Averaged Impedance Drop Estimates Conduction Gap During Pulmonary Vein Isolation in Atrial Fibrillation

Background: EnSite X provides a module to filter oscillated generator impedance signals due to cardiac beating/respiration (averaged impedance drop, AID). TactiFlex ablation catheter Sensor Enabled (TFSE) is the latest product to incorporate a contact-force sensor with a flexible tip but without lesion-estimating parameters. Methods: To examine if AID estimates lesion formation, AID were compared between ablated lesions (n = 1687) with and without conduction gaps (CGs) after first-pass pulmonary vein isolation (PVI) using TFSE in atrial fibrillation (AF) patients (n = 20). The clinical efficacy of AID-guided PVI was evaluated in another subset of AF patients (n = 30). Results: CGs were observed in 45 points (CG[+]) but not in 1642 points (CG[−]). CG[+] had lower %AID (=absolute AID/initial impedance, p < 0.0001) than CG[−]. Decrease of bipolar voltage amplitude during radiofrequency application was linearly correlated with %AID (R²= 0.785, p < 0.05). In receiver operating curve analysis, the cutoff values of %AID and unfiltered generator impedance drop for predicting CG was 9.33% and 11.0 Ω, respectively; %AID had significantly higher area under the curve than unfiltered generator impedance drop (0.761 vs. 0.627, p < 0.05). In swine heart experiments, %AID was correlated with lesion volume (R² = 0.711, p < 0.05). In another subset of AF patients, success rate of first-pass isolation in %AID (≥ 9%)-guided PVI was 90% for LPV and 90% for RPV. The 1-year event-free rate of atrial tachyarrhythmias was 82%. Conclusion: AID would improve the usability of impedance drop as an end point for RF　application. %AID-guided PVI using TFSE would be effective in AF patients.