Abstract

Background It is current practice to withhold implantable cardioverter defibrillators (ICD) from patients with severe heart failure because their deaths are judged as non-sudden and therefore assumed not to be preventable by ICD. If this was true, there should be a trend towards reduced preventability of deaths in the severe heart failure subgroups within existing randomised control ICD trials. We tested the prevailing assumption that patients with most severe heart failure would not benefit from ICD implantation.

Methods Six trials were identified enrolling 7873 patients, with 2734 patients randomly assigned to receive an ICD. Reduction in mortality in the ICD arm varied between 5.6% and 31%. All six trials provided data separated into higher and lower ejection fraction subgroups. Five trials provided data separated into higher and lower New York Heart Association (NYHA) class patient subgroups.

Results For patients subcategorised by NYHA class, there was a non-significant difference in z-score (p=0.922) between patients with mild to moderate and severe heart failure. Similarly, subgrouping by left ventricular ejection fraction (LVEF) revealed no significant difference between z-scores (p=0.170). Both observations suggest no attenuation of benefit of ICD implantation in patients with higher NYHA class or lower LVEF.

Conclusion There is no evidence within the existing trial populations of a tendency for the relative risk reduction to be smaller in patients with severe heart failure. The prevailing assumption that severe heart failure patients are less likely to benefit from ICD therapy must be questioned.