Identifying Origin of Nonpulmonary Vein Triggers Using 2 Stationary Linear Decapolar Catheters: A Novel Algorithm.

ABSTRACT

BACKGROUND: Targeting nonpulmonary vein triggers (NPVTs) of atrial fibrillation (AF) after pulmonary vein isolation can be challenging. NPVTs are often single ectopic beats with a surface P-wave obscured by a QRS or T-wave. OBJECTIVES: The goal of this study was to construct an algorithm to regionalize the site of origin of NPVTs using only intracardiac bipolar electrograms from 2 linear decapolar catheters positioned in the posterolateral right atrium (along the crista terminalis with the distal bipole pair in the superior vena cava) and in the proximal coronary sinus (CS). METHODS: After pulmonary vein isolation in 42 patients with AF, pacing from 15 typical anatomic NPVT sites was conducted. For each pacing site, the electrogram activation sequence was analyzed from the CS catheter (simultaneous/chevron/inverse chevron/distal-proximal/proximal-distal) and activation time (ie, CSCTAT) between the earliest electrograms from the 2 decapolar catheters was measured referencing the earliest CS electrogram; a negative CSCTAT value indicates the crista terminalis catheter electrogram was earlier, and a positive CSCTAT value indicates the CS catheter electrogram was earlier. A regionalization algorithm with high predictive value was defined and tested in a validation cohort with AF NPVTs localized with electroanatomic mapping. RESULTS: In the study patient cohort (71% male; 43% with persistent AF, 52% with left atrial dilation), the algorithm grouped with high precision (positive predictive value 81%-99%, specificity 94%-100%, and sensitivity 30%-94%) the 15 distinct pacing sites into 9 clinically useful regions. Algorithm testing in a 98 patient validation cohort showed predictive accuracy of 91%. CONCLUSIONS: An algorithm defined by the activation sequence and timing of electrograms from 2 linear multipolar catheters provided accurate regionalization of AF NPVTs to guide focused detailed mapping.