Prophylactic Implantation of a Defibrillator in Patients with Myocardial Infarction and Reduced Ejection Fraction (MADIT-II)

Authors, Journal, Affiliations, Type, DOI

• Arthur J. Moss, Wojciech Zareba, W. Jackson Hall, and MADIT-II Investigators
• New England Journal of Medicine, 2002; 346(12):877–883
• University of Rochester Medical Center; 76 centres (71 US, 5 Europe)
• Multicentre, open-label randomised controlled trial; industry-funded (Guidant Corporation)
• DOI: https://doi.org/10.1056/NEJMoa013474

Overview

MADIT-II was a landmark multicentre RCT enrolling 1,232 patients with prior MI and LVEF ≤30%, randomised 3:2 to ICD vs conventional medical therapy without requirement for electrophysiological testing. Over a mean follow-up of 20 months the ICD arm showed a significant 31% reduction in all-cause mortality (HR 0.69; 95% CI 0.51–0.93; P=0.016), with consistent benefit across all prespecified subgroups. By removing the EPS requirement of MADIT-I and using LVEF as the sole entry criterion, MADIT-II fundamentally simplified primary prevention ICD eligibility and is the foundational evidence base for current Class I/AUC A(8–9) recommendations in ischaemic cardiomyopathy.

Keywords

Prophylactic ICD, myocardial infarction, reduced ejection fraction, primary prevention, sudden cardiac death, survival

Key Takeaways

Study Design and Population

• n=1,232 randomised 3:2: ICD n=742 vs conventional therapy n=490
• 76 centres (71 US, 5 Europe); enrolled July 1997; stopped November 20, 2001 at pre-specified efficacy boundary
• Inclusion: prior MI ≥1 month + LVEF ≤30%; no EPS/inducible arrhythmia requirement
• Exclusion: NYHA class IV at enrolment; revascularisation <3 months; MI <1 month; prior ICD indication; advanced non-cardiac illness
• Mean age 64±10 years; 84% male; mean LVEF 23%; NYHA I/II/III: 35%/35%/25%
• Background therapy: ACEi ~70%, beta-blockers 70%, lipid-lowering statins ~65%, amiodarone ~10–13%
• Average follow-up 20 months (range 6 days to 53 months)

Primary Endpoint — All-Cause Mortality

• ICD arm: 14.2% mortality; conventional arm: 19.8% mortality
• HR 0.69 (95% CI 0.51–0.93; P=0.016) — 31% relative risk reduction
• Kaplan–Meier nominal P=0.007
• Trial stopped when efficacy boundary reached (P=0.027); final analysis P=0.016 after lag reporting

Subgroup Analyses

• Consistent benefit across all subgroups: age (<60, 60–69, ≥70), sex, LVEF (≤0.25 vs >0.25), NYHA class (I vs ≥II), QRS interval (<120ms, 120–150ms, >150ms)
• No statistically significant subgroup interactions
• Nominal (unadjusted) HR 0.66 across the full cohort

Survival Curve Pattern

• Survival benefit emerged at approximately 9 months post-implantation
• Contrast with MADIT-I (EPS-required, higher-risk population) where benefit appeared within the first few months
• Later divergence reflects lower background mortality in MADIT-II vs MADIT-I

Heart Failure Signal

• HF hospitalisation: 19.9% (ICD) vs 14.9% (conventional); P=0.09 — not significant but a concerning trend
• Interpreted as a rescue effect: patients saved from SCD by ICD live longer and thus have more time to develop progressive HF
• RV pacing from ICD backup pacing may contribute to LV dyssynchrony and worsen HF

Device Safety

• Lead problems requiring surgical intervention: 1.8% (13 patients)
• Device infections: 0.7% (5 patients)
• 54 crossovers total: 22 from control arm received ICD (4.5%); 21 from ICD arm did not receive device (2.8%)

Historical Significance

• First trial to demonstrate ICD survival benefit without requiring EPS or inducible arrhythmia
• Established LVEF as the sole eligibility criterion — fundamentally simplified primary prevention ICD prescription
• Predecessor MADIT-I (1996) required NSVT + inducible VT; MADIT-II eliminated both requirements
• Current Class I guideline threshold is LVEF ≤35% (not ≤30%) — extension extrapolated in part from SCD-HeFT (LVEF ≤35%)

Limitations of the document

• Trial stopped early for efficacy (P=0.027) — early stopping typically overestimates treatment effect magnitude
• Industry-funded by Guidant Corporation (device manufacturer); potential for publication/outcome reporting bias
• Mean follow-up only 20 months — long-term benefit trajectory and device-related complications at 5–10 years not captured
• Pre-SGLT2i, pre-ARNi, pre-CRT-D era (enrolled 1997–2001); contemporary quadruple GDMT substantially improves LVEF and may reduce arrhythmic substrate, narrowing the residual NNT
• LVEF ≤30% threshold is lower than the ≤35% threshold now used in current guidelines — the extension to ≤35% is extrapolated from SCD-HeFT, not directly from MADIT-II
• Higher HF hospitalisation trend in ICD arm (P=0.09) may indicate a long-term trade-off: preventing SCD at cost of progressive HF in a rescued population
• No sham control — open-label design with potential ascertainment bias

Key Concepts Mentioned

• concepts/VA-Risk-Stratification-DCM — LVEF-only threshold for primary prevention ICD; contrast with DCM/NICM evidence

Key Entities Mentioned

• entities/ICD — foundational primary prevention evidence in ischaemic cardiomyopathy; AUC A(8–9) basis
• entities/HFrEF — GDMT context; device therapy in ischaemic HFrEF

Wiki Pages Updated

• wiki/sources/icd-icm-maditii-nejm-2002.md — created
• wiki/entities/ICD.md — Primary Prevention CAD section enriched with MADIT-II headline data and direct citation; contradiction added; source_count 3→4
• wiki/entities/HFrEF.md — ischaemic DCM ICD bullet enriched with MADIT-II data and direct citation; contradiction added; source_count 17→18
• wiki/sourceindex.md — new entry
• wiki/wikiindex.md — ICD entry updated