Landscape of Adverse Events Related to Peroral Endoscopic Myotomy in 3135 Patients and a Risk-Scoring System to Predict Major Adverse Events.
Xinyang Liu,Lu Yao,Jing Cheng,Meidong Xu,Shiyao Chen,Yunshi Zhong,Mengjiang He,Weifeng Chen,Yiqun Zhang,Wenzheng Qin,Jianwei Hu,Mingyan Cai,Liqing Yao,Pinghong Zhou,Quanlin Li
DOI: https://doi.org/10.1016/j.cgh.2021.04.033
IF: 13.576
2021-01-01
Clinical Gastroenterology and Hepatology
Abstract:BACKGROUND AND AIMS:This article systemically describes the landscape of peroral endoscopic myotomy (POEM)-related adverse events (AEs) and compares the different grading systems; and establishes and validates a combined risk factor model and a simplified risk-scoring system to predict POEM-related major AEs.METHODS:A total of 3135 patients with achalasia treated with POEM were included and the AEs were systemically described and graded. A predictive model and risk-scoring system was developed using logistic regression and then internally validated using bootstrapping approaches.RESULTS:A total of 258 out of 3135 patients, accounting for 8.23% of the total patients, presented with 297 AEs. According to Clavien-Dindo grading, 175 (67.83%), 23 (8.91%), 56 (21.71%), 4 (1.55%), and 0 (0.00%) patients were graded as grade I-V, respectively. By American Society of Gastrointestinal Endoscopy lexicon, 175 (67.83%) patients were classified with mild AE, 66 (25.58%) were classified with moderate AE, and 17 (6.59%) were classified with severe AE, respectively. Sixty-eight (2.17%) patients were classified with major AE. Air insufflation, selective myotomy, mucosal injury, and long operation time were selected into the predictive model with an area under the curve of 0.795. They were assigned with scores of 18, 5, 3, and 5 in the risk-scoring system, respectively. By applying the risk scoring system, patients with higher scores had higher rates of major AEs. The model showed little evidence for overfitting and was well-calibrated.CONCLUSIONS:Based on a systematic landscape analysis, POEM is a safe procedure with low rates of severe AEs. Our prediction model and risk-scoring system demonstrated good performance in predicting major AEs.