Errors in pulmonary vein identification and ostia location in the absence of pulmonary vein imaging

Abstract

Background: A key point in atrial fibrillation (AF) ablation is the ability to identify the pulmonary vein (PVs) and locate their ostia. Objectives: The purpose of this study was to assess the error margin of PV identification and ostia location in the absence of previous PV imaging. Methods: This study was performed in patients referred for catheter ablation of AF. PVs were reconstructed before ablation using the CARTO system. The operator tagged the superior and inferior edges of the PV ostia before and after examining the corresponding PV angiograms. The distances between the tagged PV ostia were measured using CARTO software. Results: A total of 105 location estimations of 54 PVs were analyzed. The location of PV ostia without angiography deviated from the angiographic PV ostia by a median of 13 mm (95% confidence interval = 11-14 mm; P < .0001). In 84 of the 105 estimations (80%), wrong tagging was performed inside the PV. A multiple logistic regression revealed that, at sites displaying PV potentials, the left atrial potential amplitude was an independent predictor of location at the angiographic PV ostium (odds ratio 24 [95% confidence interval = 3.7-227] per 1-mV increase). Receiver operator characteristic analysis set the optimal cutoff level at 0.7 mV. Use of this criterion improved the accuracy of PV ostium location by 4 mm (95% confidence interval = 1-6 mm; P = .005). Conclusion: Attempts at PV identification and ostia location in the absence of previous PV imaging are subject to a broad error margin