Tobacco Documents Online

186 E. Rapiti et al. /Lung Cancer 23 (1999) 183-189 Table 2 Odds ratios (ORs) and 95% confidence intervals (Cls) for lung cancer and environmental tobacco smoke exposure during childhood, Chandigarh, India All subjects Case/control ORa 95 % Cl Ener exposed in childhood Women only Case/control OR° 95% CI No 27/93 11/51 Yes: 31/30 3.9 1.9-8.2 30/16 12 4.3-32 Cigarettes 20/9 12 4.2-34 19/4 40 9.2-171 Bidis 8/7 4.5 1.4-15 8/4 14 3.1-60 Chilum 9/21 1.3 0.5-3.5 9/11 4.2 1.2-14 To father 31/24 4.5 2.1 9.6 30/11 17 5.7 17 To mother 5/3 4.0 0.7-21 5/3 5.6 1.0-32 Weighted number of smakers >0-50.75 13/16 3.4 1.3-8.6 12/8 11 3.0-37 >0.75 18/14 4.5 1.8-11 18/8 13 4.0-41 Duration (years) 1-30.7 20/23 3.3 1.4-7.4 19/12 9.6 3.2-29 >30.7 11/7 6.0 1.9-19 11/4 18 4.3-80 ° Odds ratios adjusted for gender, age, residence and religion. ° Odds ratios adjusted for age, residence and religion. exposure did not appear to confound the associa- tion between childhood ETS exposure and lung cancer (OR of childhood exposure, after adjust- ment for adult exposure, 4.0, 95% CI =1.9-8.6). Table 3 shows the risk estimates for exposure to ETS from the spouse. The OR for lung cancer for living with a smoker spouse was 1.1 (95% CI = 0.4-2.6). When looking at the different tobacco types smoked by the spouse, however, cigarettes showed a strong association with lung cancer (OR = 5.1; 95% CI = 1.5-17), while no associa- tion was seen for bidi or chilum. A slightly lower risk was found among subjects whose ETS expo- sure had ceased ten or more years prior to the study. There was a decreasing though not statisti- cally significant trend in risk with increasing dura- tion of exposure and cumulative exposure. Restricting the analysis to women revealed the same pattern. Exposure to ETS at the workplace was associ- ated with an OR of 1.1 (95% CI = 0.3-4.1). There was an increased risk with increasing years of exposure, however the trend was not statistically significant. The results of the analysis for ETS in vehicles showed an OR of 5.2 (95% CI = 1.9-14). No trend with duration of exposure was seen. The analysis performed according to the histo- logical type of the cancer showed similar risk estimates for adenocarcinoma and squamous or small cell carcinomas. The OR associated with ETS exposure during childhood was 4.7 (95% CI = 1.7-13) for adenocarcinomas and 4.1 (95% CI = 1.6-11) for squamous/small cell carcinomas. No increased risk for adenocarcinomas or squamous or small cell carcinomas was observed with spousal ETS exposure (OR = 1.0; 95% CI = 0.3-3.2 and OR = 1.2; 95% CI = 0.4-2.5, respectively). 4. Discussion This study found that exposure to environmen- tal tobacco smoke may account for a sizable part of the lung cancer cases among non-smokers in the population of Chandigarh. The association between ETS and lung cancer was found to be particularly strong for exposure during childhood, co increasing with both the number of smokers and w duration of exposure. The observed risk arose ~, ~ Lti O %10

E. Raprtr et al. /Lung Cancer 23 (1999) 183-189 185 pipe made of clay. Tobacco is burnt along with molasses and coal and smoked from the other end either directly at the mouth or through a long pipe with the smoke passing through a water container. For the purpose of the study, we ap- plied a weight of I g of tobacco to cigarette, 0.5 to bidt and 4 to chilum. Subjects were considered lifetime non-smokers if they had never consumed more than 400 cigarette-equivalents over their lifetime. To assess the effect of individual variables un- conditional logistic regression models were used. Exposure to ETS was assessed as 'never' versus `ever' exposed in different settings as well as using the quantitative variables divided in categories defined according to the 75th percentile of the distribution among the controls. The odds ratios (ORs) were adjusted for sex, age, religion and residence. Models adjusted for age, religion and residence were also calculated restricting the anal- ysis to women. Some of the analyses were re- peated according to the histological type of cancer. 3. Results A total of 58 non-smoker cases (17 males and 41 females) were included in the data analysis, after exclusion of three patients with unknown age. There were 16 (28%) squamous cell car- cinomas, 11 (19%) small cell carcinomas, 29 (51%) adenocarcinomas, and one case of other cell type. After exclusion of one control with unknown age, 123 controls (56 males and 67 females) were in- cluded in the data analysis. The distribution of the selected characteristics of the study population is presented in Table 1. Due to the lack of individual matching between oases and controls, the distribu- tion by gender was not even, with 71% of cases being females and only 55% of controls. The two groups had similar age distribution. Approxi- mately 50% of the lung cancer cases and 25% of the controls were residents in a rural area. The majority of both cases and controls were Hindus. Regarding occupation, the majority of cases (70%) were not working (housewives, unem- ployed, etc.), compared with only 41% of con- trols. Cases tended to be engaged in agricultural and clerical work while controls had primarily clerical or professional jobs. ORs for lung cancer associated with ETS dur- ing childhood are shown in Table 2 for the whole study population and for women only. Elevated lung cancer risks were observed for lifetime non- smokers exposed to passive smoking during child- hood (OR = 3.9; 95% CI = 1.9-8.2). The risk was particularly high for those ever exposed to cigarettes (OR = 12; 95% CI = 4.2-34), but those exposed to bidis also showed an increased risk (OR = 4.5; 95% CI= 1.4-15). High risk estimates were shown with childhood exposure to ETS from either the father or the mother. Elevated risk was associated with an increasing number of smokers in the household and duration of exposure. When the analysis was restricted to women stronger associations were observed (Table 2). Adult ETS Table I Distribution by gender, age at admission, religion, residence and occupation of cases and controls Charaeteristics Cases Controls N % N % Gender Male 17 29.3 56 45.5 Female , 41 70.7 67 54.5 Age at admission (years) <49 21 36.2 47 38.2 50-59 21 36.2 32 26.0 60-69 12 20.7 23 18.7 70+ 4 6.9 21 17.1 Religion Hindu 43 74.1 112 91.1 Muslim 3 5.2 4 3.2 Sikh 12 20.7 7 5.7 Residence Rural 28 48.3 31 25,2 Urban 30 51.7 91 74.0 Unknown - - 1 0.8 Occupation Professional - . 3 5.2 17 13.8 Clerical, sales and services 8 13.8 47 38.2 Agricultural 7 12.1 7 5.7 Non-occupied 40 69.0 51 41.5 Unknown -- - - - 1 0.9

E. Rapiti et al. /Lung Cancer 23 (1999) 183-189 187 from exposure to cigarette and bidt smoke. ETS exposure from spouses during adulthood was as- sociated with an increase in risk for those exposed to cigarettes. Exposure in vehicles also resulted in an increased risk of lung cancer. A positive association between exposure to pas- sive smoking at young ages and lung cancer has been reported previously [10-12], although never showing an effect of comparable magnitude. A relative risk of 1.36 for lung cancer associated with maternal smoking was found by Correa et al. [10]. In the study conducted by Janerich et al. [11], the risk associated with household exposure to 25 or more smoker-years during childhood was 2.07. Stockwell et al. [12] reported a risk estimate of 2.4 for women exposed to ETS during childhood for more than 22 years. In contrast, other studies failed to find any relation [13-15]. This lack of consistency has been partly explained by the difficulties in recalling past exposure, particularly from surrogate respondents [16,17]. In this study, in contrast, all of the interviews were conducted with the subjects themselves, thereby possibly in- creasing the reliability of the information. The discrepancy between the results of this study and those conducted in industrialized countries could in part be explained by cultural and social differ- ences which result in different ETS exposure. The tar and nicotine content of smoke of Indian cigarettes, including filter cigarettes, and bidi is higher than the content of Western cigarettes [18,19]. Household conditions, personal practices and habits differ substantially between India and Western countries. The presence of extended or joint families in the same household, furthermore, could result in an early exposure to ETS from more than one family member [20]. This study did not find an association between risk of lung cancer and exposure to a smoking spouse during adulthood when considering all type of smoking products together. However, the risk estimate for those exposed to cigarette smoke only was high and consistent with results from other studies [21,14,22]. The small size of the study population did not permit a more detailed analysis of the subgroup of subjects who were only exposed to cigarette smoking. The results showing an increased lung cancer risk following Table 3 Odds ratios (ORs) and 95% confidence intervals (CIs) for lung cancer and environmental tobacco smoke exposure from spouse, Chandigarh, India All subjects Women only Case/control OR' 95% CI Ever exposed from spouse Case/control ORb 95% CI No 45/101 28/46 Yes: 13/22 1.1 0.5-2.6 13/21 1.2 05-2.9 Cigarettes 11/5 5.1 1.5-17 11/5 5.3 1.6-18 Bidis 1/12 0.1 0.01-1.2 1/12 0.1 0.02-1.2 Chilum 1/4 0.3 0.02-3.0 l/4 0.3 0.03-3.5 Duration (h/day x years) 1-80 11/15 1.4 0.5-3.6 11/15 1.5 0.6-3.8 > 80 2/4 0.7 0.1-5.1 2/4 0.7 0.1-4.4 Cumulatiue exposure (pack years) 1-128 1019 2.3 0.8-6.9 10/9 2.6 0.9-7.9 >128 3(10 0.4 0.1-1.9 3/10 0.4 0.1-1.8 Time since last exposure (years) 1-9 10/15 1.3 0.5-3.4 10115 1.3 0.5-3.5 10+ 3/4 0.9 0.2-4.7 3/4 1.0 0.2-4.9 00 "Odds ratios adjusted for gender, age, residence and religion. W -~1 ^ Odds ratios adjusted for age, residence and religion. ~ d ~10 co

188 E. Aapiti et aL /Lung Cancer 23 (1999) 183-189 exposure in vehicles are in agreement with the data from the study conducted by Fontham et al., reporting an increased risk in social settings [14]. Some methodological issues should be consid- ered in evaluating this study. First, the possibility of a bias due to misclassification of smokers as non-smokers must always be considered in studies on ETS and lung cancer. In this study, no direct validation of smoking status was conducted. However, the decision to set the cutoff of 400 lifetime cigarette-equivalents as the limit for inclu- sion of the subjects in the study as non-smokers reduced the risk of such misclassification. Second, no information was collected on other potential confounders such as educational level or diet. However, the confounding effect alone is unlikely to explain the observed results. Third, the catch- ment area of the Chandigarh hospital was par- tially different for cases and controls. While patients with lung cancer come to the Department of Pulmonary Medicine from both rural and local urban areas, the population from which the con- trols were selected is largely urban. In order to reduce the effect that such a selection process could have on the results, residence as well as religion were always controlled for in the analysis. Fourth, the use of hospital controls might have introduced a bias in the results. We attempted to control this problem by using a mixed group of controls, including hospital patients and visitors. The use of each group of controls separately resulted in similar risk estimates for exposure to childhood ETS, while the risk from exposure to spousal ETS was increased when hospital patients were used as controls (OR = 2.1, 95% CI = 0.7- 6.4), but not when visitors were used as controls (OR - 0.6, 95% CI = 0.2-1.7). Historically, the Indian population has shown a low risk for lung cancer as compared with that of other countries. The incidence rates range from 4.7 to 15.2 cases/100 000 population and 1.1 to 2.6 cases/100000 population, for males and females, respectively [23]. The low prevalence of smoking that characterizes India could partly explain this low incidence of lung cancer, but also differences in the mode of smoking, or inherent differences in the population's genetic susceptibility might play a role. Smoking of both cigarettes and bidi has been demonstrated to be a strong risk factor for lung cancer among the Indian population [24,20,25-28]. While active bidi smokers appear to have the same if not a higher risk of lung cancer than cigarette smokers [25,26], in this study exposure to bidt smoke seemed to exert a some- what lower risk of lung cancer among non-smok- ers. A bidi is smaller in size than a cigarette and the smoke emitted from it is also less than that from a cigarette. ETS exposure from bidis might, therefore, be small. We have no information on sources of indoor air pollution other than ETS. High particulate levels have been reported from India after use of wood or charcoal for cooking [29]. However, natural gas and kerosene are the main fuels used for cooking and heating in Chandigarh, and coal, wood and biomass are rarely used. Furthermore, sources of indoor air pollution other than ETS would confound the results of our study only if they were associated with ETS exposure. In conclusion, this study demonstrated that ex- posure to environmental tobacco smoke, long reo ognized as a risk factor for lung cancer in countries with high prevalence of smoking, may also be an important cause of lung cancer among nonsmokers in India. In comparison with other studies, in this study the risk appeared particu- larly high when the exposure occurred early in life. Further studies need to be carried out in India and in other similar settings to confirm and elucidate the role of exposure to ETS during childhood and adolescence in causation of lung cancer. References [t] US Department of Health and Human Services, Environ- mental Protection Agency. Respiratory Health Effects of Passive Smoking: Lung Cancer and Other Disorders (Smoking and Tobacco Control Monograph 4). Washing- ton DC: DHHS (NIH Publication No. 93-3605), 1993. [2] Hackshaw AK, Law MR, Wald NJ. The accumulated evidence on lung cancer and environmental tobacco smoke. Br Med J 1997;315:980-8. [3] Koo CL, Ho JHC. Worldwide epidemiological patterns of lung cancer in non smokers. Int J Epidemiol 1990:19(Suppl. l):S 14-23. co W -L~- CN O \0 \0

Lung Cancer 23 (1999) 183-189 Passive smoking and lung cancer in Chandigarh, India Elisabetta Rapiti a•b, Surinder K. Jindal °, Dheeraj Gupta °, Paolo Boffetta " a Unit of Environmental Cancer Epidemiology International Agency for Research on Cancer, 150 cours Albert-Thomas, 69008 Lyon, France ' Department of Epidemiology, Lazio Region, Rome, Italy ° Postgraduate Institute of Medical Education and Research, Chandigarh, India Received 3 August 1998; received in revised form 10 February 1999; accepted 18 February 1999 y Elsevier 307. Lung ease send icas, New Abstract The aim of this study is to assess the relationship between exposure to environmental tobacco smoke (ETS) and lung cancer in non-smokers, a case-control study among lifetime non-smokers was conducted in Chandigarh, India. Cases consisted of 58 non-smoking histologically confirmed lung cancer patients; two controls for each case were selected, one among other patients admitted to the wards and one among the visitors to hospital patients. Subjects were asked about ETS exposure from different tobacco products in childhood and in adulthood at home, at the work place and in vehicles. Multivariate logistic regression analysis was used to assess the effects of the ETS exposure variables on lung cancer. Exposure to ETS during childhood was strongly associated with lung cancer (odds ratio (OR) = 3.9; 95% confidence interval (CI) = 1.9-8.2), the effect mostly arising from exposure to cigarettes smoke. The excess risk was observed with either a smoking father or mother. An increasing risk was found with increasing number of smokers and duration of exposure. Restricting the analysis to women produced higher estimates of the risk. No increased risk was found with exposure to a smoking spouse, except for those exposed only to cigarette smoke (OR-5.1; 95% CI=1.5-17). A weak association was seen between lung cancer and ETS exposure at the workplace, which increased with the number of years of exposure. Exposure in vehicles also was detected as a risk factor for lung cancer in non-smokers. This study suggests that ETS exposure may be a strong risk factor for lung cancer also in India, a country with low prevalence of smoking and, therefore, low rates of lung cancer. Other studies need to be conducted in similar settings to confirm the role played by ETS exposure early in life in the causation of lung cancer. 0 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Lung cancer; Case-control studies; Environmental tobacco smoke; India * Corresponding author. Tel.: +33-0-7273-8441; fax: +33-0-7273 8342. E-mail address: boffetta@iarc.£r (P. BofPetta) 0169-5002/99/$ - see front matter ® 1999 Elsevier Science Iretand Ltd. All rights reserved. Pn:S0169-5002(99)00013-6 \10

184 1. Introduction E. Rapiti et af. /Lung Cancer 23 (1999) 183-189 Two non-smoker controls were selected for each patient. One control was selected from among patients admitted to the hospital. Patients with prolonged hospitalization (greater than a month), or with a disease associated with active or The role of environmental tobacco smoke (ETS) in the causation of lung cancer has been repeatedly evaluated over the last years. The US Environmental Protection Agency in 1993 classified ETS as a known human lung carcinogen (group A) [1]. Recently, a pooled analysis of published studies estimated an excess risk of lung cancer of 24% (95% confidence intervals (CIs) 13-36%) in non-smokers who lived with a smoker as compared to those who lived with a non- smoker [2]. Despite the large number of studies showing that ETS is a risk factor for lung cancer, the vast majority of these studies has been conducted in developed and developing countries with high prevalence of smoking, and overall high incidence of lung cancer [3,4]. In the present paper we describe the results of a case-control study on lung cancer in non-smokers undertaken in Chandigarh, India. The study utilizes the same methodology as an international collaborative case-control study on lung cancer in non-smokers in various countries coordinated by the Interna- tional Agency for Research on Cancer (IARC) [5]. Although lung cancer incidence in India is lower than in East Asian or Western countries (3/10 000 person-years [3]), it is of interest to look at passive smoking in a country characterized by a different pattern of tobacco consumption [6]. The historical Indian smoking products are bidis and chilum, cigarette smoking being a relatively more recent habit, although steadily increasing over the past 40 years [7]. The prevalence of smoking in the Indian population, however, is still among the lowest in the world [3]. 2. Materials and methods Lung cancer cases were identified at the Depart- ment of Pulmonary Medicine of the Postgraduate Institute of Medical Education and Research, in Chandigarh, Northern India, from newly diag- nosed primary lung cancers during the period July 1991-June 1992. Lung cancer cases were eligible for the study only if confirmed through histology or cytology. passive smoking, alcohol or diet were excluded. The second control was chosen from among the patient's visitors. No matching procedures were conducted between cases and controls. Subjects were interviewed by trained personnel in the hospital, using a questionnaire designed to gather information on demographic factors, life- time passive smoking and occasional active smok- ing, residence, and occupation [5]. The ability of this questionnaire to accurately ascertain ETS exposure, had previously been validated in 13 centers, including Chandigarh [8]. ETS exposure was assessed by inquiring about residential expo- sure to tobacco smoke from cohabitants in child- hood and in adulthood, including the amount, intensity, duration, and types of tobacco products smoked by each smoking cohabitant. In addition, ETS exposure at each workplace, in vehicles, and other indoor exposures were assessed. The age of 18 years was used as the cutoff between childhood and adulthood. Quantitative variables used to express child- hood ETS exposure included the number of smokers in the household and the duration of exposure, expressed as the number of years of exposure weighted for each type of smoker (mother = 1, father = 0.75, other adults = 0.25) [9]. Quantitative variables of ETS exposure from the spouse included: (i) the product of the number of years and the number of hours per day of exposure; (ii) the cumulative exposure, expressed as pack-years; and (iii) the number of years since last exposure to spousal smoke, for subjects whose spouse stopped smoking. All of these vari- ables were calculated by including exposure to all types of tobacco products expressed as cigarette equivalents. The most common smoking habits among Indi- ans are cigarettes, bidis and chilum. The bidi con- tains a relatively small amount of locally grown tobacco (0.2-0.3 g), which is sun-dried and -cured and then loosely wrapped in a dried temburni leaf (Dyospyros melanoxylon). The chilum resembles a 00 W ~ I W O ~ CJl

E. Aapiti et al. JLung Cancer 23 (1999) 183-189 [4] Wu All. Environmental tobacco smoke II: Lung cancer. In: Steenland K, Savitz D, editors. Topics in Environmen- tal Epidemiology. New York: Oxford University Press, 1997:227-55. [5] Boffetta P, Agudo A, Ahrens W, at al. Multicenter case- control study of exposure to environmental tobacco smoke and lung cancer in Europe. J Natl Cancer Inst 1998;90:1440-50. [6] Bhonsle RB, Murti PR, Gupta PC. Tobacco habits in India. In: Gupta PC, Hamner III JE, Murti PR, editors. Control of Tobacco Related Cancers and Other Diseases (International Symposium, 1990, Bombay). Oxford: Ox- ford University Press, 1992:25-46. [7] Directorate of Tobacco Development, Ministry of Agri- culture and Irrigation, Government of India. Tobacco in India: A Handbook of Statistics, 1983. [8] Riboli E, Preston-Martin S, Smacci R, at al. Exposure of nonsmoking women to environmental tobacco smoke: a 10-country collaborative study. Cancer Causes Control 1990;1:243-52. [9] Jarvis MJ, Mc Neill AD, Bryant A, Russel MA. Factors determining exposure to passive smoking in young adults living at home: quantitative analysis using saliva cotinine concentrations. Int J Epidemiol 1991;20:126-31. [10] Correa P, Pickle LW, Fontham E, Lin Y, Haenszel W. Passive smoking and lung cancer. Lancet 1983;2:595-7. [11] Janerich DT, Thompson WD, Varela LR, et al. Lung cancer and exposure to tobacco smoke in the household. New Eng] J Med 1990;323:632-6. [12] Stockwell HG, Goldman AL, Lyman GH, et al. Environ- mental tobacco smoke and lung cancer risk in nonsmok- ing women. J Natl Cancer Inst 1992;84:1417-22. [I3] Brownson RC, Alavanja MCR, Hock ET, Loy TS. Pas- sive smoking and lung cancer in nonsmoking women. Am J Publ Health 1992;82:1525-30. [14] Fontham ETH, Correa P, Reynolds P, et al. Environmen- tal tobacco smoke and lung cancer in nonsmoking women. J Am Med Assoc 1994;271:1752-9. [15] Pershagen G, Hrubec 4 Svensson C. Passive smoking and lung cancer in Swedish women. Am J Epidemiol 1987;125:17-24. [16] Coultas DB, Peake GT, Samet JM. Questionnaire assess- ment of lifetime and recent exposure to environmental tobacco smoke. Am J Epiderruol 1989;130:338-47. [17] Pron GE, Burch JD, Howe GR, Miller AB. The reliabil- ity of passive smoking histories reported in a case-control study of lung cancer. Am J Epidemiol 1988;127:267-73. [18] Hoffmann D, Sanghvi LD, Wynder EL. Chemical analy- sis of Indian bidi and American cigarette smoke. Int J Cancer 1974;14:49-53. [19] Sanghvi LD, Jayant K, Pakhale SS. Tobacco use and cancer in India. World Smoking Health 1980;5:4-10. [20] Jindal SK, Malik SK, Datta BN. Lung cancer in North- ern India in relation to age, sex and smoking habits. Eur J Resp Dis 1987;70:23-8. [21] Cardenas VM, Than MJ, Austin H, et al. Environmental tobacco smoke and lung cancer mortality in the American Cancer Society's Cancer Prevention Study II. Cancer Causes Control 1997;8:57-64. [22] Trichopoulos D, Kalandidi A, Sparros L. Lung cancer and passive smoking: conclusion of Greek study. Lancet I983;2:677-80. [23] Parkin DM, Whelan SL, Ferlay J, Raymond L, Young J, editors. Cancer Incidence in Five Continents, Vol. VII (IARC Scientific Publications No. 143). Lyon: Interna- tional Agency for Research on Cancer, 1997. [24] Jindal SK, Behera D. Clinical spectrum of primary lung cancer-review of Chandigarh experience of 10 years. Lung India 1990;8:94-8. [25] Jussawalla DJ, Jain DK. Lung cancer in Greater Bom- bay: correlations with religion and smoking habits. Br J Cancer 1979;40:437-48. [26] Notani P, Sanghvi LD. A retrospective study of lung eancer in Bombay. Br J Cancer 1974;29:477-82. [27] Notani PN, Shah P, Jayant K, Balakrishnan V. Occupa- tion and cancers of the lung and hladden a case-control study in Bombay. Int.J Epidemial 1993;22:185-91. [28] Sankamnarayanan R, Varghese C, Duffy SW, Padmaku- mary G, Day NE, Nair MK. A case-control study of diet and lung cancer in Kerala, South India. Int J Cancer 1994;58:644-9. [29] Smith KR, Liu Y. Indoor air pollution in developing countries. In: Samet JM, editor. Epidemiology of Lung Cancer. New York: Marcel Dekker, 1994:151-84.