Angiotensin–Neprilysin Inhibition in Heart Failure with Preserved Ejection Fraction

Authors, Journal, Affiliations, Type, DOI

• Authors: Scott D. Solomon, John J.V. McMurray, Inder S. Anand, and colleagues for the PARAGON-HF Investigators and Committees
• Journal: New England Journal of Medicine, 2019;381:1609–1620
• Affiliations: Brigham and Women's Hospital / Harvard (lead); University of Glasgow; multinational (848 centres, 43 countries)
• Type: Randomised, double-blind, active-comparator phase 3 RCT
• DOI: 10.1056/NEJMoa1908655
• Funding: Novartis

Overview

PARAGON-HF tested whether sacubitril–valsartan (ARNI) reduces the composite of total HF hospitalizations and CV death compared with valsartan in 4,796 patients with HFpEF (LVEF ≥45%, NYHA II–IV, elevated natriuretic peptides, structural heart disease) over a median 35 months. The primary endpoint did not reach statistical significance (RR 0.87, 95% CI 0.75–1.01; P=0.06), making this technically a negative trial. Meaningful secondary signals favoured sacubitril–valsartan: NYHA class improvement (OR 1.45), halved risk of worsening renal function (HR 0.50), and modestly better symptom scores (KCCQ +1.0 point). Pre-specified subgroup analysis suggested possible differential benefit in patients with LVEF 45–57% and in women, providing biological plausibility for benefit in the lower boundary of HFpEF (overlapping HFmrEF).

Keywords

Heart failure, preserved ejection fraction, sacubitril–valsartan, angiotensin receptor–neprilysin inhibitor, ARNI, natriuretic peptides, NYHA, KCCQ, hospitalisation, valsartan

Key Takeaways

Background and Rationale

• HFpEF is associated with substantial morbidity and mortality, and no therapy had convincingly reduced morbidity or mortality at the time of the trial.
• Sacubitril–valsartan had already demonstrated efficacy in HFrEF (≤40%) in PARADIGM-HF (HR 0.80 for CV death/HF hospitalisation vs enalapril). A phase 2 signal in HFpEF (lower NT-proBNP, smaller LA, improved NYHA vs valsartan) motivated PARAGON-HF.
• The trial used valsartan as active comparator because most patients were already on RAAS inhibitors, making a placebo-controlled design impractical.

Methods

• Population: Age ≥50 years, NYHA II–IV, LVEF ≥45%, elevated natriuretic peptides (thresholds varied by recent HF hospitalisation and AF status), structural heart disease, on diuretics.
• Run-in: Sequential single-blind valsartan (half-dose) then sacubitril–valsartan (half-dose) before randomisation — excluding those with unacceptable side effects or abnormal labs.
• Intervention: Sacubitril–valsartan (target: sacubitril 97 mg / valsartan 103 mg BID) vs. valsartan (target: 160 mg BID).
• Randomisation: 1:1, 4,822 patients enrolled July 2014–December 2016; 4,796 in efficacy analysis (26 excluded: major GCP violations at one site).
• Follow-up: Median 35 months (IQR 30–41) in each arm; trial completed April 2019.
• Primary outcome: Total (first + recurrent) HF hospitalisations + CV death (semiparametric model of Lin et al.).
• Secondary outcomes (hierarchical): KCCQ clinical summary score change at 8 months; NYHA class change at 8 months; renal composite (eGFR fall ≥50%, ESRD, or renal death); all-cause death.

Results — Primary Outcome

• 894 primary events in 526 patients (sacubitril–valsartan) vs. 1009 primary events in 557 patients (valsartan).
• Rate ratio 0.87 (95% CI 0.75–1.01; P=0.06) — did not meet pre-specified alpha of 0.048.
• HF hospitalisations: 690 vs. 797 (RR 0.85, 95% CI 0.72–1.00) — borderline non-significant.
• CV death: 8.5% vs. 8.9% (HR 0.95, 95% CI 0.79–1.16) — no difference.
• All-cause death: 14.2% vs. 14.6% (HR 0.97, 95% CI 0.84–1.13) — no difference.
• Nine pre-specified supportive and sensitivity analyses were consistent with the primary analysis.

Results — Secondary and Exploratory Outcomes

• KCCQ clinical summary score (8 months): Mean change −1.6 vs. −2.6 points (between-group difference +1.0 point, 95% CI 0.0–2.1). Proportion improving ≥5 points: 33.0% vs. 29.6% (OR 1.30, 95% CI 1.04–1.61).
• NYHA class improvement (8 months): 15.0% vs. 12.6% (OR 1.45, 95% CI 1.13–1.86) — significant despite hierarchical testing not being triggered.
• Worsening renal function: 1.4% vs. 2.7% (HR 0.50, 95% CI 0.33–0.77) — significant renal protection signal.
• Blood pressure: mean SBP 4.5 mmHg lower in sacubitril–valsartan group at 8 months; not correlated with treatment effect.

Results — Subgroup Analysis

• 12 pre-specified subgroups: most showed no significant heterogeneity.
• Two subgroups with suggestion of heterogeneity (multivariable model):
  1. Lower ejection fraction (LVEF 45–57%, at or below median): Possible benefit — biological plausibility given overlap with HFmrEF where systolic dysfunction is subtle.
  2. Women: Possible benefit — well-represented in trial; women have higher prevalence of HFpEF.
• Authors highlight that other treatments effective in HFrEF also show post-hoc benefit in LVEF 40–55% (spironolactone, candesartan, beta-blockers, digoxin).
• ATTR amyloid cardiomyopathy may explain reduced responsiveness at higher EF.

Safety

• Drug discontinuation (non-death): 25.3% vs. 26.7%; due to adverse events: 15.4% vs. 16.2%.
• Target dose achieved at final visit: 82.0% vs. 85.1%.
• Higher in sacubitril–valsartan: Hypotension (mean SBP 4.5 mmHg lower); angioedema (14 vs. 4 patients; no airway compromise).
• Lower in sacubitril–valsartan: Hyperkalemia; elevated creatinine.

Limitations of the Document

• Trial missed primary endpoint — all subsequent secondary outcomes were exploratory (hierarchical testing not triggered); confidence intervals unadjusted for multiplicity.
• Active comparator (valsartan) rather than placebo; potential ARB benefit in HFpEF (CHARM-Preserved) may have attenuated treatment difference.
• Run-in period likely excluded higher-risk patients and those intolerant to study drugs, reducing generalisability.
• Subgroup signals (lower EF, women) are hypothesis-generating only — not pre-specified for formal testing.
• LVEF ≥45% includes a heterogeneous population; ATTR amyloidosis may dilute response at higher EF.
• PARADIGM-HF-like entry criteria except for LVEF threshold — not directly comparable given different comparator (valsartan vs enalapril) and population.

Key Concepts Mentioned

• entities/HFpEF — primary population; sacubitril–valsartan vs. valsartan
• concepts/Sacubitril-Valsartan-ARNI — drug under study; mechanism (neprilysin inhibition + AT1R blockade)
• concepts/Pulmonary-Hypertension-Classification — HFmrEF boundary and subgroup signal

Key Entities Mentioned

• entities/Sacubitril-Valsartan — ARNI under investigation
• entities/Heart-Failure — primary disease context

Wiki Pages Updated

• wiki/sources/arni-paragon-hf-nejm-2019.md — created
• wiki/entities/Sacubitril-Valsartan.md — created
• wiki/entities/HFpEF.md — PARAGON-HF data added to management section and contradictions
• wiki/sourceindex.md — entry added
• wiki/wikiindex.md — Sacubitril-Valsartan entity added; HFpEF entry updated