Surgical or catheter ablation for longstanding persistent atrial fibrillation: a game of rock paper scissors?

This editorial refers to ‘Catheter ablation vs. thoracoscopic surgical ablation in long-standing persistent atrial fibrillation: CASA-AF randomized controlled trial’†, by S. Haldar et al., on page 4471. Patients with symptomatic paroxysmal atrial fibrillation (AF) despite antiarrhythmic drug (AAD) treatment respond very well to pulmonary vein isolation (PVI) by catheter ablation (CA) with high freedom from arrhythmia rates.1 However, once the AF episodes become longer in duration or even persistent beyond 1 year [longstanding persistent AF (LSPersAF)], the freedom from AF rates by PVI go down rapidly. Strategies of adding ablation lines at the left atrial (LA) roof or LA debulking of abnormal electrical signals did not improve outcomes.2 In the meantime, cardiac surgeons embraced a simplified minimally invasive surgical ablation (SA) by video-assisted thoracoscopic high-energy radiofrequency (RF) ablation of the bilateral PVs and ganglionated plexi, with exclusion of the LA appendage (LAA). The FAST randomized controlled trial (RCT) was the first to ever directly compare CA with SA, showing superiority of SA in AF patients with low chance of success for CA or failed prior CA, albeit at the expense of more adverse events.3 This has led to a recommendation for SA in patients that are likely to fail or have failed CA, especially when they have LSPersAF. Beyond the FAST, RCTs have been few and underpowered as they are hard to perform, and usually patients and physicians intuitively favour one or the other type of therapy based on emotion rather than on hard scientific data. In the current issue of the European heart Journal, we are lucky to find an RCT again to directly compare the relative merits of CA and SA4 in patients with LSPersAF. Although a pilot performed by the same authors suggested a superior outcome for SA, the full study found that SA and CA freedom from AF outcomes without AAD at 1 year were no different {26% vs. 28% [odds ratio (OR) 1.128, 95% confidence interval (CI) 0.46–2.82, P = 0.84]}, as well as clinical success defined as 75% or more reduction in arrhythmia burden [67% vs. 77% (OR 1.13, 95% CI 0.67–4.08, P = 0.3)]. SA did come at a higher price of all-cause procedural adverse events [40% vs. 15% (OR 3.78, 95% CI 1.55–9.21, P = 0.003)]. This could only have been an acceptable trade-off if there would have been a clearly superior overall efficacy. In contrast, freedom from arrhythmias was equal for CA and SA, while the quality of life scores improved more for CA than SA, which may even reflect the higher adverse event rate and longer recovery from the more invasive surgical procedure. Finally, in this trial, SA was also found to be more expensive, with fewer quality-adjusted life years (QALYs) than CA (0.78 vs. 0.85, P = 0.02) Before diving into an appraisal of why the principal hypothesis of superiority of SA was not proven, there are a few methodological issues to be raised. The authors powered their study with only 120 patients based on a prior RCT in patients with less advanced AF3 and their own 51 patient pilot study that observed a 30% difference in outcome in favour of SA (77% vs. 44% freedom from AF),5 which may not have been the optimal basis for power calculation. In fact, both treatment arms showed a disappointing huge underperformance with <30% freedom from arrhythmias. In part this may be explained by the very strict guideline-driven endpoint of freedom from AF in this very tough AF population that has defied any treatment so far, while the introduction of implantable loop recorders (ILRs) has exposed that invasive procedures are much less successful than we thought in the early days. On the other hand, arrhythmia burden reductions of 67–76% achieved by both treatments are quite an impressive clinical effect, and maybe a more realistic endpoint than the all or none effect electrophysiologists expect from past experience with accessory pathway and cavo-tricuspid isthmus elimination. Another limitation was the fact that in the 60 patient SA arm, five withdrew consent when they were randomized to SA, one died, in two PVI was incomplete, and five crossed over to the CA arm, so that 65 patients actually underwent CA and only 47 patients underwent the intended complete SA (of which 9 patients did not even have complete follow-up). It is questionable whether this allows for a robust appraisal of SA efficacy. In fact, as with prior trials that directly compared SA with CA, the enrolment numbers were too (s)low, and there is an urgent need for more collaboration around the world to perform much higher volume studies of both treatments, as is customary for drug trials. Coming back to the more basic reasons why SA may not have achieved superiority, there are several issues to consider: pathophysiology, strategy, and ablation tools. While at the time of their introduction epicardial SA tools were much more powerful than endocardial RFCA tools, the tides have turned with the introduction of irrigated RF ablation, and lesion durability may improve further with contact-force guiding and novel single-shot CA devices with ultralow cryo-ablation or irreversible electroporation as energy sources.6 Since the basic strategy for both CA and SA was to create a posterior box incorporating the PVs, it is not surprising that outcomes would be the same. Although in the SA group ganglionated plexi ablation and LAA exclusion were added, two RCTs7 , 8 already demonstrated that this does not benefit freedom from arrhythmia, although LAA clipping may reduce the long-term stroke rate. The added benefit of cavo-tricuspid and mitral isthmus ablation in the CA group is also not apparent from the low freedom from arrhythmia observed. This brings us back to square one; what is the pathophysiology of LSPersAF, and how do we improve outcomes beyond a standardized anatomical approach? Surgeons and electrophysiologists have started working together in (staged) hybrid approaches to extend lesion sets and check for completeness and durability of such lesions.9 Whether this results in more Cox-MAZE III-like results in the modern age of scrutinous ILR monitoring remains to be seen, and cost-effectiveness by stacking all these therapies on top of each other may be even more difficult to achieve. Novel mapping technologies10 , 11 reveal that the AF substrate may not be limited to the posterior wall and PVI, calling for patient-specific ablation of anatomical and functional targets. So, the appropriate question may actually be where to ablate, rather than how to ablate. From the current study it would appear that SA in its present form is not ready to become first-line therapy, given the fact that CA is easier and readily available, just as efficacious, safe, less invasive, and cheaper. However, with novel ablation tools, stronger energy sources, and enhanced imaging of individual AF pathophysiology, it seems that this is a game of rock paper scissors, where many different combinations of tools and targets will have to be tested in sufficiently large trials to come up with a strategy that always wins. Conflict of interest: L.V.A.B. is a consultant to Medtronic and Boston Scientific; fees are paid to the Cardiology Department The opinions expressed in this article are not necessarily those of the Editors of the European Heart Journal or of the European Society of Cardiology.