Background: Atrial fibrillation (AF) is a mechanistically heterogeneous disorder, and the ability to identify sub-phenotypes ("endophenotypes") of AF would assist in the delivery of personalized medicine. We used the clinical response to pulmonary vein isolation (PVI) to identify a sub-group of patients with non-PV mediated AF and sought to define the clinical associations.
Methods: Subjects enrolled in the Vanderbilt AF Ablation Registry who underwent a repeat AF ablation due to arrhythmia recurrence were analyzed on the basis of PV reconnection. Subjects who had no PV reconnection were defined as "non-PV mediated AF". A comparison group of subjects were identified who had AF that was treated with PVI-only and experienced no arrhythmia recurrence >12 months. They were considered a group enriched for "PV-mediated AF". Univariate and multivariable binary logistic regression analysis was performed to investigate clinical associations between the PV and non-PV mediated AF groups.
Results: Two hundred and twenty nine subjects underwent repeat AF ablation and thirty three (14%) had no PV reconnection. They were compared with 91 subjects identified as having PV-mediated AF. Subjects with non-PV mediated AF were older (64 years [IQR 60,71] vs. 60 [52,67], P = 0.01), more likely to have non-paroxysmal AF (82% [N = 27] vs. 35% [N = 32], P<0.001), and had a larger left atrium (LA) (4.2cm [3.6,4.8] vs. 4.0 [3.3,4.4], P = 0.04). In univariate analysis, age (per decade: OR 1.56 [95% CI: 1.04 to 2.33], P = 0.03), LA size (per cm: OR 1.8 [1.06 to 3.21], P = 0.03) and non-paroxysmal AF (OR 8.3 [3.10 to 22.19], P<0.001) were all significantly associated with non-PV mediated AF. However, in multivariable analysis only non-paroxysmal AF was independently associated with non-PV mediated AF (OR 7.47 [95% CI 2.62 to 21.29], P<0.001), when adjusted for age (per decade: OR 1.25 [0.81 to 1.94], P = 0.31), male gender (OR 0.48 [0.18 to 1.28], P = 0.14), and LA size (per 1cm: 1.24 [0.65 to 2.33], P = 0.52).
Conclusions: Non-paroxysmal AF was the only clinical variable found to be independently associated with non-PV mediated AF. We demonstrated that analysis of AF ablation outcomes data can serve as a tool to successfully identify a sub-phenotype of subjects who have non-PV mediated AF.
Clinical trial registration: ClinicalTrials.gov ID # NCT02404415.