Dapagliflozin in Heart Failure with Mildly Reduced or Preserved Ejection Fraction (DELIVER)

Authors, Journal, Affiliations, Type, DOI

• Authors: Scott D. Solomon, John J.V. McMurray, Brian Claggett, Rudolf A. de Boer, et al., for the DELIVER Trial Committees and Investigators
• Journal: New England Journal of Medicine
• Lead Affiliation: Cardiovascular Division, Brigham and Women's Hospital, Harvard Medical School, Boston, MA
• Sponsor: AstraZeneca
• Type: Phase 3, international, multicenter, parallel-group, event-driven, double-blind, randomized, placebo-controlled trial
• DOI: https://doi.org/10.1056/NEJMoa2206286
• Published: August 27, 2022

Overview

The DELIVER trial randomized 6,263 patients with HFmrEF or HFpEF (LVEF >40%) to dapagliflozin 10 mg daily or placebo in addition to usual therapy across 353 centers in 20 countries. Over a median 2.3 years, dapagliflozin reduced the primary composite of worsening heart failure or CV death by 18% (HR 0.82; 95% CI 0.73–0.92; P<0.001), driven by a reduction in worsening HF events (HR 0.79); CV death alone was not significantly reduced (HR 0.88; 95% CI 0.74–1.05). Benefits were consistent across the full LVEF spectrum (including LVEF ≥60%), in patients with and without T2DM, in those enrolled during or after hospitalization, and in those with previously reduced LVEF that had improved to >40%. Together with EMPEROR-Preserved, DELIVER established SGLT2 inhibition as Class I therapy for HFpEF in the ESC 2023 focused update.

Keywords

Heart failure, preserved ejection fraction, mildly reduced ejection fraction, SGLT2 inhibitors, dapagliflozin, cardiovascular death, worsening heart failure, DELIVER, randomized controlled trial

Key Takeaways

Background and Rationale

• SGLT2 inhibitors reduce risk of hospitalization and CV death in HFrEF (LVEF ≤40%) and CKD; their benefit in HFpEF/HFmrEF (LVEF >40%) was less certain at trial inception.
• EMPEROR-Preserved (empagliflozin) had shown benefit in LVEF >40% HF, but a potential attenuation in the highest EF range (≥65%) prompted the need for confirmatory data across the full LVEF spectrum.
• DELIVER was specifically designed to address three evidence gaps: benefit at the highest EF range (≥60%), benefit when initiated during/soon after hospitalization, and benefit in patients with previously reduced EF that improved to >40%.

Methods

• Population: Patients aged ≥40 years with stabilized HF, LVEF >40%, structural heart disease, and elevated natriuretic peptides (NT-proBNP >300 pg/mL in sinus rhythm; >900 pg/mL in AF). Previous LVEF ≤40% permitted if EF improved to >40% at enrollment.
• Intervention: Dapagliflozin 10 mg once daily vs. matching placebo, added to usual therapy.
• Primary endpoint: Time-to-first composite of worsening HF (unplanned hospitalization or urgent visit for HF) or CV death.
• Key secondary endpoints (hierarchical): Total worsening HF events + CV deaths (recurrent events); KCCQ total symptom score change at month 8; CV death; all-cause death.
• Statistical design: Dual primary analysis — overall population (alpha 0.024) and LVEF <60% subgroup (alpha 0.038); 93% power to detect HR 0.80 with ≥1,117 events; 6,100 patients followed 13.5–39 months.
• Enrollment: Aug 2018 – Dec 2020; 6,263 randomized (3,131 dapagliflozin, 3,132 placebo); median follow-up 2.3 years (IQR 1.7–2.8).

Primary Outcome

• Overall population: Primary composite occurred in 16.4% (dapagliflozin) vs. 19.5% (placebo); HR 0.82 (95% CI 0.73–0.92; P<0.001).
• LVEF <60% subgroup: HR 0.83 (95% CI 0.73–0.95; P=0.009) — consistent with overall.

Secondary Outcomes

• Total worsening HF events + CV deaths (recurrent): Rate ratio 0.77 (95% CI 0.67–0.89; P<0.001) — 23% reduction in total event burden.
• Worsening HF (first event): HR 0.79 (95% CI 0.69–0.91) — statistically significant.
• CV death: HR 0.88 (95% CI 0.74–1.05) — not statistically significant. Placebo CV death rate 3.8 events/100 patient-years (vs. 7.9/100 patient-years in DAPA-HF); DELIVER was not powered for CV death alone.
• All-cause death: HR 0.94 (95% CI 0.83–1.07) — not significant.
• KCCQ total symptom score (month 8): Win ratio 1.11 (95% CI 1.03–1.21; P=0.009); mean placebo-corrected difference +2.4 points (95% CI 1.5–3.4) — statistically and clinically meaningful symptom benefit.

Subgroup Analyses

• LVEF ≥60% vs. <60%: No evidence of heterogeneity — benefit consistent across entire EF spectrum. Contrasts with EMPEROR-Preserved (signal of possible attenuation at LVEF ≥65%).
• T2DM vs. no T2DM: Consistent benefit — SGLT2i effect in HFpEF is diabetes-independent.
• Enrolled during/≤30 days post-hospitalization vs. outpatient: Consistent benefit — supports initiation in the vulnerable post-discharge period.
• Previously reduced LVEF (≤40%) improved to >40% at enrollment (HFrecEF subgroup): Consistent benefit — supports SGLT2i use in the recovered EF phenotype.
• All prespecified subgroups favored dapagliflozin; no heterogeneity detected.

Safety

• Serious adverse events: 43.5% (dapagliflozin) vs. 45.5% (placebo) — numerically lower in dapagliflozin arm.
• Adverse events leading to discontinuation: 5.8% in both groups.
• Safety profile consistent with established dapagliflozin data; no new signals identified.

Contextual Significance

• Together with EMPEROR-Preserved, DELIVER completed the SGLT2i evidence base across the full HF EF spectrum.
• Pooled meta-analysis (DELIVER + EMPEROR-Preserved): 20% reduction in CV death or first HF hospitalization (HR 0.80); 26% reduction in HF hospitalization; CV death HR 0.88, P=0.052 (non-significant).
• Led to ESC 2023 Class I, Level A recommendation for SGLT2i in HFmrEF and HFpEF.
• AHA 2022 (published May 2022, before DELIVER) assigned COR 2a based on EMPEROR-Preserved alone; subsequent guidelines have upgraded this in light of DELIVER.

Limitations of the document

• Less than 5% of enrolled patients were Black — limits generalizability to this population, though proportion was reported as consistent with regional demographics.
• COVID-19 pandemic restricted KCCQ assessment (8-month data available primarily in pre-pandemic enrollees); sensitivity analyses confirmed consistent findings.
• All subgroups (including LVEF ≥60%, HFrecEF) were underpowered individually; within-subgroup results should be interpreted cautiously.
• Trial was not powered to detect a reduction in CV death alone — lower CV event rate in HFpEF vs. HFrEF population (3.8 vs. 7.9 events/100 patient-years) means a mortality signal would require a larger or longer trial, or pooled analysis.
• Specific inclusion criteria (natriuretic peptide thresholds, LVEF >40%, structural disease requirement) limit applicability to patients with very low natriuretic peptides or borderline phenotypes.

Key Concepts Mentioned

• entities/HFpEF — primary population; DELIVER is one of two pivotal trials establishing SGLT2i Class I
• entities/Heart-Failure — HFmrEF/HFpEF treatment evidence

Key Entities Mentioned

• Dapagliflozin (SGLT2 inhibitor, AstraZeneca) — investigational drug
• EMPEROR-Preserved — parallel empagliflozin trial in same population; contextual comparator
• DAPA-HF — companion dapagliflozin HFrEF trial; provides contrast for CV death event rates

Wiki Pages Updated

• wiki/sources/dapagliflozin-deliver-nejm-2022.md — created
• wiki/entities/HFpEF.md — updated with direct DELIVER citations
• wiki/entities/Heart-Failure.md — updated with DELIVER source reference
• wiki/sourceindex.md — new entry added
• wiki/wikiindex.md — existing HFpEF and Heart-Failure entries updated