Frequency analysis of atrial fibrillation surface and intracardiac electrograms during pulmonary vein isolation

Abstract

Frequency analysis of atrial electrograms from patients diagnosed with persistent atrial fibrillation (AF) appears to be crucial in its clinical diagnosis. This work explores the fibrillatory frequency properties of both surface and intracardiac electrograms before and after pulmonary vein isolation (PVI) using three time-frequency techniques. Surface electrocardiograms (ECGs) of 21 patients diagnosed with persistent AF undergoing PVI were recorded. Three methods, Fourier, ensemble average, and wavelet analysis, were used to identify the dominant frequency (DF) in surface ECGs. Dominant frequency was also computed in electrograms recorded within the coronary sinus (CS). Dominant frequency measured within the CS was best estimated in surface lead V1 using both Fourier (relative error: 10.94 ± 10.37%, correlation: 0.58) and wavelet analysis (relative error: 10.97 ± 11.08%, correlation: 0.53). Ensemble average gave highest relative error (21.29 ± 18.07%) and lowest correlation (0.10). Dominant frequency decreased after right PVI. This decrease was significant (P< 0.05) in most of the patients (13, 14, and 7 out of 14 when Fourier, wavelets, and ensemble average was used; 14 in CS). Further isolation of the left pulmonary veins (PVs) yielded a significant (P< 0.05) decrease in only a few of them (3, 4, and 2 out of 14 when Fourier, wavelets, and ensemble average was used; 4 in CS). Wavelet and Fourier analysis are good tools for estimating the atrial fibrillatory rate from surface ECG. A drop was observed in the DF value after isolation of the right PV. However, after left PVI this decrease was smaller.