Po-02-169 patterns of pulmonary vein activity recorded during persistent af may help identify their contribution to the mechanism of af

Pulmonary vein isolation (PVI) is the cornerstone interventional treatment for AF but the success of this approach is more limited in persistent AF (PsAF). In these pts, the main challenge is to discriminate those who will respond to PVI only, from those who may require an extended ablation strategy. We sought to characterize the patterns of the electrical activity recorded within the PVs based on their postulated contribution to the mechanism of PsAF. Three subgroups of consecutive pts who underwent PVI for PsAF were considered based on the mechanism of AF: 1) definitely PV-related: procedural AF termination during PVI (“PVI-term” group) 2) likely PV-related: absence of AF recurrence > 18 months off-AAD after PVI only (“PVI-success” group) 3) definitely/likely PV-unrelated: PsAF recurrence despite isolated PV or at least one redo, respectively (“PVI-failure” group). Each PV and the left atrial appendage were evaluated before ablation. An automatic detection algorithm was developed to extract atrial activations from intra-cardiac electrograms in order to perform further analyses. Epochs of 60-sec were used. Different methods based on the atrial cycle length (CL) variability were assessed to best characterize the potential markers of PV activity such as the occurrence of bursts of rapid activity. A total of 40 pts were included (mean age: 60 ± 10 years, mean AF duration: 11 ± 10 months): PVI-term group (n=10), PVI-success group (n=17), PVI-failure group (n= 13). The dynamics of PV activation intervals was assessed using the envelope of the PV intervals time series. Among all features evaluated, the best feature that could discriminate similarly both the PVI-term and the PVI-success group from the PVI-failure group, was the standard deviation of the inferior envelope relative to the median CL of the PV (p <0.001 and p = 0.005, respectively) (Fig. A-B). Cut-off selected by ROC analysis provided a 100% sensitivity and 85% specificity to discriminate the PVI-term from the PV-failure group (AUC 0.931) and a 88% sensitivity and 77% specificity to discriminate the PVI-success from the PV-failure group (AUC 0.805). This feature performed better than the recently reported “Fastest PV run” (mean CL of the fastest 10-beat run) (AUC = 0.762 and AUC = 0.738, respectively) (Fig. C). Patterns of the electrical activity recorded within the PV differ according to the mechanism of PsAF. Their evaluation may provide the key to discriminate pts who may be cured by PVI only.