| Background: pulmonary carcinoma is one of the most familiar malignant tumors in the world nowadays. Recent years, the incidence rate and death rate significantly increased in our country, which drew the attention of many researchers to the pathogeny of lung cancer. They processed plenty of research work and found out that many risk factors involve in the happening of lung cancer. As the results are not consistent, further investigation is needed to settle the relationship between those factors and lung cancer.Objection: To investigate the correlation and evaluate the correlation degree between pulmonary carcinoma and the following factors including smoking, passive smoking, respiratory system disease, psychological factor, lampblack pollution indoors, coom pollution indoors, kitchen ventilation, bedroom ventilation and polymorphism of gene GSTM1 in Chinese population using meta-analysis method.Method: After generally collecting the epidemiology literatures published domestic and overseas from 1994 to 2002 that reported on correlations between pulmonary carcinoma and the following factors including smoking, passive smoking, respiratory system disease history, psychological factor, indoor lampblack pollution, indoor coom pollution, kitchen ventilation, bedroom ventilation and polymorphism of gene GSTM1, Meta-analysis and cumulative meta-analysis are processed using Review Manager 4.2 program, selecting fixed or random model based on the heterogeneity test. Sensitivity analysis is done to the results from different "including standers", the publication bias is evaluated by funnel plots, linear regression and fail-safe number.Results: 52 literatures are selected for Meta-analysis based on our including and excluding standers. Factors which have the significant difference are as followings: smoking, passive smoking, mental repression, psychic trauma, respiratory system disease history, indoor lampblack pollution, indoor coom pollution, kitchen ventilation and polymorphism of gene GSTM1, the combined-OR and 95%confidence region are separately as following: 3.22 [2.72,3.81], 1.53 [1.05,2.23],2.74 [2.35,3.18], 3.80 [2.56, 5 3.23 [2.50, 4.18 2.83 [2.06, 3.88] 2.25 [1.54, 3.28], 1.86 [1.28, 2.69], 1.73 [1.47, 2.04]. There is no significant difference between bedroom ventilation and pulmonary carcinoma.The population attributable risk percentage(PARP) of smoking, passive smoking, respiratory system disease history, polymorphism of gene GSTM1 and mental repression to pulmonary carcinoma separately is 51.24%,18%,18.38%,26.48% and 7.37%.According to the characters of the literatures, sensitivity analysis is carried out to different factors. Except for "kitchen ventilation", other sensitivity analysis results are similar to the original results, which prove that the conclusion of this research work is stable and reliable. In the Meta-analysis to "kitchen ventilation", the result became stable after kicking out a small sample, and this can be regarded as the final conclusion.Funnel plots of the "polymorphism of gene GSTM1" is symmetry measured by eyes and also by linear regression on the stander of a =0.05, but it seems moderately asymmetrical on the stander of a =0.1, the fail-safe number is 200 expressing that at least 200 unpublished literatures without statistical significance is needed to overthrow the meta-analysis outcome, which means that the publication bias is under control because 200 is above 75(5k+1). The runnel plots for others risk factors are all symmetrical by the stander of a =0.1 and the fail-safe number is more than(5k+l) showing that the publication bias is under control. But it doesn't mean that the publication bias is thoroughly eliminated to some research without enough including literatures even if funnel plots are proved to be symmetrical by linear regression test or the fail-safe number is large. The cumulative meta-analysis results showed significant difference in each point between lung cancer and the following factors including smoking, mental repression, psychic trauma, indoor coom po...
|
PDF Full Text Request