Mediation analyses of the effect of ertugliflozin on hospitalisation for heart failure in patients with type 2 diabetes and atherosclerotic cardiovascular disease from the VERTIS CV trial

Abstract Introduction Sodium-glucose cotransporter 2 (SGLT2) inhibitors reduce risk of hospitalisation for heart failure (HHF) in outcome trials, but the biological mediators underlying the therapeutic benefit are not well established. Purpose To identify potential biological mediators through which ertugliflozin reduces risk of HHF. Methods In VERTIS CV, 8246 patients with type 2 diabetes and atherosclerotic cardiovascular disease were randomised to ertugliflozin 5 or 15 mg (observations pooled as prospectively planned) or placebo. Cox regression models were used to evaluate the associations between changes in 26 potential mediators with outcomes. Potential mediators were selected based on proposed mechanisms and/or differential change from baseline with SGLT2 inhibitors. Mediation criteria required 1) significant (P<0.05 for change from baseline) effects of ertugliflozin vs placebo on each potential mediator; and 2) significant (P<0.05) association of change in post-randomisation levels of the potential mediator with risk of HHF when added to an unadjusted model of randomised treatment assignment. Percent mediation was determined by comparing the unadjusted hazard ratio and hazard ratio adjusted for change in the potential mediator of interest. Each covariate was tested individually, such that percent mediation across the analyses summed to >100%. Time-dependent models were used to evaluate associations between early (change from baseline for the first post-baseline measurement) and average (weighted average of change from baseline using all post-baseline measurements) changes in covariates with clinical outcomes. Results Over a mean of 3.5 years, the incidence rate of HHF was 0.7 and 1.1 per 100 patient-years with ertugliflozin and placebo, respectively. Among 26 candidate mediators, 9 and 13 met the mediation criteria based on early and average changes, respectively. The 3 covariates with the largest mediating effects of early changes included haematocrit (40%), haemoglobin (27%) and HDL-C (23%) (Table); other significant biomarkers included urine albumin/creatinine ratio, and serum albumin, uric acid, chloride, protein and sodium. The 3 biomarkers with the largest mediating effects in average changes included haemoglobin (63%), albumin (50%) and uric acid (47%) (Table); other significant biomarkers included haematocrit, urine albumin/creatinine ratio, body weight, serum protein and chloride, systolic blood pressure, ALT, BUN, eGFR and heart rate. Conclusions In these analyses from the VERTIS CV trial, potential markers of volume status and haemoconcentration and/or haematopoiesis were the strongest mediators of the effect of ertugliflozin on reducing risk of HHF in the early and average change periods. Other potential mediators included uric acid, lipid markers and kidney parameters. These findings provide insights into potential mechanisms through which ertugliflozin, and potentially the SGLT2 inhibitor class, may prevent HHF. Funding Acknowledgement Type of funding sources: Other. Main funding source(s): Sponsored by Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Kenilworth, NJ, USA, and Pfizer Inc., New York, NY, USA.