Effects of metformin on pancreatic cancer survival among Medicare beneficiaries.

Abstract

e14691 Background: Evidence from preclinical models suggests that metformin may have antitumor activity against pancreatic adenocarcinoma. The effect in humans has not been well-described. A recent single-institution retrospective analysis suggested improved survival in diabetic patients who had been exposed to metformin. The current study was designed to determine whether metformin was independently associated with improved pancreatic cancer (PC) survival when compared to non-metformin controls using newly available national pharmacy data from Medicare Part D. Methods: Patients with diagnosed with AJCC Stage II-IV PC during 2007 were identified in the SEER-Medicare linked database. Part D events for oral diabetic medications were identified from the date of diagnosis until death or 12/31/2008, and divided into metformin-containing or non-metformin. Patients who did not receive an oral medication for diabetes, lacked continuous Part D prescription drug coverage, underwent surgical resection, or survived less than 30 days were excluded. Patient characteristics were compared by chi-square and Wilcoxon. Overall survival was compared by Kaplan-Meier log rank. Results: From an initial cohort of 899 patients, 100 met inclusion criteria and received oral diabetic medication. Of these, 31 received metformin and 69 non-metformin. Patient characteristics were similar among the groups, including median age (76.6 vs. 76.2, p=0.75), race distribution, Charlson comorbidity score and disease stage (p>0.05 for each). There was no statistically significant difference in median survival between the two groups: 5.6 months for metformin vs. 6.1 for non-metformin, p=0.93. Conclusions: This study using national Medicare Part D prescription drug claims revealed no detectable survival benefit for unresectable pancreatic cancer patients who received metformin when compared to non-metformin oral hypoglycemics. Sufficient power to detect modest benefits, however, will require additional subjects, which we anticipate in the near future. Additional studies are warranted and ongoing to determine the effect of metformin on pancreatic cancer survival.