Spectral turbulence and late potentials in the signal-averaged electrocardiograms of patients with monomorphic ventricular tachycardia versus resuscitated ventricular fibrillation.

Signal-averaged electrocardiograms (SAECG) were analyzed for late potentials and spectral turbulence in 208 patients with ischemic heart disease with a history of sustained monomorphic ventricular tachycardia (MVT) (n = 62), resuscitation from ventricular fibrillation (VF) (n = 64) or no ventricular tachyarrhythmia (n = 82). Receiver operating characteristic curves were utilized to optimize cut-off values for prediction of MVT and VF. Patients with MVT had a lower ejection fraction (mean = 0.37) than patients with VF (0.44; p = 0.01) and controls (0.48; p < 0.0001). The mean FQRSD in MVT patients (126 ms) was longer than in VF and controls (113 ms; p = 0.005 and 102 ms; p < 0.0001, respectively). The RMS40 was lower in MVT (19 mu V) than in VF and controls (29 mu V; p = 0.0003 and 28 mu V; p < 0.0001, respectively); 81% of the MVT patients were spectral turbulence-positive vs 47% of VF patients and 31% of control patients (p < 0.0001 for both differences). With optimized reference values, FQRSD, TQRSD and ISCSD contributed significantly to the identification of MVT patients and FQRSD to VF patients. The sensitivity of combined time-domain and spectral turbulence analysis was 90% for MVT and 58% for VF, with 63% specificity. MVT patients had a lower ejection fraction and were more often late potential and spectral turbulence positive than VF and control patients. These findings indicate that a large electroanatomic substrate is required in MVT. A long FQRSD was a risk marker for both MVT and VF. Spectral turbulence analysis added independent information, and the combination of time-domain and spectral turbulence analysis was superior to either method alone in identifying the MVT patients. Neither method of analysis, singly nor in combination, performed satisfactorily in identification of VF risk.