Abstract

Background

Luminal esophageal temperature (LET) monitoring does not prevent esophageal injury following pulmonary vein isolation (PVI) but LET excess (i.e. > 41°C) is associated with increased risk for esophageal lesions and fistula.

Objective

Detailed analysis of LET data may better predict the risk for any esophageal injury (mucosal lesions, periesophageal edema and food retention), and thus to streamline follow-up/therapy to prevent progression to fistula.

Methods

LET data (S-Cath Circa, 12 thermocouples ) of consecutive patients with radiofrequency (RF) PVI (30 Watt at the left atrial posterior wall) were analyzed offline. Ablation was stopped when LET exceeded 41.0°C. LET-maxima, duration of LET rise, area under the maximum LET curve of any thermocouple (AUCmax) and summed-up AUC of all individual LET-curves of the 12 thermocouples (3D-AUC) were calculated and correlated with the results of endoscopy (EGD [mucosal lesions, food retention]) and endoscopic ultrasound (EUS [edema]). The calculated parameters were analyzed with regard to their ability to predict (sensitivity) or exclude (specificity) esophageal injury.

Results

Following RF-PVI, 81 patients were studied by EGD+EUS (n=73) or EGD alone (n=8). Any esophageal injury was observed in 33/73 (mucosal lesions: 12/81, edema: 27/73, food retention: 13/81). Procedural data of PVI were not different between patients with/without esophageal injury.

1. LET > 41.8°C was associated with mucosal lesions but raw temperatures were not able to predict esophageal injury.
2. Duration of any LET-rise above baseline differed significantly with/without esophageal injury and performed even better for duration above 38.0°C (median time [interquartile range] exceeding 38.0°C 542 [298/668]s with vs. 289 [143/414]s without injury).
3. Summed-up area under the curve for LET exceeding 38.0°C (3D-AUC > 38.0°C) was the best predictive parameter and significantly increased in patients with esophageal injury (882 [448/1181]K.s with vs. 421 [170/691]K.s without injury [sensitivity 70%, specificity 83%, Youden-index 0.52, p<0.001]).

Conclusion

AUC of the LET curves best predicted esophageal injury following RF-PVI with sensitivity and specificity ranging from 70-80% (Fig. 1D). Whether online-assessment of AUC is useful to guide PVI strategy and to prevent esophageal injury remains to be seen.