Risk factors for intraoperative hemorrhage during endoscopic submucosal dissection for colorectal lesions / 中华消化内镜杂志

ABSTRACT

Objective:To investigate the risk factors for intraoperative hemorrhage during endoscopic submucosal dissection (ESD) for colorectal lesions.Methods:Data of 386 patients with colorectal lesions, who underwent ESD at The Third People's Hospital of Datong and its cooperative hospital, Nanjing Drum Tower Hospital, from December 2019 to August 2021 were retrospectively analyzed. The patients were divided into the hemorrhage group ( n=85) and the non-hemorrhage group ( n=301) according to intraoperative hemorrhage. The correlationship of patients'basic information, lesion-related factors and hemorrhage during colorectal ESD was analyzed. Univariate and multivariate logistic regression were used to identify the risk factors for intraoperative hemorrhage during ESD. The risk predictive model of intraoperative hemorrhage during ESD was established according to the screened risk factors, and receiver operator characteristic (ROC) curve was used to evaluate the predictive model. Results:Univariate logistic regression showed that a history of diabetes ( OR=2.340, P<0.05), a history of coronary atherosclerotic heart diseases ( OR=3.100, P<0.05), the lesion located in the rectum ( OR=3.272, P<0.05), longer lesion ( OR=1.093, P<0.05), wider lesion ( OR=1.057, P<0.05), larger lesion ( OR=1.126, P<0.05), depressed lesion ( OR=6.128, P<0.05), the laterally spreading lesion ( OR=2.651, P<0.05), the lesion infiltrated into the SM-S layer ( OR=0.088, P<0.05), the lesion infiltrated into the SM-D layer ( OR=0.174, P<0.05), the diameter of hemorrhage vessels 0.5~<1.0 times of the diameter of incision knife ( OR=246.854, P<0.05), the postoperative pathology as early cancer ( OR=7.000, P<0.05) were risk factors for intraoperative hemorrhage during ESD. Considering the quantitative relationship between the length, the width and the area of lesions, multi-factor models were constructed using the length and area of lesions respectively. Forward stepwise regression was used to screen variables and determine the final model, and the results showed that a history of coronary atherosclerotic heart diseases, the depressed lesion, the longer lesion, the larger lesion, the diameter of hemorrhage vessels 0.5~<1.0 times of the diameter of the incision knife were independent risk factors for intraoperative hemorrhage during ESD. The two modeling results of the lesion length and the lesion area were very similar. Therefore, lesion length was recommended to describe lesions in clinical practice. Conclusion:A history of coronary atherosclerotic heart disease, the depressed lesion, the longer lesion, the larger lesion, the diameter of vessels 0.5~<1.0 times of that of the incision knife are independent risk factors for intraoperative hemorrhage during ESD.