A Meta-analysis of Diabetes Mellitus and the Risk of Prostate Cancer

J. S. Kasper, E. Giovannucci
2006 Cancer Epidemiology, Biomarkers and Prevention  
Studies investigating the association between diabetes mellitus and prostate cancer have reported inconsistent findings. We examined this association by conducting a detailed meta-analysis of the studies published on the subject. Methods: MEDLINE and EMBASE databases and bibliographies of retrieved articles were searched. Studies investigating the relationship between diabetes mellitus and prostate cancer were included in the meta-analysis. Potential sources of heterogeneity between studies
more » ... between studies were explored and publication bias was evaluated. Pooled relative risk (RR) was calculated using the random-effects model. Numerous relevant subgroup analyses were also done. Results: We included 19 studies, published between 1971 and 2005, in the meta-analysis and found an inverse association between diabetes mellitus and prostate cancer [RR, 0.84, 95% confidence interval (CI), 0.76-0.93, P for heterogeneity V 0.01]. For cohort studies alone, the RR was 0.81 (95% CI, 0.71-0.92, P for heterogeneity V 0.01) and for case-control studies alone, the RR was 0.89 (95% CI, 0.72-1.11, P for heterogeneity = 0.02). The significant heterogeneity was mitigated in some of the subgroup analyses. For studies conducted before prostate-specific antigen screening was introduced as a common procedure, the RR was 0.94 (95% CI, 0.85-1.03, P for heterogeneity = 0.15), and for studies conducted after this time, the RR was 0.73 (95% CI, 0.64-0.83, P for heterogeneity = 0.10). For studies that adjusted for three or more potential confounders, the RR was 0.74 (95% CI, 0.65-0.85, P for heterogeneity = 0.06) and for studies that adjusted for less than three potential confounders, the RR was 0.93 (95% CI, 0.86-1.02, P for heterogeneity = 0.18). Conclusion: This study suggests an inverse relationship between diabetes and prostate cancer. Potential biological mechanisms underlying this association, as well as possible biases, are discussed. (Cancer Epidemiol Biomarkers Prev
doi:10.1158/1055-9965.epi-06-0410 pmid:17119028 fatcat:2nc5myldx5c3lje4khgnzb5i2q