PURPOSE The ablation index (AI) is a recently developed marker for ablation lesion quality that incorporates contact force (CF), time, and power in a weighted formula. There is a paucity of information on whether AI-guided pulmonary vein isolation (PVI) could improve the outcome in patients with atrial fibrillation (AF). We evaluated the optimal AI threshold for avoiding acute pulmonary vein reconnection (PVR), and to compare the efficacy of optimal AI-targeted PVI with that of conventional CF-guided PVI.

METHODS Seventy patients with AF (paroxysmal, 67%) were enrolled. In a phase 1 study, the patients underwent conventional CF-guided PVI (CON group), and the optimal AI threshold for avoiding acute PVR was identified. In phase 2, the patients underwent AI-guided PVI (OAI group). We compared the acute PVR rate between the CON group and the OAI group to demonstrate the efficacy of AI-guided PVI.

RESULTS In phase 1 (n = 38), acute PVR was observed in 57 of 532 (10.7%) segments. AI values of ≥ 450 at the anterior/roof segments and of ≥ 350 at the posterior/inferior/carina segments were identified as the optimal AI thresholds for avoiding acute PVR. In the phase 2 study targeting those AI values, the OAI group (n = 32) showed a significantly lower acute PVR rate than the CON group (4.2% vs. 10.7%, p < 0.001). The OAI group showed a higher minimum AI and smaller variations in AI values than the CON group.

CONCLUSIONS Optimal AI-targeted PVI is feasible and could improve the acute outcome in patients with AF.

TRIAL REGISTRATION URL: https://www.clinicaltrials.gov. Unique identifier: NCT03389074.