乌鸦传媒

Prophylactic implantable cardioverter defibrillator (ICD) therapy less than 40 days after acute myocardial infarction does not reduce overall mortality, a trial of high-risk patients showed.

Implanting ICDs within the 40-day window did reduce sudden cardiac death compared with optimal medical therapy (P=0.049), but that benefit was offset by a significant increase in nonsudden cardiac death -- 68 versus 39 (P =0.001), Gerhard Steinbeck, M.D., of the University of Munich, and colleagues wrote in the Oct. 8 issue of the New England Journal of Medicine.

Those findings from the IRIS (Immediate Risk Stratifications Improves Survival) study were first reported last March at the American College of Cardiology meeting in Orlando.

The IRIS findings confirm the results of DINAMIT (Defibrillator in Acute Myocardial Infarction Trial), which also found that early ICD therapy following MI did not extend survival. "The substantial reduction in the number of sudden cardiac deaths was offset by an excess of nonsudden cardiac deaths," wrote Alan M. Garber MD, PhD, and Mark A. Hlatky, MD, both of Stanford University, in an accompanying editorial in the New England Journal of Medicine.

When DINAMIT was reported, the findings were greeted with skepticism because they contradicted a number of earlier trials of ICD therapy, although those earlier studies did not investigate early ICD implantation.

Garber and Hlatky hailed the IRIS study as an exemplar of comparative effectiveness research since it confirmed an earlier trial and provided information "worth knowing. It can direct efforts away from an expensive yet ineffective procedure toward either new or established alternatives."

IRIS randomized 445 patients to ICDs (415 actually received the devices) and 453 to optimal medical therapy.

Entry criteria included impairment of left ventricular function (40% or less) on day five to 31 following MI, plus a heart rate of more than 90 beats per minute, and/or nonsustained ventricular tachycardia at a rate of 150 bpm or faster during Holter-ECG monitoring on day five to 31 following MI.

ICDs were implanted an average of 8.8 days after randomization.

The patients were enrolled from June 1999 through October 2007 and were followed for a mean of 37 months.

Overall, the mortality rate was almost identical for patients randomized to ICD and those in the optimal medical therapy group with 117 deaths in the control arm versus 116 in the ICD arm.

The unexpected finding was the excess number of nonsudden cardiac deaths in the ICD arm.

"This may indicate that [ventricular tachycardia/ventricular fibrillation] in the early period following MI is not a cause of sudden death, but an indicator of the patient's risk of dying anyway, but death takes another door," Steinbeck said when he reported the IRIS results at ACC.

In the NEJM paper, the IRIS investigators offered a longer list of possible explanations including "complications related to ICD implantation, the development of heart failure due to ventricular pacing of the ICD, study-group crossovers, untoward effects of defibrillator shocks and antitachycardia pacing; the substrate of acute myocardial infarction studied (i.e., rapid ventricular tachycardia and fibrillation due to pump failure, which led to death anyway); other, unidentified side effects of ICD; and between group differences in the use of concomittant therapies."

But Steinbeck and colleagues concluded -- as they did last March -- that teasing out which of those possibilities "might have contributed to the increased risk of nonsudden cardiac death in the ICD group in our study, as well as the extent of the contribution, remains unsolved."