Effects of empagliflozin according to background angiotensin receptor-neprilysin inhibitor and mineralocorticoid receptor antagonist use in heart failure with reduced ejection fraction (SUGAR-DM-HF)

Abstract

The sodium-glucose cotransporter 2 inhibitor empagliflozin reduces left ventricular (LV) volumes in heart failure with reduced ejection fraction (HFrEF). Whether this effect is additive to optimal background guideline-recommended therapy is unknown. To evaluate the effects of empagliflozin according to key baseline therapies in SUGAR-DM-HF, a multicentre, randomised, double-blind, placebo-controlled trial in patients with HFrEF and type 2 diabetes or prediabetes. Overall, 91% of participants were treated with a beta-blocker, 61% with an ACE inhibitor/ARB, 34% with an ARNI and 60% with an MRA. We determined whether the effects of empagliflozin were modified by background treatment with a) an ARNI and b) an MRA. Treatment effects from baseline to week 36 were assessed according to baseline ARNI and MRA use, with interaction testing for co-primary (LV end-systolic volume index and LV global longitudinal strain) and secondary outcomes, using analysis of covariance adjusted for treatment group, age, diabetes status, and baseline value. Among 105 randomised patients, 36 were receiving an ARNI and 63 were receiving an MRA at baseline. Patients receiving an ARNI were more likely to have type 2 diabetes and a higher urinary albumin-to-creatinine ratio and were less likely to have coronary artery disease. Patients receiving an MRA were younger, had higher body mass index, lower systolic blood pressure, and less frequently had type 2 diabetes. From baseline to week 36, compared with placebo, empagliflozin reduced LV end-systolic volume index in the overall study by 6.0 mL/m2, with a reduction of 9.9 mL/m2 in patients receiving an ARNI vs a reduction of 2.8 mL/m2 in those not receiving ARNI (p-interaction=0.16) (Figure 1), and with a reduction of 3.4 mL/m2 in patients receiving an MRA vs a reduction of 9.8 mL/m2 in those not receiving MRA (p-interaction=0.15) (Figure 2). LV end-diastolic volume index showed a similar pattern: -8.2 mL/m2 overall and -12.2 mL/m2 in those prescribed an ARNI vs -4.3 mL/m2 in those not (p-interaction=0.17), and -4.5 mL/m2 in those prescribed an MRA vs -13.8 mL/m2 in those not (p-interaction=0.09). LV global longitudinal strain showed no significant overall change, with no evidence of interaction by baseline ARNI or MRA use. Across the overall study population, uric acid and NT-proBNP decreased and haematocrit increased, with no evidence of effect modification according to baseline ARNI (all p-interaction ≥0.31) or MRA use (all p-interaction ≥0.44). Even when added to an ARNI or an MRA (on top of a background of a beta-blocker), empagliflozin consistently reduced LV volumes. These actions provide mechanistic support for the beneficial effects of SGLT2 inhibitors when added to contemporary evidence-based therapy in patients with HFrEF.Figure 1.Effects by baseline ARNI useFigure 2.Effects by baseline MRA use