Background: although implantable cardioverter-defibrillators (ICDs) lower mortality in stable patients with low ejection fraction late after myocardial infarction, randomized trials of ICD versus control subjects implanted early after myocardial infarction do not show mortality benefit. Our objective was to investigate possible mechanisms underlying the lack of mortality benefit in the Defibrillation in Acute Myocardial Infarction Trial (DINAMIT).
Methods and results: this is a secondary analysis of the prospective randomized clinical trial. Outpatients with recent (6 to 40 days) acute myocardial infarction, left ventricular dysfunction (ejection fraction <35%), and low heart rate variability were randomized to ICD (n=311) or to standard medical therapy (n=342). In a competing-risks analysis, those factors that increased the risk of arrhythmic death also increased the risk of nonarrhythmic deaths. After adjustment for these factors, receiving an ICD was associated with a decreased risk of arrhythmic death (hazard ratio, 0.33; 95% confidence interval, 0.15 to 0.71) but an increase in nonarrhythmic death (hazard ratio, 1.70; 95% confidence interval, 1.00 to 2.80). In an adjusted time-dependent analysis, patients receiving an ICD and having appropriate ICD therapy had a 15.1% yearly hazard of mortality compared with 5.2% in ICD patients with no appropriate therapy (P<0.001). The reduction in sudden death in ICD patients was completely offset by increased nonarrhythmic deaths, which were greatest in patients receiving ICD shock therapy (hazard ratio, 6.0; 95% confidence interval, 2.8 to 12.7).
Conclusions: in patients receiving ICDs early after myocardial infarction, those factors that are associated with arrhythmia requiring ICD therapy are also associated with a high risk of nonsudden death, negating the benefit of ICDs in this setting.