HFrEF — Heart Failure with Reduced Ejection Fraction

Details

HFrEF is the phenotype of heart failure defined by LVEF ≤40%, characterised by reduced myocardial contractility, eccentric LV dilatation, and neurohormonal activation. It accounts for ~60% of HF in outpatient registries and carries 5-year mortality of ~53–67%. Among the three HF phenotypes (HFrEF, HFmrEF, HFpEF), HFrEF has the largest evidence base for pharmacological and device therapy: the four-pillar GDMT regimen (ARNi/ACEi/ARB + beta-blocker + MRA + SGLT2i) reduces estimated all-cause mortality by ~73% relative to no treatment. A subtype — HFimpEF — describes patients with prior LVEF ≤40% who recover to >40% on GDMT; structural risk and relapse potential persist despite LVEF normalisation, requiring indefinite continuation of GDMT. See entities/Heart-Failure for the full HF syndrome overview including HFmrEF, HFpEF, acute HF, and special populations.

Epidemiology

• ~60% of outpatient HF registry phenotype distribution (ESC). (sources/HF-ESC-2021, rating: very high)
• 1-year mortality ~20%; 5-year mortality ~53–67% in population studies. (sources/HF-ESC-2021, rating: very high)
• HF hospitalisation averages once per year; 63% of admissions are for non-CV causes. (sources/HF-ESC-2021, rating: very high)
• HFimpEF: ~40% relapse within 6 months of GDMT withdrawal even with normalised LVEF (≥50%). (sources/HF-AHA-2022, rating: very high)

Pathophysiology

• Neurohormonal activation: Reduced cardiac output triggers RAAS (angiotensin II, aldosterone) and sympathetic nervous system (norepinephrine) activation. Initially compensatory, chronic excess drives maladaptive remodeling — myocyte hypertrophy, apoptosis, interstitial fibrosis, and progressive chamber dilatation. Natriuretic peptides (ANP, BNP/NT-proBNP) are released proportional to wall stress and serve as haemodynamic load markers and therapeutic targets. (sources/HF-ESC-2021, rating: very high)
• Eccentric LV remodeling: Volume overload pattern — sarcomere replication in series → chamber dilatation with proportional wall thinning; Frank-Starling curve shifts right and flattens. GDMT reverses eccentric remodeling in HFimpEF, but structural risk persists despite LVEF normalisation. (sources/HF-AHA-2022, rating: very high)
• Aldosterone escape despite ACEi therapy: ACE inhibitors only transiently suppress aldosterone — aldosterone "escapes" via angiotensin-independent pathways including direct potassium stimulation of adrenal zona glomerulosa, ACTH, and locally generated angiotensin II within myocardial tissue. Chronically elevated aldosterone drives myocardial and vascular fibrosis, baroreceptor dysfunction, sympathetic activation, and impaired norepinephrine reuptake — mechanisms additive to, and distinct from, ACEi haemodynamic effects. This aldosterone escape is the pathophysiological basis for MRA benefit on top of ACEi, as demonstrated by RALES. (sources/mra-hfref-rales-nejm-1999, rating: very high)
• Electrical remodeling: HF causes comprehensive ionic remodeling — ↑INaLate (CaMKII-driven), ↑If (HCN4 upregulation in ventricular myocytes), ↑NCX1, ↓IKr/IKs/Ito/IK1 — increasing arrhythmic susceptibility and SCD risk independently of LVEF. This profile elevates TdP risk with QT-prolonging drugs and reduces repolarisation reserve. Amiodarone is the only antiarrhythmic agent without evidence of harm in this ionic context (SCD-HeFT neutral result). See concepts/Electrical-Remodeling. (sources/membrane-potential-physrev-2021, rating: very high)

Diagnosis

• LVEF ≤40% on echocardiography or CMR confirms HFrEF when clinical HF criteria are met (symptoms + signs + cardiac structural/functional abnormality). (sources/HF-ESC-2021, sources/HF-AHA-2022, rating: very high)
• Natriuretic peptides: Elevated BNP or NT-proBNP supports diagnosis and guides severity assessment; values proportional to wall stress and haemodynamic load. In obesity, BNP is physiologically lower — normal BNP does not exclude HFrEF in obese patients. (sources/HF-ESC-2021, sources/HF-AHA-2022, rating: very high)
• HFimpEF (AHA 2022): Prior LVEF ≤40% with follow-up LVEF >40% on GDMT. Distinct from "recovered EF" — structural abnormalities and relapse risk persist. Continue GDMT indefinitely (COR 1, LOE B-R). (sources/HF-AHA-2022, rating: very high)

Pharmacological Management

GDMT — Four Pillars

The four foundational drug classes each carry Class I / COR 1A recommendations (ESC 2021, AHA 2022). Estimated 73% reduction in all-cause mortality with all four classes vs no treatment. (sources/HF-AHA-2022, rating: very high)

• ARNi (sacubitril-valsartan): AHA COR 1A, ESC Class I — preferred over ACEi for NYHA II–III. PARADIGM-HF (McMurray et al. 2014, NEJM; n=8,399 ITT; LVEF ≤40%, amended to ≤35%; NYHA II–IV; vs enalapril 10 mg BID; median 27 months; stopped early at third interim for overwhelming benefit): primary composite (CV death + first HF hospitalisation) HR 0.80 (95% CI 0.73–0.87; P<0.001; NNT=21); CV death HR 0.80 (NNT=32); HF hospitalisation HR 0.79; all-cause death HR 0.84 (P<0.001). KCCQ +1.64 points (P=0.001 — significant but below ~5-point MCID). Safety advantages vs enalapril: less cough, less creatinine ≥2.5 mg/dL, less potassium >6.0 mmol/L, fewer renal-impairment discontinuations (0.7% vs 1.4%; P=0.002); more symptomatic hypotension but rare discontinuation. (sources/arni-paradigm-hf-nejm-2014, sources/HF-AHA-2022, sources/HF-ESC-2021, rating: very high)
• ACEi: AHA COR 1A, ESC Class I — if ARNi not feasible. (sources/HF-AHA-2022, sources/HF-ESC-2021, rating: very high)
• ARB: AHA COR 1A — if intolerant to ACEi (cough/angioedema) and ARNi not feasible. (sources/HF-AHA-2022, rating: very high)
• Beta-blocker (bisoprolol, carvedilol, or metoprolol succinate): AHA COR 1A, ESC Class I — reduces mortality and HF hospitalisation. (sources/HF-AHA-2022, sources/HF-ESC-2021, rating: very high)
  • In HFrEF+COPD (Class I, Level A): β-blockers remain strongly recommended. β1-selective agents preferred — neutral effect on airway function and FEV1 response to β-agonists. Bisoprolol reduces mortality and incidence of both HF and COPD exacerbations in a dose-dependent manner (superior to carvedilol and metoprolol in this comorbid population). Carvedilol reduces FEV1/FVC ratio and DLCO (via β2 blockade) — avoid in hypoxic/reduced-DLCO patients; may be considered only with exaggerated VE/VCO2 slope at low dose with PFT monitoring. Agent selection: hypoxia/reduced DLCO → nebivolol or bisoprolol; exaggerated VE/VCO2 → carvedilol (low dose); frequent AECOPD → bisoprolol; metoprolol second choice. (sources/hf-copd-hfreview-2025, rating: high)
• MRA (spironolactone or eplerenone): AHA COR 1A, ESC Class I — reduces all-cause mortality and HF hospitalisation. RALES (Pitt et al., NEJM 1999; n=1,663; LVEF ≤35%; NYHA III/IV with required recent NYHA IV; ACEi + loop diuretic background; no BB requirement — pre-BB era; 195 centres, 15 countries; spironolactone 25 mg OD; mean 26 mg achieved; mean 24 months; stopped early; Searle-funded): all-cause mortality RR 0.70 (95% CI 0.60–0.82; P<0.001) — 30% RRR; 46% → 35%; progressive HF death RR 0.64 (P<0.001); SCD RR 0.71 (P=0.02); HF hospitalisation RR 0.65 (P<0.001); NYHA improved P<0.001; consistent across all 6 prespecified subgroups. Serious hyperkalemia 2% vs 1% (P=NS). Gynecomastia/breast pain 10% vs 1% in men (P<0.001). Benefit in NYHA II later established by EMPHASIS-HF (eplerenone). (sources/mra-hfref-rales-nejm-1999, sources/HF-AHA-2022, sources/HF-ESC-2021, rating: very high)
• SGLT2i (dapagliflozin or empagliflozin): AHA COR 1A, ESC Class I — fourth pillar of HFrEF GDMT. Reduces CV death and worsening HF regardless of diabetes status; effect independent of HF aetiology (ischaemic ~50% of both DAPA-HF and EMPEROR-Reduced). Estimated 17% RRR all-cause mortality (guideline synthesis across trials). (sources/HF-AHA-2022, sources/HF-ESC-2021, rating: very high)
  • DAPA-HF (McMurray et al., NEJM 2019; n=4,744; LVEF ≤40%; NYHA II–IV; median 18.2 months; 20 countries; AstraZeneca-funded): Dapagliflozin 10 mg OD vs placebo on background ARNi/ACEi/ARB + beta-blocker ± MRA. Primary composite (worsening HF or CV death): HR 0.74 (95% CI 0.65–0.85; P<0.001; NNT=21). Worsening HF: HR 0.70 (0.59–0.83). CV death: HR 0.82 (0.69–0.98). All-cause death: HR 0.83 (0.71–0.97). KCCQ ≥5-point improvement: 58.3% vs 50.9% (OR 1.15; P<0.001). Serious renal AEs lower (1.6% vs 2.7%; P=0.009). Benefit identical in T2DM and non-T2DM (55% without T2DM) — established glucose-independent mechanism. Post hoc ARNi subgroup: HR 0.75 (95% CI 0.50–1.13) — consistent. (sources/dapagliflozin-hfref-nejm-2019, rating: very high)
  • EMPEROR-Reduced (Packer et al., NEJM 2020; n=3,730; LVEF ≤40%; NYHA II–IV; median 16 months; 520 centres; Boehringer Ingelheim/Eli Lilly-funded): Enriched for more severe disease than DAPA-HF — 73% LVEF ≤30%, 79% NT-proBNP ≥1,000 pg/mL, 48% eGFR <60, ~20% on ARNi. Primary composite (CV death or HF hospitalisation): HR 0.75 (95% CI 0.65–0.86; P<0.001; NNT=19). Total HF hospitalisations: HR 0.70 (0.58–0.85; P<0.001). CV death: HR 0.92 (0.75–1.12) — not significant. All-cause death: HR 0.92 (0.77–1.10) — not significant. eGFR decline: −0.55 vs −2.28 mL/min/1.73m²/year (difference +1.73; P<0.001). Composite renal outcome: HR 0.50 (0.32–0.77). Post-discontinuation eGFR measured 23–45 days after stopping drug: −0.93 vs −4.21 mL/min/1.73m² — persistent renal benefit confirming disease-modifying (not haemodynamic) mechanism. Diabetes-independent. ARNi subgroup: HR 0.64 (0.45–0.89). (sources/empagliflozin-hfref-nejm-2020, rating: very high)

Additional Agents

• Loop diuretics: ESC Class I — for fluid retention (symptom relief, not mortality). (sources/HF-ESC-2021, rating: very high)

• H-ISDN (hydralazine + isosorbide dinitrate): AHA COR 1A — self-identified Black patients with NYHA III–IV already on optimised GDMT (A-HeFT); AHA COR 2a / ESC Class IIa if intolerant to all RAAS therapies. (sources/HF-AHA-2022, sources/HF-ESC-2021, rating: very high)

• Ivabradine: AHA COR 2a (B-R), ESC Class IIa — stable symptomatic HFrEF (LVEF ≤35%) + sinus rhythm + resting HR ≥70 bpm on maximally tolerated beta-blocker; reduces HF hospitalisation (SHIFT trial). (sources/HF-AHA-2022, sources/HF-ESC-2021, rating: very high)

• Vericiguat (soluble guanylate cyclase stimulator): AHA COR 2b (B-R), ESC Class IIb — NYHA II–IV with recent worsening HF despite GDMT. VICTORIA trial (n=5,050; median NT-proBNP 2,816 pg/mL; 41% NYHA III/IV; median 10.8 months): primary composite (CV death + first HF hospitalisation) HR 0.90 (95% CI 0.82–0.98; P=0.02); NNT ~24 for 1 year; total HF hospitalisations HR 0.91 (P=0.02); no all-cause mortality benefit (HR 0.95; P=0.38). Benefit consistent including in sacubitril-valsartan subgroup; attenuated at very elevated NT-proBNP. Contraindicated with long-acting nitrates and PDE5 inhibitors (pharmacodynamic overlap on cGMP axis). No SGLT2i co-administration data (enrolled 2016–2018). See entities/Vericiguat. (sources/vericiguat-victoria-nejm-2020, sources/HF-AHA-2022, rating: very high)

• Digoxin/Digitoxin: AHA COR 2b (B-R) — may reduce HF hospitalisation; no mortality benefit. DIGIT-HF (Bavendiek/Bauersachs et al., NEJM 2025; n=1,212 modified ITT; LVEF ≤40% NYHA III/IV or LVEF ≤30% NYHA II; on GDMT ≥6 months; digitoxin 0.07 mg OD; 65 sites, Austria/Germany/Serbia; median 36-month follow-up; ARNi 39.5%, SGLT2i 19.3%, MRA 76.2%, BB ≥93%, ICD 64.3%): primary composite (all-cause death or first HF hospitalisation) HR 0.82 (95% CI 0.69–0.98; P=0.03); NNT=22; all-cause death HR 0.86 (NS individually); noninferiority of mortality confirmed (P<0.001 vs threshold 1.303 — no mortality detriment); first HF hospitalisation HR 0.85 (NS individually); benefit consistent across ARNi, SGLT2i, and quadruple GDMT subgroups. Uses digitoxin (hepatic metabolism), not digoxin (renal clearance). Underpowered (1,212 of planned 2,190); median treatment only 18 months (58.9% discontinued). (sources/digitoxin-hfref-digithf-nejm-2025, sources/HF-AHA-2022, rating: high–very high)

• Omecamtiv mecarbil (selective cardiac myosin activator): Investigational — not approved, not guideline-listed (May 2026). GALACTIC-HF (Teerlink et al., NEJM 2021; n=8,232; LVEF ≤35%; NYHA II–IV; inpatients + outpatients; >19% ARNi; only 2.6% SGLT2i — pre-DAPA-HF era; 945 sites, 35 countries; median 21.8 months; Amgen/Cytokinetics/Servier-funded): primary composite (HF event or CV death) HR 0.92 (95% CI 0.86–0.99; P=0.03); 37.0% vs 39.1%; 2.1 pp absolute reduction. All secondary outcomes NS: CV death HR 1.01 (NS); all-cause death HR 1.00 (NS); first HF hospitalisation HR 0.95 (NS); KCCQ did not meet prespecified alpha. Critical LVEF subgroup interaction (prespecified): LVEF ≤28% HR 0.84 (0.77–0.92) — benefit; LVEF >28% HR 1.04 (0.94–1.16) — no benefit. NT-proBNP 10% lower at week 24; troponin I +4 ng/L (small, below ULN, no ischemic excess). Mechanism: binds cardiac myosin → increases force-generating myosin heads without affecting calcium transients — mechanistic distinction from classical inotropes. See entities/Omecamtiv-Mecarbil. (sources/omecamtiv-galactichf-nejm-2021, rating: high)

Drugs to Avoid in HFrEF

AHA COR 3 Harm: non-dihydropyridine CCBs (negative inotrope), Class IC antiarrhythmics + dronedarone (pro-arrhythmic in structural disease), thiazolidinediones (fluid retention), saxagliptin/alogliptin (HF hospitalisation signal), NSAIDs. (sources/HF-AHA-2022, rating: very high)

Hypertension Management in HFrEF

• Uptitrate GDMT (ARNi/ACEi/ARB + BB + MRA + SGLT2i) to maximally tolerated doses as first-line BP strategy — serves dual purpose of HF benefit and BP reduction. (sources/HT-AHA-2025, rating: very high)
• Add dihydropyridine CCB (e.g., amlodipine) only if BP remains elevated despite optimised GDMT. (sources/HT-AHA-2025, rating: very high)
• Non-dihydropyridine CCBs: COR 3 Harm in HFrEF — negative inotropic effects. Consistent across 2025 HT and HF guidelines. (sources/HT-AHA-2025, rating: very high)
• Goal SBP <130 mmHg; SPRINT: HF incidence reduced 38% (HR 0.62) with SBP <120 mmHg target. (sources/HT-AHA-2025, rating: very high)

Antiarrhythmic Drugs in HFrEF

• Amiodarone: Only antiarrhythmic without evidence of net mortality harm in HFrEF overall — SCD-HeFT (Bardy et al., NEJM 2005; n=2,521; LVEF ≤35%; NYHA II/III; median 45.5 months): overall HR 1.06 (P=0.53) vs placebo — no benefit. Significant NYHA class interaction (P=0.004): NYHA II HR 0.85 (NS) — neutral; NYHA III HR 1.44 (97.5% CI 1.05–1.97) — potential harm. Amiodarone does not substitute for ICD in primary prevention. The "acceptable in HFrEF" designation applies primarily to NYHA II; use with caution in NYHA III where the harm signal (though biologically uncertain) is a prespecified significant interaction. (sources/icd-hfref-scdheft-nejm-2005, sources/amiodarone-cvdrug-2020, rating: high; sources/HF-AHA-2022, rating: very high)
• Class IC agents (flecainide, propafenone) and dronedarone: actively harmful in structural heart disease / HFrEF — COR 3 Harm. (sources/HF-AHA-2022, rating: very high)
• See entities/Amiodarone for full pharmacology, dosing, toxicity monitoring, and drug interactions.

Iron Deficiency

Applies to HFrEF and HFmrEF (most evidence LVEF ≤45%); IV iron not currently recommended in HFpEF. Iron deficiency: ferritin <100 ng/mL OR ferritin 100–300 ng/mL with TSAT <20%. See concepts/Iron-Deficiency-in-HF for full evidence synthesis.

• To improve symptoms and QoL: IV iron (ferric carboxymaltose or ferric derisomaltose) ESC Class I, Level A (upgraded from IIa in 2021; based on FAIR-HF, CONFIRM-HF, EFFECT-HF). (sources/HF-update-ESC-2023, rating: very high)
• To reduce HF hospitalisation risk: Ferric carboxymaltose or ferric derisomaltose ESC Class IIa, Level A (based on AFFIRM-AHF + IRONMAN). (sources/HF-update-ESC-2023, rating: very high)
• HEART-FID (Mentz et al., NEJM 2023; n=3,065; ambulatory LVEF ≤40%; median 1.9 yr): The largest IV iron RCT to date. Hierarchical primary composite (death/HF hospitalizations/6MWT): win ratio 1.10 (99% CI 0.99–1.23; P=0.02) — failed to meet pre-specified p<0.01 threshold. Main secondary (CV death or HF hospitalisation): HR 0.93 (96% CI 0.81–1.06) — NS. All-cause death: HR 0.90 (95% CI 0.78–1.05) — NS. Safety acceptable. Contrasts with AFFIRM-AHF (acute HF at discharge; rate ratio 0.79; near-significant) likely due to lower-risk ambulatory population, higher baseline TSAT, COVID-19 pandemic hospitalization attenuation, and stringent p<0.01 threshold. Supports continued safety of repeated IV iron but adds uncertainty about hospitalization benefit in stable ambulatory HFrEF. (sources/fe-hf-heartfid-nejm-2023, rating: high)

Ischaemic Cardiomyopathy — Revascularisation

Revascularisation Strategy (sources/CCS-AHA-2023, rating: very high)

• CABG for LVEF ≤35% (AHA COR 1/B-R): STICH trial (n=1,212; LVEF ≤35%; CAD amenable to CABG): CABG + OMT vs OMT alone — CV death lower (28% vs 33%; P=0.05); all-cause death lower at 10-year follow-up. Patient selection: multi-vessel disease, viable myocardium, acceptable surgical risk.
• PCI for LVEF ≤35% — no survival benefit: REVIVED-BCIS2 (Perera et al. 2022, NEJM; n=700; LVEF ≤35%; BCIS jeopardy score ≥6; viability in ≥4 dysfunctional segments; 40 UK centres; median 41 months): PCI + GDMT vs GDMT alone — no difference in all-cause death or HF hospitalisation (HR 0.99; P=0.96; 37.2% vs 38.0%). LVEF did not differ at 6m (−1.6 pp; NS) or 12m (+0.9 pp; NS). Early KCCQ benefit (+6.5 pts at 6m, above MCID) eroded to non-significance by 24m as OMT group improved. ICA can still diagnose HF aetiology and direct lipid/medical therapy. (sources/pci-hf-revived-bcis2-nejm-2022, rating: very high)
• Viability imaging: LGE-CMR, PET-FDG, SPECT, and dobutamine echo all predict segmental contractile recovery after revascularisation. However, no trial has demonstrated that viability-guided revascularisation improves survival vs GDMT alone — including STICH viability substudy (SPECT/dobutamine echo; n=618; viability did not discriminate who benefited from CABG) and REVIVED-BCIS2. See concepts/Myocardial-Viability for full imaging performance data and clinical algorithms. (sources/imaging-viability-aha-2020, rating: very high)
• SGLT2i in ischaemic HFrEF: CV death and worsening HF reduction is independent of aetiology (ischaemic ~50% of DAPA-HF and EMPEROR-Reduced populations). (sources/HF-AHA-2022, rating: very high)

Device Therapy

ICD

• Secondary prevention (prior VT/VF/SCA): Class I regardless of aetiology; AUC 2025 A(9). (sources/HF-ESC-2021, sources/icd-crt-auc-2025, rating: very high)
• Primary prevention — ischaemic DCM: LVEF ≤35% after ≥3 months OMT, >40 days post-MI; ESC Class I; AUC 2025 A(8–9). MADIT-II (n=1,232; prior MI + LVEF ≤30%; no EPS required; mean LVEF 23%; mean 20 months): all-cause mortality HR 0.69 (95% CI 0.51–0.93; P=0.016); 14.2% vs 19.8%; consistent subgroups; benefit emerged ~9 months; HF hospitalisation trend higher ICD arm. SCD-HeFT (Bardy et al., NEJM 2005; n=2,521; 52% ischaemic/48% non-ischaemic; LVEF ≤35%; NYHA II [70%]/III [30%]; ACEi + BB background; median 45.5 months): ICD HR 0.77 (97.5% CI 0.62–0.96; P=0.007); 23% RRR; absolute −7.2% at 5 years; no aetiology interaction (P=0.68). The LVEF ≤35% guideline threshold derives directly from SCD-HeFT (MADIT-II used ≤30%). NYHA II HR 0.54 (absolute −11.9% at 5 years); NYHA III HR 1.16 (NS) — ICD × NYHA interaction P<0.001. (sources/icd-hfref-scdheft-nejm-2005, sources/icd-icm-maditii-nejm-2002, sources/HF-ESC-2021, sources/icd-crt-auc-2025, rating: very high)
• Primary prevention — non-ischaemic DCM: LVEF ≤35% after ≥3 months GDMT; ESC Class IIa; AUC 2025 A(7–9). SCD-HeFT established that ICD benefit applies equally to non-ischaemic HF (48% of population; no aetiology interaction P=0.68). DEFINITE (Kadish et al., NEJM 2004; n=458; NICM; LVEF <36%; NSVT or ≥10 PVCs/hour; ACEi + beta-blocker background; mean 29 months): first large NICM-specific primary prevention RCT; all-cause mortality HR 0.65 (P=0.08 — NS); sudden arrhythmic death HR 0.20 (P=0.006); subgroups: men HR 0.49, NYHA III HR 0.37; underpowered because only ~1/3 deaths were arrhythmic. DANISH trial (n=1,116; LVEF ≤35%; median 67.6 months; 58% CRT): no significant all-cause mortality benefit (HR 0.87; 95% CI 0.68–1.12; P=0.28); SCD halved (HR 0.50; P=0.005); CV death HR 0.77 (NS); all-cause mortality benefit significant only in age <68 subgroup (HR 0.64; P=0.01). Enrolled 2008–2014 before SGLT2i/ARNi era; 31% of deaths were non-CV. ESC Class IIa (not Class I) reflects modern GDMT-era uncertainty — a caveat absent in SCD-HeFT which predated ARNi/SGLT2i. (sources/icd-nicm-definite-nejm-2004, sources/icd-hfref-scdheft-nejm-2005, sources/HF-ESC-2021, sources/icd-crt-auc-2025, sources/icd-nicm-danish-nejm-2016, rating: high–very high)
• High-risk pathogenic variants (LMNA/C, desmosomal proteins, phospholamban, FLNC): ICD consideration even with LVEF >35% — arrhythmia risk disproportionate to EF. (sources/HF-AHA-2022, rating: very high)
• ICD in CKD: Benefit confirmed at eGFR >60 (HR 0.49) but not eGFR <60 (HR 0.80) in meta-analysis (MADIT I/II, SCD-HeFT); Cleveland Clinic registry: benefit in eGFR 30–59 but not eGFR <30. Higher complication rates in CKD. S-ICD is a reasonable alternative — no lead complications; CKD is an independent predictor of appropriate therapy for polymorphic VT/VF (HR 2.10). (sources/cardiorenal-aha-2019, rating: very high)
• See entities/ICD for full AUC 2025 scenarios, S-ICD considerations, and generator replacement appropriateness.

CRT

• Class I (ESC), A(9) (AUC 2025): LVEF ≤35% + LBBB + QRS ≥150 ms + NYHA II–IV on GDMT. (sources/HF-ESC-2021, sources/icd-crt-auc-2025, rating: very high)
• Class IIa (ESC), A(7) (AUC 2025): LBBB + QRS 120–149 ms. (sources/HF-ESC-2021, sources/icd-crt-auc-2025, rating: very high)
• Non-LBBB QRS ≥150 ms + NYHA III–IV: AUC A(7). Non-LBBB QRS <150 ms: AUC M(4–6). Narrow QRS: Class III / AUC R(1) — no benefit, possible harm (ECHO-CRT, RethinQ). (sources/icd-crt-auc-2025, rating: high)
• Conduction system pacing (HBP or LBBAP) vs RVP in CRT-indicated patients: AUC A(7). CSP as CRT alternative (instead of LV lead): AUC M(5–6) — pending PROTECT-HF and Left vs Left trials. (sources/icd-crt-auc-2025, rating: high)
• CRT in CKD: NYHA, LVEF, and functional MR improvements across eGFR 30–59 (MIRACLE post-hoc); eGFR improvement in CKD3 subgroup. Higher all-cause mortality post-CRT with eGFR <60 vs ≥60 (HR 1.66; meta-analysis). Weigh HF benefit against mortality risk in advanced CKD. (sources/cardiorenal-aha-2019, rating: very high)
• CARE-HF (Cleland et al., NEJM 2005) — foundational CRT mortality evidence: n=813; NYHA III/IV; LVEF ≤35%; QRS ≥120ms + echo dyssynchrony criteria; CRT-P (no defibrillator) vs medical therapy; 82 European centres; mean 29.4 months. Primary composite (death/unplanned CV hospitalisation): HR 0.63 (P<0.001); 39% vs 55%. All-cause mortality: HR 0.64 (P<0.002); NNT≈9; 20% vs 30%. HF hospitalisations 18% vs 33%. Core-lab echo confirmed LV reverse remodeling (↑LVEF, ↓LVEDVI, ↓mitral regurgitation, ↓NT-proBNP at 18m). First definitive CRT mortality RCT — COMPANION had shown CRT-P mortality trend only (P=0.06 at 16 months). Residual SCD 7% (29 patients) despite CRT-P — basis for preferring CRT-D in ICD-eligible patients. (sources/crt-carehf-nejm-2005, rating: very high)
• MADIT-CRT (Moss et al., NEJM 2009) — CRT extended to NYHA I/II: n=1,820; LVEF ≤30%; QRS ≥130ms; NYHA I/II ischaemic or NYHA II non-ischaemic; sinus rhythm; CRT-D vs ICD alone; mean 2.4 years; stopped early. Primary composite (death/nonfatal HF event): HR 0.66 (P=0.001); 17.2% vs 25.3%. Benefit driven by 41% reduction in HF events; no mortality difference at 2.4 years (3%/year each arm). Pivotal QRS interaction: QRS ≥150ms HR 0.48 (strong benefit) vs QRS 130–149ms HR 1.06 (no benefit; P=0.001 interaction) — basis for modern QRS ≥150ms Class I threshold. Women: HR 0.37 vs men HR 0.76 (P=0.01 interaction). Aetiology-independent benefit (ischaemic = non-ischaemic). Significant LV reverse remodeling at 1 year (↓LVEDV, ↓LVESV, ↑LVEF). (sources/crt-maditcrt-nejm-2009, rating: very high)
• See entities/CRT for full AUC 2025 scenarios, AF considerations, and RV pacing-induced CMP upgrade criteria.

TEER for Secondary MR

• ESC 2025 Class I A: TEER recommended for haemodynamically stable symptomatic patients with LVEF 20–50%, persistent severe ventricular SMR despite optimised GDMT + CRT, LVESD ≤70 mm, ≥1 HF hospitalisation/year or elevated BNP, SPAP ≤70 mmHg, no severe RV dysfunction. Based on COAPT, RESHAPE-HF2, and meta-analysis. (sources/vhd-esc-2025, rating: very high)
• Benefit driven by disproportionate MR (COAPT); proportionate MR (MITRA-FR) showed no benefit — anatomy-guided patient selection critical.

HFimpEF

• Definition (AHA 2022): Prior LVEF ≤40% with follow-up LVEF >40% on GDMT. Distinct from "recovered EF" — structural abnormalities and relapse risk persist. (sources/HF-AHA-2022, rating: very high)
• Continue GDMT (COR 1, LOE B-R): Withdrawal → 40% relapsed within 6 months; only 50% withdrew successfully long-term. Do not stop GDMT on the basis of LVEF normalisation alone. (sources/HF-AHA-2022, rating: very high)
• Highest LVEF reverse remodeling rates in TTNtv-associated DCM (~70%). See entities/TTN and entities/DCM.

Advanced HF

See entities/Heart-Failure for full advanced HF criteria, LVAD candidacy, heart transplantation listing, and cardiorenal considerations in LVAD/HTx.

Contradictions / Open Questions

Device Therapy

• MADIT-CRT: QRS 130–149ms no benefit; no mortality benefit; LBBB not required; pre-quadruple GDMT: MADIT-CRT enrolled QRS ≥130ms but the prespecified QRS interaction showed benefit only at ≥150ms (HR 0.48) — QRS 130–149ms HR 1.06 (no benefit). The current QRS ≥150ms Class I threshold derives directly from this interaction, not from the entry criterion. No mortality benefit was demonstrated at 2.4 years (annual mortality 3% identical both arms); long-term follow-up showed mortality benefit later. QRS morphology (LBBB vs non-LBBB) was not a primary subgroup in MADIT-CRT — subsequent LBBB-specific meta-analyses show that most CRT benefit is in LBBB; MADIT-CRT does not directly test the modern LBBB-first guideline. Pre-ARNi/SGLT2i era; enrolled 2004–2008; single manufacturer (Boston Scientific). (sources/crt-maditcrt-nejm-2009, rating: very high)
• CARE-HF: CRT-P mortality evidence limited to NYHA III/IV; 7% residual SCD; QRS not restricted to LBBB: CARE-HF established CRT-P all-cause mortality benefit (HR 0.64; P<0.002) in NYHA III/IV with QRS ≥120ms and echo dyssynchrony — not LBBB morphology specifically. Current Class I LBBB requirement is based on meta-analyses and post-hoc subgroup analyses showing attenuated non-LBBB response; CARE-HF itself did not require LBBB. 7% of CRT patients still died suddenly — the mechanistic argument for CRT-D (biventricular pacing + ICD) over CRT-P in ICD-eligible patients. NYHA II CRT benefit was established only by MADIT-CRT (2009) and REVERSE, not by CARE-HF. Pre-ARNi/SGLT2i era — modern quadruple GDMT independently promotes LV reverse remodeling, potentially narrowing the absolute CRT benefit. (sources/crt-carehf-nejm-2005, rating: very high)
• MADIT-II threshold extrapolation and trial stopping: MADIT-II used LVEF ≤30% as the entry criterion — current Class I guidelines apply ICD to LVEF ≤35%, which extends beyond the direct MADIT-II evidence and is partially supported by SCD-HeFT. MADIT-II was also stopped early at efficacy boundary (mean 20-month follow-up), which typically overestimates effect magnitude. The HF hospitalisation trend was higher in the ICD arm (19.9% vs 14.9%; P=0.09), consistent with a rescue effect: patients saved from SCD accumulate progressive HF. Enrolled 1997–2001 before SGLT2i, ARNi, and CRT-D were standard — modern quadruple GDMT substantially reduces background SCD risk, narrowing the residual NNT. (sources/icd-icm-maditii-nejm-2002, rating: very high)
• DANISH trial predates contemporary GDMT: Enrolled 2008–2014 before SGLT2i and ARNi were standard — modern quadruple GDMT likely reduces background SCD risk further, shifting the NNT for primary prevention ICD in NICM. DANISH showed SCD halved (HR 0.50) but all-cause mortality not significantly reduced (HR 0.87; P=0.28), partly because 31% of deaths were non-CV (competing mortality). A pre-specified age subgroup found significant benefit in patients <68 years (HR 0.64; P=0.01) but not ≥68 years. ESC Class IIa (not Class I) for NICM reflects this era-dependent uncertainty. (sources/icd-nicm-danish-nejm-2016, sources/icd-crt-auc-2025, sources/HF-ESC-2021, rating: high)
• SCD-HeFT ICD NYHA class interaction — no overall benefit in NYHA III: SCD-HeFT (n=2,521; LVEF ≤35%; ACEi + BB only; pre-SGLT2i/ARNi) found a prespecified ICD × NYHA interaction (P<0.001): NYHA II HR 0.54 (absolute −11.9% at 5 years) vs NYHA III HR 1.16 (NS). The attenuation in NYHA III was unexpected; investigators advised against withholding ICD in NYHA III. Current Class I guidelines recommend ICD for NYHA II–III without NYHA class stratification. The shock-only single-lead ICD design (no ATP, no CRT-D) and higher competing HF mortality in NYHA III may partly explain the finding — benefit was not tested in a modern quadruple GDMT context. (sources/icd-hfref-scdheft-nejm-2005, rating: very high)
• Amiodarone SCD-HeFT — NYHA class interaction and NYHA III harm signal: SCD-HeFT (Bardy et al., NEJM 2005; LVEF ≤35%; n=2,521) showed amiodarone HR 1.06 overall (NS) vs placebo — neutral, not beneficial. A prespecified NYHA class interaction (P=0.004) revealed: NYHA II HR 0.85 (NS) — neutral; NYHA III HR 1.44 (97.5% CI 1.05–1.97) — statistically significant increased mortality. The "acceptable in HFrEF" designation applies most defensibly to NYHA II; in NYHA III, the significant harm signal challenges this framing. Guidelines have not stratified amiodarone safety by NYHA class. ICD remains the primary prevention strategy regardless of NYHA class. (sources/icd-hfref-scdheft-nejm-2005, sources/amiodarone-cvdrug-2020, rating: very high / high)

Iron Deficiency — IV Iron

• HEART-FID vs AFFIRM-AHF divergence — ambulatory vs acute HF window: HEART-FID (n=3,065; ambulatory; median follow-up 1.9 yr) failed its primary and secondary endpoints despite a numerically favorable win ratio of 1.10 (P=0.02 vs required p<0.01). AFFIRM-AHF (n=1,132; acute HF at discharge) was near-significant (rate ratio 0.79; P=0.059) with a significant reduction in total HF hospitalizations. The key hypothesis is that the peri-discharge acute HF window represents a high-risk, iron-responsive state that is not replicated in stable ambulatory patients. However, HEART-FID's stringent p<0.01 threshold (FDA special protocol agreement) means a P=0.02 result would have been reported as significant in any standard trial. The net effect: ESC Class I (symptoms) and Class IIa (hospitalization) recommendations rest on AFFIRM-AHF/IRONMAN/older symptom trials, not on HEART-FID, which provides the largest but most uncertain single piece of evidence. Incremental benefit of IV iron in patients on contemporary quadruple GDMT + SGLT2i is untested. (sources/fe-hf-heartfid-nejm-2023, rating: high; sources/HF-update-ESC-2023, rating: very high)

GDMT Pharmacotherapy

• PARADIGM-HF stopped early + run-in enrichment: Early stopping can overestimate treatment effect. Double sequential run-in excluded ~20% of screened patients (intolerant to enalapril or sacubitril-valsartan) — enrolled population is more tolerant and potentially higher-risk than real-world HFrEF. (sources/arni-paradigm-hf-nejm-2014, rating: very high)

• PARADIGM-HF KCCQ below MCID: KCCQ improvement +1.64 points is statistically significant but below the ~5-point minimal clinically important difference — patient-perceived QoL benefit of sacubitril-valsartan may be modest at the population level. (sources/arni-paradigm-hf-nejm-2014, rating: very high)

• PARADIGM-HF pre-SGLT2i era: Enalapril comparator and background GDMT predates SGLT2i as fourth pillar — absolute benefit of sacubitril-valsartan vs contemporary quadruple GDMT is unknown. (sources/arni-paradigm-hf-nejm-2014, rating: very high)

• VICTORIA pre-SGLT2i era: Vericiguat enrolled 2016–2018; no SGLT2i co-administration data. Incremental benefit on top of modern quadruple GDMT is untested. (sources/vericiguat-victoria-nejm-2020, rating: very high)

• RALES stopped early + pre-beta-blocker era + real-world hyperkalemia divergence: RALES was stopped after mean 24-month follow-up when the efficacy boundary was crossed — early termination typically overestimates treatment effect magnitude; the 30% RRR may be inflated relative to a full-duration trial. Background therapy was ACEi + loop diuretic only; beta-blockers were not required (CIBIS-II and carvedilol published concurrently in 1999 — the BB era was just beginning). The incremental MRA benefit on top of modern quadruple GDMT (ARNi + BB + SGLT2i) is not directly established by RALES. Funded by Searle (spironolactone manufacturer). Critically, the trial's low hyperkalemia rate (2%) was not replicated in clinical practice: an observational study (Juurlink et al., JAMA 2004) found a 20-fold increase in spironolactone prescriptions and a 6.8-fold increase in hyperkalemia hospitalisation in elderly patients after RALES publication, with associated excess mortality — reflecting routine clinical failure to apply RALES's strict creatinine ≤2.5 mg/dL and potassium ≤5.0 mmol/L exclusion criteria. (sources/mra-hfref-rales-nejm-1999, rating: very high)

• DAPA-HF: low ARNi use limits modern applicability: Only ~10% of DAPA-HF patients were on sacubitril-valsartan at baseline (enrolled 2017–2018 before ARNi was widely standard). The post hoc ARNi subgroup HR 0.75 (95% CI 0.50–1.13) is consistent with overall benefit but underpowered. Whether the incremental benefit of adding SGLT2i to contemporary full quadruple GDMT (ARNi + BB + MRA + SGLT2i) matches the trial effect size is extrapolated, not directly tested. (sources/dapagliflozin-hfref-nejm-2019, rating: very high)

• DAPA-HF possible NYHA III/IV attenuation: Less benefit was observed in NYHA class III/IV vs II (the only prespecified subgroup with heterogeneity signal). However, other markers of disease severity (worse LVEF, elevated NT-proBNP, reduced eGFR) did not show attenuation, making the clinical significance uncertain. Guidelines recommend SGLT2i across NYHA classes II–IV. (sources/dapagliflozin-hfref-nejm-2019, rating: very high)

• EMPEROR-Reduced: mortality not significantly reduced despite stronger composite benefit signal: Despite enrolling a more severely ill population (73% LVEF ≤30%; ~40% higher primary event incidence than DAPA-HF), EMPEROR-Reduced showed no significant reduction in CV death (HR 0.92; NS) or all-cause death (HR 0.92; NS), compared with DAPA-HF where both were significant (CV death HR 0.82; all-cause HR 0.83). Whether this reflects the shorter median follow-up (16 vs 18.2 months), the higher background event rate, or a true drug/population difference is unresolved. The composite Class I recommendation rests on pooled HF hospitalisation reduction — the mortality pillar of GDMT evidence is more firmly established by DAPA-HF than EMPEROR-Reduced. (sources/empagliflozin-hfref-nejm-2020, sources/dapagliflozin-hfref-nejm-2019, rating: very high)

• EMPEROR-Reduced renal benefit: disease-modifying vs haemodynamic: The post-discontinuation eGFR data (−0.93 vs −4.21 mL/min/1.73m² after stopping drug) are the strongest available evidence that SGLT2i renal protection in HFrEF is structural/disease-modifying rather than purely haemodynamic. However, these data were available in only ~26% of the enrolled population (n=966 with paired measurements), and selective measurement may introduce survivorship or selection bias. A prospective renal-primary SGLT2i HFrEF trial does not exist. (sources/empagliflozin-hfref-nejm-2020, rating: very high)

• DIGIT-HF: underpowered + digitoxin ≠ digoxin + partial SGLT2i era: DIGIT-HF enrolled 1,212 of a planned 2,190 patients, and both individual components (death HR 0.86 NS; first HF hospitalisation HR 0.85 NS) were non-significant despite the significant composite. Median treatment duration was only 18 months (58.9% discontinued), introducing ITT dilution. The trial used digitoxin (hepatic metabolism, less renal accumulation) rather than digoxin (renal clearance, existing COR 2b), so the DIGIT-HF composite benefit cannot be directly attributed to guideline-listed digoxin. SGLT2i uptake was only 19.3% — incremental benefit of digitoxin/digoxin on top of full modern quadruple GDMT (ARNi + BB + MRA + SGLT2i) at complete adherence remains untested. (sources/digitoxin-hfref-digithf-nejm-2025, rating: high)

• GALACTIC-HF: primary composite significant but no secondary benefit + LVEF interaction: Omecamtiv mecarbil reduced the primary composite (HR 0.92; P=0.03) but CV death (HR 1.01), all-cause death (HR 1.00), first HF hospitalisation (HR 0.95), and KCCQ were all non-significant. A prespecified LVEF interaction showed benefit in LVEF ≤28% (HR 0.84) but not LVEF >28% (HR 1.04) — creating tension with the trial inclusion criterion (≤35%). Only 2.6% on SGLT2i (enrolled 2017–2019); incremental benefit on top of quadruple GDMT is unknown. The drug is not approved or guideline-listed. (sources/omecamtiv-galactichf-nejm-2021, rating: high)

Ischaemic CMP — Revascularisation

• REVIVED-BCIS2 CABG/PCI divergence: STICH showed 10-year CABG survival benefit; REVIVED-BCIS2 showed no PCI benefit (HR 0.99; P=0.96) in a viability-enriched population. Whether the discrepancy reflects completeness of surgical revascularisation, open-label design bias, or a true biological CABG/PCI difference in this setting is unresolved. The 2023 AHA/ACC guideline preserves CABG as COR 1; PCI does not carry a survival indication in LVEF ≤35%. (sources/pci-hf-revived-bcis2-nejm-2022, sources/CCS-AHA-2023, rating: very high)

Connections

• Related to entities/Heart-Failure — full HF hub including HFmrEF, HFpEF, acute HF, advanced HF
• Related to entities/ICD
• Related to entities/CRT
• Related to entities/S-ICD
• Related to entities/Sacubitril-Valsartan
• Related to entities/Vericiguat
• Related to entities/Amiodarone
• Related to entities/DCM
• Related to entities/TTN — highest reverse remodeling rate (~70%); TTNtv most common DCM gene
• Related to entities/LMNA — early ICD consideration with LVEF >35%
• Related to entities/Chronic-Coronary-Disease — ischaemic CMP; CABG vs PCI strategy
• Related to concepts/Myocardial-Viability — viability imaging and revascularisation evidence
• Related to concepts/Electrical-Remodeling — ionic substrate in HFrEF; CaMKII; antiarrhythmic drug safety
• Related to concepts/Conduction-System-Pacing — CSP vs RVP AUC 2025 ratings
• Related to concepts/Cardiorenal-Syndrome — ICD benefit in CKD; MRA hyperkalemia risk; diuretic resistance
• Related to entities/Omecamtiv-Mecarbil — investigational myotrope; GALACTIC-HF; LVEF ≤28% subgroup
• Related to entities/COPD — beta-blocker choice in HFrEF+COPD; bisoprolol preferred; β-blockers reduce both HF and COPD exacerbations
• Related to concepts/HF-COPD-Comorbidity — treatment algorithm, pathophysiology, and diagnostic challenges in HFrEF+COPD

Sources

• sources/arni-paradigm-hf-nejm-2014
• sources/pci-hf-revived-bcis2-nejm-2022
• sources/vericiguat-victoria-nejm-2020
• sources/icd-crt-auc-2025
• sources/HF-AHA-2022
• sources/HF-ESC-2021
• sources/HF-update-ESC-2023
• sources/CCS-AHA-2023
• sources/HT-AHA-2025
• sources/amiodarone-cvdrug-2020
• sources/membrane-potential-physrev-2021
• sources/hf-copd-hfreview-2025
• sources/imaging-viability-aha-2020
• sources/cardiorenal-aha-2019
• sources/vhd-esc-2025
• sources/dapagliflozin-hfref-nejm-2019
• sources/empagliflozin-hfref-nejm-2020
• sources/icd-nicm-danish-nejm-2016
• sources/icd-nicm-definite-nejm-2004
• sources/icd-icm-maditii-nejm-2002
• sources/digitoxin-hfref-digithf-nejm-2025
• sources/icd-hfref-scdheft-nejm-2005
• sources/mra-hfref-rales-nejm-1999
• sources/omecamtiv-galactichf-nejm-2021
• sources/crt-carehf-nejm-2005
• sources/crt-maditcrt-nejm-2009
• sources/fe-hf-heartfid-nejm-2023