Session S32.6

Stability and Correlation of Electrogram Organization and Synchronization Indices during Atrial Fibrillation

F Simón*, A Arenal, P Laguna, JP Martínez

Universidad de Zaragoza
Zaragoza, Spain

Atrial fibrillation (AF) is a very common cardiac arrhythmia due to electrical re-entry in the atria. It may lead to cardiac embolism and heart failure. When electric or pharmacological cardioversion do not succeed, ablation therapy is usually performed, with the aim of allowing only one electrical path between sinoatrial and atrioventricular nodes. A key issue in ablation is where and to what extent to ablate. Analysis of electrogram signals (EGM) may provide information about local electrical activity, which may be useful to determine the desirable approach to ablation. The aim of this work is to study the stability and correlation of indices of the regularity of a site's electrical activity, and indices of synchronization between different atrial sites.
We used a database of 30 EGM records registered in Gregorio Marańón Hospital (Madrid) during catheter ablation (average duration: 129 s). In this work we focus in the 10 bipolar EGM signals from a 20-electrode circular catheter (Lasso). EGM signals were filtered and an activation detector as well as a wavefront detector were applied. Signals were divided into non-overlapping 10-second windows, and for each one, we derived indices based on spectral analysis (dominant frequency DF, regularization index RI, organization index OI and average coherence index G), on cross-correlation analysis (maximum normalized cross-correlation ?, and lag with maximum cross correlation t) and on wavefront analysis (interquartile range of delay vectors C and median delay index µ).
The statistical analysis showed that organization indices based on spectral analysis are quite stable over time. Coefficient of variation (CV) was 0.008 for DF and 0.017 for RI and OI, while intra-record to inter-record variance ratio VR was 0.001, 0.002 and 0.003, respectively. Among the inter-lead synchronization indices, the most stable were the average coherence and the maximum cross-correlation. As expected, results show decreased synchronization and waveform consistency as the separation between electrodes increases. We found a significant correlation between RI and OI (r = 0.86, p = 0), but not between RI/OI and DF. However, both RI and OI are negatively correlated to the intra-lead standard deviation of DF (r = -0.61, p = 3•10^-4 and r = -0.66, p = 7•10^-5). Regarding the correlations between synchronization indices, we can highlight the high negative correlation (r = -0.90) observed between the averaged coherence and the interquartile range of delay vectors indices. These results show that organization and synchronization indices based on the spectral analysis (DF, OI, RI and average coherence) and cross-correlation can be reliably computed on 10-second excerpts, while those based on wavefront analysis are not stable enough to guide clinical decisions.

(Abstract Control Number: 168)