[Comparison of parameters of ventricular late potentials by using four commercial high-resolution ECG systems].

The methodologic problems of non-invasive registration of ventricular late potentials (LP) were studied. Four commercially available high resolution ECG devices were used. Method A was MAC 1 from Marquette, Method B was MAC 12 (15) from Marquette. Method C was ART 101 PC from Arrhythmia Technology Research. Method D was Fukuda VCM3000 from Fukuda Denshi in Japan. Our study showed the good correlation of the filtered QRS duration (FQRSd) between method B (fast Fourier transform filter) and method C (bidirectional digital filter). However the poor correlation between method B, C (digital filter) and D (analogue filter) was obtained. When late potentials positive patients who had high frequency and low amplitude signals were compared, the worse correlation of FQRSd among method B, C and D was obtained. FQRSd in method B was longer than that in method C, and that in method D was longer than that in method B. The differences were statistically significant. The measurements of low amplitude signals [Root mean square voltage in last 40 msec (V40) and High Frequency and low amplitude signals duration under 40 microV (HFLAD)] in various methods were much difference. We concluded that the standardization of the filter and leads to evaluate the late potentials was needed. The system specific criteria to detect the sustained ventricular tachycardia have to be redefined in VCM 3000. It was important to evaluate visually LP, as LP far from QRS offset could not be registered with vector magnitude methods in individual cases.