Introduction: Sites of complex fractionated atrial electrograms (CFAEs) and highest dominant frequency (DF) have been proposed as critical regions maintaining atrial fibrillation (AF). This study aimed to determine the minimum electrogram recording duration that accurately characterizes CFAE or DF sites for ablation without unduly lengthening the procedure.
Methods and results: Fourteen patients with AF undergoing catheter ablation had high-density (498 +/- 174 points) biatrial mapping performed during AF before ablation. At each point, 8-second electrograms were recorded. CFAE characterization using the NavX software provided a representation of electrogram complexity (CFE-mean). CFE-mean for each point from 7-, 6-, 5-, 4-, 3-, 2-, and 1-second subsamples were compared with the index 8-second CFE-mean. Offline spectral analysis defined DF as the frequency with greatest power, and DF of subsamples were compared with index DF. Index 8-second electrogram CFE-mean was 114 +/- 20 ms for right atria and 102 +/- 17 ms for left atria (P = 0.01); DF was 5.7 +/- 0.8 Hz for right atria and 6.0 +/- 0.8 Hz for left atria (P = 0.02). Means from shorter electrograms were nonsignificantly decreased for CFE-mean and overestimated for DF (P < 0.001). Mean absolute differences between subsampled and index values ranged from 3.3 to 20.1 ms for CFE-mean and 0.11 to 1.18 Hz for DF. Subsampled electrograms deviating >10% from index values ranged from 2.5 to 56% for CFE-mean and 3.5 to 41% for DF. Intraclass correlation coefficients ranged from 0.992 to 0.788 for CFE-mean and 0.897 to 0.233 for DF. Unacceptable differences from index values were found with CFE-mean and DF from electrograms <5 seconds.
Conclusion: Electrograms of >or=5-second duration are required to accurately characterize CFAE and DF sites for ablation.