New England Journal of Medicine September 29, 2016

Defibrillator Implantation in Patients with Nonischemic Systolic Heart Failure (DANISH)

Lars Køber, Jens J. Thune, Jens C. Nielsen, et al.

Bottom Line

In patients with nonischemic systolic heart failure, prophylactic ICD implantation significantly reduced sudden cardiac death but did not significantly improve overall survival compared to usual clinical care.

Key Findings

1. Over a median follow-up of 67.6 months, the primary endpoint of death from any cause occurred in 21.6% (120/556) of the ICD group versus 23.4% (131/560) of the control group, which was not statistically significant (HR 0.87; 95% CI, 0.68 to 1.12; P=0.28).
2. ICD implantation significantly reduced the secondary endpoint of sudden cardiac death, occurring in 4.3% (24/556) of the ICD group compared to 8.2% (46/560) of the control group (HR 0.50; 95% CI, 0.31 to 0.82; P=0.005).
3. There was no significant difference in cardiovascular death between the ICD group (13.8%) and the control group (17.0%) (HR 0.77; 95% CI, 0.57 to 1.05; P=0.10).
4. A prespecified subgroup analysis indicated a significant age interaction: younger patients (age <=68 years) experienced a significant reduction in all-cause mortality with an ICD (HR 0.64; 95% CI, 0.45 to 0.90), whereas older patients did not (HR 1.18; 95% CI, 0.82 to 1.69).

Study Design

Design
Randomized Controlled Trial
Open-Label
Sample
1,116
Patients
Duration
67.6 mo
Median
Setting
Multicenter, Denmark
Population Patients with symptomatic systolic heart failure (NYHA class II-IV, or class I if previously symptomatic), left ventricular ejection fraction <=35%, heart failure not caused by coronary artery disease, and elevated NT-proBNP.
Intervention Implantable cardioverter-defibrillator (ICD) plus usual clinical care (including CRT if clinically indicated).
Comparator Usual clinical care (including CRT if clinically indicated).
Outcome Death from any cause.

Study Limitations

The trial may have been underpowered to detect a modest but potentially clinically meaningful difference in all-cause mortality, as the 95% confidence interval for the primary endpoint included up to a 32% relative risk reduction.
A very high proportion of patients (58%) received cardiac resynchronization therapy (CRT), which independently reduces mortality and sudden cardiac death, thereby potentially diluting the incremental absolute benefit of an ICD.
The study population reflected a heterogeneous group of nonischemic etiologies; it did not differentiate between specific genetic or structural subsets (e.g., lamin A/C mutations, cardiac sarcoidosis) that inherently carry a disproportionately high arrhythmic risk.
The trial was open-label, though the primary endpoint (all-cause mortality) is objective and immune to ascertainment bias.

Clinical Significance

The DANISH trial fundamentally challenged the routine, universal use of primary prevention ICDs in all patients with nonischemic HFrEF. By demonstrating that an ICD halves sudden cardiac death but fails to significantly reduce all-cause mortality in an era of highly effective background medical therapy and CRT, it highlighted the competing risk of death from progressive pump failure and non-cardiac causes. The findings have prompted a paradigm shift toward individualized shared decision-making, emphasizing that younger patients with fewer comorbidities are more likely to realize a survival benefit from an ICD, whereas older, frailer patients may not.

Historical Context

Prior to the DANISH trial, the mortality benefit of implantable cardioverter-defibrillators (ICDs) for primary prevention was firmly established in patients with ischemic cardiomyopathy (e.g., MADIT-II). While the landmark SCD-HeFT trial also included patients with nonischemic etiology, subgroup analyses from that trial and the DEFINITE trial showed borderline or non-significant mortality benefits for the nonischemic population. Consequently, guidelines broadly recommended ICDs for both ischemic and nonischemic HFrEF patients, largely extrapolating from ischemic data. Over the subsequent decade, background medical therapy for heart failure improved significantly, and the use of cardiac resynchronization therapy (CRT) became widespread. The DANISH trial was initiated to evaluate the efficacy of primary prevention ICDs in a contemporary, strictly nonischemic heart failure population optimized on modern guideline-directed therapies.

Guided Discussion

High-yield insights from every perspective

Med Student
Medical Student

How does the pathophysiology of sudden cardiac death differ between ischemic and nonischemic cardiomyopathy, and why might an implantable cardioverter-defibrillator (ICD) be less effective in improving overall survival for the nonischemic population?

Key Response

Ischemic cardiomyopathy often features a distinct fibrotic scar from prior infarction that serves as an anatomic substrate for fatal reentrant ventricular tachyarrhythmias. In contrast, nonischemic cardiomyopathy has a more heterogeneous pathophysiology (diffuse fibrosis, toxic, genetic) where death is more frequently due to progressive pump failure. Therefore, while an ICD successfully aborts sudden arrhythmic death, these patients often still die from progressive heart failure, preventing a significant improvement in overall all-cause mortality.

Resident
Resident

A 72-year-old patient with nonischemic cardiomyopathy (LVEF 30%) on optimal guideline-directed medical therapy presents to your clinic. Based on the DANISH trial, how should you counsel this patient regarding the risks and benefits of a primary prevention ICD?

Key Response

The resident should recognize that DANISH showed no overall survival benefit for prophylactic ICDs in nonischemic cardiomyopathy, particularly in older patients. Counseling must involve shared decision-making, explaining that while the device halves the risk of sudden cardiac death, it does not prevent death from progressive heart failure or non-cardiac age-related causes. For a 72-year-old, competing risks of mortality are high, making an all-cause survival benefit from the ICD unlikely.

Fellow
Fellow

The DANISH trial demonstrated an age-dependent effect of ICD implantation, with a potential all-cause mortality benefit in younger patients (under 68 years) but not older patients. How do competing risks explain this finding, and how does the high baseline rate of Cardiac Resynchronization Therapy (CRT) in this trial affect generalizability?

Key Response

Competing risks, such as non-cardiac death and progressive pump failure, increase significantly with age, diluting the mortality benefit of preventing sudden cardiac death in older adults. Furthermore, 58 percent of patients in DANISH received a CRT device (CRT-P or CRT-D). Because CRT itself induces reverse remodeling, improves left ventricular function, and reduces both sudden and non-sudden death, this high baseline utilization likely lowered the overall risk of sudden cardiac death, making the incremental benefit of an ICD much harder to prove than in older trials.

Attending
Attending

The DANISH trial challenged a long-held dogma that all patients with LVEF under 35 percent should receive an ICD. How do we synthesize the negative primary outcome of DANISH with the positive outcomes of older landmark trials like SCD-HeFT to teach trainees about the evolution of optimal medical therapy?

Key Response

Attending physicians must emphasize that baseline risk changes as medical therapy evolves. SCD-HeFT was conducted before the widespread use of modern guideline-directed medical therapy and CRT. Modern therapies (e.g., ARNIs, SGLT2 inhibitors, MRAs) have drastically reduced the background rate of both sudden and non-sudden cardiac death. DANISH perfectly illustrates the teaching point that as foundational medical therapy improves, the absolute risk reduction provided by device therapies diminishes, necessitating highly individualized, risk-stratified shared decision-making rather than blanket recommendations.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD
PhD

The DANISH trial was powered for a primary outcome of all-cause mortality, but the observed event rate was lower than anticipated due to modern background therapies. How does this lower-than-expected event rate impact the statistical power, and what alternative time-to-event models would better isolate the ICDs effect in the presence of competing risks?

Key Response

Lower event rates reduce statistical power, increasing the risk of a Type II error (failing to detect a true, modest difference in all-cause mortality). While the prespecified subgroup analysis suggested a benefit in younger patients, underpowered subgroups are prone to multiplicity and false conclusions. A rigorous methodological critique would advocate for the use of competing risk frameworks, such as the Fine-Gray model, to appropriately evaluate the cumulative incidence of sudden cardiac death versus death from worsening heart failure, providing a more accurate estimation of the ICDs specific treatment effect.

Journal Editor
Journal Editor

From an editorial perspective evaluating trial integrity, what are the primary threats to internal validity in the DANISH trial regarding the open-label design and crossover rate, and how might these factors bias the hazard ratio for the primary endpoint?

Key Response

An editor would flag that in an unblinded device trial, crossover from the control arm to the active treatment arm is a major threat. In DANISH, approximately 5 percent of the control group eventually received an ICD. In a standard intention-to-treat analysis, this crossover dilutes the treatment effect, biasing the hazard ratio toward the null (1.0). A thorough peer review would demand robust per-protocol or as-treated sensitivity analyses to ensure that this crossover did not mask a true mortality benefit.

Guideline Committee
Guideline Committee

Current heart failure guidelines historically gave a Class I recommendation for ICDs in nonischemic cardiomyopathy based on older data. How does the DANISH trial influence the strength of recommendation and level of evidence for primary prevention ICDs in updated guidelines, particularly regarding the incorporation of patient age?

Key Response

The committee must weigh DANISH against older positive trials and meta-analyses. Because DANISH showed a significant reduction in sudden cardiac death and a likely survival benefit in younger patients, completely removing the ICD recommendation is unjustified. However, DANISH provides high-quality evidence (Level A) to support downgrading the routine primary prevention recommendation in nonischemic cardiomyopathy to a Class IIa (as reflected in updated ESC guidelines). The guidelines must now explicitly incorporate age, life expectancy, and comorbidities into the recommendation algorithms, acknowledging that the incremental survival benefit over modern quadruple therapy is attenuated.

Clinical Landscape

Noteworthy Related Trials

2002

MADIT-II Trial

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 resulted in a 31% reduction in the risk of all-cause mortality compared to conventional medical therapy.
2004

DEFINITE Trial

n = 458 · NEJM

Tested

Implantable cardioverter-defibrillator (ICD)

Population

Patients with non-ischemic dilated cardiomyopathy, LVEF < 36%, and premature ventricular complexes or nonsustained ventricular tachycardia

Comparator

Standard medical therapy

Endpoint

All-cause mortality

Key result: There was a non-significant trend toward reduced all-cause mortality with ICD therapy, though sudden cardiac death was significantly reduced.
2005

SCD-HeFT Trial

n = 2,521 · NEJM

Tested

Implantable cardioverter-defibrillator (ICD) or Amiodarone

Population

Patients with NYHA class II or III heart failure and LVEF <= 35% (ischemic and non-ischemic)

Comparator

Placebo

Endpoint

All-cause mortality

Key result: Amiodarone had no favorable effect on survival, whereas ICD therapy significantly reduced overall mortality by 23% compared to placebo.

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