Spectral analysis of 87-lead body surface signal-averaged ECGs in patients with previous anterior myocardial infarction as a marker of ventricular tachycardia

Abstract

Background. There were few studies on the relation between the body surface distribution of high- and low-frequency components within the QRS complex and ventricular tachycardia (VT). Methods and Results. Eighty-seven signal-averaged ECGs were obtained from 30 normal subjects (N group) and 30 patients with previous anterior myocardial infarction (MI) with VT (MI-VT[+] group, n=10) or without VT (MI-VT[-] group, n=20). The onset and offset of the QRS complex were determined from 87-lead root mean square values computed from the averaged (but not filtered) ECG waveforms. Fast Fourier transform analysis was performed on signal-averaged ECG. The resulting Fourier coefficients were attenuated by use of the transfer function, and then inverse transform was done with five frequency ranges (0-25, 25-40, 40-80, 80-150, and 150-250 Hz). From the QRS onset to the QRS offset, the time integration of the absolute value of reconstructed waveforms was calculated for each of the five frequency ranges. The body surface distributions of these areas were expressed as QRS area maps. The maximal values of QRS area maps were compared among the three groups. In the frequency ranges of 0-25 and 150-250 Hz, there were no significant differences in the maximal values among these three groups. Both MI groups had significantly smaller maximal values of QRS area maps in the frequency ranges of 25-40 and 40-80 Hz compared with the N group. The MI-VT(+) group had significantly smaller maximal values in the frequency ranges of 40-80 and 80-150 Hz than the MI-VT(-) group. These three groups were clearly differentiated by the maximal values of the 40-80-Hz QRS area map. Conclusions. It was suggested that the maximal value of the 40-80-Hz QRS area map was a new marker for VT after anterior MI.