Cardiovascular safety of metformin and sulfonylureas in patients with different cardiac risk profiles.

Objectives/background: Based on previous experiences, the Food and Drug Administration and the European Medicines Agency recommend that clinical trials for novel antidiabetic drugs are powered to detect increased cardiovascular risk. In this context, data concerning licensed drugs such as metformin and sulfonylureas are conflicting. The influence of baseline cardiovascular risk on any treatment effect appears obvious but has not been formally proven. We therefore evaluated association of metformin and sulfonylureas with cardiovascular events in patients with different cardiovascular risk profiles indicated by N-terminal of the prohormone brain natriuretic peptide (NT-proBNP) levels.

Methods: 2024 patients with diabetes mellitus were included in this observational study. The primary endpoint was defined as a combination of cardiovascular events and death. Association of metformin and sulfonylureas was assessed using Cox regression models. Possible differences of these associations in patients with different NT-proBNP levels were studied by stratifying and through interaction analysis.

Results: During a median follow-up of 60 months, the primary endpoint occurred in 522 (26%) of patients. The median age was 63 years. A Cox regression analysis was adjusted for site of treatment, concomitant medication, age, gender, body mass index, glycated haemoglobin, duration of diabetes, glomerular filtration rate, cholesterol, and history of smoking and cardiac disease. Metformin was associated with a decreased risk in the cohort with elevated NT-proBNP ≥300 pg/mL (HR 0.70, p=0.014) and a similar association was found for the interaction between metformin and NT-proBNP (p=0.001). There was neither an association for sulfonylureas nor a significant interaction between sulfonylureas and NT-proBNP.

Conclusions: Metformin is associated with beneficial cardiovascular outcomes in patients with diabetes only when (sub)clinical cardiovascular risk defined by NT-proBNP levels is present.