Prediction of Sudden Death after Myocardial Infarction: Value of Electrophysiologic Parameters

Abstract

A vast array of non-invasive and invasive tests have been evaluated for the prediction of sudden death in asymptomatic patients after myocardial infarction.1–3 These tests share a common characteristic: Using accepted criteria for a “positive” and “negative” result, the tests have excellent negative predictive value, i.e., the risk for sudden death is low in patients with a negative test. However, these tests also have a relatively poor positive predictive value. For example, a low left ventricular ejection fraction (LVEF ,40%) is a strong predictor of mortality. Yet, only 10% of patients with low LVEF will die.1 Also, although 97% of patients with normal signal averaged electrocardiogram will remain free of sustained ventricular arrhythmias, only 30% of patients with evidence of “late potentials” will experience such complication.1 Using tests with high negative predictive value is warranted when effective, safe and inexpensive prophylactic therapy is available. Unfortunately, the use of class 1 antiarrhythmic drugs for prevention of sudden death has so far been ineffective or even detrimental.1 In addition, recently completed large-scale trials of amiodarone4 (like the European Myocardial Infarction Amiodarone Trial [EMIAT] presented in a preliminary form at the American College of Cardiology Scienti~c Sessions in March 1996), have failed to con~rm the positive impression created by earlier trials involving high-risk post-infarction patients.5,6 The most important recent development in the ~eld of sudden death prevention relates to the early termination of the Multicenter Automatic De~brillator Implantation Trial (MADIT).7 This multicenter trial randomized almost 200 asymptomatic high-risk patients to implantation of automatic de~brillator (AICD) or to “best medical therapy.” Patients studied had a prior myocardial infarction and all the following: (1) LVEF #35%, (2) spontaneous non sustained ventricular tachycardia (NSVT), and (3) inducible sustained VT that was not suppressible by intravenous procainamide during electrophysiologic (EP) studies. The 5-year cumulative mortality rate7 was 39% among patients with conventional treatment (mainly amiodarone therapy). For comparison, only 16% of patients in the AICD-group died (p , 0.01). This 60% mortality reduction was a direct result of an impressive reduction in the number of arrhythmic deaths. Demonstration that AICD implantation may improve the long-term survival of high-risk post-infarction patients will have a tremendous impact on the practice of medicine in several ways: (1) Publication of MADIT—and the availability of smaller, non-thoracotomy devices—will increase the pressures on the consulting electrophysiologist in favor of AICD use. In other words, the burden of proof will be on the physician recommending against implantation. (2) Yet, AICD implantation is a very expensive treatment, which involves potentially serious complications. Thus, our screening approach will have to shift toward tests with high positive predictive value (in order to reduce the number of unnecessary implantations) without seriously compromising on safety. Publication of the results of MADIT is likely to lead to an increased number of EP studies performed in asymptomatic infarct survivors. Thus, it is appropriate to review the lessons learned on this subject over the last years. By far, the largest experience with EP studies performed in unselected consecutive patients after a myocardial infarction has been gathered in Australia.8 Over the course of 9 years, more than 1000 patients underwent EP studies within 4 weeks of an infarct and were followed without speci~c antiarrhythmic therapy. Sustained monomorphic VT was induced in only 6% of patients. Within 1 year of follow-up, 19% of patients with inducible VT died suddenly or had non-fatal VT/VF. In comparison, only 3% of patients without inducible VT experienced arrhythmic events (p , 0.0005). By increasing the number of extrastimuli used in the stimulation protocol from 2 to 3 or more, the sensitivity of EP studies could be doubled. By 28 months of followup, 25% of inducible patients had experienced a spontaneous arrhythmic event. However, because of the relatively low incidence of VT/VF in the post-infarction population, 1200 EP studies had to be performed in or-