COMPARISON AND COMBINATION OF LATE POTENTIALS AND SPECTRAL TURBULENCEANALYSIS TO PREDICT ARRHYTHMIC EVENTS AFTER MYOCARDIAL-INFARCTION IN THE POSTINFARCTION LATE POTENTIAL (PILP) STUDY

Ventricular late potentials detected at the end of the QRS complex by the signal-averaged ECG have been shown to predict arrhythmic events a fter acute myocardial infarction. Spectral turbulence analysis is a no vel technique for detecting abnormalities of cardiac electric activati on inside the QRS complex. The purpose of this study was to combine th ese two analysis methods in order to increase the predictive power of the signal-averaged ECG in post-infarction patients. The study compris ed a prospective series of 778 males under 66 years of age who survive d the acute phase of myocardial infarction. Signal-averaged ECG record ings were performed before hospital discharge 2 to 3 weeks after infar ction. The original Simson method was used for recording and analysing the time-domain signal-averaged EGG. Spectral turbulence analysis was performed using the same averaged vector magnitude QRS complexes (Del Mar Avionics). During the follow-lip period of 6 months, 33 patients (4.2%) had an arrhythmic event (sustained monomorphic ventricular tach ycardia in 13 cases, ventricular fibrillation in eight cases and sudde n cardiac death in 12 cases). The predictive power of late potentials in the time domain, spectral turbulence analysis and their combination s were tested together with clinical valuables using the Cox regressio n method The combination of late potentials and a spectral turbulence analysis score (one ol both abnormal) had the greatest influence on th e prediction of arrhythmic events: late potential+spectral turbulence analysis (regression coefficient B=0.75, P=0.0014), Selvester ECG scor e >10 (B=0.44; P=0.027), history of previous infarction (B=0.49; P=0.0 18) and non-sustained ventricular tachycardia during long-term ECG rec ording (B=0.50; P=0.035). The combination of late potentials and spect ral turbulence analysis reached a sensitivity of 76%, a specificity of 63%, a positive predictive value of 8% a negative predictive value of 98% and a total predictive accuracy of 61%. Patients with late potent ials and/or abnormal spectral turbulence analysis were significantly l ess often free of arrhythmic events compared to patients with normal f indings (92% vs 98%, <0.0001, Breslow test). The relative risk ratio o f combining late potentials+spectral turbulence analysis was 5.0 (CI 9 5% 2.1-11.5) for arrhythmic events.The combination of late potentials in the time domain and a spectral turbulence analysis score (one or bo th abnormal) is recommended for identification of patients at risk of arrhythmic events after acute myocardial infarction. Optimization of t he criteria of abnormality could further improve the rather low positi ve predictive accuracy of the signal-averaged ECG in this setting.