Aim: To compare four analysis techniques of the signal-averaged-electrocardiogram, including time-domain, spectral temporal mapping, spectral turbulence analysis and the new acceleration spectrum analysis.
Methods and results: We studied 634 subjects (77 with bundle branch block) divided into three groups. Group 1 comprised 117 post-myocardial infarction patients tested for inducibility of sustained ventricular tachycardia, and which was induced in 54 of them. Group 2 comprised 407 consecutive acute myocardial infarction survivors, followed for 1 year; 29 of them had suffered major arrhythmic events: 15 were cases of sustained ventricular tachycardia, three resuscitated ventricular fibrillation and 11 sudden cardiac death. Group 3 comprised 110 control subjects. The different analysis techniques were compared by their likelihood ratio for the prediction of ventricular tachycardia inducibility (Group 1) or major arrhythmic events (Group 2). The likelihood ratios of spectral-turbulence-analysis, acceleration spectrum analysis, spectral temporal mapping and time-domain were 8.0, 3.3, 1.7, 1.3 in Group 1, and 3.8, 2.1, 1.5, 2.6, in Group 2, while the corresponding false-positive rates in Group 3 (control) were 0.9%, 10.0%, 4.5%, and 3.6%, respectively.
Conclusion: Spectral turbulence analysis was the most accurate technique for the prediction of either ventricular tachycardia inducibility or major arrhythmic events after myocardial infarction. It also showed the highest specificity among control subjects.