USEFULNESS OF SIGNAL-AVERAGED ELECTROCARDIOGRAM IN IDIOPATHIC DILATEDCARDIOMYOPATHY FOR IDENTIFYING PATIENTS WITH VENTRICULAR ARRHYTHMIAS

In idiopathic dilated cardiomyopathy (IDC), the relation between the s ignal-averaged electrocardiogram and ventricular tachycardia (VT) rema ins unclear. In this study, conventional time domain and frequency dom ain analyses (2-dimensional, spectral temporal mapping and spectral tu rbulence analysis) of the signal-averaged electrocardiogram were perfo rmed in 64 patients with IDC. Eight patients had a history of symptoma tic sustained VT and an additional 24 had nonsustained VT recorded dur ing ambulatory electrocardiography. Conventional time domain analysis, using the 25 and 40 Hz filter, and spectral temporal map ping, detect ed late potentials within the terminal QRS in 8 (13%), 14 (22%) and 18 (28%) patients, respectively. Late potentials were seen more often in patients with than without VT, and in patients with sustained versus nonsustained VT, but these differences were not significant. The predi ctive accuracy of these techniques in detecting either form of VT were : sensitivity, 22, 25 and 31%; specificity, 97, 81 and 75%; and overal l predictive value, 59, 53 and 50%, respectively. Two-dimensional freq uency domain analysis of the signal-averaged electrocardiogram reveale d a higher energy and area ratio in patients with than without VT (ent ire QRS), and in patients with sustained versus nonsustained VT (entir e QRS and terminal QRS). Spectral turbulence analysis was abnormal in 24 patients (39%), but no differences were observed between patients w ith and without VT. During follow-up (mean duration 18 +/- 14 months) 5 patients had arrhythmic events (3 died suddenly, 1 had aborted sudde n death and 1 developed sustained VT). The signal-averaged electrocard iogram was abnormal in 1 patient using time domain analysis and spectr al temporal mapping, and in 3 using spectral turbulence analysis (scor e = 4) conclusion, time domain and spectral temporal mapping analysis of the signal-averaged electrocardiogram identifies IDC patients with VT with high specificity but low sensitivity. The application of the 2 -dimensional frequency analysis improves identification of patients wi th sustained VT. Spectral turbulence analysis of the entire QRS comple x is a promising technique for the identification of patients with IDC at high risk of sudden death.