Catheter ablation
Ablations were performed under general anesthesia with low tidal volume ventilation and uninterrupted anticoagulation. Three dimensional guidance was performed with the CARTO 3 system (Biosense Webster). Heparin infusion was utilized to maintain an activated clotting time of 350 seconds. Pacing from either the coronary sinus (CS) or right ventricle was performed to reduce the stroke volume and variability in cardiac motion. All procedures utilized irrigated radiofrequency with the STSF catheter aiming for interlesion distances of 4mm (Biosense Webster). In addition to VOM ethanol infusion, the standard lesion set for patients with persistent AF in this study consisted of PVI, left atrial posterior wall isolation (PWI), CS isolation, posterior mitral isthmus line (MIL), and a cavotricuspid isthmus line (CTI) (Figure 1). The mitral isthmus was ablated with air-filled balloon occlusion of the CS in almost all cases to improve lesion transmurality near the annulus.7 In the subset of patients both with persistent and occasionally paroxysmal AF wherein VOM ethanol was not preplanned and rather performed ad hoc (most commonly for induced mitral annular flutter in patients with myopathic low voltage regions, Table 1), MIL, CS isolation, and PWI was still systematically performed. The sought end points for CS isolation and PWI included elimination of local electrograms and lack of pace capture of the atrium with high output bipolar pacing from within the isolated structure (20mA, 2msec). The proximal third of the CS was not ablated to avoid iatrogenic atrioventricular (AV) block. For the first half of the enrollment period, luminal temperature monitoring was used for esophageal protection and occasionally, PWI was avoided in favor of a reinforced roof line if the esophageal risk was deemed too great because of esophageal heating or close proximity to the posterior wall on intracardiac echo (ICE). Towards the latter half of the study, patients underwent active esophageal cooling at 4°C (Enso ETM, Attune Medical, Chicago, IL). Rarely, CTI ablation was avoided if typical atrial flutter was not inducible and the appearance of the region on ICE appeared anatomically challenging. Adenosine was used to assess for dormant conduction and in select procedures, isoproterenol infusion was used to disclose non-PV foci.