Monitoring risk for oesophageal thermal injury during radiofrequency catheter ablation for atrial fibrillation: does the characteristic of the temperature probe matter?

This editorial refers to ‘Comparison between single- and multi-sensor oesophageal temperature probes during atrial fibrillation ablation: thermodynamic characteristics’ by C.M. Tschabrunn et al., on page 891–897. Atrioesophageal fistula as a complication of radiofrequency (RF) catheter ablation of atrial fibrillation (AF) was first described in 2004, shortly after strategies for pulmonary vein isolation moved from ostial to circumferential antral ablation.1,2 Atrioesophageal fistula was initially interpreted as a transitory technical problem that could be prevented by recognizing the oesophagus position and titrating RF energy during ablation in the proximity to the oesophagus. After more than 10 years of development, AF ablation has become the most common catheter ablation procedure performed worldwide. However, atrioesophageal fistula has remained a serious and unpredictable complication occurring in 0.03–0.2% of the ablation procedures even in experienced centres,3 without a convincing demonstration that a specific strategy might prevent it. The crucial problem is that atrioesophageal fistula is often unexpected, and when it occurs, the risk of death or severe neurologic sequelae is around 80%. Moreover, the incidence of atrioesophageal fistula may increase in the next years because technical improvements have been implemented in order to obtain ‘better RF lesions’ to prevent pulmonary vein reconnections and the increasing number of new operators. Thermal injury is the postulated mechanism for oesophageal injury during RF ablation due to the close anatomical relationship between the left atrium and the oesophagus. Radiofrequency energy delivery may extend beyond the atrial myocardium and may reach the oesophagus leading to distinctive grades of ischaemic necrosis of the mucosal layers with different consequences.4 In the past years, many interventions have being proposed aiming at reducing the thermal oesophageal injury and consequently, the atrioesophageal fistula.5–7 Monitoring the oesophageal temperature during RF ablation using an intraluminal temperature probe has been one of most popular methods, and different thermal probes have been empirically used. Since the occurrence of atrioesophageal fistula is very rare, studies to evaluate the impact of the protective measures have considered the occurrence of endoscopic oesophageal lesions, such as erythema and ulcerations as the endpoint. Halm et al.8 studied 185 patients who underwent RF ablation for paroxysmal AF or post-AF ablation flutter, using a 5 olive linear temperature probe for oesophageal temperature monitoring. The incidence of oesophageal lesions was 14.6%. They showed no occurrence of oesophageal lesions when the maximum temperature was <41°C but that for every 1°C increase in endoluminal temperature, the odds of an oesophageal lesion increased by a factor of 1.36. Sause et al.9 using a cut-off of 40°C identified lesions in 1.6% of the 184 patients evaluated. Contreras-Valdes et al.10 studied 219 patients after RF ablation and used a cut-off point of 38°C. They performed endoscopy if the luminal oesophageal temperature reached 39°C and found that the incidence of oesophageal lesions was 27%, with 38% of these lesions considered deep. Two other studies which explored whether intraluminal oesophageal temperature monitoring prevents oesophageal lesions during pulmonary vein isolation have reported conflicting results. Singh et al.11 were the first to demonstrate the association between oesophageal temperature monitoring and the reduction in oesophageal injury during point-by-point AF ablation. Limiting or interrupted RF delivery when the temperature reached 38.5°C was associated with a significantly lower incidence of oesophageal injury compared with the unmonitored group (6 vs. 36%). Deneke et al.12 showed different results with luminal oesophageal temperature monitoring in patients undergoing pulmonary vein isolation by duty-cycled phased RF ablation using a circular catheter over-the-wire. Radiofrequency ablation was stopped when oesophageal temperature reached 40°C, and endoscopy was performed in patients whose oesophageal temperature reached 39°C (56%) revealing oesophageal lesions in 19%. In the unmonitored group, endoscopy performed in all patients showed no oesophageal lesions (P = 0.0077). This study raised the hypothesis that the metal electrodes in the oesophageal probe might act as antennas during RF delivery to the left atrial posterior wall, increasing the risk of damage to the oesophagus. Single probe oesophageal temperature monitoring requires changes in its position from cranial to caudal, in order to get it closer to the ablation catheter. Multi-electrode thermometers arranged in a linear probe solve this problem; however, they do not cover all biplanar oesophageal surfaces, and some points may be missed. This could explain why some patients have oesophageal injury even with temperature monitoring. In order to solve these limitations, a sinusoidal-shaped multi-electrode thermometer was developed. Carroll et al.13 compared a multi-sensor S-shaped oesophageal probe (MSP) with a single sensor probe (SSP) in 543 patients undergoing RF ablation (455 in the SSP group and 88 in the MSP group). The maximum oesophageal temperature >39°C was detected more frequently (75%) in patients monitored by MSP than in patients under SSP monitoring (39%, P < 0.0001). In the MSP group, RF time and the total energy application were greater than those in the SSP group. The authors explained this by the need to perform ablation away from the pulmonary veins due to the oesophageal temperature increase. Surprisingly, the oesophageal injury was more frequent in the MSP group (46 vs. 29%; P = 0.021), without any difference in the incidence of severe ulceration. These results could be explained by the oesophageal surface increase due to the expandable probe and also by a heat-sink effect caused by the multi-electrode probe. However, endoscopy was not performed in the entire group in this study, but only when oesophageal temperature increased >39°C. Thus, since MSP is more sensitive, a large number of oesophageal lesions might have been missed in the SSP group. In this issue of EP-Europace, Tschabrunn et al.14 compared prospectively the thermodynamics of two models of oesophageal probes during RF ablation. One model was SSP and the other model was a 12-sensor self-expandable probe, similar to the MSP model tested by Carroll et al.13 Twenty patients were evaluated, 10 underwent dual monitoring with both SSP and MSP (the operators were blinded to the MSP measurements) and 10 control patients were monitored with SSP alone. Application of RF energy was interrupted if SSP temperature increased by 2°C. The thermodynamic analysis showed that the increase in temperature was detected earlier by MSP, with a shorter time to a rise by 1°C and a higher change in peak temperature. The mean peak temperature and the rate of temperature decay were similar between the two models. Endoscopy was performed in all 20 patients in the study, and the oesophageal lesions were found in 5 (50%) patients in the MSP group and 4 (40%) in the SSP group. The severity of lesions was also similar between the two groups. The authors performed a complementary animal study in five pigs to evaluate the hypothesis that the self-expandable probe might expand the oesophagus in contact with the left atrium and showed that the self-expanded probe did not increase the contact area between the oesophagus and left atrial posterior wall. For the heat-sink hypothesis, the similar peak in the luminal oesophageal temperature and the rate of decay suggests that this effect has no clear significance. The main limitation of this study was the small number of patients which did not allow a further investigation of the variables related to oesophageal injury in both probes. Additionally, patients selected for this study had an enlarged left atrium (56.3 ± 5.0 and 58.0 ± 5.2 mm) in the both groups. It is important to acknowledge that the objective of this study was to compare the thermodynamic between the two models, therefore, it has not been designed to evaluate whether MSP protects from the occurrence of oesophageal lesions. Additionally, temperature values obtained with MSP did not guide RF interruption, since these were not known to the operators. Anatomical comparisons of the oesophagus between the pig model and humans cannot be perfect and this observation should be confirmed in man. Despite these limitations, the results of this study may be useful and may lead to further multicentre trials aiming at evaluation better methods to minimize the oesophageal injury during ablation. In conclusion, despite recent developments in AF ablation, atrioesophageal fistula is still an issue of concern, because although rare, it is frequently lethal when it happens. Owing to its rarity, it is difficult to identify the best way of preventing this complication. Esophageal temperature monitoring is widely accepted as a tool to improve safety during ablation, but it has limitations. Although MSP could theoretically improve the safety of the procedure because of better sensitivity and wider area of temperature measurements, its real efficacy could only be assessed in a multicentre randomized trial comparing both models of probes and endoscopy performed in all patients.