MeRIT Project: Sulfonylureas and the risk of sudden death and ventricular arrhythmia -- interim results

Introduction:  Persons with diabetes mellitus (DM) are at increased risk for sudden death. In persons with type 2 DM, sulfonylureas (SUs) remain commonly used. Individual SUs differ with respect to their extrapancreatic effects, with glyburide and glimepiride potentially blocking cardiac ion channels. It is unknown whether such extrapancreatic effects are proarrhythmogenic (e.g., via QT interval prolongation, attenuation of ischemic preconditioning) or antiarrhythmic (e.g., prevented action potential duration shortening) in nature.

Objectives:  To examine the association between individual SUs and a composite outcome of outpatient-originating sudden cardiac death and ventricular arrhythmia (SD/VA)

Study Design: Retrospective cohort study, using 1999–2010 Medicaid claims from five large states

Methods:  Exposure was determined by new use of glyburide, glimepiride, or glipizide. Glipizide served as the reference exposure as its effects are very highly pancreas-specific. Outcomes were ascertained by a validated algorithm of first-listed emergency department or principal inpatient ICD-9 diagnoses indicative of SD/VA (positive predictive value ~ 85%). Since SD/VA is not necessarily fatal, we examined SD/VA events resulting in near-immediate death as a secondary outcome. Potential confounding was addressed by adjusting for multinomial high-dimensional propensity scores included as continuous variables in a Cox proportional hazards model.

Results:  In 176,998 person years (p-y) of SU exposure, we identified 632 SD/VA and 319 fatal SD/VA events for crude incidence rates of 3.6 and 1.8 per 1,000 p-y, respectively. Compared to glipizide, propensity score-adjusted hazard ratios were 0.82 (95% confidence interval: 0.69–0.98) and 0.89 (0.69–1.14) for glyburide and 1.10 (0.89–1.36) and 1.27 (0.94–1.71) for glimepiride, for SD/VA and fatal SD/VA respectively.

Conclusions:  These preliminary findings suggest that glyburide may be associated with a lower risk of SD/VA vs. glipizide, consistent with small clinical studies in humans showing that glyburide may have antiarrhythmic effects, particularly in settings of acute ischemia. Forthcoming secondary analyses will investigate SU dose-response relationships and examine effect modification by interacting drugs.