TIME-DOMAIN AND FREQUENCY-DOMAIN ANALYSES OF THE SIGNAL-AVERAGED EGG IN PATIENTS WITH VENTRICULAR-TACHYCARDIA AND ISCHEMIC VERSUS NONISCHEMIC DILATED CARDIOMYOPATHY

The value of time-domain and frequency-domain (spectral turbulence) an alyses of the signal-averaged electrocardiogram was investigated to pr edict induced sustained monomorphic ventricular tachycardia (VT). Two groups of patients with spontaneous nonsustained VT and left ventricul ar ejection fraction less than 50% were enrolled: 70 patients with idi opathic dilated cardiomyopathy (group 1) and 70 patients with ischemic heart disease (group 2). Sustained VT was induced in 9 cases (13%) in group 1 and 16 (23%) in group 2. The prevalence of abnormal time-doma in and spectral turbulence analysis was 16 and 37%, respectively, in g roup I and 27 and 51%, respectively, in group 2 (NS). In group 1, the predictive accuracy of time-domain and spectral turbulence analysis fo r induced VT was 86 and 67%, respectively (P < .01). In group 2, the p redictive accuracy of the two techniques for induced VT was, respectiv ely, 79 and 66% (NS). In both groups, the predictive accuracy of time- domain analysis was higher than that of spectral turbulence analysis i n patients with intraventricular conduction defect (IVCD): 65 versus 2 5%, respectively, in group I (P < .01), and 81 versus 44%, respectivel y, in group 2 (P < .05). However, the predictive accuracy of time-doma in and spectral analyses was similar in patients without IVCD: 94 vers us 84%, respectively, in group 1, and 77 versus 74%, respectively, in group 2. Thus, in patients with dilated cardiomyopathy, (1) the etiolo gy does not affect the predictive accuracy of time and frequency domai n of the signal-averaged electrocardiogram for VT induction; (2) both time-domain and frequency-domain analyses have high predictive accurac y in patients without IVCD; and (3) spectral turbulence analysis does not improve VT prediction in patients with IVCD.