Author + information
- Published online August 30, 2017.
- Hakan Oral, MD∗ ( and )
- Konstantinos C. Siontis, MD
- Cardiac Arrhythmia Service, Division of Cardiovascular Medicine, University of Michigan, Ann Arbor, Michigan
- ↵∗Address for correspondence:
Dr. Hakan Oral, Cardiac Arrhythmia Service, Division of Cardiovascular Medicine, University of Michigan, CVC, SPC 5853, 1500 East Medical Center Drive, Ann Arbor, Michigan 48109-5853.
First reported after surgical ablation, atrioesophageal fistula (AEF) has been recognized as an infrequent complication of catheter ablation (CA) that is associated with severe morbidity and a high fatality rate, ∼80% (1). Although infrequent, AEF remains a significant concern considering the increasing number of ablation procedures performed to eliminate atrial fibrillation (AF) worldwide.
The pulmonary venous antrum and the posterior left atrium (LA) are the usual targets during CA of AF. The variable spatial relationship between the esophagus and posterior LA is well characterized (2). The distance between the esophageal wall and posterior LA endocardium is often <5 mm, well within reach of any ablation energy source. A thin layer of adipose tissue between the atrial epicardium and esophagus has been proposed to be protective by insulating the esophageal wall. However, the adipose layer often has a patchy and unpredictable distribution, and there are no data to suggest that CA over an adipose layer is safer.
The very structure of the esophagus may also promote development of AEF. The esophagus does not have a serosa. The muscular esophageal wall is only a few millimeters thick. During CA, the vascular supply of the esophagus can be compromised, potentially leading to ischemic necrosis. Gastrointestinal secretions can impede tissue healing and recovery and can facilitate progression to AEF. Shunting of radiofrequency (RF) energy from unipolar to the indifferent electrodes through the esophageal wall has also been considered as a possible mechanism. However, AEF can develop as a result of direct thermal injury such as after cryo, laser, or ultrasound ablation.
Several approaches have been proposed to minimize the risk of AEF:
1. Localization and avoidance of the esophagus: The position of the esophagus can be determined fluoroscopically by barium swallow, which enables complete visualization of the endoluminal borders of the esophagus. Esophageal movement can also be monitored during the course of CA. The esophagus can be localized by insertion of catheters that can be tracked in 3 dimensions. However, the full width of the esophagus may not be appreciated. Intracardiac echocardiography has been used to visualize the esophagus. Once the location of the esophagus is determined, then ablation in proximity to the esophagus can either be attenuated or avoided. The downside of this approach is the risk of creating ineffective lesions and a higher risk of AF recurrence.
2. Monitoring luminal esophageal temperature (LET): It can be difficult to confirm the proximity of the sensor to the ablation site, and LET may not accurately reflect intramural temperatures. Calibration of the spatial and thermal resolution of the system to detect esophageal injury when it still is reversible can be problematic. A sensor that is too sensitive may not allow adequate delivery of energy. Multielectrode probes with 3-dimensional designs may increase the footprint of the esophagus along the posterior LA wall. Thermal sensors may provide a false sense of safety, and AEF has been reported even when there was <1°C of rise in LET (3).
3. Modification of the ablation technique and energy settings: Although short applications of RF energy using a brushing technique are often used, the upper limit of a safe application is unclear. Alternative lesion sets to avoid the esophagus, such as encircling around the posterior LA has been proposed. However, ablation will often have to be performed along a segment next to the esophagus.
4. Pharmacotherapy for mucosal protection: Reduction of acid content in gastrointestinal secretions using proton-pump inhibitors or sucralfate for mucosal coating during the periprocedural period has been used as an adjunctive prophylactic approach. There are no randomized studies establishing the efficacy of these approaches.
5. Esophageal deviation: Traction of the esophagus using an endoscope or a custom probe, separation of the esophagus by inserting an epicardial balloon percutaneously, or by insulation with saline injection have been proposed.
However, the efficacy and safety of these approaches have not been assessed in large series.
In this issue of JACC: Clinical Electrophysiology, Palaniswamy et al. (4) report their experience in 114 consecutive patients who underwent CA to eliminate AF. The investigators used an off-the-shelf malleable metal stylet within a plastic tube to deviate the esophagus away from the posterior LA wall to allow uninterrupted RF delivery. All procedures were performed under general anesthesia and mechanical esophageal deviation (MED) was performed after transseptal puncture but before the electroanatomical map was created. Concomitantly, barium was inserted into the esophagus to delineate the trailing edge of the esophagus, which was substantial in all cases and often remained at the same or close to the original position. LET was monitored continuously and guided energy delivery in the posterior wall. Placement of the MED device was graded as easy by the operators. It should be noted that all operators had prior experience with this technique. Yet, the actual esophageal deviation portion of the technique was occasionally graded as difficult.
All patients underwent pulmonary vein isolation with contact force-sensing, open-tip irrigated RF catheters. Additional linear ablation was performed in 5.3% of the patients. The effective MED (the distance between the ablation lesion set and the trailing edge of the esophagus) varied significantly among patients and was inversely associated with an increase in LET. The effective MED was 0 to 10 mm in ∼25% of the patients, and up to 75% of these patients experienced LET rise. In contrast, increase in LET was rare (4.1%) when effective MED was >20 mm, which could be achieved in only 20% of the patients.
The investigators should be commended for carefully conducting a clinical study to determine the safety and utility of this approach, which is rather simple and inexpensive. Using barium, they appropriately monitored the location of the endoluminal borders of the esophagus and demonstrated a high prevalence of a trailing edge of the esophagus close to the original position, in the ablation target zone. This critical observation underscores that the esophageal borders and full width of the esophagus must be carefully monitored during CA of AF.
Mechanical trauma to the oropharynx occurred in 3 patients with the stiffer stylet, which was then abandoned. However, the softer stylet was less effective in achieving adequate MED. Endoscopy was not routinely performed in this study. In an earlier, proof-of-concept study of this technique (5), endoscopy demonstrated esophageal trauma in 63% of the patients. Although no immediate clinical sequelae were reported, long-term effects need to be addressed. The MED device itself may displace the esophagus closer to the posterior LA. Unless the trailing edge of the esophagus is identified precisely, suboptimal MED may lead to a false sense of safety and may further predispose to AEF. The risks of stylet malposition and injury may be higher in patients with previously unrecognized anatomical variations of the esophagus, including hernias and diverticulae, and in patients with thoracic anatomical variations such as pectus excavatum, or prior thoracic surgery. Routine imaging of the esophagus and thoracic structures prior to the procedure may have to be performed. This approach may not be feasible in patients who undergo CA without general anesthesia. Whether MED, without leaving a trailing esophageal wall behind, can be developed, and truly reduce the risk of AEF without an increase in more common esophageal complications, remains to be determined in large, sufficiently powered trials.
Despite substantial technological advances in navigation, mapping, catheter, and energy source technology over the last 2 decades, there has been little progress to prevent AEF. Each of the approaches discussed has considerable limitations in efficacy and clinical utility. Some of the novel approaches may include titrating the lesion depth by adjusting bipolar to unipolar RF energy delivery, assessing the extent of lesion formation in real-time in vivo, and active cooling or heating of the esophagus. However, these will require significant efforts for development and clinical validation.
Given the very low frequency of AEF, the absence of experimental and animal models, the associated cost, and the risk of trading AEF with a different set of more prevalent complications, development of an ideal approach remains as a challenge. However, mortality and severe morbidity are not acceptable complications of CA. Therefore the heat to accelerate efforts to develop safe, effective, simple, and low-cost approaches to mitigate the risk of AEF goes on.
↵∗ Editorials published in JACC: Clinical Electrophysiology reflect the views of the authors and do not necessarily represent the views of JACC: Clinical Electrophysiology or the American College of Cardiology.
Both authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- 2017 American College of Cardiology Foundation
- Barbhaiya C.R.,
- Kumar S.,
- Guo Y.,
- et al.
- Lemola K.,
- Sneider M.,
- Desjardins B.,
- et al.
- Palaniswamy C.,
- Koruth J.S.,
- Mittnacht A.J.,
- et al.