A Cohort Study of Metformin Exposure and Survival in Patients with Stage I–III Colorectal Cancer

Abstract Background: Preclinical evidence suggests a beneficial effect of metformin in colorectal cancer. This study aimed to investigate associations between metformin exposure and colorectal cancer–specific survival using population-level data. Methods: Adult patients with stage I–III colorectal cancer diagnosed from 2001 to 2006 were identified from the National Cancer Registry Ireland. Use of metformin and other antidiabetic medications was determined from a linked national prescription claims database. Multivariate Cox regression was used to estimate hazard ratios (HR) with 95% confidence intervals (CI) for associations between prediagnostic metformin exposure (versus nonmetformin antidiabetic drugs) and colorectal cancer–specific mortality. Models were stratified by antidiabetic drug coprescription and intensity of metformin exposure. Results: The cohort included 207 diabetics who received metformin, 108 diabetics not exposed to metformin, and 3,501 nondiabetic patients. In multivariate analyses, a nonsignificant reduction in colorectal cancer–specific mortality was observed for metformin-exposed patients relative to other treated diabetics (HR, 0.61; 95% CI, 0.37–1.01). In stratified analyses, no significant association was observed for patients receiving low-intensity metformin or metformin in combination with other antidiabetic drugs. High-intensity exclusive metformin use was associated with a significant reduction in colorectal cancer–specific mortality (HR, 0.44; 95% CI, 0.20–0.95). Conclusions: Significant associations between metformin exposure and colorectal cancer–specific mortality were observed only for high-intensity exclusive metformin use in the diabetic cohort. Impact: This study provides moderate evidence of an association between metformin exposure and improved colorectal cancer survival in a diabetic population. Additional studies in larger cohorts, with detailed information on diabetes severity, are required to confirm these results. Cancer Epidemiol Biomarkers Prev; 22(8); 1364–73. ©2013 AACR.