Diabetes and mortality in patients with prostate cancer: a meta-analysis.

BACKGROUND

There are conflicting results as to the association between pre-existing diabetes and the risk of mortality in patients with prostate cancer. The purpose of this study is to estimate the influence of pre-existing diabetes on prostate cancer-specific mortality and all-cause mortality.

METHODS

We searched PubMed and Embase to identify studies that investigated the association between pre-existing diabetes and risk of death among men with prostate cancer. Pooled risk estimates and 95 % confidence intervals were calculated using fixed-effects models or random-effects models. Heterogeneity tests were conducted between studies. Publication bias was analyzed by using the Egger's test, Begg's test, and the trim and fill method.

RESULTS

Of the 733 articles identified, 17 cohort studies that had 274,677 male patients were included in this meta-analysis. Pre-existing diabetes was associated with a 29 % increase in prostate cancer-specific mortality [relative risk (RR) 1.29, 95 % CI 1.22-1.38, I(2) = 66.68 %], and with a 37 % increase in all-cause mortality (RR 1.37, 95 % CI 1.29-1.45, p < 0.01, I(2) = 90.26 %). Additionally, in a subgroup analysis that was a type specific analysis focusing on type 2 diabetes and was conducted only with three cohort studies, pre-existing type 2 diabetes was associated with all-cause mortality (RR 2.01, 95 % CI 1.37-2.96, I(2) = 95.55 %) and no significant association with prostate cancer-specific mortality was detected (RR 1.17, 95 % CI 0.96-1.42, I(2) = 75.59 %). There was significant heterogeneity between studies and no publication bias was found.

CONCLUSIONS

This meta-analysis suggests diabetes may result in a worse prognosis for men with prostate cancer. Considering heterogeneity between studies, additional studies should be conducted to confirm these findings, and to allow generalization regarding the influence that each type of diabetes has on prostate cancer mortality.