New England Journal of Medicine January 20, 2005

Amiodarone or an Implantable Cardioverter-Defibrillator for Congestive Heart Failure (SCD-HeFT)

Gust H. Bardy, Kerry L. Lee, Daniel B. Mark, Jeanne E. Poole, Douglas L. Packer, et al.

Bottom Line

SCD-HeFT demonstrated that a prophylactic single-lead, shock-only implantable cardioverter-defibrillator significantly reduces all-cause mortality in patients with symptomatic heart failure and an ejection fraction of 35% or less on optimal medical therapy, whereas amiodarone provides no survival benefit.

Key Findings

1. ICD therapy significantly reduced the risk of all-cause mortality compared to placebo, with a 23% relative risk reduction (HR 0.77; 97.5% CI, 0.62 to 0.96; P=0.007) [6.1.1].
2. At 5 years, the absolute decrease in mortality for the ICD group compared to placebo was 7.2 percentage points (22% vs. 29% overall mortality at median 45.5 months follow-up).
3. Amiodarone was associated with a similar risk of death as placebo (HR 1.06; 97.5% CI, 0.86 to 1.30; P=0.53), demonstrating no survival benefit for primary prevention.
4. The mortality benefit of the ICD was consistent across both ischemic (52% of cohort) and nonischemic (48% of cohort) etiologies of congestive heart failure.
5. Subgroup analysis revealed that the ICD survival benefit was prominent in NYHA class II patients, whereas NYHA class III patients did not show a clear survival advantage.

Study Design

Design
Randomized Controlled Trial
Double-Blind (Drug), Open-Label (ICD)
Sample
2,521
Patients
Duration
45.5 mo
Median
Setting
Multicenter, US and Canada
Population Patients aged 18 years or older with stable NYHA class II or III congestive heart failure of ischemic or nonischemic etiology and a left ventricular ejection fraction of 35% or less.
Intervention Conservatively programmed, shock-only, single-lead implantable cardioverter-defibrillator (ICD) or amiodarone.
Comparator Placebo (double-blind match for the amiodarone arm); all groups received standard conventional heart failure therapy.
Outcome Death from any cause (all-cause mortality).

Study Limitations

The trial was conducted before the widespread integration of contemporary guideline-directed medical therapies (GDMT), such as ARNIs and SGLT2 inhibitors.
Patients requiring cardiac resynchronization therapy (CRT) or dual-chamber pacing were not specifically evaluated for those device types, as the trial utilized single-lead, shock-only devices [6.1.1].
The subgroup finding suggesting no benefit in NYHA class III patients was secondary and hypothesis-generating; it may be confounded by competing risks of progressive pump failure rather than a true lack of arrhythmic protection.
Comparing an open-label device intervention against a double-blinded pharmaceutical intervention introduces inherent differences in clinical care, though the hard primary endpoint of all-cause mortality minimizes subjective reporting bias.

Clinical Significance

SCD-HeFT is a foundational trial that revolutionized the primary prevention of sudden cardiac death in heart failure. By definitively showing a 23% mortality reduction with ICDs and no benefit with amiodarone, the trial established prophylactic ICD placement as a Class I guideline recommendation for stable patients with an LVEF of 35% or less and NYHA class II or III symptoms, irrespective of whether the etiology is ischemic or nonischemic. Additionally, it reinforced the strategy of 'conservative' device programming (shock-only settings) to minimize the delivery of inappropriate shocks.

Historical Context

Prior to SCD-HeFT, the MADIT (1996) and MADIT-II (2002) trials had proven the efficacy of prophylactic ICDs in patients with ischemic cardiomyopathy. However, the benefit of primary prevention in nonischemic cardiomyopathy remained unproven and highly debated. Concurrently, empirical antiarrhythmic therapy, particularly amiodarone, was commonly prescribed to suppress asymptomatic ventricular arrhythmias in heart failure. SCD-HeFT definitively answered these controversies, expanding primary prevention ICD indications to nonischemic populations and officially removing amiodarone from the primary prevention survival algorithm.

Guided Discussion

High-yield insights from every perspective

Med Student
Medical Student

In patients with heart failure and reduced ejection fraction, why does an implantable cardioverter-defibrillator (ICD) reduce all-cause mortality while amiodarone, a potent antiarrhythmic, does not?

Key Response

Students must understand the pathophysiology of sudden cardiac death in heart failure and the mechanisms of these interventions. While amiodarone suppresses arrhythmias (Class III antiarrhythmic), it also possesses proarrhythmic effects (e.g., QT prolongation) and significant systemic toxicities (pulmonary, thyroid, hepatic) that can offset any survival benefit. An ICD does not prevent arrhythmias but effectively terminates fatal ventricular tachyarrhythmias (VT/VF) once they occur, thereby preventing sudden cardiac death without the systemic toxicity of chronic antiarrhythmic drugs.

Resident
Resident

Based on the principles established in SCD-HeFT, what are the prerequisite medical optimization steps and the required waiting period before referring a patient with newly diagnosed non-ischemic cardiomyopathy (LVEF 30%) for a primary prevention ICD?

Key Response

Residents need to focus on clinical application and appropriate patient selection. Before an ICD is indicated, the patient must be on optimally titrated guideline-directed medical therapy (GDMT) for a sufficient period—typically 3 to 6 months for non-ischemic cardiomyopathy (and at least 40 days post-MI or 90 days post-revascularization for ischemic etiology). This waiting period is crucial because optimal GDMT (e.g., beta-blockers, ARNIs, MRAs, SGLT2 inhibitors) promotes reverse remodeling, which may improve the LVEF above the 35% threshold, negating the need for an ICD.

Fellow
Fellow

Subgroup analysis in SCD-HeFT demonstrated a robust survival benefit from ICD placement in NYHA Class II patients but no significant mortality benefit in NYHA Class III patients. How does the concept of competing risks explain this finding, and how does it influence the decision to pursue a shock-only ICD versus a CRT-D in advanced heart failure?

Key Response

Fellows must grasp nuanced trial interpretations like competing risks. NYHA III patients have a higher baseline risk of dying from progressive pump failure rather than sudden arrhythmic death, which dilutes the mortality benefit of a shock-only ICD. For highly symptomatic advanced heart failure patients (especially those with a wide QRS), upgrading to a Cardiac Resynchronization Therapy Defibrillator (CRT-D) provides biventricular pacing to actively treat the progressive pump failure while also protecting against sudden cardiac death.

Attending
Attending

SCD-HeFT utilized a conservatively programmed 'shock-only' single-lead ICD strategy (VVI backup at 50 bpm) to minimize right ventricular pacing-induced cardiomyopathy. In contemporary practice, how do you balance this mortality benefit against the psychological burden of inappropriate shocks and the necessity of end-of-life deactivation discussions?

Key Response

Attendings must manage long-term complications, shared decision-making, and quality of life. The clinical pearl is that while ICDs save lives, shocks (both appropriate and inappropriate) cause significant psychological trauma and can even increase subsequent mortality risk. It underscores the importance of modern device programming (e.g., higher rate cut-offs, delayed therapies) to minimize unnecessary shocks, and highlights the ethical imperative to establish early advance directives for device deactivation when a patient enters the terminal phases of progressive pump failure or another terminal illness.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD
PhD

The SCD-HeFT trial utilized a double-blind, placebo-controlled design for the amiodarone arm, but the ICD arm was necessarily unblinded. How might this lack of masking in the ICD arm introduce performance bias regarding co-interventions, and what post-hoc statistical methods could be used to evaluate whether differences in medical management influenced the observed mortality benefit?

Key Response

PhDs should focus on methodological rigor. In an unblinded arm, providers and patients might alter their behavior—for example, ICD patients might have been monitored more closely or had their heart failure medications titrated differently. Researchers could use marginal structural models or time-dependent covariate analyses in Cox proportional hazards models to adjust for post-randomization changes in medical therapy and assess if performance bias confounded the primary mortality outcome.

Journal Editor
Journal Editor

A major threat to the external validity of older device trials like SCD-HeFT is the evolution of baseline medical therapy. If you were reviewing a contemporary manuscript attempting to replicate this trial, what specific baseline medication thresholds would you demand to see before accepting the generalizability of the trial's mortality endpoints?

Key Response

A seasoned editor evaluates how shifting baselines affect trial validity. Baseline therapy in SCD-HeFT mostly consisted of ACE inhibitors and beta-blockers. Today, optimal GDMT includes quadruple therapy (ARNI, beta-blocker, MRA, SGLT2 inhibitor), which significantly lowers the absolute baseline risk of sudden cardiac death. An editor would require >85% adherence to modern quadruple therapy to validly assess whether the marginal absolute risk reduction of an ICD still outweighs its risks in the modern era.

Guideline Committee
Guideline Committee

SCD-HeFT drove the AHA/ACC/HRS Class I recommendation for primary prevention ICDs in non-ischemic cardiomyopathy (LVEF ≤ 35%). Given the recent DANISH trial showing no overall mortality benefit for ICDs in non-ischemic etiology, how should current guidelines reconcile the historical SCD-HeFT data with contemporary evidence?

Key Response

Guideline committees must synthesize temporally conflicting data. While SCD-HeFT established the Class I (LOE A) indication for both ischemic and non-ischemic etiologies, the DANISH trial challenged the benefit in non-ischemic patients treated with modern GDMT. The committee must weigh downgrading the recommendation for non-ischemic patients (e.g., to Class IIa) or risk-stratifying based on specific factors like age (DANISH showed benefit in younger patients) and the presence of myocardial fibrosis on cardiac MRI, moving towards personalized guidelines rather than a rigid LVEF ≤ 35% cutoff.

Clinical Landscape

Noteworthy Related Trials

2002

MADIT-II

n = 1,232 · NEJM

Tested

Implantable cardioverter-defibrillator (ICD)

Population

Patients with prior myocardial infarction and LVEF <= 30%

Comparator

Conventional medical therapy

Endpoint

All-cause mortality

Key result: Prophylactic ICD implantation reduced the risk of all-cause mortality by 31% compared to conventional medical therapy.
2004

COMPANION Trial

n = 1,520 · NEJM

Tested

Cardiac resynchronization therapy with a defibrillator (CRT-D)

Population

Patients with advanced heart failure (NYHA III/IV), LVEF <= 35%, and QRS >= 120 msec

Comparator

Optimal pharmacological therapy alone

Endpoint

Composite of all-cause mortality or first hospitalization for heart failure

Key result: CRT-D significantly reduced the risk of the primary composite endpoint by 20% and significantly decreased all-cause mortality by 36% compared to optimal medical therapy.
2016

DANISH Trial

n = 1,116 · NEJM

Tested

Implantable cardioverter-defibrillator (ICD)

Population

Patients with non-ischemic systolic heart failure (LVEF <= 35%)

Comparator

Usual clinical care

Endpoint

All-cause mortality

Key result: Prophylactic ICD implantation did not significantly reduce the overall rate of death from any cause compared to usual care, although sudden cardiac death was reduced.

Tailored to your role

Want this tailored to you?

Add your specialty or training stage to get role-specific takeaways and more questions.

Personalize this analysis