Empagliflozin in Heart Failure with a Preserved Ejection Fraction (EMPEROR-Preserved)

Authors, Journal, Affiliations, Type, DOI

• Authors: Stefan D. Anker, Javed Butler, Gerasimos Filippatos, João P. Ferreira, Milton Packer, and the EMPEROR-Preserved Trial Investigators
• Journal: New England Journal of Medicine
• Affiliations: Charité Universitätsmedizin Berlin; University of Mississippi Medical Center; and 622 centres in 23 countries
• Type: Phase 3 randomised, double-blind, placebo-controlled, event-driven trial
• DOI: https://doi.org/10.1056/NEJMoa2107038
• Funding: Boehringer Ingelheim and Eli Lilly

Overview

EMPEROR-Preserved randomised 5,988 patients with HFpEF/HFmrEF (LVEF >40%, NYHA II–IV) to empagliflozin 10 mg OD or placebo. Over a median of 26.2 months it became the first large pharmacotherapy trial to demonstrate a statistically significant positive primary outcome in HFpEF: the composite of CV death or HF hospitalisation was reduced (HR 0.79; 95% CI 0.69–0.90; P<0.001; NNT=31). The benefit was driven predominantly by reduction in HF hospitalisations (HR 0.71); CV death (HR 0.91, NS) and all-cause death (HR 1.00, neutral) were not significantly reduced. eGFR decline was significantly slower. Benefit was consistent across the full prespecified LVEF spectrum (subgroup thresholds at LVEF 50% and 60%) and was diabetes-independent. EMPEROR-Preserved, together with DELIVER (dapagliflozin, 2022), established SGLT2i as the first pharmacotherapy class with Class I, Level A evidence in HFpEF, ending two decades of failed trials in this population.

Keywords

Empagliflozin; SGLT2 inhibitor; heart failure with preserved ejection fraction; HFpEF; HFmrEF; cardiovascular death; heart failure hospitalisation; eGFR; diabetes-independent; EMPEROR-Preserved

Key Takeaways

Trial Design

• Phase 3 RCT; n=5,988; 622 centres; 23 countries; enrolment March 2017 – April 2020; median follow-up 26.2 months (IQR 18.1–33.1)
• Inclusion: LVEF >40%, NYHA II–IV, NT-proBNP >300 pg/mL (sinus rhythm) or >900 pg/mL (AF)
• Enrichment for higher EF range: For LVEF 31–35% (enrolled as mildly reduced), required prior HF hospitalisation within 12 months OR NT-proBNP ≥1,000 pg/mL; LVEF 36–40% required NT-proBNP ≥2,500 pg/mL (doubled for AF)
• Baseline characteristics: ~50% T2DM; ~50% eGFR <60 mL/min/1.73m²; 2/3 of patients with LVEF ≥50%; median LVEF 54%; randomisation stratified by geographic region, diabetes status, eGFR, and LVEF (<50% vs ≥50%)
• Background therapy: All recommended HF/comorbidity medications permitted; no restrictions on co-medications

Primary Outcome

• Composite CV death or first HF hospitalisation:
  • Empagliflozin: 415/2,997 (13.8%) vs placebo: 511/2,991 (17.1%) — 6.9 vs 8.7 events per 100 patient-years
  • HR 0.79 (95% CI 0.69–0.90; P<0.001); NNT=31 (95% CI 20–69)
• HF hospitalisation component: HR 0.71 (95% CI 0.60–0.83) — 8.6% vs 11.8%
• CV death component: HR 0.91 (95% CI 0.76–1.09) — not statistically significant (7.3% vs 8.2%)

Secondary Outcomes

• Total (first + recurrent) HF hospitalisations: 407 vs 541 events; HR 0.73 (95% CI 0.61–0.88; P<0.001)
• eGFR decline rate: −1.25 vs −2.62 mL/min/1.73m²/year; P<0.001 — preserved renal function on treatment
• All-cause death: 422 (14.1%) vs 427 (14.3%); HR 1.00 (95% CI 0.87–1.15) — completely neutral

Subgroup Analyses

• Benefit consistent across all prespecified subgroups, including T2DM/no T2DM (interaction NS)
• LVEF subgroup analysis (prespecified thresholds: 50% and 60%): Hazard ratios less than 1 in each LVEF subgroup (>40–49%, 50–59%, ≥60%). However, the authors note a possible signal of attenuation in patients with LVEF ≥65%, acknowledged in HFpEF.md as a contradiction vs. DELIVER's fully consistent benefit across all EF ranges
• Benefit appears similar to EMPEROR-Reduced pattern, supporting phenotype-independent SGLT2i effects on HF hospitalisation

Historical Context — First Positive HFpEF Pharmacotherapy Trial

• Prior large HFpEF trials with near-miss primary endpoints (all P>0.05):
  • CHARM-Preserved (candesartan): modest risk reduction, borderline significance
  • TOPCAT (spironolactone): primary endpoint not met; post-hoc signal in Americas subgroup
  • PARAGON-HF (sacubitril-valsartan): RR 0.87 (P=0.06); possible benefit at lower EF and in women
• These prior trial effects were 10–15% risk reductions, mostly concentrated in LVEF 40–57% range
• EMPEROR-Preserved was the first to demonstrate a statistically significant effect across the full LVEF >40% spectrum, driven by SGLT2 inhibition's non-neurohormonal mechanism

Safety

• Uncomplicated genital and urinary tract infections: Higher with empagliflozin
• Hypotension: More frequent with empagliflozin
• Serious adverse events occurred in 47.9% (empagliflozin) vs 51.6% (placebo) — overall serious AE burden lower with active treatment
• No excess diabetic ketoacidosis, hypoglycaemia, volume depletion, or AKI
• 23% discontinued treatment (similar in both groups); 10.6% stopped for adverse events

Limitations of the Document

• CV death and all-cause death not significantly reduced; trial powered for composite — the mortality pillar of HFpEF pharmacotherapy remains unproven for any agent
• 23% treatment discontinuation rate is high and may have biased composite results toward the null; similar rates in both groups limit bias but reduce power for hard outcomes
• Post-discontinuation eGFR measurements were not performed (unlike EMPEROR-Reduced), so the disease-modifying vs. haemodynamic nature of the renal benefit cannot be confirmed from this trial alone
• Relatively short follow-up (median 26.2 months) for a population with HFpEF where outcomes accumulate more slowly than in HFrEF
• Low prevalence of sacubitril-valsartan use at baseline — incremental benefit on top of contemporary full HFpEF treatment (SGLT2i + ARNi + MRA) is not tested

Key Concepts Mentioned

• entities/HFpEF — primary disease context; EMPEROR-Preserved is the first large positive HFpEF pharmacotherapy trial; forms dual evidence base with DELIVER for Class I, Level A recommendation

Key Entities Mentioned

• entities/Heart-Failure — broader HF syndrome; first pharmacotherapy with unequivocal evidence in HFpEF phenotype

Wiki Pages Updated

• wiki/sources/empagliflozin-hfpef-nejm-2021.md — created (this file)
• wiki/entities/HFpEF.md — EMPEROR-Preserved section enriched with full trial data; direct citation added; new contradiction added; source_count updated
• wiki/entities/Heart-Failure.md — direct EMPEROR-Preserved citation added to HFpEF SGLT2i bullet
• wiki/sourceindex.md — new entry added
• wiki/wikiindex.md — HFpEF entry updated
• log.md — entry appended