[1]Fudan Universit, School of Public Healt,Shanghai,China
[2]Shanghai Jiao Tong University School of Medicine, Shanghai Cancer Institute-Renji Hospital,Shanghai,China
[3]Renji Hospital, State Key Laboratory of Oncogenes and Related Genes,Shanghai,China
[4]Renji Hospital, Department of Epidemiology,Shanghai,China
[5]Renji Hospital, SKLORG & Department of Epidemiology,Shanghai,China
Objective: To introduce the dose-response analysis of Meta-analysis. Methods: We first introduced the dose-response analysis of Meta-analysis. Then we conducted a meta-analysis to quantitatively assess the association between coffee consumption and prostate cancer risk to present the details about the dose-response analysis. Results: We illustrated the application of generalized least-squares trend estimation and restricted cubic splines in the dose-response analysis of Meta-analysis using a practical example. “GLST”, “MKSPLINE” and some other commands of Stata were used to process statistical analysis. Twelve studies involving 39 943 prostate cancer cases and 577 458 total cohort members and twelve case-control studies involving 7 909 prostate cancer cases and 9 461 controls were included in the dose-response analysis. A borderline significant influence was found between coffee consumption and prostate cancer risk, and the pooled effect size of prostate cancer was 0.992 (95% CI: 0.986-0.999) with one more cup per day. Conclusions: The dose-response analysis in Meta-analysis is a method to present the combination between the levels of continuous exposure factor and relative risk for the outcome, making the results to be more productive and persuasive in cancer or chronic disease epidemiology. ©, 2015, Fudan University. All right reserved.