Session PB1.3

A Comparative Study of Abnormal Intra-QRS Potentials and High-Frequency Components in Signal-Averaged Electrocardiogram

CC Lin, WC Hu*

National Chin-Yi University of Technology
Taiping, Taiwan

The abnormal intra-QRS potentials (AIQP) which was defined as the low-amplitude notches and slurs in the QRS complex have been proposed as a new index for evaluating the risk of ventricular arrhythmias. The current autoregressive moving average (ARMA) model built in discrete cosine transform domain has the advantage that a low order model can estimate most of the normal QRS complex and then extract the AIQP from the modeling residuals. Because of the abnormal myocardial conduction, the patients with ventricular tachycardia (VT) were expected to display an increased of AIQP. However our previous study showed that the mean AIQP parameters of VT patients surviving myocardial infarction (MI) were significantly lower than those of normal subjects. The reduced AIQP in VT patients may be related to the decreased high-frequency (HF) components induced by MI. Several previous studies have reported that MI could reduce the HF components within QRS complex. This study will be devoted to comparing the clinical performance of AIQP and HF parameters in the diagnosis of VT, and analyzing the reduced of AIQP in VT patients would be correlated to the changes in HF components.
High-resolution electrocardiograms of 42 normal subjects and 30 VT patients surviving clinically documented MI were analyzed in this study. The AIQP were estimated using the ARMA models with a low-order of (6, 6) and a high-order of (12, 12), respectively. The HF components were extracted using a fourth-order bi-directional Butterworth filter with the frequency band of 150 to 250Hz. The results show that the low-order AIQP parameters in lead Y and Z, the high-order AIQP parameter in lead Z, and the HF(150-250Hz) parameters in lead Y and Z of VT patients were significantly lower than those of normal subjects (p<0.05). The cross-correlation analysis furthermore demonstrated that both the low- and high-order AIQP parameters were significantly positively correlated with the HF parameters (p<0.001). The cross-correlation coefficients in leads X, Y and Z are 0.75, 0.56, and 0.71 between the low-order AIQP and HF parameters, respectively, and 0.74, 0.52, and 0.68 between the high-order AIQP and HF parameters, respectively. The reductions of AIQP and HF components in VT patients are consistent. The significant correlations between the AIQP and HF parameters imply that the reduced AIQP in VT patients is correlated with the reduction of HF components.

(Abstract Control Number: 24)