Propensity Score Matching to Determine the Impact of Metformin on All-Cause Mortality in Older Veterans with Diabetes Mellitus

Objectives To determine whether metformin is associated with reduced all-cause mortality in older adults with diabetes mellitus as compared with insulin or sulfonylureas, and to evaluate whether the metformin cumulative exposure followed a dose-response relation. Methods Retrospective cohort study with propensity score matching in veterans 65 years old and older with diabetes mellitus. Patients who had new prescriptions for metformin were matched for demographic and clinical factors with patients receiving new prescriptions for insulin or sulfonylureas using propensity score matching. All-cause mortality risks were compared between metformin and insulin/sulfonylureas using multivariate Cox regression models. A similar approach was used for tertiles of cumulative metformin doses. Results A sample of 174 veterans taking metformin was matched with 174 who took insulin/sulfonylureas. Most patients were men (97.4%), White (80.45%), and their mean +/- standard deviation age was 69.15 +/- 7.65 years. Metformin exposure was associated with reduced risk of all-cause mortality (hazard ratio 0.57, 95% confidence interval 0.39-0.84, P = 0.005). The upper tertile of cumulative metformin exposure was associated with lower all-cause mortality in the fully adjusted model (hazard ratio 0.28, 95% confidence interval 0.10-0.77, P = 0.013). Conclusions This propensity matching study shows that metformin exposure is associated with a lower risk of all-cause mortality. Higher metformin cumulative exposure seems to reduce the risk of all-cause mortality in older veterans with diabetes mellitus.