Prediction of serious arrhythmic events after myocardial infarction: signal-averaged electrocardiogram, Holter monitoring and radionuclide ventriculography.

Noninvasive assessment was undertaken before hospital discharge in 210 patients who had recovered from acute myocardial infarction. This comprised signal-averaged electrocardiography, Holter monitoring and radionuclide left ventriculography. An abnormal signal-averaged electrocardiogram was defined as the presence of a low voltage signal less than 20 microV in the terminal 40 ms of the filtered QRS complex or a long filtered QRS complex greater than 120 ms. During a follow-up period of 6 months to 2 years (median 14 months), 15 patients had arrhythmic events: eight died suddenly and seven presented with sustained, symptomatic ventricular tachycardia. Using univariate analysis, abnormalities in each of the three noninvasive tests were able to predict arrhythmic events. Stepwise logistic regression demonstrated that each test was independently significant in predicting outcome, with a left ventricular ejection fraction less than 40% being the most powerful variable (beta = 2.8, p less than 0.005). This process generated an algorithm that allowed assessment of combinations of variables: the finding of an abnormal signal-averaged electrocardiogram in the presence of an ejection fraction less than 40% identified patients with a 34% probability of arrhythmic events. By contrast, in patients with left ventricular dysfunction but a normal signal-averaged tracing, the risk of arrhythmic events was 4% (p less than 0.001). This combination of variables was associated with a sensitivity of 80% and a specificity of 89%. Hence, using a combination of noninvasive tests after myocardial infarction, patients can be stratified according to risk of serious arrhythmic events.