Ablation of Atrial Fibrillation: How Can Less Be More?

Introduction

AF ablation remains disappointing – despite spectacular results in individual patients –for both persistent and paroxysmal AF using contemporary catheters¹. The benefits of pulmonary vein isolation (PVI) likely extend beyond isolating triggers – since ablation may be more successful if wide areas are ablated by radiofrequency or cryoballoons², and patients without AF may have reconnected PVs³. Hence it seems intuitive that additional lesion sets should improve outcomes.

However, several meta-analyses have suggested that commonly applied ablation sets beyond PVI provide inconsistent benefit⁴. This was illustrated vividly by the STAR-AF2 trial, in which the 59% success of PVI was not improved by adding lines or CFAE⁵. However, a major criticism of STAR-AF2 is that lines were not blocked in 26% of patients.

Into this area comes the BOCA (No Benefit OF Complex Fractionated Atrial Electrogram (CFAE) Ablation in Addition to Circumferential Pulmonary Vein Ablation and Linear Ablation) trial of ablation strategies for persistent and longstanding persistent AF in this issue of Circulation: Arrhythmia and Electrophysiology⁶. BOCA addresses 2 unanswered questions – what is ablation success when lines are proven to be complete, and what is the value of supplementary widespread ablation (CFAE) if lines are blocked?

In BOCA, the investigators prospectively randomized 131 persistent AF patients in 1:1 fashion to PVI + Lines (roof and mitral, “control”) or PVI+ lines + CFAE (“CFAE”) in a single center. Conduction block was confirmed in 95% of mitral lines and 100% of roof lines and PVI. Notably, patients in the CFAE arm had no advantage over the control arm, with similar single (46.2% vs. 56.9%, p=0.20) and multi-procedure (78% vs. 80%, p=1.0) freedom from atrial arrhythmias at 12 months. Patients in the CFAE limb experienced more atrial flutters, and had longer procedural and ablation times.

The authors should be commended on designing and executing this important trial. Their success compares favorably with STAR-AF2, particularly given that many patients had LV systolic dysfunction, elevated BMI and long standing persistent AF. Limitations include the fact that follow-up using ambulatory ECGs only at 12 months could easily miss asymptomatic recurrences, that block across PVI and/or lines was not verified with adenosine and that contact force sensing was not used. However, these factors would influence both limbs similarly. Taken in context, results from the BOCA⁶ and STAR-AF2⁵ trials suggest that PVI produces at least equivalent success with less ablation than PVI+lines, PVI+CFAE or PVI+lines+CFAE in persistent AF. The overwhelming message that `less is more' for AF ablation gives us pause to reflect upon which AF mechanisms are ablated by current lesion sets.

Possible Mechanisms By Which Lines/CFAE May Not Improve PVI

If AF is due to disordered wavelets that self-sustain, as hypothesized by Moe from computer models⁷, then lines and additional lesions should dramatically reduce the tissue available for these wavelets and should thus improve outcomes.

The figure shows a conceptual schematic of AF driven by 2 fundamental mechanisms: (a) Preferred regional mechanisms (sources) driving disorganized waves or (b) disorganized wavelets that self-sustain, i.e. without preferred regional sources, such as the multiple-wavelet⁷ or endo-epi dissociation⁸ hypotheses. Ablation lesions are shown in gray shading, and may operate by hitting sources (figure 1A) or by constraining disordered wavelets (figure 1B) as shown in the table.

Figure.

Organized and disorganized mechanisms of AF. The failure of extensive ablation to improve AF outcomes in BOCA and STAR-AF2 is better explained by (A) preferred regional sources that can be missed by ablation, than (B) meandering multiple wavelets that would be constrained by greater ablation.

It is in fact questioned whether self-sustaining disorder (figure 1B) is a key mechanism for AF. Reports of AF caused simply by “complexity”⁸ often ignore data that activation shows consistent patterns in AF⁹ or that localized ablation can terminate persistent AF¹⁰, and typically examined <10% of atrial area without mechanistic interventions⁸ Since successful therapy for self-sustaining disordered AF must compartmentalize the atrium, its strongest support has come from the need for extensive ablation¹¹. One could thus argue that lines applied in BOCA and STAR-AF2 were insufficient to limit atrial mass, yet this does not explain the success of limited ablation¹⁰, or how in the Bordeaux stepwise approach¹¹ the first steps were often successful yet patients with the longest ablation times often had the worst acute and long-term outcomes – all of which argue against this mechanism.

Conversely, mounting data supports the concept that disorganized activity in AF is maintained by spatially preferred mechanisms (sources) which, in the majority of optical mapping studies, take the form of localized rotors that produce disorder. In recent human optical mapping studies, AF was sustained by stable endocardial micro-reentrant sources anchored to muscle bundles that produced unstable breakthroughs and rotations on the epicardium abolished by endocardial ablation¹². These optical data support results from wide-area contact basket mapping¹⁰, and may reconcile the instability of sources on ECG imaging of the epicardium¹³.

If AF Is Due To Localized Sources, Shouldn't Lines Hit Some Of Them?

In the CONFIRM trial¹⁰, stereotypical lines sometimes hit sources, but also often missed sources that lay in patient-specific locations with small areas of 1–2 cm² on endocardial optical¹² and basket maps¹⁰ (figure 1A). Since studies increasingly show that persistent AF is a bi-atrial disease, with a third of sources occur in the right atrium on endocardial or epicardial mapping^10,13, ablation lines in CONFIRM intersected only 40–60% of sources, and such patients had substantially higher success than those in whom sources were missed¹⁴.

Shouldn't CFAE Ablation Hit Some Sources, and Are Widespread Lesions Pro-Arrhythmic?

CFAE are increasingly felt to be mechanistically non-specific, because electrograms in AF may poorly reflect local activation¹⁵ and because CFAE-algorithms vary¹⁶. CFAE ablation is often patchy and scattered, and while its impact on clinical rotors is unclear, in animal models such lesions destabilize rotors¹⁷ possibly by creating additional attractors. This adds to the concern that widespread ablation is pro-arrhythmic, also shown by the high incidence of atrial tachycardia after CFAE ablation in BOCA and STAR-AF2.

The Path Forwards: Patient-Specific Mapping of AF

Given ample evidence that “less is more” in AF ablation, is there a way to improve the current 50–60% success? A nihilistic view – that outcomes cannot be improved – is at odds with abundant anecdotes of success when ablating AF outside the PVs. To solidify these anecdotes into “personalized ablation”, a promising approach is to target ablation to mapped AF mechanisms in each patient. Wide area basket mapping shows that AF is often maintained by stable rotors and focal sources¹⁰, with similarities to recent human optical mapping studies¹², whose ablation is promising in multicenter non-randomized studies¹⁸. While AF sources have been questioned, those studies had issues such as atrial cycle lengths of 250–500ms (dominant frequencies 2–4 Hz) in many patients ostensibly with AF and analysis of unipolar signals using indices designed for bipolar signals (Shannon's entropy)¹⁹, or mapped very small areas⁸. Recent data showing how localized ablation can terminate AF rotors, by interacting with non-uniformities in remodeled atria, explain and thus strengthen the results of AF source ablation²⁰. At the other end of the spectrum, it has been proposed that debulking should be further extended – such as by posterior LA wall ablation. Limitations of this approach include the fact that recurrences still occur, that targeting right atrial sources will require essentially bi-atrial obliteration, and a risk of sequelae including but not limited to stiff left atrial syndrome.

Conclusions

The BOCA trial shows that non-targeted debulking of atrial tissue, based on early assumptions that AF is a random process, has repeatedly failed to advance patient outcomes. A promising emerging direction, based on mounting evidence that human AF is sustained by localized sources in both atria, is to personalize ablation based on map guided mechanisms. It is time to translate decades of basic and translational science to the bedside to improve AF ablation – our field requires it and our patients demand it.