How Does Empagliflozin Reduce Cardiovascular Mortality? Insights From a Mediation Analysis of the EMPA-REG OUTCOME Trial

Abstract

OBJECTIVE In the BI 10773 (Empagliflozin) Cardiovascular Outcome Event Trial in Type 2 Diabetes Mellitus Patients (EMPA-REG OUTCOME) trial involving 7,020 patients with type 2 diabetes and established cardiovascular (CV) disease, empagliflozin given in addition to standard of care reduced the risk of CV death by 38% versus placebo (hazard ratio [HR] 0.62 [95% CI 0.49, 0.77]). This exploratory mediation analysis assesses the extent to which treatment group differences in covariates during the trial contributed to CV death risk reduction with empagliflozin. RESEARCH DESIGN AND METHODS Effects of potential mediators, identified post hoc, on the HR for CV death with empagliflozin versus placebo were analyzed by Cox regression models, with treatment group adjusted for the baseline value of the variable and its change from baseline or updated mean (i.e., considering all prior values), each as a time-dependent covariate. HRs were compared with a model without adjustment for covariates. Multivariable analyses also were performed. RESULTS Changes in hematocrit and hemoglobin mediated 51.8% and 48.9%, respectively, of the effect of empagliflozin versus placebo on the risk of CV death on the basis of changes from baseline, with similar results in analyses on the basis of updated means. Smaller mediation effects (maximum 29.3%) were observed for uric acid, fasting plasma glucose, and HbA1c. In multivariable models, which incorporated effects of empagliflozin on hematocrit, fasting glucose, uric acid, and urine albumin:creatinine ratio, the combined changes from baseline provided 85.2% mediation, whereas updated mean analyses provided 94.6% mediation of the effect of empagliflozin on CV death. CONCLUSIONS In this exploratory analysis from the EMPA-REG OUTCOME trial, changes in markers of plasma volume were the most important mediators of the reduction in risk of CV death with empagliflozin versus placebo.