Prevention ICD Survival Gains Not Tied to EF Reduction

April 03, 2011

April 3, 2011 (New Orleans, Louisiana) — The survival benefit from an implantable cardioverter-defibrillator (ICD) for primary prevention isn't dependent on the severity of left ventricular systolic dysfunction, according to a meta-analysis of many of the major trials that established LVEF as the main gauge of device eligibility [1].

The analysis suggests that, given an LVEF of 35% as the qualifying threshold for getting a device, the mortality reduction with an ICD should be about the same for someone in the 30%-to-34% range, for example, as for someone who's at <25%, first author Dr Meltiady Issa (Marshfield Clinic, WI) told heartwire .

Some of the very trials in the meta-analysis, he observed, hinted in subgroup analyses that the survival benefit in the primary-prevention setting is inversely related to LVEF, and as a result, some physicians might balk at recommending ICDs for frail patients or those with significant comorbidities if, for example, their LVEF is only a shade under 35%.

According to the current findings, which Issa presented in a poster here at the American College of Cardiology (ACC) 2011 Scientific Sessions/i2 Summit, "you don't have to hesitate if the LVEF is close to 35%; they'll get the same benefit."

He and his colleagues gathered five major ICD primary-prevention trials that provided all-cause mortality reductions by whether LVEF was severely or only moderately reduced; included were 2227 patients randomized to ICD therapy and 2009 controls.

Each trial (DINAMIT, DEFINITE, SCD-HeFT, MADIT, and MADIT-2) had different LVEF cutoffs for defining severity. SCD-HeFT, for example, defined severely reduced LVEF as <30% and moderately reduced LVEF as 31% to 35%. For MADIT-2, the respective ranges were <25% and 26% to 30%.

Across the five trials, the patients with severely reduced LVEF showed a hazard ratio for all-cause mortality (ICD vs control) of 0.71 (95% CI 0.61–0.83; p<0.001). The benefit for those with moderately reduced LVEF wasn't all that different: 0.73 (95% CI 0.55–0.95; p=0.022). Indeed, the p value for the difference in hazard ratios showed solid nonsignificance at 0.90.

Both mortality reductions are on par with the hazard ratio covering all patients in all five trials, which was 0.72 (p<0.0001).

Issa et al had no disclosures.