PURPOSE: Previous epidemiological studies reported inconsistent results regarding the association between pioglitazone use and the risk of bladder cancer (BC). We conducted a dose-response meta-analysis to assess this association. METHODS: PUBMED and EMBASE databases were searched through August 2015. Pooled results derived from a random-effects model, and the dose-response analyses were conducted for the association between cumulative dose or duration of pioglitazone use and BC risk. RESULTS: 14 studies were included. After comparing `ever use` with `never use` of pioglitazone, an increased risk of BC (HR = 1.16, 95% CI = 1.06 to 1.25) was present, and there was no significant heterogeneity between studies (P for heterogeneity = 0.54, I2 = 0.0%). Every 12 months increase (HR = 1.16, 95% CI = 1.03 - 1.30) or 10 g increase (HR = 1.05, 95% CI = 1.02 - 1.09) in pioglitazone use was marginally associated with elevated BC risk. The evidence of linear relationships was found for the cumulative dose of pioglitazone and BC risk, and a nonlinear curve for association between the duration of pioglitazone use and the risk of BC was revealed. Subgroup analyses revealed that an increased risk of BC for pioglitazone use was observed in European male subjects with more than 12 months of pioglitazone use and in the `multivariate adjusted` group. CONCLUSIONS: Pioglitazone use among subjects with diabetes mellitus increases BC risk mildly. More studies are needed to investigate the association between the cumulative dose of pioglitazone use and BC risk.
.