Non-invasive identification of drivers in persistent AF

Michel Haïssaguerre

In paroxysmal atrial fibrillation, pulmonary vein isolation is successful while the results in persistent atrial fibrillation are less satisfactory because of ill-defined atrial substrate including multiple atrial wavelets, macro-re-entries and sources. We have recently published a study in Circulation, which has demonstrated that persistent atrial fibrillation particularly in early months is maintained by drivers—most being unstable re-entries—clustered in few regions allowing high rate of atrial fibrillation termination from limited catheter ablation, writes Michel Haïssaguerre, lead author of the study.

For the optimal strategy in persistent atrial fibrillation (localised target vs. global intervention), the key question is whether the multitude of activation waves that characterise individual atrial fibrillation emanates from few periodic drivers or whether the waves are transitory, widely distributed and self-perpetuating. Published guidelines mention a variety of techniques notably targeting complex fractionated electrograms and linear lesions in addition to pulmonary vein isolation.

The ‘rotor’ saga

Recent work has thrust localised sources (focal or re-entrant= “rotor”) back to centre stage. Based on Jalife’s optical mapping work demonstrating rotor sources and prior reports of arrhythmogenic foci, attempts to identify these localised sources have been reported clinically using multielectrode catheter or epicardial surgical arrays. Wider mapping field of atrial fibrillation is currently possible from 64 electrode basket electrograms (RhythmView, Topera) or 252 electrode body surface electrograms (ECVUE, CardioInsight). These technologies use a mathematical algorithm to reconstruct phase electrograms offering new insights into the mechanism of human atrial fibrillation.

Recently, we (Haïssaguerre et al, Circulation 2014 130(7):530–538) evaluated a completely non-invasive body surface mapping to identify drivers in 103 consecutive patients with distinct categories of persistent atrial fibrillation; 63% had structural heart disease. Accurate biatrial geometry relative to the electrodes was obtained from non-contrast CT-scan. The multiple atrial fibrillation sequences acquired bedside allowed to construct a ‘statistical’ density-map of drivers that draws the roadmap for ablation.

Results of panoramic mapping

The maps showed incessantly changing beat-to-beat wavefronts and varying spatio-temporal behaviour of driver activities. Re-entries were not sustained and meandered substantially but recurred repetitively in the same region. Interestingly, most focal discharges emanated from same regions as reentries. The anatomical distribution of drivers confirmed the importance of regions of pulmonary vein antra, adjacent septum and left appendage with wide inter-individual variations at other locations including the right atrium. Although the drivers regions harbour endocardially rapid and fractionated activity, such CFAE electrograms were also present at non-driver regions thus confirming their low specificity; driver regions defined by non-invasive mapping cover about 20% of atrial surface versus 50% for CFAE areas (Ongoing study by Ammar et al, from Munich Herz centrum, Germany, participating in the AFACART study). 

Importantly, the number of driver regions increased with the duration of continuous atrial fibrillation resulting in better ablation outcomes in early persistent than longer lasting atrial fibrillation. Results were particularly favourable in those patients presenting in sinus rhythm (as they had been cardioverted previously), probably because of lesser structural remodelling.

Driver ablation alone terminated 75% of persistent atrial fibrillation using a mean radiofrequency delivery of 28±17min vs. 65±33min in the stepwise ablation control group. Despite this limited ablation, we observed a significant incidence of atrial tachycardia after atrial fibrillation termination (requiring further ablation) probably reflecting the presence of underlying abnormal atrial substrate in persistent atrial fibrillation. At 12 months, 85% patients with atrial fibrillation were free from atrial fibrillation.

The ablation of persistent atrial fibrillation under the driver-mapping guidance is clearly advantageous over conventional ablation strategies. These results are promising especially as a prospective European multicentre study (AFACART study – Knecht et al) shows similar and reproducible results in eight European participating centres. In addition, the new possibility of periprocedural remapping during ablation shows dynamic changes as extinction or emergence of drivers which will allow furthering improving the clinical results.

Michel Haïssaguerre is with Hôpital Cardiologique du Haut Lévêque, Bordeaux, France and Université de Bordeaux, IHU LIRYC L’Institut de Rythmologie et modélisation Cardiaque, France. He is stockowner of CardioInsight