LAUSR

OBJECTIVES To assess the impact of metformin on biochemical failure (BF) in localized prostate cancers (PC) treated with radical prostatectomy (RP) or radiation therapy (RT). SUBJECTS/PATIENTS AND METHODS 1,449 patients undergoing RP (n=1338, 92.3%) or RT (n=108, 7.5%) for localized PC between July 2007 and January 2020 were evaluated for metformin use, demographic/oncologic characteristics, and biochemical outcomes. Androgen deprivation therapy was utilized per NCCN guidelines. BF rates were assessed overall and at 1, 3, and 5 years. Time to BF was estimated via Kaplan-Meier; logistic regression and Cox proportionate hazards models were generated to adjust for significant differences. RESULTS Of 1,449 patients, 148 (10.2%) utilized metformin at time of diagnosis, while 1,301 (89.8%) did not. Patients on metformin were significantly older, had higher body mass indexes (BMI), and more aggressive disease (Gleason score>7). At a mean±SD follow-up of 3.6±2.6 years, patients on metformin were less likely to experience BF at later timepoints; however, univariate analysis showed no differences at 1, 3, and 5 years. In multivariate analysis, patients on metformin were significantly less likely to experience BF at 5 years and overall in both treatment groups. In Cox regression, metformin was independently associated with a 40% relative risk reduction in BF. CONCLUSION In multivariate analysis, metformin use was associated with a significant risk reduction in BF overall and at 5 years following primary treatment; this trend was not witnessed in univariate analysis. This suggests the need for future investigations of metformin's role in disease-free survival in men with localized PC.