Tobacco Documents Online

LUNG CANCER IN NONSMOKING CHINESE WOMEN: A CASE CONTROL STUDY Wang, Tianjue; Zhou, Baosen; and Shi, Jinpu China Medical University, Shenyang, China Abstract The importance of risk factors for lung cancer in lifetime nonsmoking women was investigated in a case-control study in the urban area of Shenyang, China, between April 1992 and May 1994. One hundred and thirty-five newly diagnosed lung cancer cases and an equal number of controls, matched for age and sex, were enrolled and interviewed by trained personnel who administered a standardized questionnaire. The histopathological cell type was dominated by adenocarcinoma (54.5%), followed by small cell carcinoma (20%), squamous cell carcinoma (16.4%), and others (9.1%). The data were analyzed using the Mantel-Haenszel method and by multivariate logistic regression analysis. The odds ratio (OR) and confidence interval (CI) associated with cooking oil vapors and with family history of cancer were 3.79 (95% CI, 2.29- 6.27) and 2.29 (95% CI, 1.01-5.17), respectively. No association was found between exposure to passive smoke, presence of previous lung diseases, and other variables. Introduction Study of the etiology and risk factors for lung cancer in nonsmoking women has received increasing worldwide attention in recent years (1). In China, a few studies have been performed on

the relationship between lung cancer and a number of risk factors. These studies have generated variable and equivocal results. Herein we report the results of a study aimed at exploring the influence of indoor air pollution, including exposure to passive smoke, and a number of other suspected risk factors, on lung cancer. Materials and Methods one hundred and thirty-five cases of newly diagnosed and eligible cases of primary lung cancer (according to the International Classification of Diseases, Ninth Revision, ICD-9 code 162, reference 2) were collected and identified in eighteen hospitals in the city of Shenyang between April 1992 and May 1994. All subjects, who ranged in age from 35-69 years and were lifetime nonusers of tobacco, were interviewed in person in the hospital by trained personnel within two weeks of diagnosis. Every case enrolled in this study was diagnosed by review of relevant medical records, chest x-ray and CT films, and cytologic and histologic slides (in 57.2% of the cases). All reviews were studied and confirmed by senior pathologists or clinicians. One hundred and thirty-five female controls, matched for age (+/- 5 years), were randomly selected from the general population located in urban areas of Shenyang. The number and age distribution of the controls were determined in advance based on the number and age distribution - 2 -

of reported primary lung cancer cases existing in the Liaoning provincial anti-epidemic station between 1988-1989. Two controls in the appropriate age group were randomly selected. If the first control was absent or not eligible for interview during the study, the second control was accepted as the alternative. During the interview, a structured questionnaire was administered to obtain information on demographic characteristics, exposure to tobacco, dietary and cooking practices, the type of fuel used, general medical conditions, history of previous lung diseases, history of cancer (including lung cancer) within the family, menstruation and pregnancy/child bearing, and job history. Completed questionnaires and relevant medical information were checked for accuracy by a supervisor, then coded and computerized. The Mantel-Haenszel method and multivariate logistic regression analysis were used to estimate the odds ratio (OR) for different risk factors and the statistical significance of their association with lung cancer. Population attributable risk (PAR) estimates were also computed for relevant risk factors. Results One hundred and thirty-five lung cancer patients were identified and interviewed. They were all lifetime nonusers of tobacco. None of them refused interview. More than half (57.2%) - 3 -

of the cases were diagnosed by tissue biopsy or cytology, and the other cases by medical symptoms and repeated X-ray films. Among the cases diagnosed pathologically or cytologically, the distribution of histologic cell type was as follows: adenocarcinoma was the predominant cell type accounting for 54.5% of the cases, 16.4% were squamous cell carcinoma, 20.4% were small cell or oat cell carcinoma, and 9.1% were a mixture or undifferentiated carcinoma. An equal number of controls were also interviewed. The distribution by age and marital status was generally similar between cases and controls. Exposure to Passive Smoke overall, no significant increase in risk was observed, for exposure to passive smoke or environmental tobacco smoke (ETS). Table l shows that for workplace exposure, the OR was 0.89 (95% CI, 0.45-1.77). For childhood exposure (exposure before marriage), the OR was computed to be 0.91 (95% CI, 0.55-1.49). Exposure to ETS from a spouse who smoked was not associated with a significantly increased risk of lung cancer (OR = 1.11, 95% CI = 0.65-1.88). Table 1. Association of different risk factors with lung cancer in nonsmoking women. - 4 -

Variable Case Control OR 95% CI PAR% (n=135) (n=135) Work place exposure 113 115 0.89 0.45-1.77 Childhood exposure 80 83 0.91 0.55-1.49 Adult exposure 92 89 1.11 0.65-1.88 Coal 100 107 0.75 0.43-1.31 Kang 88 89 0.97 0.64-1.48 Cooking oil vapor 77 35 3.79 2.29-6.27 42 Coal smoke 65 38 2.37 1.44-3.91 28 Previous lung disease 29 19 1.67 0.89-3.14 Family cancer history 19 9 2.29 1.01-5.17 8 To further investigate whether exposure to ETS was associated with lung cancer risk in nonsmoking women, the relative risk associated with the number of years living with a husband who smoked, or with the number of cigarettes smoked per day by a husband who smoked, was also calculated (Tables 2 and 3). No significant increase in lung cancer risk was found in either case. Table 2. Relative risk of lung cancer among nonsmoking women associated with years lived with a smoking husband - 5 -

Years lived with Case Control OR 95% CI smoking husband (n=135) (n=135) <20 65 70 20- 21 16 1.41 0.68-1.94 30- 32 32 1.08 0.58-2.00 40- 17 17 1.08 0.37-3.14 Table 3. Relative risk of lung cancer among nonsmoking women associated with the number of cigarettes smoked per day be smoking husband Number of Cigarettes Case Control OR 95% CI smoked per day (n=135) (n=135) 0 43 49 1- 4 13 0.35 0.11-1.12 10- 45 38 1.35 0.75-2.45 20- 43 35 1.40 0.76-2.56 Cooking Practices Soybean oil is the oil used most often for cooking in Shenyang. When cooking practices that frequently or sometimes - 6 -

generate cooking oil vapor were compared with those that generate little or no cooking oil vapor, a statistically significant increased risk of lung cancer was observed, OR = 3.79, 95% CI, 2.29-6.27. Likewise, exposure to coal smoke during cooking was associated with a significant increase in risk for lung cancer, OR = 2.37, 95% CI, 1.44-3.91 (Table 1). Family history of cancer Family cancer history, which refers to history of lung cancer or other forms of cancer in next-of-kin relatives, is significantly associated with an increase in lung cancer, OR = 2.29, 95% CI, 1.01=5.17 (Table 1). After stratifying subjects into a lung cancer group (group 1) and an other-than-lung cancer group (group 2), the association between family cancer history for each subgroup and risk for lung cancer in nonsmoking women was no longer statistically significant. A positive association, however, remains. The OR for group 1 was 3.64, 95% CI, 0.81-16.23 and the OR for group 2 was 2.12, 95% CI, 0.84-5.35, respectively. Previous lung disease Previous lung diseases refer to tuberculosis, chronic bronchitis, pneumonia, emphysema, lung abscess, and asthma. The overall relative risk between previous lung diseases and risks for - 7 -

lung cancer in nonsmoking women was computed to have an OR of 1. 67, 95% CI 0.89-3.14. A history of tuberculosis alone produced an OR of 1,39, 95% CI, 0.94-3.04. Thus, no statistically significant association was observed between previous lung disease and risk of lung cancer in nonsmoking women. Coal Coal was the most common fuel used in Northeast China. However, in this study, its use was not associated with an increase in risk for lung cancer in nonsmoking women (OR = 0.75, 95% CI 0.43-1.31). Rang "Kang" are made of brick and are often heated by burning coal in the winter. Prior to the modernization of China, kang were traditionally the most common form of bed as well as room heating device in the city of Shenyang. Use of kang was not significantly associated with lung cancer risk in nonsmoking women, OR = 0.97, 95% CI, 0.64-1.48 (Table 1). Multivariate nonconditional logistic regression analysis - 8 -

When multivariate analysis was applied to all variables examined in this study, only two -- oil vapor during cooking (OR = 4.02, 95% CI, 2.38-6.78) and family cancer history (OR = 3.07, 95% CI, 1.30-7.26) -- were found to be statistically significant risk factors for lung cancer in nonsmoking women. Coal smoke, which was a risk factor in the univariate analysis, no longer appears in the equation. None of the other factors appear in the equation, indicating that they are not risk factors for lung cancer in this population. Discussion The relationship between cigarette smoking and lung cancer has been well established. The risk factors for lung cancer in nonsmoking women, on the other hand, are still unclear (1, 3-7) and are the primary focus of this investigation. Possible etiologic factors being considered include: exposure to ETS, coal smoke pollution, previous lung disease, family history of cancer, and others (Table 1). The results of our study suggest that exposure to cooking oil vapors and family history of cancer are the primary risk factors. In northeast China, soybean oil is the most commonly used oil for cooking. When cooking oil is poured into a cooking utensil and heated to a high temperature, the kitchen and living room may - 9 -

become smoky. Moreover, exposure to oil smoke often becomes unavoidable because of cooking practices frequently used by Chinese females. Gao et al. (5) and Wong et al. (8) previously reported that the risk of lung cancer among women was increased by various modes of exposure to cooking oil vapors. It is therefore reasonable to suggest that cooking oil vapors play an important role in increasing the risk of lung cancer in nonsmoking women. The notion that lung cancer may be etiologically related to cooking oil vapor is biologically plausible since extracts prepared from volatile condensates of rapeseed oil and soybean oil heated at 270°C show mutagenicity based on the Ames test (9), i.e., giving positive results when tested in S9-activated TA98 strain of Salmonella. Genetic factors also appear to be involved in the occurrence of lung cancer, as suggested by the increased risk associated with family cancer history in this study. Indeed, genetic make-up as well as the fact that families often share identical environments and are exposed to similar indoor air pollutants may contribute to the tendency for lung cancer to cluster in a family (10). Whether exposure to ETS is a risk factor for lung cancer has been a subject of controversy (1, 3-7). In the present study, we found that exposure to ETS was not a risk factor for lung - 10 -