| Popis: |
Several studies showed that time domain analysis of the signal-averaged ECG may identify groups of patients with low and high risk for arrhythmic events after myocardial infarction (MI). However, the signal averaging methods were not uniform and the definition of abnormal signal-averaged ECG was empiric. To identify the best quantitative signal-averaged variable in predicting arrhythmic events (sustained ventricular tachycardia, ventricular fibrillation and witnessed, instantaneous death) 262 patients surviving acute MI were prospectively evaluated. Twelve clinical variables, left ventricular ejection fraction (LVEF), complex ventricular arrhythmias (CVA) on Holter monitoring and three conventional signal-averaged variables (either at 25-250 or 40-250 Hz) were entered in a Cox proportional hazards regression model. During a mean follow-up of 20.3 +/- 13.7 months 16 (6.1%) patients had arrhythmic events. All six signal-averaged variables were independent predictors of arrhythmic events and the filtered QRS duration (fQRSD)or = 120 ms at 40 Hz high pass filtering resulted the most predictive. In a regression analysis, including the best signal-averaged variable, LVEF and CVA, only fQRSDor = 120 ms at 40 Hz and LVEF independently predicted arrhythmic events. Sensitivity, specificity, positive predictive value and odds ratio for fQRSDor = 120 ms at 40 Hz were 63, 90, 29 and 11%, respectively, and for the combination of fQRSDor = 120 ms at 40 Hz and LVEF40%, were 73, 95, 47 and 39%, respectively. In conclusion, the fQRSDor = 120 ms at 40 Hz best predicts arrhythmic events in the post-infarction period. The combination of signal-averaged ECG and LVEF is recommended to stratify patients at risk of arrhythmic events after MI. |
