Procedural approach
Endo-/epicardial CA was performed in the electrophysiology laboratory, under general anesthesia. Patients were studied during AF; in case of sinus rhythm at the beginning of the procedure, AF/atrial tachycardia was induced using burst stimulation.
We performed an extensive ablation procedure in each patient, starting from the LA endocardium, which was accessed with a single transseptal puncture. Unfractionated heparin was administered in order to maintain activated clotting time values >300 s. Procedures were guided by electroanatomical mapping system (CARTO, Biosense Webster)-derived three-dimensional reconstructions of the atria. Biatrial endocardial maps were obtained with a circular mapping catheter (Lasso, Biosense Webster), whereas biatrial epicardial replica were obtained using the ablation catheter (Thermocool Smarttouch, Biosense Webster). CA was guided by ablation index, targeting values higher than 550 in every site, apart from the posterior wall of the LA and the epicardium, where values higher than 500 were obtained.12
The procedural approach is represented in figure 1. The endocardial phase of the procedure aimed at insulating pulmonary veins and closing LA and right atrial (RA) anatomical isthmuses. Endocardial steps included antral PVI, followed by linear lesions on the roof of the LA connecting right and left superior PVs, on the anterior wall of the LA connecting the roof line with mitral valve annulus, in the septal aspect of the posterior wall of the right atrium (RA) connecting superior and inferior venae cavae, and cavotricuspid isthmus ablation.
After completing this lesion set, access was gained to the epicardium, by creating a small (2-3 cm) pericardial window, through which a steerable introducer (Agilis EPI, Abbott) was placed into the pericardial space. The pericardial window was performed in collaboration with a cardiac surgeon. We preferred such an approach to the classic pericardial puncture described by Sosa et al.,13because of the more favorable catheter manipulability with the epicardial window.
In the epicardium, BB and the LOM were ablated first according to their anatomic location (in the floor of the transverse sinus close to the interatrial septum, and between the LA appendage and pulmonary veins, respectively), followed by the eventual completion of roof and anterior linear lesions in the proximity of the mitral valve in case of absence of lesion transmurality. In case of ongoing AF, further ablation lesions were delivered in the atria from the endocardium, targeting complex/fractionated atrial electrograms in both left and right atria. In case patients developed atrial tachycardia at any step during the procedure, the arrhythmia was mapped and ablated, using a high-density catheter (PentaRay, Biosense Webster). If sinus rhythm was not restored at this point, patients were electrically cardioverted. Bidirectional conduction block across all linear lesions and PV circumferential entrance and exit blocks were verified in sinus rhythm by pacing techniques, as described elsewhere.14 At the end, burst pacing was performed to confirm loss of arrhythmia inducibility from two different sites (coronary sinus and LA appendage), until refractoriness was reached, or an arrhythmia induced. In case of atrial tachycardia induction, the latter was mapped and ablated, aiming to restore sinus rhythm; in case of AF induction, we performed electrical cardioversion. AF termination, lack of arrhythmia inducibility, and validation of conduction block through ablation lines were the acute procedural endpoints.
Class III or Ic antiarrhythmic drugs were prescribed for the first three months after CA and suspended thereafter in case of absence of arrhythmia recurrences.