DANISH

Published

Kober L, et al. "Defibrillator implantation in patients with nonischemic systolic heart failure". The New England Journal of Medicine. 2016. epub 2016-08-28:1-10.
PubMed • Full text • PDF

Clinical Question

In patients with non-ischemic cardiomyopathy with LVEF ≤35% with symptomatic heart failure, do implantable cardioverter-defibrillators (ICDs) reduce overall mortality versus standard care?

Bottom Line

In patients with non-ischemic cardiomyopathy with LVEF ≤35% with symptomatic heart failure, ICDs did not result in an overall mortality benefit versus standard care after 5.5 years of follow up. ICD implantation did result in a 3% absolute reduction in sudden cardiac death that was somewhat mitigated by device-related morbidity including an absolute 1.5% increase in device-related infection.

Major Points

On the strength of large RCTs including SCD-HeFT and MADIT-II demonstrating a clear mortality benefit, ICD implantation carries a class I recommendation for patients with symptomatic cardiomyopathy of any cause with LVEF ≤35% for the purposes of aborting sudden cardiac death. The benefit of ICD implantation is generally considered to be similar in patients with non-ischemic vs. ischemic cardiomyopathy even though the majority of patients in SCD-HeFT and all of the patients in MADIT-II had cardiomyopathy with ischemic origin. Furthermore, it is not clear that ICD implantation carries the same benefit in the heart failure population in the era of more frequent use of optimal medical therapy for heart failure and cardiac resynchronization therapy (CRT) in patients with severe cardiomyopathy with significant conduction system disease.

The 2016 Danish Study to Assess the Efficacy of ICDs in Patients with Non-ischemic Systolic Heart Failure on Mortality (DANISH) randomized 1116 patients with symptomatic systolic heart failure with LVEF ≤35% due to nonischemic cardiomyopathy to ICD placement vs. standard care (which included CRT when indicated but not ICD). At 68 months, the primary endpoint of all-cause mortality was not significantly lower in the ICD group. However, ICD therapy did result in a 3% absolute reduction in sudden cardiac death that was mitigated at least partially by a corresponding 1.5% absolute increase in device-related infection. Importantly, the study was powered to detect a benefit of ICD therapy of 25% difference in total mortality and so a more modest benefit of ICD therapy on overall mortality cannot be ruled out (and in fact is likely given trends towards decreased overall mortality as well as cardiovsacular mortality). There was also a relatively high rate of noncardiovascular mortality in the study (31%), which may have washed out any potential benefit from ICD therapy, as ICD placement is unlikely to have effects on noncardiovascular death. Furthermore, an age-by-therapy interaction was observed whereby patients < 59 years old derived a mortality benefit from ICD therapy whereas older patients did not. Likewise, patients with less severe CHF as determined by NTproBNP level also appeared to benefit from ICD. Ultimately, the DANISH trial suggests that routine ICD therapy for all patients with symptomatic ischemic cardiomyopathy with ≤35% may not result in the strong mortality benefit previously extrapolated from data in patients with ischemic cardiomyopathy; rather, the decision to place an ICD in these patients should likely be considered on an individual basis considering severity of underlying disease and concomitant arrhythmia risk versus age/life expectancy and risks of competing illness.

Guidelines

No guidelines have been published reflecting the results of this study.

Design

• Multicenter, open-label, randomized, controlled trial
• N=1116
  • ICD (n=556)
  • Standard care (n=560)
• Setting: 5 sites in Denmark
• Period: February 7, 2008 - June 30, 2014
• Median follow-up: 67.6 months
• Analysis: Intention-to-treat
• Primary outcome: Overall mortality

Population

Inclusion Criteria

• Clinical heart failure
• Non-ischemic etiology
• Optimal medical treatment
• NYHA class II or III (IV if planned for CRT)
• LVEF ≤35%
• NT-proBNP above 200pg/mL

Exclusion Criteria

• Dysregulated permanent AF (resting HR > 100bpm)
• Uncorrected congenital heart disease or valve obstruction, obstructive cardiomyopathy, active myocarditis, constrictive pericarditis, untreated hypothyroidism or hyperthryoidism, adrenal insufficiency, active vasculitis due to collagen vascular disease
• On the urgent waiting list for a heart transplant
• Recipient of any major organ transplant
• Receiving or having received cytotoxic or cytostatic chemotherapy and/or radiation for treatment of malignancy within 6 months before randomization or clinical evidence of current malignancy
• HIV positive with life expectancy < 5 years
• Renal failure on dialysis
• Alcohol or substance use disorder within past 3 months
• Any condition that, in the investigator's opinion, could put the subject at significant risk, confound the study results, or interfere significantly with the subject's participation in the study
• Lack of informed consent

Baseline Characteristics

From the ICD group.

• Demographics: Age 64yr, male 73%
• Co-morbidities: BMI 26.8, GFR 74 mL/min, HTN 33%, DM 18%, pAF 24%
• CHF characteristics: NTproBNP 1244, QRS duration 146ms, LVEF 25%, NYHA II 53%, NYHA III 45%, CHF duration 20mo
• Ischemia rule out: Nuclear 1%, CTA 3%, Catheterization 96%
• CHF etiology: Idiopathic 76%, valvular 4%, hypertensive 11%, other 9%
• Medications: Beta blocker 92%, ACE inhibitor/ARB 96%, Mineralocorticoid receptor antagonist 59%, Amiodarone 6%
• Device: CRT implanted 58%, pre-existing CRT or pacemaker 10%

Interventions

• Patients randomized 1:1 to ICD versus standard care
  • Randomization performed in permuted blocks, with block sizes ranging from 2 to 6 patients
  • Randomization stratified according to center and whether patients were scheduled to receive CRT
• Qualifying LVEF and NTproBNP level had to be measured after the doses of ACE/ARB and beta blockers had been increased to target levels
• Non-ischemic cause of heart failure confirmed by coronary angiography, CTA, or nuclear myocardial perfusion imaging
• Implantation of ICD (or CRT) was planned to be performed no later than 4 weeks after randomization
• All patients were seen at follow-up visits after 2 months and every 6 months thereafter until the end of the trial
• An endpoint classification commmittee, the members of which were unaware of treatment assignments, used prespecified criteria to adjudicate all clinical outcomes

Outcomes

Comparisons are ICD therapy versus standard care.

Primary Outcome

All cause-mortality

120 (21.6%) vs. 131 (23.4%) (HR 0.87; 95% CI 0.68-1.12; P=0.28)

Secondary Outcomes

Cardiovascular death

77 (13.8%) vs. 95 (17.0%) (HR 0.77; 95% CI 0.57-1.05; P=0.10)

Sudden cardiac death

24 (4.3%) vs. 46 (8.2%) (HR 0.50; 95% CI 0.31-0.82; P=0.005)

Subgroup Analysis

Age (interaction p=0.009)

< 59 years: 17/167 vs. 34/181 (HR 0.51; 95% CI 0.29-0.92; P=0.02)

≥ 59 to < 68 years: 36/173 vs. 50/202 (HR 0.75; 95% CI 0.48-1.16; P=0.19)

≥ 68 years: 67/216 vs. 47/177 (HR 1.19; 95% CI 0.81-1.72; P=0.38)

NT-proBNP (interaction p=0.06)

< 1177 pg/mL: 32/266 vs. 74/268 (HR 0.59; 95% CI 0.38-0.91; P=0.02)

≥ 1177 pg/mL: 57/292 vs. 88/290 (HR 0.99; 95% CI 0.73-1.36; P=0.96)

Adverse Events

Device infection (No CRT)

12/234 (5.1%) vs. 2/237 (0.8%) (HR 6.35; 95% CI 1.38-58.87; P=0.006)

Inappropriate shocks

33 (5.9%) vs. 0 (0%)

Criticisms

• Trial was powered to detect a 25% difference in mortality with ICD therapy, which is likely an overestimate of the expected effect given wide use of guideline-based therapy including CRT. A smaller benefit with ICD therapy cannot be ruled out.
• There was a relatively high 31% incidence of noncardiovascular mortality, which may cause underestimation of the effect of ICD therapy on mortality by introducing a prevalent competing risk for which ICD therapy would be expected to have no benefit.

Funding

• Study supported by unrestricted grants from Medtronic, St. Jude Medical, TrygFonden, and the Danish Heart Foundation