Recent results of a randomized trial showed that prophylactic placement of an implantable cardioverter defibrillator (ICD) in patients with non-ischaemic heart failure (HF) does not reduce all-cause mortality whether or not the patients are candidates for cardiac resynchronization therapy (CRT).

Findings from the DANISH trial (Danish study to Assess the Efficacy of ICDs in Patients with Non-ischaemic Systolic Heart Failure on Mortality), presented recently at the European Society of Cardiology Congress 2016, showed that all-cause mortality did not differ significantly between patients in the ICD (n=556) and control (n=560) groups who were followed up for a median of 67.6 months (hazard ratio [HR], 0.87; p=0.28). [N Engl J Med 2016, doi: 10.1056/NEJMoa1608029]

This serves as an admonition on recent guidelines recommending prophylactic ICD in patients with non-ischaemic HF. “The level of evidence is highly questionable because [they are] based on small trials or sub-studies from larger trials… and because the management of HF has changed since the earlier trials, especially in pharmacotherapy and CRT, the impact of ICD on mortality in non-ischaemic cardiomyopathy has to be re-evaluated,” explained Dr Christophe Leclerq of the Centre Cardio-Pneumologique in Rennes, France, who discussed the DANISH trial.

The DANISH trial included 1,116 patients with heart failure with documented non-ischaemic aetiology, and left ventricular ejection fraction of ≤35 percent who were on optimal medical treatment. Fifty-eight percent of the patients had an indication for CRT.

In addition to the primary outcome of all-cause mortality, the secondary outcome of cardiovascular mortality also did not differ significantly between the ICD and control groups (HR, 0.77; p=0.10).

“However, ICD placement reduced sudden cardiac death by 50 percent [p=0.005],” reported investigator Dr Lars Kober of the Copenhagen University Hospital, Denmark. “Mortality in younger patients was also reduced [p=0.009], with patients below 68 years of age benefiting the most from the procedure [HR, 0.64; p=0.01].”

Interestingly, there was no interaction with CRT in the subgroup analysis. “Some may believe that if you are scheduled for CRT, you might not need an ICD… but there were no interactions,” noted Kober.

There was no difference in device infection rate overall (odds ratio [OR], 1.38; p=0.29). However, in patients with no indications for CRT, the rate of infection was higher with ICD vs control (5.1 vs 0.8 percent; OR, 6.35; p=0.006). Rare adverse events included bleeding requiring intervention (0.2 vs 0 percent), pneumothorax (2 vs 1.1 percent; p=0.22) and inappropriate shocks (5.9 vs 0 percent).

“The DANISH trial raises questions on guidelines recommending the use of ICD in non-ischemic HF and suggests the need for more accurate criteria for patient selection prior to ICD implantation,” said Leclerq.