Risk Stratification of Patients After Myocardial Infarction Based on Wavelet Decomposition of the Signal‐Averaged Electrocardiogram

Abstract

Objective: The objective of this study was to investigate the prognostic significance of the wavelet decomposition (WD), when applied to single leads of the signal‐averaged ECG (SAECG) of postinfarction patients, as well as to compare it with conventional time‐domain (TD) analysis in order to optimize the utility of high resolution ECG techniques in risk stratification. Comparisons with clinical variables were also performed.Methods: SAECGs of 769 patients were analyzed (mean age 53 ± 7 years). A total of 42 arrhythmic events occurred within 54 months. For numerical calculations of WD, relative energies of the QRS complex were obtained in seven frequency bands. Conventional parameters of late potential analysis in TD were calculated at 25 and 40 Hz high pass filter.Results: No significant correlation was found between parameters of both methods. Several parameters of the WD of single leads showed slightly higher sensitivity (e.g., 52% vs 36%) and positive predictive values (e.g., 13.3% vs 10.2%) compared with those obtained with the QRS duration of TD analysis. In multivariate analysis, combination of QRS duration with WD parameters (relative risk [rr] =4.55, P < 0.0001), followed by the WD of single leads (rr = 4.19, P < 0.0001), nonsustained ventricular tachycardia in the long‐term ECG (rr = 3.72, P = 0.0005) and the history of previous infarction (rr = 2.72, P = 0.0035) were the most significant predictors in risk stratification.Conclusion: Time domain and wavelet transform of SAECG resulted in an improvement in postinfarction risk stratification compared with the clinical variables included in this study. Wavelet parameters may not be vastly superior to QRS duration, but they are a new index measuring intra‐QRS phenomena.