Endoscopic versus surgical treatment for the management of oesophageal fistula after atrial fibrillation ablation – a subanalysis of the worldwide POTTER-AF study

Background: Oesophageal fistula (OF) is considered the most dreadful complication of atrial fibrillation (AF) catheter ablation. The endoscopic and surgical treatment of OF has been proposed to improve the prognosis of these patients, but the data on the individual impact of each treatment technique are scarce. 

Methods: The POTTER-AF study is a worldwide, multi-centre registry conducted under the auspices of the Working Group of Cardiac Electrophysiology of the German Cardiac Society (AGEP, DGK) which evaluated the incidence, management, and outcome of post-procedural OF following catheter ablation of AF.  A total of 553 729 patients underwent ablation procedures for AF or atrial tachycardia (AT) in 214 electrophysiological centres from 35 countries between 1996 and 2022. Of them, 138 (0.025%) patients experienced postprocedural OF and data regarding the therapeutical approach were available in 116 patients. Of them, 78 underwent surgical or interventional treatment. A detailed description of the two techniques was not available. Herein we conducted a therapy- based analysis of the patients exhibiting OF. 

Results: A total of 23 (29.5%) patients were treated using endoscopic interventions only (Endoscopy group), while 55 (70.5%) patients underwent surgical procedures (Surgery group). Of the latter, 8 (14.5%) underwent additional endoscopic treatment.  There was no difference between the Endoscopy and Surgery groups regarding the demographic characteristics (13 (59.1%) vs. 21 (38.2%) females, p=0.129; median age 66.5 (60.0; 71.5) vs. 62.0 (52.0; 71.75) years, p=0.170; median BMI 27.5 (23.1; 31.3) vs. 26.4 (24.0; 28.7) kg/m2, p=0.664). The two populations were similar in terms of comorbidities. The median left ventricular ejection fraction was 55.0 % (46.5; 62.25) in the Endoscopy group and 60.0 % (53.0; 65.0) % in the Surgery group (p=0.279). In the Endoscopy group 22 (95.7%) patients underwent radiofrequency-based ablations, no one cryoballoon-based ablation and 1 (4.3%) laser-based procedures, compared to 52 (94.5%; p=1), 3 (5.5%; p=0.551) and no patient (p=0.295) respectively in the Surgery group. There was a trend towards a higher use of an oesophageal thermal probe among the patients undergoing exclusively endoscopic treatment (43.5% vs. 21.8%; p=0.096). The duration until symptoms onset was 10 (5.5; 18.5) days in the Endoscopy group and 18.0 (10.0; 22.0) days in the Surgery group (p=0.081), while the time to OF diagnosis was 18.0 (10.0; 26.0) days and 20.0 (15.0; 29.25) days respectively (p=0.100). The symptoms at presentation were similar between the two groups.  Endoscopy was the diagnostic method in 45.5% of the patients in the Endoscopy group and in 16.7% in the Surgery group (p=0.017). Similar use of echocardiography, computed tomography and magnetic resonance tomography was reported between groups. The subsequent complications of the OF were also similar between groups. The mortality was 56.5% in the Endoscopy group and 51.9% in the Surgery group (p=0.805), while minor sequelae occurred in 8.7% patients in the former group and 16.7% in the latter (p=0.490) and major sequelae in 4.3% and 16.7% respectively (p=0.266). 

Conclusion: There was no clinical or paraclinical parameter associated with the choice of the therapeutical approach. Patients diagnosed by means of endoscopy were more likely to undergo exclusively endoscopic interventions. The mortality, incidence of minor and major sequelae were similar between the groups.