Prevalence, characteristics, and predictors of endocardial and nonendocardial conduction gaps during local impedance-guided extensive pulmonary vein isolation of atrial fibrillation with high-resolution mapping.

Abstract

Local impedance (LI) drops during radiofrequency ablation can predict lesion formation. Some conduction gaps during pulmonary vein isolation (PVI) can be associated with nonendocardial connections. This study aimed to investigate the incidence, characteristics, and predictors of endocardial and nonendocardial conduction gaps during an LI-guided PVI. We prospectively enrolled 157 consecutive patients undergoing an initial LI-guided extensive PVI of atrial fibrillation (AF). After the first-pass encirclement, the residual conduction gaps and reconnected gaps were mapped using Rhythmia (Boston Scientific) and a mini-basket catheter. Right and left PV (RPV/LPV) gaps were observed in 22.3% and 18.5% of the patients, respectively: 27 endocardial and 49 nonendocardial gaps. The carina regions were common sites for the gaps (51 carina-related vs. 25 noncarina-related). The carina-related gaps consisted of more nonendocardial gaps than endocardial gaps (RPVs: 90.0% vs. 10.0%, p = .001; LPVs: 76.2% vs. 23.8%, p < .001). A univariate analysis revealed that paroxysmal AF and the left atrial (LA) volume index for RPV endocardial gaps (odds ratio [OR]: 8.640 and 0.946; p = .043 and 0.009), minor right inferior PV diameter for RPV nonendocardial gaps (OR: 1.165; p = .028), and major left inferior PV diameter for LPV endocardial gaps (OR: 1.233; p = .028) were significant predictors. During the LI-guided PVI, approximately two-thirds of the conduction gaps were nonendocardial. The carina regions had more conduction gaps than noncarina regions, which was due to the presence of nonendocardial connections. Paroxysmal AF, a lower LA volume index, and larger inferior PV diameters may increase the risk of conduction gaps.