MP86-07 ASSOCIATION OF METFORMIN USE WITH PROSTATE CANCER INCIDENCE IN A PROSPECTIVE SCREENING TRIAL (ERSPC AARAU)

Abstract

INTRODUCTION AND OBJECTIVES: There is conflicting data regarding the effect of the oral antidiabetic drug metformin on prostate cancer (PCa) incidence and mortality. We aimed to assess the effect of metformin on PCa incidence and mortality in men participating in a prospective, population-based screening trial. METHODS: Data of men who underwent PSA-screening in our population-based screening trial (ERSPC Aarau) were analysed. Information on metformin exposure before biopsy was obtained by a selfadministered questionnaire. Cox regression analysis was used to examine the relationship between covariates and PCa incidence. RESULTS: Overall, 4,314 men were observed during a mean follow-up time of 7.3 ( SD2.3) years. Mean age at baseline was 65.5 years (65.5 SD 4.4). Overall, n1⁄4150 (3.5%) men used metformin [metfþ] at baseline. Of those, 80% (88 of 110) who were actively screened during follow-up visit were still on metformin after 4 years whereas 105 men had new [metfþ] exposure. Mean baseline PSAlevels were comparable between both groups ([metfþ] 1.6ng/ml 2.4 vs. [metf-] 1.8ug/l 2.2, p1⁄40.4) while f/t-ratio was slightly higher in metformin users ([metfþ] 30.7% 10.9 vs. [metf-] 27.3% 10.9, p1⁄40.01). Overall, n1⁄4372 (8.6%) PCa cases were detected. Neither cumulative PCa incidence (n1⁄411; 7.3% [metfþ] vs. n1⁄4361 8.7% [metf-]; p1⁄40.5) nor d’Amico risk groups were significantly different between both groups. On multivariate analysis, only increasing baseline PSA (HR 1.11, 95%CI 1.09-1.13; p<0.0001) achieved an independent predictor status for PCa detection. An increasing freetotal PSA ratio was an independent predictor for a decreased risk for PCa detection during follow-up (HR 0.95, 95%CI 0.94-0.96; p<0.0001). Among 150 [metfþ] and 4164 [metf-] men, 1 (0.7%) and 1 (0.02) died from PCa (OR 27.94, 95%CI 1.73-448.84; p<0.0001) whereas 26 (17.3%) and 329 (7.9%) (unadjusted OR 2.50, 95%CI 1.59-3.82; p<0.0001) died from other causes during follow-up, respectively. In a competing risk analysis, [metfþ] exposure among men who died during follow-up had no influence on PCa incidence (SHR 0.91, 95%CI 0.50-1.66; p1⁄40.77). CONCLUSIONS: No significant differences in PSA levels or PCa incidence and grade were observed between metformin users and non users. The slightly higher f/t-ratio did not result in lower PCa detection rate. However, metformin users were at significantly higher risk to die from any cause. This underlines the idea of a risk-based strategy of early PCa detection with recognizing diabetes mellitus as an important competing comorbidity.