Localization of complex fractionated atrial electrograms by approximate entropy in a 3D model of human atria

Atrial fibrillation (AF) is the most common tachyarrhythmia seen clinically. Radiofrequency ablation has become an accepted therapeutic alternative in patients refractory to pharmacologic management. Different ablation strategies ranging from targeting triggers originating in the vicinity of the pulmonary veins (PV) to mapping areas with high dominant frequency (DF) or complex fragmented atrial electrograms (CFAE) have been proposed to identify potential sites for ablation. Mapping of high DF has not proven to be entirely effective on chronic AF. To determine the behavior of approximate entropy (ApEn) and its ability to detect CFAE regions, we simulated an episode of chronic AF in a 3D model of human atria. A circular area of the posterior wall of the left atrium was selected to record electrograms. The following indexes were calculated: ApEn in short time intervals, DF and regularity index (RI). Wavefront activation maps were plotted for each of the above indexes. CFAE region was viewed as the tip of the rotors, and correlated with high ApEn values analyzed in short periods of time. DF and RI maps did not identify CFAE in short time intervals. These findings suggest that targeting ApEn areas may potentially be effective in identifying sites for ablation in chronic AF.