Looking beyond PVI for persistent AF termination

Michel Haissguerre, MD

Michel Haissguerre, MD

When the arrhythmia is persistent, PVs are not the only dominant drivers. Therefore, PV antral ablation is efficient in 20% to 50% of persistent AF patients.

Presenting recent work on the elimination of driver domains in persistent AF this morning is Michel Haissaguerre (Hopital Cardiologique du Haut Leveque, Pessac Cedex, France), as part of a broader session on the latest strategies being adopted in persistent cases that reach beyond PVI alone.

Professor Haissaguerre and colleagues applied noninvasive signal processing to identify driver domains in 103 patients, demonstrating that driver sites can be mapped accurately, leading to shorter RF times than conventional ablation techniques. 1 He discussed the work in an interview with AF Symposium News.

“Persistent AF—Is PVI sufficient”? How would you answer this question?

PVI remains the backbone of AF ablation. One its own, it suffices as the treatment of the majority of paroxysmal AF cases. When the arrhythmia is persistent, PVs are not the only dominant drivers. Therefore, PV antral ablation is efficient in 20% to 50% of persistent AF patients (as published in the literature). In other words, additional ablation is necessary in the majority of patients with persistent AF in order to achieve high acute AF termination and favorable clinical outcome. The results of the recently published Star AF II study are not contradictory, since they only indicate that nonselective atrial ablation does not add to the efficacy of PVI, which is not the same strategy as selective driver ablation. 2

In the AFACART study, the Munich group has made an interesting observation with regard to the overlap between noninvasively identified driver areas and endocardial CFAE areas: the mean atrial surface covered by CFAEs was 49% of the total atrial surface, i.e. nearly half of atrial tissue; while the one occupied by the driver regions was 19%, indicating a better specificity of targets using a focal/re-entrant driver mapping strategy. 3

Could you describe the sorts of patterns of driver activity you have seen in your work? How variable are these patterns from one patient to the next?

The extent of sources gets wider with the increasing duration of AF, such that the sources are spread over wider atrial substrates in long lasting persistent or so-called ‘permanent’ AF. Two types of driver activity were observed while mapping persistent AF noninvasively using phase mapping1: (1) focal—originated from a point or an area (for example, a PV) and propagated centrifugally; (2) re-entry, a tern that I prefer to the more sexy ‘rotor,’ as we are unable to demonstrate a wave fully rotating around a functional center point. All re-entrant driver activities that we mapped were intermittently active and spatially unstable but recurred periodically in the same regions.

The regional distribution of drivers provided further evidence supporting the notion of targets outside the PV. I would like to emphasize that driver regions are multiple; we observed a single region responsible for persistent AF in only 9% of patients. In addition to PV antra, the drivers were found located in the adjacent septum, any part around the left appendage and inferior left atrium (including the coronary sinus) with wide inter-individual variations at non-PV locations including the right atrium. These regions were already known as critical sites responsible for AF perpetuation, so this is not a surprise. But the combination of all specific parts of these regions harboring drivers generates an individualized driver-map which allows delivery of personalized therapy.

Hence, do you think that we could avoid having to treat persistent cases by more accurate mapping of AF in the first instance, or by simply performing ablation earlier on? I.e. avoiding or minimizing the tissue remodeling that leads to an increasing extent of the biatrial surface that is occupied by reentrant and focal drivers?

These questions are very pertinent. Tissue remodeling is fundamental to the AF disease process. The electrical component of remodeling reverses completely within weeks after restoration of sinus rhythm. Conversely, structural remodeling—the outcome of complex alteration in cellular and tissue composition—results in fibrosis, which reverses slowly, if at all.

Avoiding or minimizing tissue remodeling should probably prevent the increase in the extent of biatrial surface that is occupied by reentrant and focal AF drivers.

At this stage, the only clinical evidence that we have is that persistent AF in early months or in patients presenting in sinus rhythm (where AF is induced by pacing) displays a more limited driver domain extent and therefore we obtain higher success in terms of AF termination with limited amount of (targeted) ablation than in longer-duration persistent AF. A median of two regions has to be targeted in early persistent AF and in patients presenting in sinus rhythm, resulting in mean RF delivery for a lower rate of AF termination.

Therefore, avoiding or minimizing tissue remodeling should probably prevent the increase in the extent of biatrial surface that is occupied by reentrant and focal AF drivers even though, at present, we lack a specific study testing this hypothesis.

What can you say about the patient population in whom AF could not be terminated after driver and linear ablation in the recent study published in Circulation? Are there any clues as to why their AF could not be terminated?

The factors predictive of nontermination are similar to those published previously: the termination rate decreases with the longer duration of continuous AF and shorter AF cycle length. In the current multicenter AFACART Study involving eight European centers, a high rate of acute AF termination (65%) by driver-only ablation has been reported despite the fact that none of the centers had any practical experience with the noninvasive mapping system before the study. Sebastian Knecht (Brugmann University Hospital, Brussels, Belgium), PI of this study, reported that a wide diffusion of driver regions correlates with nontermination, which is logical. It is likely that long-lasting persistent AF and short AF cycle length (AFCL) indicate a higher degree of atrial remodeling and therefore an involvement of multiple wavelets and/or sources, which are difficult to ablate owing to their dissemination.

How can we reduce the impact of these factors? In practice, when we see a patient with persistent AF lasting continuously for, say, six months, we will propose cardioversion to restore sinus rhythm if we cannot give him or her an early date for the ablation procedure. In patients with persistent AF and short AFCL, (less than 140 ms) as determined from the rate of “f” waves in V1, a pre-ablation treatment by amiodarone or other anti-arrhythmic drug will help prolong AFCL, which may facilitate the achievement of procedural end point of AF termination. WE know that the latter is a not consensual end point, but to our knowledge there are 17 studies in the literature, with 14 of them demonstrating better clinical outcome when acute AF termination is achieved than otherwise.

One of the advantages of mapping is the shorter procedure time. What can you say about the outcomes when compared to the control group?

Besides shorter procedure time as AF mapping is performed before the procedure, RF duration to achieve AF termination is also shorter in the driver group versus controls. The control group was actually treated conventionally with PVI followed by CFAE ablation and optionally linear lesions.

This clinical outcome (freedom from AF) at the end of one year was similar between the two groups (85% vs. 87%) but with a much lower RF delivery when AF drivers were mapped and targeted.



  1. Haissaguerre M et al. Driver domains in persistent atrial fibrillation. Circulation. 2014 Aug 12;130(7):530-8.
  2. Substrate and Trigger Ablation for Reduction of Atrial Fibrillation Trial—Star AF II Study. https://clinicaltrials.gov/ct2/show/NCT01203748 (retrieved January 2015).
  3. Noninvasive mapping before ablation for atrial fibrillation: the AFACART study. https://clinicaltrials.gov/ct2/show/NCT02113761 (retrieved January 2015).
Tammy Griffin-Kumpey