Radiofrequency ablation or cryoablation has become the first-line therapy for many patients with atrial fibrillation (AF). While the vast majority of patients undergo these procedures with no adverse events, the proximity of the esophagus to the heart can be problematic when ablating endocardially. In 40% of the population, the distance between the esophagus and the left atrial posterior wall is less than 5 mm.1 Collateral injury to the esophagus can result in a spectrum of disorders: esophageal dysmotility, gastroparesis, esophageal ulcers, and atrioesophageal fistulas. Fortunately, the occurrence of fistulas is rare. The incidence of atrioesophageal fistula ranges from 1 in 500 to 1 in 10,000 with radiofrequency ablation,2 and <1 in 10,000 when cryoballoon is used.3
The esophagus is a muscular tube, approximately 25 cm long and 2.0 cm wide, which travels posterior to the left atrium and trachea. (Figure 1) The cervical part of the esophagus is connected to the trachea, which lies anterior to it, by fibroelastic membranes and sparse muscle fibers. In the thorax, the esophagus usually does not have connective tissue anchoring the esophagus except in situations where an inflammatory process has occurred. For example, adhesions may exist between the esophagus and pleura or lymph nodes in a patient with a history of tuberculosis or radiation therapy. The thoracic esophagus is supplied by the bronchial and esophageal branches of the thoracic aorta, and drains into a submucosal plexus. The esophagus is innervated by the left and right vagus nerves and recurrent laryngeal branches of the vagus nerve. The left vagus courses anteriorly, while the right vagus courses posteriorly along the esophagus. The right and left vagal trunks form an anterior plexus 34% of the time, posterior plexus 19% of the time, and both anterior and posterior plexi 44% of the time.4 The vagal nerve plexi synapse in the esophagus wall in the myenteric plexus (Auerbach’s plexus) and submucous plexus (Meissner’s plexus). Below the level of the carina, the esophagus is relatively free and mobile as the vascular supply and nervous plexus supplying the esophagus is elastic.
In relation to AF ablation, the course of the esophagus is more frequently towards the left pulmonary veins than it is midline or rightward. The esophagus is most frequently closest to the left inferior pulmonary vein ostium versus the other ostia. The average distance between the left inferior pulmonary vein and the esophagus was 7.2 mm compared to 27 mm for the right inferior pulmonary vein from CT scans.5 Rolf et al found that the esophagus was midline in 23.4% of patients, leftward in 54.2%, and rightward in 22.4%.6 A leftward esophagus may course near the coronary sinus, where ablations performed in the coronary sinus for a mitral isthmus line should be performed with caution. A very rightward esophagus may even course near the posterior right atrium.7
The thickness of the left atrial posterior wall is approximately 1.9-2.4 mm, while the thickness of the anterior esophageal wall is approximately 2.7-3.6 mm.6 (Figure 2) The average distance from the left atrial endocardium to the esophageal lumen is approximately 4.4-4.7 mm in CT studies; however, it is even less (2.5-4.5 mm) in real-time fluoroscopy probably due to tissue compression from the catheter tip. Given the proximity of the esophagus to common areas of ablation during wide antral pulmonary vein isolation, particular attention is required to avoid esophageal injury.
There are two proposed mechanisms for esophageal injury following AF ablation: thermal injury and ischemic injury. Direct thermal energy can cause transmural injury to the esophagus. Esophageal ulcerations can occur in 6% to 26% of patients undergoing AF ablation, even with esophageal temperature monitoring.8 Thermal injury can induce an inflammatory reaction and granulation of the esophagus which can lead to perforation. Injury to the esophageal vasculature likely compounds the direct thermal injury and creates a progressive ischemic lesion. There appears to be a bimodal temporal distribution of injury. Early direct thermal injury, causing gastroparesis or esophageal ulceration within the first 5 days, usually resolves without permanent sequela. Late ischemic injury, presenting after 5 days, appears to have a higher incidence of serious injury such as atrioesophageal fistula formation.9 While 12 hours to 3 days is sufficient to develop gastrointestinal dysmotility, it can take up to 2-6 weeks post procedure to develop an atrioesophageal fistula. Inflammation and acid reflux on an ischemic esophagus is theorized to transform esophageal ulcerations into fistulas. Unfortunately, atrioesophageal fistulas are usually fatal. The high rate of mortality is likely due to a combination of clinician unawareness, delayed presentation, and complex surgical approach required for treatment.
In addition to direct injury to the esophageal wall, functional disorders may arise due to collateral injury to the vagus nerve and its components. The esophageal vagal plexus innervates the esophageal sphincters as well as the myenteric plexus of the esophagus to regulate esophageal peristalsis. The left vagus nerve continues inferiorly to innervate the stomach. Although not life-threatening, upper gastrointestinal disorders such as esophageal dysmotility and delayed gastric emptying are fairly common among patients undergoing AF ablation via vagus nerve injury. In a small study of 27 patients, Lakkireddy et al found new-onset esophageal dysmotility in 48% of patients and delayed gastric emptying in 48% of patients.10 While a third of patients had persistent symptoms at 3 months, symptoms generally resolved by 6 months.
Patients with atrioesophageal fistulas can present with a spectrum of signs and symptoms. Patients may initially present with gastric hypomotility, esophagitis, or acid reflux, which may progress over days to weeks. Fever, chest pain, and dysphagia 2-6 weeks after an AF ablation are concerning findings that would warrant immediate investigation. Pleural effusion, pericardial effusion, altered mental status, sepsis, hematemesis, and symptoms of a cerebrovascular accident are serious manifestations of an impending poor outcome.
Expedient diagnosis is of critical importance. Neuroimaging with CT or MRI may reveal septic, air emboli, or pneumocephalus. The preferred imaging modality is CT angiography of the chest with IV contrast (without PO contrast), which may show a connection between the left atrium and esophagus and extravasation of contrast from the left atrium into the esophagus via a fistulous tract or into the mediastinum. If there is no active fistulous connection, contrast may not be seen in the left atrium. A barium swallow x-ray or CT scan with PO contrast would aid diagnosis of an esophageal rupture. Air may also be found in the mediastinum between the esophagus and the left atrium, as well as within the pericardium.
Finally, an esophago-gastro-duodenoscopy (EGD) and endoscopic ultrasound (EUS) may be considered to assess for esophageal injury as well as diagnosing injury to adjacent mediastinal structures. EGD is the gold standard for evaluating upper gastrointestinal bleeds; however, it should be cautiously used in evaluating suspected cases of atrioesophageal fistulas given the possibility of massive air embolism. As an alternative, capsule endoscopy may be performed to evaluate for luminal evidence of esophageal injury.
Practices and Preventive Measures
Anesthesia Versus Conscious Sedation
Di Biase et al report that the use of general anesthesia was associated with more esophageal injury than conscious sedation in a study of 88 patients.11 The authors attributed the risk of injury in the general anesthesia group to the lack of swallowing, fixation of the esophagus by the nasogastric tube, and reduced esophageal peristaltic movement. Active peristalsis and swallowing may result in esophageal cooling and inconsistent transfer of heat to the esophagus. Additionally, an awake patient may be able to signal to the physician when there is pain and injury to the esophagus while it is occurring, thereby aiding in reducing ablation time and collateral damage. While conscious sedation may theoretically have the above advantages, monitoring of esophageal temperatures is more difficult without general anesthesia. Additionally, data from Martinek et al did not find any significant difference in the incidence of esophageal injury with the general anesthesia arm versus the conscious sedation arm (2.7% vs 2.2%, respectively; P=0.86).12
Luminal Esophageal Temperature Monitoring
Monitoring of luminal esophageal temperatures (LET) can alert the operator when esophagus heating and injury is occurring. For RF ablation, what temperatures are too high? Luminal esophageal temperature was recorded in one study to be >39°C in almost 40% of the patients, 10% of which developed esophageal injury.13 Using an ablation strategy of interrupting ablation when LET surpassed 39°C, Di Biase et al found a 17% incidence of esophageal injury with capsule endoscopy.11 Similarly, asymptomatic esophageal lesions were found in 15% of patients in a study by Halm et al with LET less than 41°C.14 These authors also reported that the odds of mucosal injury increased with every 1°C rise in LET over 41°C. Using a lower cutoff of 38.5°C to interrupt ablation, Sing et al reported a lower incidence of esophageal injury versus no temperature monitoring (6% vs 36%; P<0.006).15 Operators may choose to utilize absolute temperature cutoffs such as 38.5°C versus relative rises in temperatures. For patients with lower baseline esophageal temperatures, one may choose to interrupt energy if LET increases by 2°C, or more conservatively, 1°C.
In addition to peak temperature rise, the rate of temperature rise may be a predictor of transmural esophageal heating and injury. Rates of >0.05-0.1°C rise per second were associated with higher risk. Repeated radiofrequency energy applications can also prevent the esophagus from cooling, thus leading to temperature stacking and higher risk of mural heating.16 If heating is noted, an operator can focus ablation on the contralateral set of veins to allow for adequate time for cooling. The esophagus should return to baseline temperatures within 3 minutes.
Unfortunately, LET may be unreliable if the temperature probe is not in proximity to the region of ablation. Single-sensor probes are frequently employed for temperature monitoring; however, they require constant repositioning to ensure that the temperatures are measured near the ablation catheter. Multisensor probes such as the CIRCA S-CATH™ Esophageal Temperature Monitoring System (CIRCA Scientific) can be utilized to monitor a wide territory of the esophagus without repositioning. (Figure 3) The CIRCA S-CATH is a soft, expandable probe with a diameter of 18 mm that conforms to the width of the adult esophagus. In comparison to single-sensor probes, this multisensor probe has been demonstrated to be more sensitive in reflecting earlier rises in temperature (13.4 sec vs 30.5 sec; P<0.001), shorter time to 1.0°C rise (18.5 sec vs 32.1 sec; P<0.001), and higher changes in peak temperature (1.6°C vs 0.6°C; P<0.001).17
In the case of cryoablation, what esophageal temperatures are too cold? There is no clear consensus for what temperature cutoff would prevent esophageal damage; however, John et al report an occurrence of atrioesophageal fistula with an LET of 20.9°C and recommend discontinuation of cooling when esophageal temperature falls to 30°C.3 They also reported that the most common site of fistula formation was near the left inferior pulmonary vein. To reduce the chance of esophageal injury, consider a single inflation limited to 180 seconds if isolation is obtained within the first 40 seconds of cryoablation.
Cooling of the Esophagus
One approach to mitigate radiofrequency ablation esophageal damage has been to utilize a closed-loop water-irrigated intraesophageal balloon; however, the device is somewhat complex to set up, and it runs the risk of increasing exposure to esophageal injury due to the balloon expansion of the esophagus towards the left atrium and atrial cooling, resulting in ineffective ablation.18 An alternative approach of using ice water injections through a gastric tube has been utilized by others. Kuwahara et al injected 5 cc of ice water if esophageal temperatures reached 42°C during RF ablation in patients undergoing AF ablation with conscious sedation.19 Disappointingly, the operators found that the incidence of esophageal lesions at 1 day was the same whether or not ice water was utilized, although the severity of the esophageal lesions were slightly milder in the cooled patients.
AF ablation power and technique (point-by-point versus continuous drag) can vary across operators. Power application usually ranges between 20W to 40W, with lower wattage along the posterior wall. Given that the posterior wall is also one of the thinnest regions of the left atrium, a reduction of power or time in this region is recommended from a safety perspective. Posterior RF ablation is frequently performed at 20-25W. When LET is utilized, power can be titrated as needed to achieve pulmonary vein isolation while aiming to minimize esophageal heating. Lowering of power from 25-30W to 20-25W on the posterior wall also appears to reduce the incidence of periesophageal vagal nerve injury when ablation transecting the esophagus during linear ablation is required.20
Real-time Imaging of the Esophagus During Ablation
While in many studies, CT or MRI has been used to assess the relationship of the esophagus to the left atrium prior to an ablation, the distance of the esophagus in relationship to the left atrium was highly variable, and up to 15 mm change was recorded at time of procedure.21 Real-time imaging of the esophagus using electromagnetic mapping or intracardiac echocardiography to avoid left atrial ablation within 3-5 mm of the esophagus is possible; however, it requires active monitoring and mapping.
Since the esophagus will frequently course behind a desired ablation path, temperature rises may not allow for adequate ablation and preclude pulmonary vein isolation or linear block. A transesophageal echocardiography probe may be utilized to deviate the esophagus. In a study of 704 patients, TEE was used to displace the esophagus by a mean of 5.9 cm where temperature rise was 0.11°C in the displaced group vs 1.1°C control group (P<0.01).22 Only 3.4% of patients had an esophagus with reduced mobility, which limited esophageal deviation or allowed for only one-way displacement.
Should mechanical displacement with a TEE probe be utilized, the probe should be adequately lubricated and gently inserted into the esophagus. (Figure 4) Manipulation of the transducer to laterally displace the esophagus should be performed gently and maintained for short periods of time to reduce the possibility of esophageal ischemia. Advancement and retraction of the esophagus should never be performed with the probe in a hard, flexed position. A small amount of barium contrast in the esophagus can help identify the degree of esophageal shift away from the region of ablation.
A commercially available nitinol stylet designed to work with an 18 French orogastric tube may also be used to deviate the esophagus. The EsoSure (NE Scientific, Inc.) nitinol stylet is temperature sensitive. (Figure 5) At room temperature, the stylet is softer to allow for easier introduction into the orogastric tube and placement within the esophagus. As the stylet warms to body temperature, it takes on a greater curve to allow for displacement of the esophagus. Rotation, advancement, or retraction of the curve allows for ~3 cm displacement of the esophagus, allowing the operator to safely ablate. (Figure 6)
There is no solid evidence supporting the use of prophylactic therapy with stomach acid suppression or sucralfate. However, given the potentially catastrophic outcome of an atrioesophageal fistula and the not uncommon occurrence of esophageal ulceration, a 2-4 week course of proton pump inhibitors, an H2 blocker, or oral sucralfate solution may be prescribed post ablation to reduce the risk of esophageal ulceration progression.
Management of Esophageal Complications
Esophageal complications can range from dysmotility to ulceration, to perforation and esophageal fistulas. Fortunately, most occurrences of esophageal dysmotility are temporary and usually resolve within the first few months. Asymptomatic and symptomatic ulcerations usually resolve within 2-4 weeks with use of acid suppression therapy, and therefore, proton pump inhibitors and H2 blockers may be prescribed for a month. In patients diagnosed with ulcers, oral sucralfate solution 10 ml four times a day may be prescribed for mucosal protection.
Atrioesophageal fistulas and perforations carry a high mortality rate, and require early diagnosis and intervention. A non-surgical approach involving esophageal stenting and pericardiocentesis have been reported in patients with esophageal fistula without atrial involvement, but a definitive open repair is preferred due to the possibility of fistula progression despite treatment.23 Unfortunately, atrioesophageal fistulas often have an inflammatory or necrotic component that makes simple closure of the atrium and esophagus difficult or impossible. Surgical closure of the atrial aspect of the fistula with a pericardial flap and primary repair of the esophagus with a muscular flap was superior to the non-surgical approach with decreased mortality.24 A multidisciplinary discussion between the cardiologist, gastroenterologist, and cardiothoracic surgeon is critical in managing such patients.
Although technological advances have allowed electrophysiologists to produce more effective ablation lesions, the risk for collateral injury remains. The posterior wall ablation required in a pulmonary vein isolation may be in close proximity to the esophagus and its vascular supply and nervous system. Complications can range from esophageal dysmotility to life-threatening atrioesophageal fistulas. Careful titration of power along the posterior wall, in combination with luminal esophageal temperature monitoring and judicious use of esophageal deviation, help operators achieve a successful ablation while minimizing risk to the esophagus.
Acknowledgment: A special thanks to thoracic surgeon Dr. Jason Shaw for his input on esophageal anatomy.
Disclosure: The authors have no conflicts of interest to report regarding the content herein.
- Sanchez-Quintana D, Cabrera JA, Climent V, et al. Anatomic relations between the esophagus and left atrium and relevance for ablation of atrial fibrillation. Circulation. 2005;112(10):1400-1405.
- Barbhaiya CR, Kumar S, Guo Y et al. Global Survey of Esophageal Injury in Atrial Fibrillation Ablation: Characteristics and Outcomes of Esophageal Perforation and Fistula. JACCCEP. 20162(2):143-150.
- John RM, Kapur S, Ellenbogen KA, et al. Atrioesophageal fistula formation with cryoballoon ablation is most commonly related to the left inferior pulmonary vein. Heart Rhythm. 2016;pii:S1547-5271(16)30936-5.
- Doubilet H, Shafiroff BG, Mulholland JH. The anatomy of the peri-esophageal vagi. Ann Surg. 1948:127-135.
- Miyazaki S, Nakamura H, Taniguchi H, Takagi T, et al. Esophagus-related complications during second-generation cryoballoon ablation: insight from simultaneous esophageal temperature monitoring from 2 esophageal probes. J Cardiovasc Electrophysiol. 2016;27:1038-1044.
- Rolf S, Boldt LH, Parwani AS, et al. Findings and outcome of fluoroscopic visualization of the esophageal course during catheter ablation of atrial fibrillation. Europace. 2011;13:796-802.
- Tsai WK, Koruth J, Reddy VY. Esophageal Heating is Not Limited to Left Atrial Ablation. Circulation: Arrhythmia and Electrophysiology. 2014;7:178-179.
- Liu E, Shehata M, Liu T, et al. Prevention of esophageal thermal injury during radiofrequency ablation for atrial fibrillation. JICE. 2012;35:35-44.
- Barbhaiya CR, Kumar S, John RM, et al. Global Survey of Esophageal and Gastric Injury in Atrial Fibrillation Ablation. JACC. 2015;65(13)1377-1378.
- Lakkireddy D, Reddy YM, Atkins D, et al. Effect of Atrial Fibrillation Ablation on Gastric Motility: The Atrial Fibrillation Gut Study. Circ Arrhythm Electrophysiol. 2015;8:531-536.
- Di Biase L, Dodig M, Saliba W, et al. Capsule endoscopy in examination of esophagus for lesions after radiofrequency catheter ablation: a potential tool to select patients with increased risk of complications. J Cardiovasc Electrophysiol. 2010;21:839-884.
- Martinek M, Meyer C, Hassanein S, et al. Identification of a high-risk population for esophageal injury during radiofrequency catheter ablation of atrial fibrillation: procedural and anatomical considerations. Heart Rhythm. 2010;7:1224-1230.
- Contreras-Valdes F, Heist E, Danik S. Severity of esophageal injury predicts time to healing after radiofrequency catheter ablation for atrial fibrillation. Heart Rhythm. 2011;8:1862-1186.
- Halm U, Haspar T, Zachaus M. Thermal esophageal lesions after radiofrequency catheter ablation of left atrial arrhythmias. Am J Gastroenterol. 2010;105:551-556.
- Singh S, d’Avila A, Doshi S, et al. Esophageal injury and temperature monitoring during atrial fibrillation ablation. Circ Arrhythm Electrophysiol. 2008;1:162-168.
- Teplitsky L, Perzanowski C, Durrani S, et al. Radiofrequency catheter ablation for atrial fibrillation produces delayed and long lasting elevation of luminal esophageal temperature independent of lesion duration and power. Heart Rhythm. 2005;2:S8-S9.
- Tschabrunn CM, Silverstein J, Berzin T, et al. Comparison between single- and multi-sensor oesophageal temperature probes during atrial fibrillation ablation: thermodynamic characteristics. Europace. 2015;17:891-897.
- Tsuchiya T, Ashikaga K, Nakagawa S, et al. Atrial fibrillation ablation with esophageal cooling with a cooled water-irrigated intraesophageal balloon: a pilot study. J Cardiovasc Electrophysiol. 2007;18:145-150.
- Kuwahara T, Takahashi A, Okubo K, et al. Oesophageal cooling with ice water does not reduce the incidence of oesophageal lesions complicating catheter ablation of atrial fibrillation: randomized controlled study. Europace. 2014;16:834-839.
- Miyazaki S, Taniguchi H, Kusa S, et al. Factors Associated with Periesophageal Vagal Nerve Injury After Pulmonary Vein Antrum Isolation. J Am Heart Assoc. 2014;3:e001209.
- Orlov M, Hoffmeister P, Chaudhry G, et al. Three-dimensional rotational angiography of the left atrium and esophagus — a virtual computed tomography scan in the electrophysiology lab? Heart Rhythm. 2007;4:37-43.
- Mateos J, Mateos E, Pena T, et al. Simplified method for esophagus protection during radiofrequency catheter ablation of atrial fibrillation — prospective study of 704 cases. Rev Bras Cir Cardiovasc. 2015;30(2):139-147.
- Singh S, d’Avila A, Singh S, et al. Clinical outcomes after repair of left atrial esophageal fistulas occurring after atrial fibrillation ablation procedures. Heart Rhythm. 2013;10(11):1591-1597.
- Mohanty S, Santangeli P, Mohanty P, et al. Outcomes of atrioesophageal fistula following catheter ablation of atrial fibrillation treated with surgical repair versus esophageal stenting. J Cardiovasc Electrophysiol. 2014;25(6):579-584.