Dr. Arthur Moss presented results from the Multicenter Automatic Defibrillator Implantation Trial (MADIT) II. The manuscript reporting the study was also published the same week in the New England Journal of Medicine (Moss A, et al. N Engl J Med. 2002;346:877-883). The Multicenter Automatic Defibrillator Implantation Trial II (MADIT II) was a follow-up study to the original MADIT trial. MADIT I enrolled patients with prior myocardial infarction (MI), nonsustained ventricular tachycardia, and low ejection fractions and randomized them to either ICD therapy or "conventional" therapy if they had an inducible ventricular tachycardia at electrophysiologic study that was not suppressible with intravenous procainamide. Over time, it has been recognized that electrophysiologic studies may not be as sensitive or as specific as initially thought, and a beneficial role of antiarrhythmic drugs has become increasingly questioned. Therefore, MADIT II was designed to see whether patients with prior MI identified only by a low ejection fraction would benefit from prophylactic implantable cardioverter defibrillator (ICD) therapy.

Patients of either sex were eligible for the study if they had had a MI 1 month or more before entry and had an ejection fraction of 0.30 or less. Patients were excluded from enrollment if they had a history of sustained VT or syncope with inducible ventricular tachycardia, if they were in the New York Heart Association Class IV at the time of enrollment, had undergone revascularization within the preceding 3 months, or if they had a MI within the previous month. Initially, patients were also required to have frequent or repetitive ventricular premature beats during monitoring but subsequently this requirement was eliminated. Patients were randomly assigned in a 3:2 ratio to receive either an ICD or conventional medical therapy. Appropriate use of beta blockers, angiotensin-converting inhibitors and lipid lowering drugs was encouraged in both study groups. The use of antiarrhythmic drugs was at the discretion of the investigators. The primary end point was death from any cause. A triangular sequential design was used for monitoring with analyses performed weekly.

MADIT II randomized 1232 patients. The mean age was 65 and 85% were male. The ICD and conventional therapy groups were equally matched for New York Heart Association functional class, diabetes, and other important clinical characteristics. Follow-up averaged 20 months at the time the trial was stopped on November 20, 2001.

During the trial, there were 105 deaths (14.2%) in the defibrillator group and 97 deaths (19.8%) in the conventional therapy group for a hazard ratio of 0.69 (P = 0.016). Actuarial survival curves began to diverge approximately 9 months after entry and continued to diverge thereafter. The risk reductions for total mortality were 12% at 1 year, 28% at 2 years, and 28% at 3 years. Age, sex, ejection fraction, New York Heart Association class, or QRS interval did not substantially affect the hazard ratio.

There were no deaths related to ICD implantation. Thirteen patients experienced lead problems (1.8%), and there were 9 nonfatal infections (0.07%) that required surgical intervention. The incidence of new or worsened heart failure was higher in the defibrillator group. Patients in the ICD group were hospitalized at a rate of 11.3 per 1000 patient months of follow-up compared to 9.4 in the conventional therapy group. They conclude that their study demonstrates that ICD implantation improves survival in patients with prior MI and advanced left ventricular dysfunction and they recommend prophylactic implantation of a defibrillator in these patients.

Comment by John P. DiMarco, MD, PhD

MADIT II is an important study that raises significant questions about how we should manage patients with advanced heart failure. It shows that the ICD can decrease total mortality in patients without prior sustained arrhythmias to the same extent that it reduces mortality in cardiac arrest survivors. It has long been recognized that patients with low ejection fractions have high mortality rates and that approximately 50% of all cardiac deaths in these patients are sudden. Therefore, it should not be surprising that an ICD, which should effectively reverse most life-threatening ventricular tachyarrhythmias, would decrease mortality in these groups.

The real questions that arise from MADIT are economic and ethical. If we use Medicare payment levels as a standard, each new ICD implant represents an immediate cost of between $40,000 and $50,000. Additional costs for follow-up care, device problems, and associated therapy for cardiac and noncardiac problems are not included in this figure. If we estimate that there are well more than 1 million patients with congestive heart failure and a low ejection fraction in the United States alone, the first year cost to implant an ICD in each of these patients would be staggering $40 to 50 billion. This is a tremendous expense that would certainly unbalance the current Medicare budget. However, if we decide that we cannot afford this sum, how are we going to ethically decide who receives this therapy? If one takes the approach of trying to identify those who are most likely to die in a short period of time, then the oldest patients with the most comorbidities and the lowest ejection fractions would be the ideal candidates. However, here the ICD becomes palliative therapy that may extend the duration of life but not necessarily enhance its quality. If younger patients with slightly better ejection fractions and fewer comorbidities are selected, the number needed to treat for each life-year saved will increase and the cost in the short term will be high.

My personal opinion is that it is still too early for society to decide whether a defibrillator is standard therapy for all patients with low ejection fractions. Several other randomized trials are now in progress and they may help us get a better estimate of the benefits likely to accrue with more widespread use of ICD therapy. The Sudden Cardiac Death-Heart Failure Trial randomized 2500 patients and is comparing ICD insertion to amiodarone and placebo. If amiodarone has even a small benefit, the relative cost-effectiveness of an ICD will be much less. In addition, this trial includes patients with both ischemic and nonischemic myopathies and we desperately need data about the latter patients. We also should expect additional data from MADIT II. MADIT II performed a number of noninvasive tests, including measurements of T-wave alternans and heart rate variability. Perhaps these tests will identify subgroups in whom ICD therapy will be more cost effective. Finally, it should be remembered, that the ICD alone does not improve a patients symptoms of heart failure. In fact, some believe that heart failure may be worsened if right ventricular pacing is frequently used. New trials that are looking at the role of an ICD with biventricular pacing are now in progress. If these studies show improvements in both heart failure and mortality, it will place further pressure on society to decide if, and for whom, it can afford these new treatments.

Dr. DiMarco is Professor of Medicine, Division of Cardiology, University of Virginia, Charlettesville.