Tobacco Documents Online

I I I I I I I I I I t I I i I I I I I A COMPARATIVE STUDY OF THE RISK FACTORS FOR LUNG CANCER IN GUANGDONG, CHINA Wang Shen¢-vone*, Hu Yi-ling*, Wu Yi-long**, Li Xin*, Chi Gui-bo*, Chen Ying*** and Dai Wen-shan* * Medical College of Jinan University, Guangzhou, China ** Tumor Hospital of Sun Yat-sen Medical University, Guangzhou, China *** People's Hospital of the Yue-xiu District of Guangzhou, China Abstract To compare the risk factors for lung cancer of different histopathologic types in both sexes, a 1:1 matched case-control study of 390 pairs was carried out in Guangdong Province. Female lung cancers appear to differ from male lung cancers in epidenriological characteristics, pathologic types and risk factors. The 291 male lung cancer cases were predominantly squamous cell lung carcinomas (squamous cell carcinoma:adenocarcinoma = 1:0.5), whereas the 99 female lung cancer cases were predominantly adenocarcinomas (squamous cell carcinoma:adenocarcinoma = 1:2.7). The age at which lung cancer was first diagnosed was lower for females than for males (P < 0.0001). Single-factor conditional logistic regression analysis showed an association of lung cancer with family history of tumor, family history of lung cancer, history of chronic bronchitis/emphysema, history of tuberculosis, history of other lung disease, smoking, passive smoking in the home, passive smoking in the workplace, being professional drivers, use of oral contraceptives, and consumption of pickled and cured food (P < 0.05). Further multi-variate logistic regression analysis showed that family history of tuberculosis, history of chronic bronchitis/emphysema, family history of tumor, smoking, passive smoking in the home, passive smoking in the workplace and consumption of pickled and cured food were independent risk factors for lung cancer. By log-linear model analysis, it was confirmed that lung cancer had significant interactions with chronic bronchitis/emphysema, passive smoking, history of tuberculosis and smoking. Smoking, however, could explain only 115 of the incidence of female lung cancers. Since 93% of the female lung cancer cases had a history of passive smoking, passive smoking was considered one of the important risk factors of female lung cancer. Family history of lung cancer and the use of oral contraceptives were somewhat related to lung cancer in women. Except for a weak relationship with history of chronic bronchitis/emphysema, adenocarcinoma was found to have no association with the other risk factors. Introduction Lung cancer has been on the rise every year in China since the 1980s. From 1982 to 1989, the 58 National Disease Detection Centers reported an increase of 14.1 %(urban increase 43.7%) in the lung cancer death rate, with an average annual increase of 2%. From 1990 to 1992, the 145 National Disease Detection Centers reported another increase of 9.3% (rural increase 21.3%), with an average annual This study was financially supported by the Science Committee of Guangdong Province, Department of Health of Guangdong Province and Overseas Chinese Scientific Research Foundation of the Ministry of Foreign Affairs N 0 00 i V tb W O l0 a

increase of 4.7 %. The lung cancer mortality rate for Guangzhou is the third highest in China, lower only than that for Chongquing and Shanghai. The lung cancer death rate is equal to the death rates for liver, stomach, and laryngeal cancers combined, making it the highest among all cancers.(1) Now that smoking is established as one of the most important risk factors of lung cancer(2), the other risk factors warrant our further attention. Consequently, the authors undertook a comparative analysis of the lung cancer risk factors of different cell types in both men and women in a case-control study from 1990 to 1993. Materials and Methods 1. Study Subjects. The cases were drawn from current inpatients in the city of Guangzhou from the five affiliated hospitals of Sun Yat-sen Medical University, Ji-nan Medical College and Guangzhou Medical College, whose diagnoses of primary lung cancers had been pathologically confirmed. The controls consisted of non-malignant tumor patients who were hospitalized during the same time period at the same hospitals as the cases, and who were found to be free of any lung diseases. The cases and controls were matched for sex, place of residence, educational level, age (± 5 years) and were required to be either Guangzhou natives or residents of Guangzhou for over 20 years. 2. Survey Items. Items included in the survey were general health conditions, history of lung disease, family history of tumors, smoking status, passive smoking (at home and at work), dietary habits, kitchen ventilation, use of oral contraceptives, socio-psychological factors (Eysenck's Personality- Stress Inventory items). 3. Methods of Analysis. The data were subjected to single-factor logistic regression analysis with X2 test, conditional multi-variate logistic regression analysis, and analysis by log-linear model. The analyses of data were performed on the AST-386 computer with the SAS 6.03 software. Results 1. Gender and Patholo¢v Tvoe. The 390 primary lung cancer cases, aged 32-78 with an average age of 56.6, had a male to female ratio of 2.9:1. The overall squamous cell carcinoma to adenocarcinoma ratio was 1.2:1, with a predominance of squamous cell carcinomas among the male cases (squamous cell carcinoma:adenocarcinoma = 1:0.5) and a majority of adenocarcinomas among the female cases (squamous cell carcinoma:adenocarcinoma = 1:2.7). (See Table 1) -2- I I I I 1 I I I I I I I I I I

I I I I I I I I I 1 1' I Table 1. Comparison of Pathology Types in Male and Female Lung Cancer Sex Squamous Cell Carcinoma Adenocarcinoma Large and Sma1I Cel1 Cancer - Other Tbtai Male 150 (51.6) 81 (27.8) 29 (10.0) 31 (10.6) 291 Female 21 (21.2) 57 (57.6) 5 (5.1) 16 (16.1) 99 Total 171 (43.8) 138 (35.4) 34 (8.7) 47 (12.1) 390 * Percentages are given in parentheses. 2. Comparison of Single-Factor Analysis Results. By single-factor conditional logistic regression analysis, at the single-side a = 0.05 level, the following 11 risk factors were identified as risk factors for lung cancer: family history of tumors, family history of lung cancer, history of chronic bronchitis/emphysema, history of tuberculosis, history of other lung diseases, smoking, passive smoking at home, passive smoking in the workplace, being professional drivers, use of oral contraceptives, and consumption of pickled and cured food. Six or seven of the above items were identified as risk factors for both male lung cancer and lung squamous cell carcinoma. Female lung cancer was found to be associated with four risk factors, i.e. family history of lung cancer, smoking, passive smoking in the home, and the use of oral contraceptives (XZ = 5.00-6.56, P < 0.05). The only significant risk factor for adenocarcinoma was a history of chronic bronchitis/emphysema (X2 = 4.69, P < 0.05). (See Table 2) -3- I

I I Table 2. Lung Cancer Risks By Sex and Pathology Type (OR value) Sex Pathology Type Risk Factors. TohaF mate Female Squamous . Adenocarcinoma Cell . Catciuoma.. Family History of Tumor 2.66* 2.90* 2.00 Family History of Lung Cancer 3.796 2.00 739* History of Chronic Bronchitis/ Emphysema 3.64* 3.24* 4.45* 2.27" History of Tuberculosis _ 3.06* 3.46* 1.33 13.50* History of Other Lung Disease 2.87* 3.70* 1.20 6.50* 1.02 Smoking 3.56* 3.47* 4.006 4.66* 2.22 Passive Smoking. in the Home 1.91* 1.02 2.54A 3.50* Passive Smoking in the Workplace 1.90 2.10 1.78* Professional Driver 4.00" 3.67A Use of Oral Contraceptives 3.00" 3.00" 1.79~ Diet of Pickled and Cured Food 1.614 * P < 0.01 A P < 0.05 3. Primary Effects Model. The risk factors screened out by single-factor analysis were again fit to the primary effects model for a multi-variate logistic regression analysis. The final coefficients brought into the equation were, in sequence, X18 (history of tuberculosis), X20 (history of chronic bronchitis/emphysema), X27 (family history of tumor), X30 (smoking), X48 (passive smoking in the home), X54 (passive smoking in the workplace), X71 (diet including pickled and cured food). (See Table 3) The combined OR for all the risk factors were: ln OR = 0.94 X18 + 1.24 X20 + 0.94 X27 + 1.06 X30 + 0.58 X48 + 0.52 X54 + 0.55 X71. I I I I I i I ~ I I I I I N G ~ ~ ~ -4 ~ O CD ~ -4- -P ~ t

a I I I 1 I I I I I I I I I I I I I Table 3. Results of Conditional Multi-variate Logistic Analysis Data Regression Coefficient Standardized Regression Coefficient OR 95% Cl X? p X18 0.94 2.96 2.57 1.37-4.80 8.76 <0.01 X20 1.24 5.00 3.45 2.12-5.61 25.00 <0.005 X27 0.94 2.47 2.57 1.21-5.43 6.10 <0.05 X30 1.06 4.10 2.88 1.74-4.77 16.81 <0.005 X48 0.58 2.26 1.79 1.08-2.97 5.12 <0.05 X54 0.52 2.58 1.68 1.13-2.48 6.66 <0.01 X71 0.55 2.31 1.73 1.09-2.75 5.34 <0.05 4. Log-Linear Model Analysis. The results of two-factor, one-stage analysis showed significant interaction between history of chronic bronchitis/emphysema, history of tuberculosis, passive smoking, smoking and lung cancer. Interactions were also found between history of chronic bronchitis/emphysema and history of tuberculosis, between history of chronic bronchitis/emphysema and smoking or passive smoking, between history of tuberculosis and passive smoking. No significant interactions, however, were found in the three-factor or more-than-three-factor, two-stage analysis, indicating the interactions among them were not enough to affect the interactions between history of chronic bronchitis/emphysema, history of tuberculosis, smoking, passive smoking and lung cancer. (See Table 4) -5- I

Table 4. Log-Linear Model Analysis of Primary and Interactional Effects R.isk Factors df X2 P a (Lung Cancer) 1 4.35-5.32 0.0370-0.0211 b (History of Tuberculosis) 1 53.49-57.00 0.0001 c (History of Chronic 1 27.46-31.33 0.0001 Bronchitis/Emphysema) d (Smoking) 1 22.18-22.90 0.0001 e (Passive Smoking) 1 20.50-21.08 0.0001 a b 1 5.70-8.92 0.0168-0.0028 a c 1 8.94-22.70 0.0028-0.0001 a d 1 4.91-8.14 0.0267-0.0043 a e 1 11.92-12.02 0.0006-0.0005 c b 1 13.83-17.13 0.0002-0.0001 c d 1 11.83-13/27 0.0006-0.0003 c e 1 4.17-4.54 0.0411-0.0330 b c 1 7.19-7.87 0.0073-0.0050 a b c 1 0.31 0.5770 a b c d 1 0.27 0.1317 a b c d e 1 0.01 0.9292 G (Test of Likelihood) 25 28.04-30.32 0.3060-0.2124 -6- I I I I I I I I I I I

I I I I I I I .1 I Discussion Differences were found between male and female lung cancers in respect to epidemiologic characteristics, histopathologic types or risk factors. The age of diagnosis for lung cancer in females was obviously lower (X ± SE = 53.6 ± 1.0) than in males (X ± SE = 57.7 ± 0.5 P< 0.001). The average age of lung cancer occurrence in both men and women in China is 7-8 years earlier than what has been reported from abroad(3,4). Since the cases of female lung squamous carcinoma and the male lung adenocarcinoma, in proportion to all lung cancers had increased marked by from 1/10 in 1964 to 1/4 in 1985, and because the ratio of male lung squamous carcinomas has shown a slight decrease, while the ratio changes in female lung adenocarcinomas have fluctuated, the ratio of occurrence of lung squamous carcinoma and adenocarcinoma in males and females have changed from the original 9:1 and 1:10 to 1.6:1 and 1:1.5(5). The male to female ratios of lung squamous carcinomas and lung adenocarcinomas in this study are 2.4:1 and 1:2.1 respectively. (See Table 5) Table 5. Ratio of Lung Cancer Pathologic Types in Men and Women (Male:Female) Reference Counny Lwng Cancer Pathologic Type - : Squsmous Cell _Adenpcarcinpma .SSne/l.Cell - _Largr Cell Carcinoma Catcinoma Can:inOn>a 1 E1-Torky(5) U.S.A.* 9.3:1 1:10.3 1:1.5 0.8:1 1.6:1 1:1.5 1:1.8 4.1:1 2 Osann(6) U.S.A. 1.5:1 1:1.2 1:1.2 1.1:1 3 Sobue(7) Japan 2.3:1 1:1.7 1:0.7 1.6:1 4 Du, Y.X.(1) China 2.0:1 1:2.0 1:1.2 2.7:1 5 Wang (this China 2.4.t 1:2.1 1:0.5 2.0:1 report) * 1964 and 1985 data. ~ Smoking is a common risk factor for lung cancer in both men and women, but at a different level of significance (male OR = 3.5, 95% CI 2.1-5.8, X2 = 25.99 P < 0.01; female OR = 4.0, 95% CI 1.3-12.0, X2 6.05 P < 0.05). According to reports from abroad, about 9-22% of female lung cancer ~ patients and 1-13% of male lung cancer patients are nonsmokers(7). In our study, 83% of the female cases and 10% of the male cases are never-smokers. A Chinese researcher is of the opinion that "since smoking can only explain 1/4 of the female lung cancer cases, as a result, it cannot explain the increase ~ of female lung cancer incidence." (Gao, Y.T., Shanghai Tumor Research Institute, 1990). Another researcher thinks that "the rapid rise of female lung cancer in Taiwan cannot be explained by smoking alone and female lung cancer seems to be unrelated to smoking." (Yang, S.B., Medical College of ' Taiwan University, 1993). In Guangdong, very few smokers are women (4-8%), but as many as 70% N , of the men over 30 years of age are smokers. (Sun Yat-sen Medical University, 1994). As a result, it o~~o V CO ' 7 O <O V I

I is estimated that 93% of the female lung cancer patients have a history of passive smoking (87% are exposed to environmental tobacco smoke (ETS) at home and 63% are exposed to ETS in the workplace). Passive smoking should not be overlooked as a risk factor for female lung cancer (OR = 2.5 95% Cl 1.3-5.1, X2 = 6.56 P < 0.05). Family history of lung cancer appears to be more closely associated with female lung cancer (OR = 1.4 P< 0.05) than with male lung cancer (OR = 2.0 P> 0.05). Studies of family clusters of lung cancer have revealed that women have a higher susceptibility to the genetic factors of lung cancer(8,9). Our study has found some association of female lung cancer with the use of oral contraceptives (OR = 3.0 95 % Cl 1.1-8.3, X2 = 5.0 P< 0.05); but due to the small sample size of oral contraceptive users, (17.2% of cases, 7.1 % of controls), conclusions cannot yet be drawn. In this regard, we are attempting to increase the sample base for further investigation. Since 61.5% of the male lung cancer cases and 64.9% of the lung squamous carcinoma cases have had chronic bronchitis/emphysema, tuberculosis or other lung diseases, a history of lung disease was found to have a closer relationship with male lung cancer or lung squamous carcinoma (OR = 3.2-13.5, X2 = 14.71-29.58 P< 0.01). In regard to lung adenocarcinoma, which accounts for 35.4% of all lung cancers, besides a weak association with chronic bronchitis/emphysema, no associations with other risk factors have been confirmed. -8- I I I I I I

 , I I 1. I 2. ' 3. I 4. ' 5. I 6. 7. I 8. , 9 I . References Du, Y.X.; Zha, Q.; Chen; Y.L. et al. "An epidemiologic study of risk factors of lung cancer in Guangzhou," Guangzhou 3rd symposium on lung cancer, Guangzhou, 1992, pp. 1-23. Zhao, N. and Yu, S.C. "Meta-analysis of smoking and lung cancer in China," Chinese J. of E12idemiol. 14(6): 350, 1993. McDuffie, H.H.; Klaassen, D.J. and Dosman, J.A. "Men, women and primary lung cancer - a Saskatchewan personal interview study," J. Clin. Epidemiol. 44(6): 537-544, 1991. Sobue, T.; Suzuki, T. and Naruke, T. "A case-control study for evaluating lung-cancer screening in Japan," Int. J. Cancer 50: 230, 1992. El-Torky, M; El-Zeky, F. and Hall, J.C. "Significant changes in the distribution of histologic types of lung cancer," Cancer 65(10): 2361-2367, 1990. Osann, K.E.; Anton-Culver, H.; Kurosaki, T. and Taylor, T. "Sex differences in lung-cancer risk associated with cigarette smoking," Int. J. Cancer 54(1): 44-48, 1993. Alavanja, M.C.; Brownson, K.C.; Boice, J.D. and Hock, E. "Preexisting lung disease and lung cancer among nonsmoking women," Am. J. Epidemiol. 136(6): 623-632, 1992. Jin, Y.T. and He, X.Z. "An analysis of lung cancer family clusters in Xuan-wei," Chinese J. of Preventive Medicine 27(6): 329, 1993. Shaw, G.L.; Falk, R.T.; Pickle, L.W. et al. "Lung cancer risk associated with cancer in relatives," J. Clin. Epidemiol. 44(4-5): 429-437, 1991.  ~ I O ~ ~ -4 ' 00 . 9 W O cD cfl I