439 Risk of Air Leak and Mucosal Injury in Esophageal POEM After Prior Myotomy

Abstract

INTRODUCTION: Peroral endoscopic myotomy (POEM) is deemed to be more challenging with more periprocedural adverse events. We aimed to assess the risk of air leak and mucosal injury in POEM after prior myotomy (Heller’s myotomy [HM] or POEM), the Redo POEM, in addition to extubation time when air leak occurs. METHODS: A retrospective study comparing Redo POEM to an index POEM at a tertiary center from March 2016 to August 2018. Table 1 summarizes the data collected. Air leak was defined by subcutaneous emphysema, pneumomediastinum, pneumoperitoneum, and/or pneumothorax documented by physical exam and/or imaging. Difficult procedures were determined by endoscopist impression as challenging or based on difficulty creating submucosal tunnel and/or difficulty performing myotomy. RESULTS: We identified 21 cases of Redo (1st myotomy: HM in 8/21, POEM in 13/21) and 56 cases of index POEM. One case had two prior HMs and two cases had HM and POEM each before Redo POEM. The results are summarized in Table 1. The Redo group was younger with more difficult procedures encountered, shorter myotomy length (due to difficulty with myotomy) and longer procedure time. There were more mucosal injuries with more clips used to close them in the Redo group. Short myotomy (≤5 cm) was performed in 5 cases in the Redo group (23.8%) and 5 cases in the index POEM (8.9%). There was a trend toward more air leak in the Redo group, but it was not statistically significant. Air leak was seen in 14/77 (18.2%) cases in the study. Extubation time was longer in cases with air leak (Table 2). Postprocedural symptoms did not predict air leak. Procedure difficulty did not predict air leak or mucosal injury. No intervention was required for air decompression and there was no ICU admission pertinent to air leak or mucosal injury in either group. CONCLUSION: Redo POEM appears to be more challenging, takes longer time to perform with a shorter than intended myotomy. Mucosal injuries are more common and more clips are used to close them. Air leak tends to be higher in Redo POEM, but not statistically significant. The incidence of air leak was likely underestimated as imaging was not routinely obtained for all procedures. Air leak is associated with a slight delay in postprocedural extubation time, but it does not appear to affect the overall procedural outcomes or prolong hospitalization. Postprocedural symptoms and procedure difficulty did not predict air leak or mucosal injury.