PARADIGM-HF: Angiotensin–Neprilysin Inhibition versus Enalapril in Heart Failure

Authors, Journal, Affiliations, Type, DOI

• Authors: McMurray JJV, Packer M, Desai AS, Gong J, Lefkowitz MP, Rizkala AR, Rouleau JL, Shi VC, Solomon SD, Swedberg K, Zile MR; for the PARADIGM-HF Investigators and Committees
• Journal: New England Journal of Medicine. 2014;371:993–1004. Published August 30, 2014 (updated September 11, 2014)
• Affiliations: University of Glasgow (McMurray); UT Southwestern (Packer); Brigham and Women's Hospital (Desai, Solomon); Novartis Pharmaceuticals (Gong, Lefkowitz, Rizkala, Shi); Université de Montréal (Rouleau); University of Gothenburg / Imperial College London (Swedberg); Medical University of South Carolina (Zile)
• Type: Double-blind, randomized controlled trial (Phase 3); 1,043 centers; 47 countries; industry-funded (Novartis)
• DOI: 10.1056/NEJMoa1409077

Overview

PARADIGM-HF is the landmark RCT demonstrating that sacubitril–valsartan (LCZ696; ARNI) is superior to enalapril — the then gold-standard ACE inhibitor at full mortality-reducing doses — for reducing cardiovascular mortality, HF hospitalization, and all-cause mortality in patients with HFrEF. Among 8,399 randomized patients (LVEF ≤40%, amended to ≤35%; NYHA II–IV) followed for a median of 27 months, the primary composite of CV death or first HF hospitalization was reduced by 20% (HR 0.80, P<0.001), CV death by 20% (HR 0.80), all-cause death by 16% (HR 0.84), and HF hospitalization by 21% (HR 0.79). The trial was stopped early at the third planned interim analysis after crossing the pre-specified overwhelming benefit boundary. These results established sacubitril–valsartan as a Class I pillar of HFrEF therapy in all major guidelines (AHA 2022 COR 1A; ESC 2021 Class I) and directly replaced ACE inhibitors/ARBs as the preferred RAAS strategy in symptomatic HFrEF.

Keywords

LCZ696, sacubitril, valsartan, ARNI, neprilysin inhibitor, angiotensin receptor blocker, HFrEF, heart failure with reduced ejection fraction, enalapril, ACE inhibitor, cardiovascular mortality, HF hospitalization, natriuretic peptides, neurohormonal activation, KCCQ, renal function, angioedema

Key Takeaways

Background

• ACE inhibitors (enalapril in CONSENSUS/SOLVD) had been the cornerstone of HFrEF management for ~25 years, reducing mortality by ~16% vs placebo. Beta-blockers and MRAs added further incremental mortality reductions of 30–35% and 22–30% respectively.
• Neprilysin degrades natriuretic peptides (ANP, BNP), bradykinin, and adrenomedullin. Neprilysin inhibition alone causes RAAS counter-activation, requiring concomitant AT1R blockade. Prior combined ACE+neprilysin inhibition (omapatrilat) caused serious angioedema. LCZ696 (sacubitril + valsartan) avoids ACE inhibition, reducing angioedema risk.
• Small earlier trials showed LCZ696 superior to ARB alone for haemodynamic and neurohormonal endpoints in HFpEF (Lancet 2012) and hypertension, justifying the large Phase 3 trial.

Methods

• Design: Double-blind RCT; sequential single-blind run-in (enalapril 10 mg BID for ~2 weeks → LCZ696 titrated to 200 mg BID for 4–6 weeks); double-blind randomization 1:1 to LCZ696 200 mg BID vs enalapril 10 mg BID
• Population: LVEF ≤40% (amended to ≤35% December 2010); NYHA II–IV; BNP ≥150 pg/mL or NT-proBNP ≥600 pg/mL (or BNP ≥100/NT-proBNP ≥400 if HF hospitalisation within prior 12 months); on stable beta-blocker + ACEi/ARB ≥ equivalent of enalapril 10 mg/day ≥4 weeks
• Exclusion: SBP <100 mmHg (screening) or <95 mmHg (randomization); eGFR <30 mL/min/1.73m²; K⁺ >5.2 mmol/L at screening; history of angioedema; current ACE inhibitor use (36-hour washout required)
• Run-in result: 10,513 entered enalapril run-in; 9,419 entered LCZ696 run-in; 8,442 randomized; 43 excluded for protocol violation or GCP issues → 8,399 ITT population (4,187 LCZ696; 4,212 enalapril). Run-in excluded ~12% with adverse events (cough, hypotension, hyperkalemia, renal dysfunction)
• Follow-up: Median 27 months; visits every 2–8 weeks initially then every 4 months; stopped early at third interim analysis (March 31, 2014)
• Background therapy at randomization: Beta-blocker 93%; MRA 54%; diuretic 80%; ICD 15%; CRT 7%

Results — Primary and Secondary Endpoints

• Primary composite (CV death + first HF hospitalisation):
  • LCZ696: 21.8% (914 events); Enalapril: 26.5% (1,117 events)
  • HR 0.80 (95% CI 0.73–0.87; P<0.001; exact P=4.0×10⁻⁷)
  • NNT = 21 (to prevent one primary event over trial duration)
• CV death:
  • 13.3% vs 16.5%; HR 0.80 (95% CI 0.71–0.89; P<0.001)
  • NNT = 32 (to prevent one CV death)
• First HF hospitalisation:
  • 12.8% vs 15.6%; HR 0.79 (95% CI 0.71–0.89; P<0.001); 21% RRR
• All-cause death:
  • 17.0% vs 19.8%; HR 0.84 (95% CI 0.76–0.93; P<0.001)
• Consistent benefit across all 18 prespecified subgroups — no significant heterogeneity; nominally significant NYHA class interaction for primary endpoint (P=0.03) but not for CV death (P=0.76)

Results — Secondary Endpoints

• KCCQ clinical summary score (change from baseline to 8 months):
  • LCZ696: –2.99 points; Enalapril: –4.63 points
  • Between-group difference: +1.64 points (95% CI 0.63–2.65; P=0.001)
  • Without death imputation: +0.95 points (95% CI 0.31–1.59; P=0.004) — still significant
  • Note: MCID for KCCQ is ~5 points; the 1.64-point difference is statistically significant but below MCID
• New-onset atrial fibrillation: 84 vs 83 patients (P=0.84) — no difference
• Renal function decline (ESRD or ≥50% eGFR decrease or eGFR decrease >30 mL/min to <60 mL/min): 94 vs 108 patients (P=0.28) — no significant difference; ESRD: 8 vs 16 (P=0.11)
• Systolic blood pressure at 8 months: 3.2 mmHg lower in LCZ696 (P<0.001); however, BP as a time-dependent covariate did NOT explain the incremental outcome benefit

Results — Safety

• Symptomatic hypotension: Higher in LCZ696 (reported but discontinuation rates similar: 0.9% vs 0.7%, P=0.38)
• Cough: Lower in LCZ696 (statistically significant, P<0.05)
• Creatinine ≥2.5 mg/dL: Lower in LCZ696 (P<0.05)
• Potassium >6.0 mmol/L (hyperkalemia): Lower in LCZ696 (P<0.05)
• Renal impairment discontinuations: 0.7% vs 1.4% (P=0.002) — significantly less in LCZ696
• Angioedema (adjudicated): 19 vs 10 patients (P=0.13) — not statistically different; no airway compromise in either group
• Overall discontinuation for adverse event: 10.7% vs 12.3% (P=0.03) — lower in LCZ696

Limitations of the Document

• Sequential double run-in enrichment: Both an enalapril run-in and LCZ696 run-in were performed, excluding ~12% of patients who were intolerant to either drug. This enriches the randomized population for tolerators of both agents, reducing external validity — the benefit in unselected real-world populations (including those intolerant to ACEi or ARNI) may differ
• Stopped early: Terminated at the third interim analysis for overwhelming benefit, which by definition means the trial did not run to its planned completion. Trials stopped early for benefit typically overestimate effect size
• Industry funding and data management: Novartis designed the trial, collected and managed data, and performed primary analysis; academic statistician performed independent replication; first draft was by first two authors only. These are standard safeguards, but industry involvement in data management remains a limitation
• GDMT of the 2009–2014 era: No SGLT2 inhibitors or other subsequently proven agents were included in background therapy. The incremental benefit of ARNI on top of modern quadruple GDMT (including SGLT2i) is not tested in PARADIGM-HF
• EF threshold change: The LVEF cutoff was amended mid-study from ≤40% to ≤35% (December 2010). Patients enrolled at LVEF 36–40% prior to the amendment remain in the ITT population, creating heterogeneity around the EF boundary
• NYHA II predominance: ~70% of patients were NYHA class II; NYHA III–IV were underrepresented, limiting generalizability to more symptomatic patients
• KCCQ below MCID: The 1.64-point between-group difference in KCCQ clinical summary score is statistically significant but below the minimal clinically important difference (~5 points); QoL benefit is real but modest
• BP confound: LCZ696 caused 3.2 mmHg greater SBP reduction. Although BP was modeled as non-significant in covariate analysis, modest BP difference may contribute to outcome difference via mechanism independent of RAAS/neprilysin inhibition

Key Concepts Mentioned

• entities/HFpEF — LCZ696 also had earlier Phase 2 data in HFpEF (PARAMOUNT, Lancet 2012), which motivated PARAGON-HF

Key Entities Mentioned

• entities/Sacubitril-Valsartan — primary entity; PARADIGM-HF is the foundational HFrEF evidence
• entities/Heart-Failure — HFrEF pillar therapy basis
• entities/Atrial-Fibrillation — no increase in new-onset AF with LCZ696

Wiki Pages Updated

• wiki/sources/arni-paradigm-hf-nejm-2014.md — created (this page)
• wiki/entities/Sacubitril-Valsartan.md — updated (PARADIGM-HF source citation added; detail expanded; source_count 1→2)
• wiki/entities/Heart-Failure.md — updated (PARADIGM-HF source citation added to HFrEF section; source_count 22→23)
• wiki/sourceindex.md — updated
• wiki/wikiindex.md — updated