Atrial fibrillation type characterization and catheter ablation acute outcome prediction: Comparative analysis of spectral and nonlinear indices from right atrium electrograms

Atrial fibrillation (AF) is the most common arrhythmia in clinical practice. As catheter ablation (CA) is increasingly becoming the preferred treatment, identifying predictors of CA outcome is important to assist clinical decision making. Previous studies have explored spectral and nonlinear indices. However, a comparative analysis of indices from preprocedural intracardiac recordings is lacking. The aim of this study was to present a comparative analysis of spectral and nonlinear indices derived from a simple threshold-based segmentation of intracardiac electrograms (EGM) to characterize AF type (paroxysmal vs. persistent) and predict AF termination by CA. Preprocedural 1 minute recordings of right atrium EGM from 54 AF patients (age 58±9 years, 37 male; 27 paroxysmal AF, 27 persistent) were used EGM were band-pass filtered (F-EGM). f-wave activation fiducial points were identified using a threshold-based segmentation. The AF cycle length (AFCL) time series was built calculating the distance between consecutive fiducial points. The instantaneous fibrillatory rate (IFR) time series was calculated as AFCL inverse. Nonlinear indices were calculated: i) median AFCL (MAFCL); ii) localization index (LI) (concentration of 1FR histogram around the mode); iii) sample entropy (SampEn); iv) root mean square error (RAISE) of IFR vs. Gaussian distribution fit. Standard methods were used to calculate spectral indices: i) dominant frequency (f<;sub>p<;/sub>); ii) organization index (OI) (ratio of area under f<;sub>p<;/sub> and its harmonics to total power) of F-EGM Persistent AF was associated with higher f<;sub>p<;/sub> (p<;0.005), lower MAFCL (p<;0.01) and LI (p<;0.05), higher SampEn (p<;0.05) and RAISE (p<;0.01). 01 was not associated with AF type. Nonlinear indices: MAFC