Complications of Atrial Fibrillation Ablation

When Prevention Is Better Than Cure

Antonio Sorgente; Gian-Battista Chierchia; Carlo de Asmundis; Andrea Sarkozy; Lucio Capulzini; Pedro Brugada


Europace. 2011;13(11):1526-1532. 

In This Article

Esophageal Injury/Atrio-oesophageal Fistula

Atrio-oesophageal fistula is a very rare complication of AF catheter ablation. Described for the first time in two very experienced centres in 2004,[14,15] this complication is the most dreadful and lethal among all the others related to AF catheter ablation. Its clinical presentation is extremely variable: patients with an atrio-oesophageal fistula could present with a variety of signs and symptoms such as chest pain, heartburn, dysphagia, anorexia, and haematemesis immediately after or also late after the index procedure. Usually death occurs because of cerebral or myocardial air embolism, endocarditis, massive gastrointestinal bleeding and septic shock.[16] Surgical cardiac and oesophageal repair are mandatory to increase the survival in these patients; therefore, a rapid diagnosis is highly desiderable. To this aim, cardiac computed tomography is the most indicated while endoscopy has to be avoid for the high risk of worsening the damage on the oesophageal walls. Interestingly, Bunch et al.[17] have proposed a temporary oesophageal stenting as a possible non-surgical treatment of this complication, able to warrant the complete healing of the damaged oesophageal wall.

Even more desiderable is avoiding the occurrence of an atrio-oesophageal fistula. To this aim, a lot of studies sought to investigate the physiopathologic mechanism underlying this complication. As demonstrated by computed tomography,[18] cardiac magnetic resonance,[19] intracardiac echocardiography,[20] the strict anatomic relationship between the left atrium and the oesophagus together with the delivery of radiofrequency energy on the posterior wall of the left atrium are the principal causes leading to the occurrence of atrio-oesophageal fistula or, more generally, of oesophageal injury. Interestingly, Meng et al.[21] have demonstrated that new oesophageal late gadolinium enhancement is present in almost one-third of patients after PVI and this finding is irrespective of the type of catheter ablation (irrigated vs. not-irrigated tip) used during the procedure, of ablation time, of anatomical location of the oesophagus compared with the left atrium, of the size of left atrium cavity or of the timing of cardiac magnetic resonance study after PVI.

Different strategies can be adopted to avoid or reduce the incidence of this dreadful complication. First of all, localization of the region of contact between left atrium and oesophagus can be obtained before the procedure itself by means of computed tomography after a barium swallow or magnetic resonance after a barium plus gadolinium diglutamate swallow[22] or during the ablation procedure with intracardiac echocardiography.[23] Moreover, electroanatomical mapping systems such as CARTOTM (Biosense Webster, Diamond Bar, CA, USA) and NavXTM (St Jude Medical, Sylmar, CA, USA) allow the superimposition of the oesophageal imaging obtained with the Carto SoundStarTM probe or with the NavXTM itself with the real-time electroanatomical map of the left atrium.[24,25] Secondly, it is strongly advisable to avoid delivery of high levels of radiofrequency energy on the posterior wall of the left atrium or on the posterior aspect of the PV antra, usually areas of presumable contact with the oesophagus. To this aim, since theoretically radiofrequency energy exerts a rise in local temperatures, it's common practice in many centres to monitor the oesophageal temperature with an oesophageal probe[26,27] to titrate the radiofrequency energy application on the areas at potential risk of oesophageal injury and to stop radiofrequency energy delivery when a rapid elevation of the oesophageal internal temperature is recorded. In centres which do not perform cardiac computed tomography or cardiac magnetic resonance before the procedure, the use of oesophagram with water soluble contrast may represent a valid trick to avoid ablation on portions of the left atrium in close vicinity of the oesophagus.

However, the main issue related to this practice is the lack of knowledge of what has to be considered the 'safest' amount of radiofrequency energy. Indeed, the accuracy of luminal temperature monitoring to estimate the oesophageal heating and then in anticipating the formation of the oesophageal injury is uncertain.[28] Poor correlation between oesophageal internal temperature and total radiofrequency energy delivery has been demonstrated in initial clinical studies.[20,29] Interindividual variability in oesophageal and posterior left atrial wall thickness may explain this finding along with a presumable lack of fidelity of the luminal probe in measuring the heating of the oesophageal wall. Furthermore, as brilliantly evidenced by Nakagawa,[30] oesophageal tissue heating and consequent injury seem to be related more to the catheter tip-tissue contact force rather than to the total amount of radiofrequency energy delivered for PVI.

There's no compelling evidence that any strategy employed can reduce or even avoid this complication although many are used. The most common measures adopted in this context are reducing the power and the duration of radiofrequency energy delivery near or over the oesophagus (25 W is usually considered safe) as well as monitoring the oesophageal temperature by means of an endoscopic probe. Another very simple and intuitive option is to avoid any radiofrequency delivery at the posterior left atrial wall, as to the best of our knowledge, there is no conclusive evidence that linear lesions in this region would improve late outcome as compared with accurate, stable (i.e. long-lasting) PV isolation in whatever AF form (i.e. paroxysmal, persistent, and permanent). The empiric use of proton-pump inhibitors after ablation to mitigate reflux that has been observed in many cases and correlated with pH studies after ablation.[31] Questionable is also the clinical utility of a cooled saline-irrigated balloon inside the oesophageal lumen during AF catheter ablation.[32] Indeed, if on the one hand cooling of the internal lumen of the oesophagus should limit the transmural rise of the temperature, on the other hand inflation of a balloon device inside the oesophagus could increase the area of contact with the left atrium, enhancing paradoxically the heat transfer and the chances of thermal injury.

The majority of atrio-oesophageal fistulas described in the literature has been related to radiofrequency energy delivery. A possible and well-known alternative to radiofrequency is cryothermal energy. Used first for surgical treatment of ventricular arrhythmias and then for endocardial treatment of specific supraventricular arrhythmias, very recently cryothermal energy indication has been enlarged to include also AF or left atrial tachycardias.[23] Cryothermal energy can be delivered focally through traditional deflectable ablation catheters or circumferentially through innovative inflatable balloons with proven efficacy.[1] Unfortunately, as radiofrequency energy, cryothermal energy induces a conductive heat transfer on the oesophagus, resulting in an equivalent amount of oesophageal injury compared with radiofrequency.[33] There's evidence in animal models that cryothermal ablation could determine a transmural injury of the oesophagus but, unlike radiofrequency energy which produces necrosis and ulcers, it preserves the architecture of the cells reducing at great extent the probability of a fistula or of ulceration.[28] Even though contrasting data have been produced on the risk of oesophageal ulcerations after cryoballoon ablation,[34,35] cryothermal energy could represent at the moment a valid alternative to radiofrequency energy when ablation of the posterior wall is needed and hybrid approaches to AF catheter ablation has already been proposed.[36]