Tobacco Documents Online

a I I I I I I I I I I I I I I I I I ANALYSES OF SEX DIFFERENTIALS IN RISK FACTORS FOR PRIMARY LUNG ADENOCARCINOMA Shen Xiao-bin¢*, Wang Guo-xiong*, Xiang Long-sheng* and Huang Yuan-zhu** * Nanjing Railway Medical College, Nanjing, China ** Nanjing Medical University, Nanjing, China Abstract To analyze potential sex differences in risk factors for primary lung adenocarcinoma, a case- control study was carried out in Nanjing, China. One hundred and eighty cases (100 males, 80 females) involving Nanjing residents who had lived in an urban area for at least 20 years and who had been diagnosed with primary lung adenocarcinoma were analyzed. Age- and sex-matched controls were identified from healthy neighbors of cases. Information on possible exposure to risk factors for the past 20 years before diagnosis was obtained by trained interviewers and included: cigarette smoking index (average number of cigarettes smoked per day times number of years smoked), passive smoking, family history of lung cancer and other tumors, chronic bronchitis, pulmonary tuberculosis, occupational exposure to cooking oil fumes, home exposure to cooking oil fumes, cooking practices, housing conditions, types of domestic fuel (quantified by an index with a weighted average), heating from coal stoves in the winter, etc. Conditional logistic regression analyses were performed in order to identify risk factors and to estimate the relative risks (RR) of selected factors. Population attributable risk (PAR) estimations for various risk factors were also computed. The data suggest that exposure to cooking oil fumes in the home, chronic bronchitis, and family history of tumors are the most common risk factors for lung adenocarcinoma in both men and women. The relative risks (RR) of the above three factors for male adenocarcinoma were 2.84, 2.30 and 4.89, respectively. In females, the same three factors had RR of 3.20, 3.23, and 4.23, respectively. PAR of these factors were 42.78%, 20.36%, and 17.5%, in males, and 53.41 %, 14.68%, and 19.0%, in females. In addition to the three most common risk factors, cigarette smoking index is also a risk factor for male lung adenocarcinoma, with a RR of 1.01 and a PAR of 27.69%. For females, another risk factor is the use of a coal stove for winter heating, which produced a RR of 2.29 for female adenocarcinoma and a PAR of 17.59%. These data suggest that exposure to cooking oil fumes is a major risk factor for lung adenocarcinoma in the city of Nanjing and could conceivably, in part, account for the high incidence of lung adenocarcinoma in Chinese women. In separate studies, the mutagenicity of cooking oil fumes has been demonstrated. Taken together, it seems reasonable to propose that lung adenocarcinoma may be primarily induced by exposure to cooking oil fumes. Additional studies must be performed to further test this hypothesis. I

I Introduction The incidence of lung cancer in China has been on an apparent increase in recent years. According to data published in many parts of the world, ethnic Chinese women, known to have a low smoking rate, also have high incidence of lung cancer, primarily pulmonary adenocarcinoma(1,2). Numerous epidemiological studies on risk factors for adenocarcinoma have been conducted in different regions of China. To examine the risk differences in lung cancer due to sex difference, we conducted a pair-matched case-control study of 180 primary pulmonary adenocarcinoma cases in Nanjing during 1986-1993. Materials and Methods This study was based on the retrospective pair-matched case-control study method. Selection of Cases: Primary adenocarcinoma cases, confirmed by analysis of pathological sections or exfoliated cells, were drawn from Nanjing municipal hospitals during 1986-1993. All 180 cases were Nanjing residents of over 20 years, including 100 men and 80 women. Selection of Controls: Healthy controls were residents of Nanjing for 20 years and were randomly selected from the same neighborhoods as cases. They were matched 1:1 with cases by sex, age (± five years), and street address. Data Collection: The standardized questionnaire utilized information in the published literature, and also incorporated distinct local conditions and population characteristics. Indices that appeared in the questionnaire and their quantitation are shown in Table 1. Method of Analysis: The information was first analyzed by the conditional logistic regression model. The identified risk factors were further analyzed for relative risks and population attributable risk(3-5). -2- 1 I I I I I I I I I I I I I I I I

I I I I I I I I I I I I I I I I I I Table 1. Data Index and Assigned Value Variable Factor Assigned Value X1 Smoking index Amount of smoking (cigarettes/day) x years of smoking/20 X2 Smoke inhalation* No: 0; Shallow: 2; Medium: 3; Deep: 4 X3 Passive smoking exposure No: 0; Yes: 1 X4 Occupational exposure No: 0; Yes: 1 X5 History of chronic bronchitis No: 0; Yes: 1 X6 History of tuberculosis No: 0; Yes: 1 X7 Family tumor history No: 0; Yes: I X8 Crowded living conditions No: 0; Yes: 1 X9 Fuel use in the home (fuel Non-solid fuel: 0; Solid fuel (coal, charcoal, index)** etc. ): 1 X10 Coal stove used for winter No: 0; Yes: heating 1 X11 Amount of cooking oil used Fat consumption per person per month X12 Kitchen cooking fume pollution No: 0; Yes: 1 X13 Regular consumption of fried No: 0; Yes: 1 food X14 Cooking index Average times of cooking per week Note: * Shallow: exhale by mouth; Medium: exhale by nose; Deep: swallow smoke. ** Based on use in the last 20 years. The index represented average fuel used/year. Results and Analysis 1. Analysis of risk factors for pulmonary lung adenocarcinoma in men. Fourteen indexed variables were subjected to single-factor analysis by conditional logistic regression. Using a one-sided test with a = 0.05. Smoking index, occupational factors, history of chronic bronchitis, cooking fumes, and family tumor history were identified as the five significant variables for men. These were then further analyzed by the multi-variate conditional logistic regression -3- I

model, at a = 0.05. The four variables shown in Table 2 appeared as factors which were associated with adenocarcinoma. Table 2. Results of Analysis of Pulmonary Adenocarcinoma in Males by Conditional Logistic Regression Multi-Factor Model (a = 0.05) actor Regression Coefficient Standard Error of Regression Coefficient Value Relative Risk (RR) Smoking Index 00.0135 0.0069 0.0260 1.0136 Chronic bronchitis 0.8338 0.4111 0.0213 2.3021 Cooking vapors 1.0448 0.4235 0.0068 2.8428 Family cancer history 1.586 0.6338 0.0062 4.8856 2. Analysis of risk factors for female adenocarcinoma of the lung. Using the same method for females, at a = 0.05, single-factor analysis identified six variables, which were chronic bronchitis, history of tuberculosis, heating by coal stove, cooking fumes, fried food, and family tumor history. They were subjected to further multi-variate analysis, at a = 0.05. Four risk factors were identified (Table 3). Table 3. Results of Analysis of Female Pulmonary Adenocarcinoma By Conditional Logistic Multi-Factor Model (a = 0.05) actor Regression Coefficient Regression Coefficient Standard Error Value Relative Risk (RR) Chronic bronchitis 1.1736 0.6702 0.0399 3.2336 Coal Stove for heat 0.8278 0.4476 0.0322 2.2883 Cooking fumes 1.1625 0.4208 0.0029 3.1979 Family tumor history 1.4415 0.7235 0.232 4.2270 3. Estimate of attributable risks from male and female pulmonary adenocarcinoma risk factors. Attributable risks were calculated by multi-variate analysis of identified adenocarcinoma risk factors (Table 4). -4- I 1 I I I 1 I I I I I I I I I I

I I I I I ' I I I I I I I Table 4. Attributable Risks from Male and Female Lung Adenocarcinoma Risk Factors Mzle Fewl. F~ ~I Retxtlw Ruk Number of - C9sa Ai¢ibuuble R6t' ReWirx Risk Number of hles Attrlbuublc Risk smokfrg iMea Meuurement 1.0136 t0o 0.2769 Chmnic brorchius Caking fume Fsmily .vwr M1urory HeuinB by cml ame 0 1 0 1 0 1 0 1 1 2.3021 1 2.8428 1 4.8956 6E 36 34 66 78 22 ' 0].¢36 0.4278 0.1750 I 3]]36 1 3.I979 I 4,2270 I 2.4881 63 17 19 61 60 20 55 25 0.1468 0.5241 0.1909 0.1759 I CambiaA populs[ion .IVibuuble risk _..0.7898 0.76B7 Discussion The results of this study show the incidence of pulmonary adenocarcinoma in persons both sexes in Nanjing to have similar risk factors. These are cooking fumes, chronic bronchitis, and family tumor history. Among these, exposure to cooking fumes is most harmful. Cooking fumes are the product of pyrolysis resulting from cooking oil or food at very high temperatures. The Chinese traditionally cook with high heat, and the resulting cooking fumes are one of the most common indoor pollutants found in Chinese living quarters. We have studied the chemical composition of cooking fumes and its mutagenicity. Our results show that cooking fumes contain benzo(a)pyrene, benz(a)anthracene, etc. The mutagenicity of cooking fumes has also been shown in toxicology studies. The attributable risk of cooking fumes on the incidence of pulmonary adenocarcinoma in Nanjing residents were calculated to be 42.78 % for males and 52.41 % for females. Thus, exposure to cooking fumes may be one of the reasons for the high incidence of lung adenocarcinoma in Chinese women. The PAR of family cancer history on the incidence of adenocarcinoma was 0.175 for men and 0.1909 for women. Thus, people with a family history of tumor are in the high-risk group for lung cancer and should, if possible, avoid exposure to other risk factors. The respective PAR of chronic bronchitis was 0.2036 for men and 0.1468 for women, suggesting that people should seek early treatment of their respiratory disease in order to minimize the chance of lung cancer. The incidence of adenocarcinoma in males was related to smoking, with a PAR of 0.2769. However, no effect of passive smoking was found in this study. Lung adenocarcinoma in females was ~ associated with the use of coal-burning stoves. Since this is a traditional method used for winter heating, N the indoor air pollution it causes warrants attention. o 00 V t'to W W ' S_ 4 != O> I

I I References I I. Waterhouse, J. et al. "Cancer Incidence in Five Continents," (4): 1982. I 2. IARC, Monoeraph on the Evaluation of the Carcinogenic Risk of Chemicals to Humans (38): 1986. I 3. Whittemore, A.S. "Estimating Attributable Risk From Case-Control Studies," Am. J. Enidemiol. 117 76 85 1983 : - , . t 4. Walter, S.D. "Effects of Interaction, Confounding and Observational Error on Attributable Risk Estimation," Am. J. Epidemiol. 117: 598-604, 1983. I 5. Bruzzi, P. et al. "Estimating the Population Attributable Risk for Multiple Risk Factors Using Case-Control Data," Am. J. Enidemiol. 122: 904, 1985. ' I I I I -6- 1 I I I I I I