Wavelet decomposition analysis of the signal averaged electrocardiogram used for risk stratification of patients with hypertrophic cardiomyopathy.

Abstract

To study the predictive value of wavelet decomposition, as demonstrated by the signal-averaged ECG, in order to identify patients with hypertrophic cardiomyopathy at increased risk for malignant ventricular arrhythmias or sudden death. Two hundred and forty-six patients with hypertrophic cardiomyopathy were studied. During a mean follow-up of 68 +/- 17 months 32 patients died, of whom 17 died suddenly. Patients with sudden death, together with eight patients with a history of ventricular fibrillation (sudden death/ventricular fibrillation group) were analysed and compared to the other 221 patients as well as to a subgroup of 82 patients without a history of syncope, ventricular arrhythmias on a long-term ECG recording or a family history of sudden death. There were no differences in mean values of the four wavelet decomposition parameters among patients in the sudden death/ ventricular fibrillation group, those without sudden death/ ventricular fibrillation or patients in the low risk group. There were, however, significant differences between patients dying non-suddenly and patients being alive at the end of follow-up. Eighty-seven patients (35%) demonstrated evidence of non-sustained ventricular tachycardia on a long-term ECG. Analysis of wavelet decomposition resulted in abnormal findings in these patients more often than in those without ventricular arrhythmias. The usefulness of wavelet decomposition analysis in predicting sudden death or ventricular fibrillation is limited in patients with hypertrophic cardiomyopathy. It may, however, play a role in identifying patients at risk of dying non-suddenly and of patients with non-sustained ventricular tachycardia.