Combined Analysis of Case-Control Studies of Smoking and Lung Cancer in China

SZ Yu,N Zhao

DOI: https://doi.org/10.1016/s0169-5002(96)90221-4

IF: 6.081

1996-01-01

Lung Cancer

Abstract:Fifteen case-control studies on the relationship between smoking and lung cancers in China were analyzed and evaluated by the method of meta-analysis, using both the fixed-effect and the random-effect models. The cumulative cases and controls were 5703 and 5669, respectively. Calculations based on the fixed-effect model resulted in a combined odds ratio (OR) of 2.19, 95% CI, 2.03–2.73, with a population attributable risk (PAR) of 33.64%. A dose-response relationship between the amount of smoking, the duration of smoking (years), the age at which smoking started and the OR for developing lung cancer was found. The pooled OR for squamous cell carcinoma was 4.79, 95% CI, 4.02–5.70, with a PAR of 65.44%. The combined OR for adenocarcinoma was 1.02, 95% CI, 0.87–1.20, with a PAR of 0.99%. To further investigate the relationship between smoking and lung cancer in Chinese women, twelve case-control studies were analyzed by meta-analysis, using the fixed-effect model. The cumulative cases and controls were 2168 and 2496, respectively. The combined OR was 2.19, 95% CI, 1.93–2.48, and the PAR was 30.34%. The combined OR for female squamous cell carcinoma was 7.45, 95% CI, 5.21–10.67, with a PAR of 53.97%. The combined OR for female adenocarcinoma was 1.09, 95% CI, 0.82–1.94, with a PAR of 1.65%. The combined OR for exposure to environmental tobacco smoke (ETS) was 1.004, 95% CI, 0.74–1.35 and the PAR was only 0.16%. The possible bias and confounding factors for this analysis are also discussed.