Effect of metformin on survival of diabetic patients with pancreatic adenocarcinoma.

4063 Background: There is increasing epidemiological and experimental evidence that the common antidiabetic drug, metformin, has antitumor activity. METHODS We conducted a retrospective cohort study to examine the survival benefit of metformin in diabetic patients with pancreatic cancer (PC). A total of 190 patients who carried the diagnosis of diabetes mellitus and PC at the time of their initial visit were identified from a case control study conducted at MD Anderson Cancer Center (MDACC) during 2000 to 2009. An additional 112 cases were identified from the MDACC tumor registry during the same study period. Information on diabetes history was obtained by personal interview and medical record review for the two groups, respectively. Because most patients received combination therapies for glycemic control, we compared metformin ever-users and never-users regardless other antidiabetics in this study. Multivariate Cox proportional hazard regression model was used to evaluate the influence of metformin therapy on overall survival (OS). RESULTS There was no significant difference in the distribution of age, sex, race, stage, tumor size, resection status, ECOG performance status (PS), serum CA19-9, or cancer treatment between metformin ever-users (n=117) and never-users (n=185). Prevalence of tumors involving pancreatic tail, however, was higher in metformin ever-users than never-users (20.9% vs. 11.5%, p=0.029). Patients with lower surgical stage, metformin use, low serum CA19-9 (< population median), better PS, resected tumor, negative surgical margins, and smaller tumor, had longer OS. The median survival was 16.6 vs. 11.5 months for metformin ever-users vs. never-users, respectively (p=0.0044). Metformin users had a 33% decrease in risk of death; hazard ratio (HR) = 0.67 (95% CI: 0.51-0.88, p=0.005). The HR remained significant even after adjusting all other clinical predictors (HR=0.63, 95%CI=0.46-0.85, p=0.003) or exclusion of insulin user (HR=0.58, 95%CI=0.40-0.83, p=0.003). CONCLUSIONS Our study suggests that metformin may improve the outcome in diabetic patients with PC independent of other known prognostic factors. Further prospective studies are required to validate results of this study.