association between exposure to ETS and the risk of lung cancer in life-time non-smoking women was investigated by means of a hospital based case-control study in Moscow, Russia. The main importance of our study is that it was conducted on a population with a specific smoking pattern from which no information is available on health effects of ETS. A total of 189 incident cases of histologically confirmed lung cancer were identified in 2 principal cancer treatment hospitals in Moscow. A total of 358 female oncology patients from the same hospitals were selected as controls. The controls matched by the hospitals to the cases were similarly restricted to never-smokers. Women diagnosed with cancer of the upper respiratory organs were ineligible for selection as controls. Personal interviews of cases and controls were conducted in the hospital wards, using a closed-form struc- tured questionnaire. An elevated risk of lung cancer was observed in women whose husbands smoked. The odds ratio (OR) adjusted by age and education for husband's smoking was 1.53 (95% Cl, 1.06-2.21). Smoking by other members of the family, by colleague's, or by fathers in the women's childhood do not affect the risk of lung cancer. The risk is higher for women whose husbands smoke "papirosy" (OR 2.12; 95% Cl, 1.32-3.40), a special Russian type of cigarettes with a long mouthpiece, and usually very high levels of tar (>30 mg/cig) and nicotine (> 1.8 mg/cig). Our study suggests that thb association between exposure to ETS of the spouse and risk of lung cancer in non-smoking women is somewhat stronger for squamous-cell carcinoma (OR, 1.94; 95% Cl, 0.99-3.81) than for adenocarcinoma (OR, 1.52; 95% Cl, 0.96- 2.39). Int. f. Cancer 75:335-338, 1998. © 1998 4i91ey-Liss, Inc. Lung cancer is the most common malignant neoplasm in developed countries (Parkin et al., '1992), including Russia (Zaridze and Gurevicius, 7.986). Tobacco smoking being the main cause of this fatal disease, responsible for an estimated 90 to 95% of lung-cancer cases in men and 80 to 859o in women (TARC, 1985; Petoetal., 1994). Research on the hazards of smoking has turned recently to the association between exposure to environmental tobacco smoke (ETS) and adverse health events, itlcluding lung cancer. The ubiquitousness of tobacco smoke in homes, work places and public areas makes exposure to ETS virtually unavoidable. According to different studies, 39 to 63% of the non-smoking population is exposed to ETS (Friedman et al, 1983; Lebowitz et al., 1992). Epidemiological studies have provided evidence of a causal association between ETS and lung cancer ("Iredaniel et al., 1994; Pershagen, 7994). The increase in relative risk of lung cancer associated with ETS was typically observed in lifelong non- smoking women living with a smoking spouse. The results suggest that non-smokers exposed to ETS have an estimated 35% increased risk of lung cancer, as compared with non-smokers not exposed to ETS. The available studies, though consistent in the overallresultc, are in trtost cases too small in size to precisely quantify the magnitude of the effect and to address the dose-response relationship and the interaction with other risk factors for lung cancer. Moreover, no studies are available from Eastern Europe, and specifically from Russia, where the features of tobacco smoking are different from countries in which the effects of ETS on the risk of lung cancer have been studied. We here report the results of a hospital-based case-control study of lung cancer in lifetime non-smoking women residents of Moscow city. MATERIAL AND METHODS The sources of subjects for this study were the Moscow City Oncology Hospital and Cancer Research Center, 2 principal cancer treatment hospitals in Moscow. Cases consisted of all women with histologically confirmed primary carcinoma of the lung who were admitted to the hospital or who were treated at outpatient clinics at either of these facilities. An uniform histopathology review was undertaken to confirm diagnosis of primary lung cancer and to assess histology. For this study, cases were restricted to women who had never smoked cigarettes and who lived inMnscow city. A total of 189 incident cases were identifierl Because of weekly hospital contacts, all cases were interviewed within 2 or 3 days of admission to the hospital, before the start of any treatment. A total of 358 female oncology patients from the same hospitals were selected as controls. The control women matched by hospital to the cases were similarly restricted to never-smokers and Moscow residents. Women diagnosed with cancer of the upper respiratory organs were ineligible for selection as controls. In-person interviews of cases and controls were conducted in the hospital wards using a closed form, structured questionnaire. Data were collected on demographic characteristics (education and marital status), residential history, housing characteristics, oecupa- taoual history and ETS exposure, Occupational history included data on location, duration and potential deleterious exposures for all jobs within the last 20 years. Information on family history of cancer was also obtained. The conrpliance rate for both cases and controls was 100%. Information on ETS was derived from questions on smoking habits of the husband, other members of the household, and smoking habits of collegues at the work place. In addition, information on ETS exposure in childhood was sought by asking about parental smoking habits. Initially, occupational exposure was assessed by job title and listing possible carcinogenic exposures. However, few women were exposed. For the current analysis, we used only a crude job classification; office worker or manual worker. The odds ratio (OR) was used as a measure of association between exposure to factors of interest and lung cancer. The OR compares the odds of exposure of cases with that of controls. Multivariate logistic regression methods were used to compute OR and confidence intervals (CI) and to adjust for potential confound- ing variables (Breslow and Day, 1980). The Maitel-extension test was used to test for trend in the OR with exposure, using category ranks as the quantitative trend variable (Mante2 and Haenszei, 1959). All trend tests were 2-sided. *Correspondence to: Institute of Carcinogenesis, Kashirskoye shosse 24, IvIoscow, Russia. Fax: 7-095-324-12-05. E-mail: maz L~epidem.msk.su Received 18 June 1997; Revised 17 September J997 99035512

338 ZARIDZ.B FTALL cancer in women whose husbands smoked filter cigarettes, either for all histological types of lung cancer combined, or for squamous- cell cancer or adenocarcinoma. We identilied some potential methodological problems in our study, concerning some aspects of design of the study, and in particular the seleetion of controls. Although, in general, hospital- based studies are more prone to selection bias than population- based studies (Breslow and Day, 1980), they may offer less opportunity to recall bias and therefore differential exposure misclassification (1}edaniel et al., 1994). Misclassification of the non-smoking status of cases and controls (i.e., confounding by active smoking) is an important potential source of bias in studies of the health effects of ETS (9}edaniel et al., 1994). This problem is more likely to be present in studies including both lifetime non-smokers and long-term quitters (Brown- son et aL, 1992). The misclassification of smoking status is less likely to he present among never-smokers, though both cases and controls in our study could include a small fraction of light smokers. Misclassification of exposure to ETS is another important potential source of bias (']tedaniel et al„ 1994). We have not used a method of biochemical validation of ETS exposure, for example, measurement of urinary cotinine. However, these methods could only validate a recent history of exposure, which is clearly far less important than Iong-term exposure to ETS. Finally, differential misclassification of exposure (i.e., cases over-reporting ETS exposure as compared with controls) is likely to be less pronounced in our hospital-based study, since cases and controls alike were patients and were interviewed in their hospital wards. Au important potential problem that may arise in studies on ETS are potential confounders other than active smoking. We found no evidence that occupational exposure or diet affects the risk of lung cancer. However, in our smdy cases were less educated than controls. In the statistical analyses, therefore, adjustments were made for age and education. In addition, the results of our study suggest the importance of different indices of air pollution as risk determinants of lung cancer in non-smoking women. These results will be reported elsewhere. In conclusion, our study provides estimates of the effect of ETS on lung-cancer risk in Russian non-smoking women who are residents of Moscow. We found no increased risk for childhood exposure, a result that is in line with most available data. There was no association between exposure to ETS at the workplace and risk of lung cancer. The risk of lung cancer from exposure to ETS from the spouse was elevated. The use of several quantative indicators of ETS exposure reinforced these conclusions- There was evidence of a higher risk from ETS of cigarettes, without filter and "papirosy". Detailed analyses showed a stronger association of ETS from husbands' smoking with squamous-cell carcinoma. ACKNOWII:UGEMENTS The final statistical analysis was undertaken during D.M.'s one-month visit to the IARC Unit of Environmental Cancer Epidemiology, supported by UICC ICRETT fellowship 506. REFERENCES BLOT. W,J. and ERAmmNI, I IC, Passive smoking and lung cancer. J. nat. Concer Inst, 77, 993 10(]0 (1996). BRESLOw, N. and DAY, N., Statistical methods in cancer research-the analysis of case control studies. lARC Scientific Publication, 32, Lyon (1980). BROwnsoN, R.C., .ALAVANtA, M,C.R_. Hlx'K, E.T and Loy, T.S., Passive smoking and lung cancer in non-smoking women. Amer, J. publ. Hith, 82, 1525-153O(1992). CARDENAS, V.M., THIT, M,S., AUSTIN, H.. LALLY. C.A., CLARK, W.S.. GREENBERG. R.S. and HEATH. C.W., Environmental tobacco smoke and lung-cancer mortality in the American Cancer Society's Cancer Prevention study II. Cancer Causes Control, 8, 57-64 (1997), FoNTHAM, E.T.H., CORREA.. P, Wu-WILLIAMs, A., REYNOLns, P., GREEN- aeRG, R.S., Bt1FFP.R, PA., CHEN, V.W-, BOYn, P., AFTERMAF, T., Al15rIN, D.R. L[Pr, YY and GRSEN-eERG, S.D. Environmental tobacco smoke and lung cancer in non-smoking women. A multiecnter study. J. Am. med. A,us., 271, 1752-1759 (1994). FRtEnHAN, G.D., PcTrrrL D. and RAwoL, R.D., Prevalence and cnrrelates of passive smoking-Ame:.L publ. H1th, 73, 401-405 (1983). IARC MONOGRAPHS ON THE EVALlIAT1ON OF THE CARCINUI:ENIC RISK OF CHennCALS To HLMANS. Tobrrcco smnking. IARC, Lyon (1985). KALANDIDI, A., KATSl1lIYANNt, K.. VpROPJOLOII, N., SASTAS. G., SARACCI, R. and TRlnnorouLOS, D., Passive smoking and diet in the etiology of lung cancer in non-smokers. Cancer Causes Control, 1, U- 21 (1990). LE(IwITZ, M,D., QUACKEN903S, J.J., K22YZANOWSKL M., G'ROIRKE. M.K. and HAYCS, C., Multipollutant exposures and health responses to particulate matter. Arch, environ. Hith, 47, 71-7511992). MaNreL, N- and HAESSZm-, W.. Statistical aspects of the analysis of data from retrospective studies of disea.-ea J nor. Cancer lnrt, 22, 719-748 (1959). NATIONAL RESEARCH CoUNCIL, Enrironmvnral tobacl'll smoke: measurin,Q exposures and assessing healrh effects. National Academy Press, Washing- ton. DC (1986). PARKIV, D.M., Mom. C.S.. WHELA6, S.L., GAO, Y-T.. FERLAY, J. and POwELL, 1., Canaer incidence in five eontinents. VI. IARC Scientific Publication I2t1, JARC, Lyon (1992)- PcRSH.4GE.v. G., HmrnrC, Z. and SvENssos C., Passivce smoking and lung cancer in Swedish women. Aceeq J. Epidemiol., 125, 17-24 (1987), PER.vwcev, G„ Passive smoking and lung cancer. In.: J.M. Samet (ed), Epidemiology nf lung cancep pp. 109-130. Marcel Dekker, New York (1994)- PeTO, R.. LOPEZ, A,D., B(1RE11AM, J., THOM, 11. and HEATII, C., Mprtality from smoking in deveGtped countries, 1950-2000, Indirect Fatimates frorn National Vital Statistics, /CRFand WHO, Oxford University Press, Oxford. New York, Tokyo (1994). SARACCI. R. and Rtuou, E.. Passive smoking and lung cancer: current evidence and ongoing studies at the Jnternational Agency for Research on Cancer. Muration Res., 222, 1I7-127 (1989). 'IteEUAnEL, J., BoFFETTA, P, SARACCL R. and HLescn, A., Exposure to envimnmental tobacco smoke and risk of lung cancer: the epidemiological evidence. Furop. mspic J.,7,IH77-I888(1994). TRrcxorouLOS, D., KArt,NDtDI, A. and SPAnaos, I.., Lung cancer and passivesmoking: conclusion of Greek study. Il.ncet, H, 677-678 (1983). US DHHS, ENvIRONME4rAL PROTECIlON AGENCY, Respiratory health eJfects of passive smoking: lung cancer and other disorders. Smnking and Tobacco Control Monograph 4, NIH Publication 93-3605, Washington DC (1993). US Pcuuc I1eA1TH SERVICE, The health consequences of involuntary smoking. A report of the Surgeon General DHHS Publication (CDC) 87-N398, US Government Printing Office, Washington, DC (1986). WALD, N.J., NANCHAHAL, K., THOMFSON, $.G. and CCCKLE, H.S., Does breathing other people's tobacco smoke cause lung cancer? Brit. med. J., 293,1217-1222 (1986). ZARmzF D.G. and Guuevlclus, R., Lung-canecr pattern and trends in the USSR. In: D. Zaridze D. and R- Peto (eds.), Tobac[o-mar international health ha~mds, pp. 87-101, 1ARC Scientific PubGcation (1986d) 74, TARC, Lyan . ZARmzE, D.G., DvotRm, V.V.. KoeuArcov, VA, and PtsKLOe, V.V., Smoking pattern in Ihc USSR. In: Zarirl~e D. and R. Peto (eds.), '!'obacco -major brternational health im¢ard, pp. 75-80, IARC Scientific Publication 74, IARC, Lyon (19866). 99035515