Time- and frequency-domain characteristics of atrial electrograms during sinus rhythm and atrial fibrillation

Abstract

Background: Complex fractionated atrial electrograms (CFAE) have been considered to be helpful during catheter ablation of atrial fibrillation (AF). The purpose of this study was to analyze the characteristics of CFAEs recorded during sinus rhythm (SR) and AF, and to determine their relationship to perpetuation of AF and clinical outcome.

Methods and results: Antral pulmonary vein isolation (APVI) was performed in 34 consecutive patients (age = 59 ± 10 years) with paroxysmal AF who presented in SR. Time- and frequency-domain characteristics of electrograms recorded from the same sites in the coronary sinus (CS) were analyzed during SR and AF, before and during isoproterenol infusion. There was a modest correlation in fractionation index (FI: change in the direction of depolarization, r = 0.40, P = 0.001) and complexity index (CI: change in the polarity of depolarization, r = 0.41, P = 0.001), but not in the dominant frequency (DF) between SR and AF. There was no relationship between the DF and CI or FI during AF. Isoproterenol was associated with an increase in DF during AF (6.6 ± 0.9 vs 5.1 ± 0.6 Hz, P < 0.001) but had no effect on CI or FI (P = 0.6). A higher CI (58.3 ± 21.0/s vs 38.0 ± 21.0/s, P < 0.01), and FI (123.5 ± 44.8/s vs 75.6 ± 44.6/s, P < 0.01) during AF were associated with a lower likelihood of termination of AF during APVI and a higher probability of recurrent AF after ablation. Ratio of FI during AF to SR was also higher when AF persisted than terminated after APVI (29.7 ± 12.4 vs 19.1 ± 9.7, P = 0.002). However, time- or frequency-domain parameters during SR were not predictive of termination or clinical outcome.

Conclusions: Structural and functional properties of the atrial myocardium during AF contribute to electrogram complexity, which may indicate the presence of extra-PV mechanisms of AF that are not eliminated by APVI. Mapping of complex electrograms in SR is not likely to be sufficient to identify drivers of AF.