Cardiac-Resynchronization Therapy for the Prevention of Heart-Failure Events (MADIT-CRT)

Authors, Journal, Affiliations, Type, DOI

• Authors: Arthur J. Moss, W. Jackson Hall, David S. Cannom, Helmut Klein, Mary W. Brown, James P. Daubert, N.A. Mark Estes III, Elyse Foster, Henry Greenberg, Steven L. Higgins, Marc A. Pfeffer, Scott D. Solomon, David Wilber, Wojciech Zareba; for the MADIT-CRT Trial Investigators
• Journal: New England Journal of Medicine, 2009;361:1329–1338
• Published: September 1, 2009 (online); October 1, 2009 (print)
• Type: Multicentre, randomised, open-label, controlled trial; 3:2 allocation (CRT-D vs ICD-only)
• Funding: Boston Scientific grant to University of Rochester; sponsor not involved in data collection or analysis; authors vouch for accuracy
• DOI: 10.1056/NEJMoa0906431
• Sites: 110 centres (88 US, 2 Canada, 20 Europe); enrolled December 2004 – April 2008

Overview

MADIT-CRT randomised 1,820 patients with LVEF ≤30%, QRS ≥130ms, and NYHA class I/II (ischaemic) or NYHA II (non-ischaemic) to CRT-D vs ICD alone over a mean 2.4 years. The primary composite of death or nonfatal HF event was reduced 34% (HR 0.66; P=0.001), driven entirely by a 41% reduction in HF events — with no mortality difference at this follow-up duration. The defining subgroup finding: benefit was confined to QRS ≥150ms (HR 0.48) and was absent or potentially harmful at QRS 130–149ms (HR 1.06; P=0.001 interaction), establishing the evidence base for the modern QRS ≥150ms Class I threshold. Women showed disproportionate benefit (HR 0.37 vs men HR 0.76). Together with CARE-HF (NYHA III/IV) and RAFT (NYHA II/III), MADIT-CRT extended guideline-supported CRT indications into the mild-symptom population.

Keywords

Cardiac resynchronization therapy, CRT-D, biventricular pacing, MADIT-CRT, QRS duration, LBBB, heart failure, NYHA class I/II, LV reverse remodeling, HF events, prevention

Key Takeaways

Background

• Prior CRT evidence (CARE-HF, COMPANION) confined to NYHA III/IV — symptomatic HF
• A preliminary study (REVERSE; Linde 2008 JACC) suggested CRT improved cardiac structure/function in mildly symptomatic patients — provided rationale for MADIT-CRT
• ICD-only therapy prolongs life in HFrEF but is associated with increased HF events (MADIT-II data) — hypothesis that adding CRT could prevent this HF progression
• 2008 ACC/AHA/HRS guidelines: CRT Class I for LVEF ≤35% + QRS ≥120ms + NYHA III/ambulatory IV; MADIT-CRT designed to test extension to NYHA I/II

Population

• n=1,820; CRT-D: 1,089; ICD-only: 731; 3:2 allocation ratio
• Ischaemic cardiomyopathy: NYHA class I or II
• Non-ischaemic cardiomyopathy: NYHA class II only (class I non-ischaemic excluded)
• LVEF ≤30% (stricter than CARE-HF ≤35%)
• QRS ≥130ms; sinus rhythm
• All met guideline ICD indications
• Excluded: existing CRT indication (NYHA III/IV); prior CABG/PCI/enzyme-positive MI within 3 months; AF within 1 month; existing pacemaker/ICD/CRT device
• Background therapy: optimal pharmacological HF therapy in both groups; pre-ARNi, pre-SGLT2i era (enrolled 2004–2008)
• Mean follow-up 2.4 years; trial stopped at 9th of 20 planned analyses (prespecified efficacy boundary reached June 22, 2009)

Device

• Commercially available transvenous CRT-D (Boston Scientific); standard implant technique
• CRT-D programmed: DDD, lower rate 40 bpm, hysteresis off
• ICD-only: VVI (single-chamber) or DDI (dual-chamber), lower rate 40 bpm

Primary Endpoint — Death or Nonfatal HF Event

• Composite of death from any cause or nonfatal HF event (whichever came first)
• HF event definition: signs/symptoms of HF responsive to IV decongestion (outpatient) or augmented oral/parenteral decongestive regimen (inpatient); adjudicated by blinded HF committee
• CRT-D: 17.2% (187/1,089) vs ICD-only: 25.3% (185/731)
• HR 0.66 (95% CI 0.52–0.84; P=0.001) — 34% reduction
• Curves diverge within first 2 months and maintain separation throughout
• Benefit consistent across ischaemic and non-ischaemic cardiomyopathy (no interaction)

Primary Endpoint Components

• HF events alone: CRT-D 13.9% vs ICD-only 22.8% — 41% reduction in HF risk
  • Hospitalised HF: 136 (CRT-D) vs 140 (ICD-only); outpatient HF: 15 vs 27
• Mortality: 36 deaths (3.3%) CRT-D vs 18 deaths (2.5%) ICD-only; annual mortality ~3% in each group — no significant mortality difference at mean 2.4 years

Prespecified Subgroups — Critical QRS and Sex Interactions

QRS duration (P=0.001 for interaction):

• QRS ≥150ms: HR 0.48 (95% CI 0.37–0.64) — strong benefit; 52% reduction
• QRS 130–149ms: HR 1.06 (95% CI 0.74–1.52) — no benefit, point estimate favours ICD alone

Sex (P=0.01 for interaction):

• Women: HR 0.37 (95% CI 0.22–0.61) — 63% reduction; marked benefit
• Men: HR 0.76 (95% CI 0.59–0.97) — modest but significant benefit

Other subgroups:

• Ischaemic vs non-ischaemic: no significant interaction — benefit similar in both
• Exploratory: patients with prior NYHA III/IV (>3 months before enrolment; n=182) and prior AF (>1 month; n=213) had less CRT-D effect than those without these histories
• No significant interaction by centre size or BUN level

LV Reverse Remodeling (1-Year Echocardiography)

• 746 paired studies in CRT-D group (CRT on at 1 year); 620 in ICD-only
• CRT-D showed significantly greater:
  • ↓ LV end-diastolic volume
  • ↓ LV end-systolic volume
  • ↑ LVEF
• All changes significantly greater than ICD-only group at 1 year

Adverse Events (30-Day Peri-Implant)

• 1 peri-implant death (pulmonary embolus at autopsy) — CRT-D group
• Pneumothorax: 1.7% (CRT-D) vs 0.8% (ICD-only)
• Infection: 1.1% (CRT-D) vs 0.7% (ICD-only)
• Pocket haematoma requiring evacuation: 3.3% (CRT-D) vs 2.5% (ICD-only)
• Coronary venous dissection with pericardial effusion: 0.5% (CRT-D)
• LV coronary-vein lead repositioned within 30 days: 4.0%
• Long-term device-related adverse events: 4.5 per 100 device-months (CRT-D) vs 5.2 per 100 device-months (ICD-only)
• Crossovers: 12.4% of ICD-only patients received CRT-D during trial (61 after HF event, 30 at physician discretion); 7.5% of CRT-D patients received ICD-only (technical failure to position LV lead)

Limitations of the Document

• Stopped early at 9th of 20 analyses — prespecified efficacy boundary reached after mean 2.4 years; early stopping typically overestimates effect magnitude; no mortality difference had emerged — longer follow-up needed
• Open-label: investigators aware of treatment assignment; HF event diagnosis made by aware investigators, though adjudication was by blinded committee — potential ascertainment/management bias inflating HF events in ICD-only arm
• QRS ≥130ms entry criterion but no benefit at QRS 130–149ms: The critical subgroup interaction (HR 1.06 for QRS 130–149ms) means the entry criterion was broader than the population that benefits — enrolling patients unlikely to benefit; modern guidelines restrict Class I CRT to QRS ≥150ms based partly on this finding
• LVEF ≤30% (not ≤35%) — more restrictive than CARE-HF; generalisation to LVEF 31–35% is extrapolation
• Non-ischaemic NYHA I excluded — NYHA I ischaemic included; asymmetric enrolment across aetiology-NYHA combinations
• No mortality benefit at 2.4 years — driven entirely by HF event reduction; mortality benefit at longer follow-up was subsequently shown in MADIT-CRT long-term follow-up analysis (Goldenberg et al., 2014) and meta-analyses
• QRS morphology not stratified: LBBB vs non-LBBB not reported as a primary subgroup — the QRS ≥150ms interaction does not distinguish whether benefit is LBBB-specific; subsequent analyses and meta-analyses (including MADIT-CRT LBBB vs non-LBBB substudy) showed benefit largely confined to LBBB
• Single manufacturer (Boston Scientific): lead and device characteristics may not generalise
• Pre-ARNi/SGLT2i era — background GDMT did not include quadruple therapy; modern GDMT independently promotes LV reverse remodeling, potentially narrowing absolute CRT benefit
• Women sex interaction was a prespecified subgroup but multiple testing; the sex × CRT benefit interaction has been replicated in subsequent meta-analyses

Key Concepts Mentioned

• concepts/Cardiac-Dyssynchrony — mechanistic target; QRS duration as surrogate
• concepts/LV-Reverse-Remodeling — primary mechanism of CRT benefit; echo endpoint

Key Entities Mentioned

• entities/CRT — primary subject; NYHA I/II evidence; QRS threshold
• entities/HFrEF — population; HF prevention; CRT section

Wiki Pages Updated

• Created wiki/sources/crt-maditcrt-nejm-2009.md
• Updated wiki/entities/CRT.md — MADIT-CRT added to Randomised Trial Evidence; QRS interaction contradiction added; source_count 2→3
• Updated wiki/entities/HFrEF.md — MADIT-CRT data added to CRT section; source_count 23→24
• Updated wiki/sourceindex.md
• Updated wiki/wikiindex.md
• Updated log.md