Effect of metformin use on the survival outcomes in diabetic patients with resectable pancreatic cancer: A single-institutional experience and meta-analysis.

465 Background: Observational studies have shown that metformin use in diabetic patients decreases both cancer incidence and mortality. The aim of this study was to determine whether metformin use had a survival benefit in patients with resected pancreatic cancer. Methods: All patients with diabetes who underwent pancreatic resections for adenocarcinoma between 12/1/1986 and 4/30/2013 at our institution were identified. Patients were categorized by metformin use. Survival analysis was done using the Kaplan Meier method, with log rank test, and Cox proportional hazards multivariable regression models. The comprehensive Meta-Analysis version 2.2 was used to perform a meta-analysis using data from our study and the only other published study. Results: The study from our institution was comprised of 19 patients receiving treatment with metformin, and 25 patients not receiving metformin. There were no significant differences in major clinical and demographic characteristics between metformin and non-metformin users. The estimated median survival for metformin users was 35.3 months (95% CI 10.7 - not estimable) vs. 20.2 months (95% CI 14.4-58.6) for nonusers. The estimated 2, 3 and 5 year survival rates for non-metformin users were 42% (95% CI 22-62%), 28% (95% CI 7-55%) and 14% (95% CI 1-43%) respectively. Metformin users fared better with corresponding rates of 68% (95% CI 41-84%), 34% (95% CI 12-59%) and 34% (95% CI 12-59%) respectively. Metformin users had a 46% lower risk of death; hazard ratio 0.54 (95% CI 0.16-1.86) in a multivariate analysis, p = 0.3275. For the meta-analysis, there were a total of 111 patients from both studies. There were 46 metformin users and 65 nonusers. Metformin users had a 33% lower risk of death; the HR 0.668 (95 % CI 0.397-1.125), in a multivariate analysis p= 0.129. Conclusions: Although the findings of the current study suggests that metformin use does not significantly impact survival outcomes in diabetic patients with resected pancreatic cancer, there appears to be a clinically meaningful improvement in survival. The potential benefit of metformin should be investigated in adequately powered prospective studies.