Background: To evaluate whether metformin (Met) a widely-used, nontoxic oral antidiabetic drug with putative anticancer properties leads to improvements in prostate cancer (PC) outcomes in the CHAARTED trial. Methods: In the CHAARTED database where metformin use at baseline was recorded prospectively, we identified patients with metastatic PC who underwent either ADT alone or ADT and docetaxel (D) chemotherapy. Cox proportional hazards models were used to determine the effect of Metformin on outcomes. Results: A total of 788 patients (median age, 63 y) had complete data after randomization. Comparison of ADT+D+Met (n = 39) to ADT+D (n = 357) and ADT+Met (n = 29) to ADT alone (n = 363) revealed similar clinicopathologic characteristics. Cause of death was PC in 13(81%) of ADT+D+Met, 72(85%) ADT+D, 9(82%) ADT+Met and 105(84%) ADT alone groups. See table for PC outcomes and overall survival by metformin use. Cox regression analysis for overall survival stratified by stratification factors at randomization demonstrates Met use was associated with a trend for worse overall survival (HR 1.47 95%CI: [0.95,2.26], p = 0.08) with adjustment for treatment arm and prior local therapy. In contrast, ADT+D use (HR 0.62; 95%CI: [0.47,0.81]) and prior local therapy with surgery or radiation (HR 0.56; 95% CI: [0.38, 0.82]) were associated with improved survival. Conclusions: In this study, baseline metformin did not improve PC outcomes. Partial support and drug supply by Sanofi. Clinical trial information: NCT00309985

¹Wilcoxon

²Fisher’s exact

³Stratified Cox regression model