Prophylactic Defibrillator Implantation in Patients with Nonischemic Dilated Cardiomyopathy (DEFINITE)

Authors, Journal, Affiliations, Type, DOI

• Authors: Alan Kadish, Ann Dyer, James P. Daubert, Richard Quigg, Neil E. Estes III, Stephen C. Hammill, Heather T. Banchs, Richard Bricker, Ira Molter, Charles Schaechter, Douglas P. Zipes (for the Defibrillators in Non-Ischemic Cardiomyopathy Treatment Evaluation [DEFINITE] Investigators)
• Journal: New England Journal of Medicine, May 20, 2004; Vol 350(21):2151–2158
• Affiliations: Northwestern University (lead); 37 US centres
• Type: Randomised controlled trial (open-label); parallel groups; multicentre; US-only
• Funding: St. Jude Medical; independent data analysis at Northwestern University
• DOI: https://doi.org/10.1056/NEJMoa033088

Overview

DEFINITE was the first large prospective RCT of primary prevention ICD specifically in non-ischaemic cardiomyopathy (NICM). 458 patients with LVEF <36%, ambient arrhythmias (NSVT or ≥10 PVCs/hour), and symptomatic HF on ACEi and beta-blockers were randomised to single-chamber ICD (St. Jude Medical) plus standard therapy versus standard therapy alone; mean follow-up was 29.0 months. The primary endpoint — all-cause mortality — was not statistically significant (HR 0.65; P=0.08), because the trial was underpowered: only ~1/3 of deaths were arrhythmic versus >50% assumed at design, attributable to high background ACEi/beta-blocker compliance. However, sudden arrhythmic death was significantly reduced (HR 0.20; P=0.006), establishing the mechanistic proof-of-concept that ICD halves SCD in NICM. Pre-specified subgroup analyses showed survival benefit in men (HR 0.49; P=0.018) and NYHA III (HR 0.37; P=0.02). DEFINITE preceded SCD-HeFT (2005) and DANISH (2016), contextualising the pattern that ICD reduces SCD in NICM but all-cause mortality benefit depends on the arrhythmic fraction of competing deaths.

Keywords

Nonischemic dilated cardiomyopathy, prophylactic defibrillator, sudden cardiac death, primary prevention, ventricular arrhythmia, ICD, LVEF, NSVT, ventricular ectopy

Key Takeaways

Study Design

• Population: NICM (non-ischaemic confirmed by clinical criteria), LVEF <36%, symptomatic HF, ambient ventricular arrhythmia (NSVT on 24-hour Holter OR ≥10 PVCs/hour), aged 21–80 years
• Randomisation: 1:1; standard medical therapy ± single-chamber ICD (St. Jude Medical Photon or equivalent); open-label
• ICD programming: VVI backup pacing 40 bpm; VF detection zone 180 bpm; no ATP zone programmed in this protocol (single-chamber shock-only)
• Enrolment: July 1998 – June 2002; 37 US centres; n=458 (229 each arm)
• Background therapy: ACEi 85.6%, beta-blocker 84.9%, diuretic 86.7%; notably no aldosterone antagonist (RALES not yet established for NYHA I–III at design); no ARNi, no SGLT2i
• Primary analysis endpoint: 68th death (required for 80% power at 2-sided alpha based on revised sample size); mean follow-up 29.0±14.4 months

Primary Endpoint — All-Cause Mortality (NOT Significant)

• ICD arm: 28/229 deaths (12.2%); Standard therapy arm: 40/229 deaths (17.5%)
• HR 0.65 (95% CI 0.40–1.06; P=0.08) — did NOT reach statistical significance
• 2-year mortality: 7.9% ICD vs 14.1% standard therapy
• The trial was underpowered because only ~1/3 of deaths were arrhythmic (vs >50% assumed at design); driven by high ACEi + beta-blocker compliance substantially reducing background SCD risk

Secondary Endpoint — Sudden Arrhythmic Death (Significant)

• ICD arm: 3/229 sudden deaths; Standard therapy arm: 14/229 sudden deaths
• HR 0.20 (95% CI 0.06–0.71; P=0.006) — statistically significant
• Confirms ICD mechanistically eliminates ~80% of sudden arrhythmic deaths in NICM
• HF deaths: 9 (ICD) vs 11 (standard) — similar, no redistribution effect
• Noncardiac deaths: 26 total — 10 cancer, 7 pneumonia, 5 stroke; high competing non-cardiac mortality reflects enrollment of broader NICM population

Subgroup Analyses (Prespecified; Cox Model)

• Men: HR 0.49 (95% CI 0.27–0.89; P=0.018) — significant benefit
• NYHA III: HR 0.37 (95% CI 0.17–0.80; P=0.02) — significant benefit (consistent with higher arrhythmic death fraction in more symptomatic patients)
• NYHA I/II: trend only, not significant (lower background SCD risk)
• Women: no significant benefit (too few events)
• No significant aetiology subgroup — by design (NICM-only trial)

Device Performance and Safety

• Appropriate shocks: 41 patients received 91 appropriate shocks (18% of ICD arm); predominantly VF/fast VT
• Inappropriate shocks: 49 patients — predominantly AF (most common) and sinus tachycardia
• Implant complications (1.3%): hemothorax, pneumothorax, cardiac tamponade
• Follow-up complications (4.4%): lead dislodgements, venous thrombosis, 1 device infection
• 10% crossover: standard therapy patients who received ICD during follow-up (reduces power to detect primary endpoint effect)

Why the Trial Was Underpowered

• Original power calculation assumed >50% of deaths would be arrhythmic (based on older registry data with minimal beta-blocker use)
• Actual arrhythmic fraction was ~1/3 — reflecting better-than-expected medical therapy compliance
• Sample size was extended mid-trial from a one-sided 85% power design to two-sided 85% requiring 68 events — but still insufficient for all-cause mortality given the arrhythmic fraction
• The same mechanism (competing non-arrhythmic mortality attenuating all-cause mortality benefit) would later be replicated in DANISH (2016)

Design Choices and Their Implications

• No amiodarone arm by design: Investigators deliberately excluded amiodarone — concern it would limit beta-blocker uptitration, a primary goal of concurrent trials
• No aldosterone antagonist: RALES was published 1999 (during enrolment); RALES required NYHA III/IV; NYHA I/II patients in DEFINITE not yet covered by RALES evidence at design
• Single-chamber ICD with shock-only programming: No ATP programmed; all therapy was high-energy shocks — does not reflect modern programming (ATP-first as per HRS 2015 consensus)
• St. Jude Medical funding: Independent data analysis at Northwestern mitigated but does not eliminate industry sponsor bias

Limitations of the Document

• Underpowered for all-cause mortality (primary endpoint not significant; ~1/3 arrhythmic deaths vs >50% assumed)
• Open-label design — no blinding of patients or investigators; placebo-controlled comparison not possible with active ICD
• US-only (37 centres) — geographic generalisability limited
• Industry-funded (St. Jude Medical; sponsor of single-chamber device used)
• 10% crossover (standard therapy → ICD) attenuates ITT estimate of ICD benefit
• Single-chamber, shock-only ICD design (no ATP) — not reflective of modern ICD programming; inappropriate shock rate (49 patients) may have been reduced with ATP-first programming
• Pre-SGLT2i, pre-ARNi, pre-aldosterone antagonist era — absolute background mortality risk in NICM has declined further with modern quadruple GDMT; NNT in 2026 clinical practice will be larger
• No LBBB/non-LBBB stratification — some DEFINITE patients were CRT-eligible and the benefit of concurrent CRT was not explored
• Short follow-up (mean 29 months) — DANISH (67.6 months) with longer follow-up still showed no significant all-cause mortality benefit in NICM

Key Concepts Mentioned

• concepts/Sudden-Cardiac-Death — primary mechanism of benefit; 80% SCD reduction despite NS primary endpoint
• concepts/VA-Risk-Stratification-DCM — NSVT/PVCs as entry arrhythmia criteria; relationship to LVEF in risk stratification

Key Entities Mentioned

• entities/ICD — device tested; DEFINITE primary evidence for NICM primary prevention
• entities/HFrEF — population context; NICM with LVEF <36% and symptomatic HF

Wiki Pages Updated

• wiki/sources/icd-nicm-definite-nejm-2004.md — created (this file)
• wiki/entities/ICD.md — DEFINITE added to Primary Prevention — NICM; contradiction added; source_count 6→7
• wiki/entities/HFrEF.md — DEFINITE added to ICD primary prevention NICM section; source_count 24→25
• wiki/sourceindex.md — new entry added
• wiki/wikiindex.md — ICD and HFrEF entries updated
• log.md — ingest entry appended