Tobacco Documents Online

Combined analysis of case-control studies of smoking and lung cancer in China Yu, S.T. and Zhao, N. Department of Epidemiology, Shanghai Medical University INTRODUCTION: Despite the fact that adverse health effects of smoking have been well known for many years, smoking continues to increase in China. According to a survey of 0.5 million persons, the smoking rate for males _20 years old was 68.9%. Thirty million people smoke and 1,400 billion cigarettes are consumed annually. In some cities and counties lung cancer is the leading cause of death. METHODS: Using literature screening criteria, we selected 15 case- control studies for active and 3 for passive smoking analyses. The total numbers of lung cancer cases were 6,085 and controls were 6,328. Using meta-analysis, we calculated the fixed and random effects models and tested their heterogeneity (Peto, DerSimonian , and Laird), the pooled Odds Ratios (ORs), and 95% Confidence Intervals (95%CI). The pooled Population Attributable Risk (PAR) was calculated by the method of Levin and Bruzzi, and the Mantel test was used for trend.

RESCTLTS: 1. The proportion of smokers among the lung cancer cases and controls was 69.09% and 31.15% respectively. 2. The pooled OR (smoking vs non-smoking) was 2.19 (95%CI 2.03- 2.37) and the pooled PAR was 33.64%. There were no significant differences between males (OR=3.01, 95%CI 2.63-4.46) and females (OR=2.32, 95%CI 2.02-2.66). According to exposure rates, PAR were 56.84% for males and 33.10% for females. 3. The larger quantities smoked, the longer the smoking period, and the earlier the beginning age, the higher the Odds Ratios. There was a significant trend for amount of cigarette consumption: for example the OR=1.00 for non-smoking, OR=1.24 for <10 cig./day, OR=2.19 for 10-19 cig./day, and OR=4.47 for ?20 cig/day. Chi- square for trend was 223.13 (P<0.01). 4. smoking is associated with squamous cell carcinoma (OR=4.79, 95%CI 4.02-5.70) but not adenocarcinoma (OR=1.02, 95%CI 0.87-1.20). 5. Although passive,smoking has been suggested to be important, the OR was 1.004 (95%CI 0.74-1.85) and, therefore, not statistically significant, and the PAR only 0.16% in this data set. CONCLUSION: If the situation is not changed, chronic disease deaths, including lung cancer, will increase rapidly. - 2 -