Cardiac-Resynchronization Therapy with or without an Implantable Defibrillator in Advanced Chronic Heart Failure

Authors, Journal, Affiliations, Type, DOI

• Authors: Michael R. Bristow, Leslie A. Saxon, John Boehmer, Steven Krueger, David A. Kass, Teresa De Marco, Peter Carson, Lorenzo DiCarlo, David DeMets, Bill G. White, Dale W. DeVries, Arthur M. Feldman, for the COMPANION Investigators
• Journal: New England Journal of Medicine, 2004;350:2140–2150
• Affiliations: University of Colorado Health Sciences Center (lead); University of Southern California; Johns Hopkins; UCSF; University of Wisconsin–Madison; 128 US centres
• Type: Prospective, randomised, controlled trial (open-label; 3-arm; 1:2:2 allocation)
• Funding: Guidant Corporation (device manufacturer); independent statistical centre at University of Wisconsin
• DOI: Not provided in raw document

Overview

The COMPANION (Comparison of Medical Therapy, Pacing, and Defibrillation in Heart Failure) trial enrolled 1,520 patients with NYHA class III/IV heart failure (LVEF ≤35%; QRS ≥120ms) due to ischemic or non-ischemic cardiomyopathy, randomised 1:2:2 to optimal pharmacologic therapy (OPT) alone, OPT plus CRT-P, or OPT plus CRT-D. Both CRT arms reduced the primary composite endpoint of death or any-cause hospitalisation by approximately 20%; only CRT-D achieved a statistically significant all-cause mortality reduction (HR 0.64; P=0.003), while CRT-P yielded a borderline non-significant trend (HR 0.76; P=0.059). The trial established that biventricular pacing reduces hospitalisations and improves functional status, and that adding an ICD to CRT provides incremental survival benefit beyond resynchronisation alone.

Keywords

Cardiac resynchronisation therapy, biventricular pacing, intraventricular conduction delay, left bundle branch block, implantable cardioverter-defibrillator, advanced heart failure, dilated cardiomyopathy, QRS interval, dyssynchrony

Key Takeaways

Background

• Intraventricular conduction delays (≥120 ms QRS) occur in 15–30% of patients with HF due to dilated cardiomyopathy, causing dyssynchronous LV contraction that reduces systolic function and increases systolic volume
• Short-term studies had shown CRT improves symptoms, quality of life, exercise tolerance, and partially reverses maladaptive remodelling; long-term mortality effects were unknown
• ICD reduces mortality in ischaemic cardiomyopathy (MADIT-II); unclear whether additive benefit existed with CRT in advanced HF with severe LV dysfunction or non-ischaemic CM

Methods (Trial Design)

• Population: NYHA class III/IV HF; LVEF ≤35%; QRS ≥120ms; PR >150ms; sinus rhythm; ≥1 HF hospitalisation in preceding 12 months; ischaemic or non-ischaemic CM
• Exclusion: no bradycardia indication for pacemaker; no documented VT/VF qualifying for secondary prevention ICD; imminent transplant candidacy
• Randomisation: 1:2:2 → OPT alone (N=308) : CRT-P (N=617) : CRT-D (N=595); 128 US centres; enrollment January 2000 – November 2002
• Background OPT: diuretics + ACEi (or ARB) + beta-blockers + spironolactone; all groups
• Devices: CRT-P — Contak TR model 1241 (Guidant); CRT-D — Contak CD model 1823 (Guidant); LV lead via coronary sinus over-the-wire; AV delay calculated from proprietary algorithm; VDD pacing mode
• Blinding: patients, physicians, independent statisticians, and data-management team were NOT blinded; steering committee, end-points committee, and sponsor were blinded
• Primary endpoint: composite of death from any cause or first hospitalisation for any cause (from randomisation)
• Secondary endpoint: death from any cause; also death/hosp for CV causes; death/hosp for HF (post-hoc)
• Stopping: stopped November 2002 on recommendation of DSMB when pre-established stopping boundaries crossed in CRT-D group (primary + mortality endpoints) and CRT-P group (primary endpoint); 1,520 randomised at stopping

Results — Study Population

• No clinically significant baseline differences across three groups (Table 1; Wilcoxon/chi-square comparisons)
• Protocol-mandated OPT achieved in all groups at similar rates

Results — Implantation Success

• CRT-P: successful in 87% (539/617); median procedure time 164 minutes
• CRT-D: successful in 91% (541/595); median procedure time 176 minutes
• Procedural deaths: 0.8% (CRT-P) and 0.5% (CRT-D)
• 30-day mortality: 1.0% CRT-P vs 1.8% CRT-D vs 1.2% OPT (P=0.34 and P=0.79) — no significant procedural mortality increase
• Moderate/severe implant-related adverse events: 10% CRT-P; 8% CRT-D
• Notable implant complications: coronary venous perforation (~1%), coronary venous tamponade (~0.5%), coronary venous dissection (~0.3–0.5%)

Results — Withdrawal and Follow-Up

• 26% withdrawal in OPT group vs 6–7% in CRT groups — patients crossed over to receive commercially available CRT-P/CRT-D, particularly ischaemic CM patients post-MADIT-II publication
• "Reconsent" approach obtained for patients who withdrew before reaching an endpoint; vital status complete for 96% of OPT and 99% of CRT patients
• Median follow-up: OPT 14.8 months (mortality); CRT-P 16.5 months; CRT-D 16.0 months

Results — Primary Endpoint (Death or Any-Cause Hospitalisation)

• 12-month primary endpoint rate: 68% OPT vs 56% CRT-P vs 56% CRT-D
• CRT-P: HR 0.81 (95% CI 0.69–0.96; P=0.014; adjusted P=0.015) — 19% risk reduction
• CRT-D: HR 0.80 (95% CI 0.68–0.95; P=0.010; adjusted P=0.011) — 20% risk reduction
• Equivalent reduction in primary composite between CRT-P and CRT-D

Results — Secondary Endpoint (All-Cause Mortality)

• OPT: 77/308 deaths (25%) during study period; 1-year mortality 19%; 75% of deaths classified cardiac
• CRT-P: HR 0.76 (95% CI 0.58–1.01; P=0.059; adjusted P=0.06) — NOT statistically significant (24% reduction trend)
• CRT-D: HR 0.64 (95% CI 0.48–0.86; P=0.003; adjusted P=0.004) — 36% significant mortality reduction
• 12-month mortality: 19% OPT vs 15% CRT-P vs 12% CRT-D

Results — HF and CV Composite Endpoints

• Death or hosp for HF (post-hoc):
  • CRT-P: HR 0.66 (34% reduction; P=0.002)
  • CRT-D: HR 0.60 (40% reduction; P<0.001)
  • 12-month rate: 45% OPT vs 31% CRT-P vs 29% CRT-D
• Death or hosp for CV causes:
  • CRT-P: HR 0.75 (25% reduction; P=0.002)
  • CRT-D: HR 0.72 (28% reduction; P<0.001)
  • 12-month rate: 60% OPT vs 45% CRT-P vs 44% CRT-D

Results — Functional Status

• NYHA class, 6-minute walk distance, and Minnesota Living with Heart Failure questionnaire score were all significantly better in both CRT groups vs OPT at 3 and 6 months
• Systolic blood pressure changes from baseline significantly better in both CRT groups vs OPT — consistent with improved systolic function

Results — Subgroup Analyses

• Hazard ratios for primary endpoint consistently <1.0 across all baseline characteristic subgroups for both devices
• Progressive QRS widening → greater primary endpoint benefit for both CRT-P and CRT-D
• Ischaemic and non-ischaemic CM both benefited for primary endpoint
• Beta-blockers or spironolactone + CRT → lower hazard ratios than other agents
• Mortality subgroup (CRT-D vs OPT):
  • Non-ischaemic CM: HR 0.50 (95% CI 0.29–0.88; P=0.015) — significant 50% reduction
  • Ischaemic CM: HR 0.73 (95% CI 0.52–1.04; P=0.082) — 27% reduction, NOT significant (P=0.082); interaction test NS
  • CRT-P mortality: non-ischaemic HR 0.91 (NS); ischaemic HR 0.72 (P=0.058 borderline)

Results — Adverse Events

• Moderate/severe adverse events from any cause: 61% OPT vs 66% CRT-P (P=0.15) vs 69% CRT-D (P=0.03)
• No significant difference in device-related moderate/severe adverse events between CRT-P and CRT-D (P=0.42)

Discussion Key Points

• Biventricular pacing improves major clinical outcomes (hospitalisation, functional status) in NYHA III/IV with prolonged QRS and moderate-to-severe LV systolic dysfunction
• The even larger reduction in HF-specific endpoints (40% HF death/hosp with CRT-D) supports that most reduction derives from favourable device effect on HF syndrome, with augmented systolic function as a mediator
• Addition of a defibrillator to CRT does not appreciably affect combined hospitalisation/death endpoints (heavily driven by hospitalisations) but provides incremental mortality benefit
• Non-blinded design → high OPT withdrawal; reconsent strategy substantially reduced effect on ITT analysis
• 50% mortality reduction in non-ischaemic CM with CRT-D provides the first large-scale evidence for ICD utility in non-ischaemic cardiomyopathy (DEFINITE was published simultaneously in same NEJM issue)

Limitations of the Document

• Open-label design — patients and physicians not blinded; OPT arm had 26% withdrawal rate, driven by MADIT-II and commercial availability of CRT-D during trial; ischaemic CM patients disproportionately withdrew (68% vs 55%), potentially biasing subgroup comparisons
• Reconsent approach for withdrawn OPT patients — post-withdrawal data collection is methodologically non-standard; introduces uncertainty in ITT analysis completeness
• Stopped early — both CRT-P and CRT-D arms crossed pre-specified stopping boundaries before planned events accrued; early stopping typically overestimates treatment effects
• Short follow-up — median ~12–16 months, shorter than CARE-HF (29 months) or MADIT-CRT (2.4 years); long-term mortality trajectory unknown
• CRT-P mortality endpoint borderline (P=0.059) — CARE-HF (2005) subsequently confirmed CRT-P mortality benefit with longer follow-up; COMPANION's shorter duration explains the borderline result
• Industry funding (Guidant) — several investigators were consultants to or equity holders in Guidant; device company employee (DeVries) was a co-author; data held independently at University of Wisconsin
• No QRS morphology sub-analysis — LBBB vs non-LBBB not separately reported; modern guidelines stratify response by LBBB morphology, which COMPANION did not examine
• Pre-modern GDMT — spironolactone was mandated but SGLT2i, ARNi, and dapagliflozin were not available; modern GDMT further reduces arrhythmic substrate and LV dysfunction; NNT may differ in contemporary practice
• Mortality interaction test NS — non-ischaemic vs ischaemic CM mortality subgroup difference (HR 0.50 vs 0.73) was not statistically significant on interaction testing; subgroup effect should be interpreted cautiously

Key Concepts Mentioned

• concepts/Conduction-System-Pacing — biventricular CRT as the original coronary sinus-lead approach; trial predates conduction system pacing era
• concepts/VA-Risk-Stratification-DCM — incremental mortality benefit of CRT-D in advanced HF supports ICD component utility

Key Entities Mentioned

• entities/CRT — primary entity; COMPANION as foundational CRT-D mortality RCT in NYHA III/IV
• entities/ICD — CRT-D component; 36% mortality reduction in NYHA III/IV advanced HF
• entities/HFrEF — population; NYHA III/IV; LVEF ≤35%; QRS ≥120ms

Wiki Pages Updated

• wiki/sources/crt-companion-nejm-2004.md — created
• wiki/entities/CRT.md — COMPANION trial evidence added to Randomised Trial Evidence; contradiction added; source added; source_count updated
• wiki/entities/ICD.md — COMPANION cross-reference added in Primary Prevention — NICM section; source added
• wiki/sourceindex.md — new entry added
• wiki/wikiindex.md — CRT and HFrEF entity entries updated