Value of signal-averaged electrocardiography, radionuclide ventriculography, Holter monitoring and clinical variables for prediction of arrhythmic events in survivors of acute myocardial infarction in the thrombolytic era

OBJECTIVES: This study assessed the ability of signal-averaged electrocardiography, radionuclide ventriculography and Holter electrocardiographic (ECG) monitoring and clinical variables to identify patients at risk of serious arrhythmic events after myocardial infarction in the thrombolytic era. BACKGROUND: Most studies of signal-averaged electrocardiography, radionuclide ventriculography and Holter ECG monitoring in risk stratification after myocardial infarction preceded the introduction of thrombolytic therapy. METHODS: A consecutive series of 301 survivors of myocardial infarction, 205 (68%) of whom received thrombolytic agents, underwent signal-averaged electrocardiography (1st 48 h, day 6 and discharge), Holter ECG monitoring (days 6 to 7) and radionuclide left ventriculography (days 7 to 14). Median follow-up time was 1.03 years. RESULTS: Thirteen patients (4.3%) had an arrhythmic event (sudden death in 11, sustained ventricular tachyarrhythmia in 2). The 25-Hz high pass filtered signal-averaged ECG at discharge was 64% sensitive (95% confidence intervals CI] 36% to 92%) and 81% specific (95% CI 76% to 86%). High grade ventricular ectopic activity on the Holter ECG was only 38% sensitive (95% CI 12% to 64%) and 74% specific (95% CI 71% to 77%). Left ventricular ejection fraction < 0.4 was the best test for prediction of arrhythmic events (sensitivity 75% 95% CI 50% to 100%] and specificity 81% 95% CI 76% to 85%]). In multivariate analysis, in rank order, digoxin therapy at discharge, an abnormal 25-Hz signal-averaged ECG before discharge, absence of angina before index infarction and previous infarction were predictive of arrhythmic events. With digoxin therapy excluded, ejection fraction was an independent predictor. Discriminant analysis identified a high risk group (12% of the study patients) with an event rate of 26%. CONCLUSIONS: The signal-averaged ECG and left ventricular ejection fraction are each independently predictive of arrhythmic events after myocardial infarction, but the Holter ECG is not. A combination of clinical and investigative variables, including the signal-averaged ECG, best identifies patients at highest risk.