Medical Student Faculty of Medicine and Health Sciences, McGill University, Montréal, QC, Canada Montreal, Canada
Background: The cardiovascular safety of sulfonylureas is controversial, with several studies identifying higher rates of all-cause and cardiovascular mortality among patients using sulfonylureas. One possible mechanism behind this increased risk of cardiovascular death is an increased risk of ventricular arrhythmias.
Objectives: Our objective was to determine whether the use of sulfonylurea monotherapy, compared with metformin monotherapy, is associated with an increased risk of VA among patients initiating pharmacotherapy for type 2 diabetes.
Methods: We conducted a population-based cohort study using electronic health data extracted from the United Kingdom’s Clinical Practice Research Datalink (CPRD) Aurum. Using the active comparator, new-user cohort design, we compared rates of VA among patients aged ≥ 18 years using sulfonylurea monotherapy with those using metformin monotherapy as their initial pharmacological treatment for type 2 diabetes between April 1998 and December 2019. We used a Cox proportional hazards model with inverse probability of treatment weighting by propensity score to estimate the adjusted hazard ratio (aHR) and a corresponding bootstrap 95% confidence interval (CI) for VA with sulfonylurea monotherapy versus metformin monotherapy. We pre-specified 15 sensitivity analyses to assess the robustness of our results to potential biases, including residual confounding, exposure misclassification, outcome misclassification, prevalent user bias, informative censoring, and competing risks.
Results: The cohort included 92,638 new users of sulfonylurea and 506,882 new users of metformin. A total of 279 VA events occurred among sulfonylurea users (incidence rate per 10,000 person-years: 25.5, 95% CI: 22.7 to 28.7) and 1,537 VA events occurred among metformin users (incidence rate per 10,000 person-years: 18.5, 95% CI: 17.6 to 19.5). Compared to metformin, sulfonylureas were associated with an increased risk of VA (aHR: 1.42, 95% CI: 1.18 to 1.69). Our finding of an increased risk of VA was robust to several sensitivity analyses. When stabilized weights were used, an aHR of 1.45 (95% CI: 1.21 to 1.74) was obtained. Our primary analysis produced an E-value of 2.19.
Conclusions: Sulfonylureas are associated with an increased risk of VA when used as first-line therapy for type 2 diabetes relative to metformin use. This increased risk should be considered when prescribing sulfonylureas as an initial treatment for type 2 diabetes.
