Reconnecting to the endpoint of atrial fibrillation ablation: should we mind the gaps?

This editorial refers to ‘Minimal invasive surgery for atrial fibrillation: an updated review’ by M. La Meir et al., on page 170–182and ‘Circumferential pulmonary vein isolation: wait or stop early after initial successful pulmonary vein isolation?’ by D. Bänsch et al., on page 183–188 Circumferential pulmonary vein isolation (PVI) is the current treatment of choice for drug refractory symptomatic atrial fibrillation (AF). Although PVI is safe, recurrences of AF may occur, and are almost always associated with reconnection of one or more pulmonary veins.1 Therefore, there is consensus that complete isolation of all the pulmonary veins is the cornerstone of the ablation procedure. However, we can never be sure that the PVs remain isolated over time, because myocardial oedema might cause acute but reversible conduction block. Additionally, the decision to plan a redo procedure depends on recurrent AF, and not on reconnection of the pulmonary veins. Pratola et al.2 showed in 20 asymptomatic patients undergoing electrophysiological evaluation after successful PVI that reconnection of one or more PVs was present in as many as 62.5%, and voltage abatement was not complete around 60% of PVs. Remarkably, these rates did not differ from 10 patients undergoing a redo ablation because of recurrent AF, who served as control. In other words, the durability and meaning of acute conduction block during PVI are questionable. One could argue that waiting long enough for acute reconnection to appear would allow treatment of any reconnection of the pulmonary veins, and confer a better outcome, and there are data supporting this strategy.3 The questions how long a wait is worth waiting for, and whether that is feasible in terms of procedure duration and the electrophysiology laboratory logistics remain unanswered. In this issue of the Journal, Bänsch et al.4 tested the hypothesis that waiting after successful isolation of the pulmonary veins and treating reconnection that occurred during the waiting time would favourably affect long-term freedom from AF.4 They randomized 107 consecutive patients undergoing PVI, the majority of whom had paroxysmal AF (65.4%), to either an additional 1h waiting period after successful PVI was achieved and subsequent ablation of any reconnection (‘wait’ group), or immediate abortion of the procedure once the procedural endpoint was reached (‘stop’ group). Ablation was randomly started at the septal or the lateral veins to prevent bias in all patients. During the waiting period, 36 gaps occurred in 27 (out of 54) patients, which were subsequently ablated. Surprisingly, repairing the gaps that appeared during the 1h waiting period did not result in a better long-term effect of the procedure: after a little more than 9 months atrial arrhythmias recurred in 51.9% in the ‘wait’ group and 56.6% in the ‘stop’ group. Patients in the ‘stop’ group had a higher prevalence of previous stroke, structural heart disease, and previous heart surgery. Seventeen and 21 patients had recurrent paroxysmal AF, 4 and 6 patients had recurrent persistent AF in the ‘wait’ and ‘stop’ group, respectively. It should be noted that these numbers are reported without taking a blanking period of 3 months into account, which might negatively affect the outcome. Interestingly, in patients who underwent a redo procedure because of recurrent AF (n = 16 in both groups), the number of gaps ablated was higher in the ‘wait’ than in the ‘stop’ group, although this difference did not reach statistical significance. So it seems that the number of additionally ablated reconnections, performed during the 1h waiting period after the initial PVI procedure, cannot be subtracted from the number of future gaps in the ablation lines, and therefore cannot be regarded as an investment to reduce the number of future ablation lesions or ablation procedures. Even when the absence of a 3-month blanking period is taken into account, the efficacy of the PVI procedure, either directly terminated or after a waiting period, must have been disappointing for the authors too, although they explicitly stated that redo procedures should be regarded as an integral part of the ablation procedure. These results may seem to contradict with the previous PVI studies, but it should be noted that in the recent HRS/EHRA/ECAS consensus document, a 1-year, single procedure freedom of AF rate of 50% in paroxysmal AF and 40% in persistent AF, is considered ‘reasonable’.5 It can be inferred from that statement that many of us have the same experience. It is possible that techniques other than point-by-point ablation, such as epicardial ablation of PVs, are less vulnerable to the occurrence of reconnection. Surgical treatment of AF, the Cox-Maze III, and IV operation in particular, is highly efficacious, but its use as a stand-alone treatment has never gained widespread acceptance due to its invasive nature, surgical complexity, and associated complications including frequent pacemaker implantations and 1–2% mortality.6 Therefore, in an attempt to make the surgical approach less invasive while pertaining the excellent efficacy, a minimal invasive approach using bilateral thoracotomy was described in 2005,7 and subsequently developed into a totally thoracoscopic procedure. In this issue of the Journal, La Meir et al.8 provide an extensive overview of data currently available on minimally invasive surgery for AF.8 They reviewed a total of 1051 patients, enrolled in 28 studies, using various energy sources among which radiofrequency (RF) was most frequently used. In contrast to an earlier systematic review on this topic where, only studies using RF and reporting outcome in a systematic manner were included,9 La Meir and colleagues present all the available data to the reader, and report a freedom from AF, off antiarrhythmic drugs, of 65–92% in paroxysmal AF and 67–80% in persistent AF. They note that due to the heterogeneity among the publications reviewed, the absence of prospective, randomized data and the difference in follow-up, the comparison of such studies might be difficult.8 Despite these uncertainties, they have demonstrated that minimally invasive surgery for AF provides satisfactory results, and speculate that these might improve further when using a hybrid approach. They have shown that periprocedural, endocardial confirmation of ablation lesions results in an excellent 1 year freedom of AF, and off antiarrhythmic drugs of 83%.10 This number is very similar to the 86% freedom of AF recently reported after a fully epicardial hybrid approach.11 Hence, in the setting of thoracoscopic surgery for AF it seems that looking for incomplete ablation lines and subsequently ablating the gaps seems to pay off. There are limited data on reconnection after minimally invasive surgery for AF, but at least in selected cases undergoing redo catheter ablation after failed surgery, reconnection of the pulmonary veins or gaps in left atrial lesions play an important role.12 The main difference between many studies reviewed by La Meir et al. and the study of Bänsch et al. is that Bänsch et al. waited after the demonstration of an initially successful procedure, that is, they demonstrated initial PVI in all their patients. It is unknown whether PVI and bidirectional conduction block across left atrial lines was reliably detected during non-hybrid surgical ablation, and whether surgical instruments used for epicardial ablation were capable of doing that. It cannot be excluded that the failed surgical procedures resulted from the absence of conduction block across all ablation lines in the first place rather than from the occurrence of reconnection of initially complete ablation lines. Should we encourage the surgeons to wait longer, after initial success of their (hybrid) thoracoscopic ablations, to be able to treat any reconnection that appears? That might certainly add to the understanding of the nature of the procedure, but it seems more important to realize that, regardless of what procedure is used, the ultimate and only clinically relevant outcome of AF treatment is not the periprocedural isolation of the PVs, but the absence of symptomatic AF recurrence. The author is supported by a personal grant from the Netherlands Heart Foundation, 2009T021, and received speakers/consultancy honoraria from Atricure Inc. Sébastien Krul and Wouter Berger are acknowledged for critical reading of the manuscript. Conflict of interest: none declared.