摘要
目的:探讨经内镜逆行胰胆管造影(ERCP)术后胰腺炎预测模型的建立与评价。方法:收集2021年8月至2022年8月经解放军总医院第一医学中心消化内科医学部行ERCP术治疗300例患者临床资料，根据是否发生术后胰腺炎的不良事件分为发生组(43例)、未发生组(257例)，采用单因素分析、多因素Logistic回归分析ERCP术后不良事件发生的影响因素，并构建风险预测模型，采用受试者工作特征曲线(ROC曲线)分析该模型的预测价值。结果:43例发生术后不良事件患者中，35例急性胰腺炎、5例感染、2例出血、1例消化道穿孔，均予以对症治疗后明显缓解。多因素Logistic回归分析结果显示，胰腺疾病史、Oddi括约肌功能障碍(SOD)史、插管困难、ERCP手术时间≥60 min、进入胰管次数≥2次为ERCP术后不良事件发生的独立危险因素(P<0.05)。基于以上5个指标构建的风险预测模型的AUC为0.901，敏感度、特异度为86.0%、93.8%，均高于模型中各单项指标，且该风险预测模型拟合程度较好(Hosmer-Lemeshow χ2＝1.501，P＝0.913)。结论:胰腺疾病史、SOD史、插管困难、ERCP手术时间、进入胰管次数为ERCP术后不良事件发生的独立危险因素，基于以上5个指标构建的风险预测模型对ERCP术后不良事件发生具有较高预测价值，能为预防ERCP术后不良事件发生提供理论依据。
Abstract
Objective:To investigate the establishment and evaluation of the risk prediction diagram model of complications after endoscopic retrograde cholangiopancreatography (ERCP).Method:The clinical data of 300 patients treated by ERCP from August 2021 to August 2022 were collected, which were divided into occurrence group(43 cases) and no group (257 cases), univariate analysis and multivariate Logistic regression were used to analyze the influencing factors of postoperative adverse events after ERCP, and the predictive value of the model was analyzed by the receiveroperating characteristic curve (ROC curve).Result:Among the 43 patients with postoperative adverse events, 35 patients had acute pancreatitis, 5 infections, 2 bleeding, and 1 gastrointestinal perforation, which were significantly relieved after symptomatic treatment. The results of multivariate Logistic regression analysis showed that the history of pancreatic disease, history of SOD, difficult intubation, ERCP operation time of 60 min, and two visits into the pancreatic duct were independent risk factors for adverse events after ERCP (P<0.05). The AUC of the risk prediction model constructed based on the above five indexes was 0.901, and the sensitivity and specificity were 86.0% and 93.8%, which were all higher than the single indexes in the model, and the degree of fitting of the risk prediction model was good (Hosmer-Lemeshow χ2=1.501, P=0.913).Conclusion:History of pancreatic disease, history of SOD, difficultyin intubation, operation time of ERCP, and number of entry into thepancreatic duct are independent risk factors for postoperative adverseevents in ERCP.The risk prediction model constructed based on theabove five indicators has high predictive value for the occurrence of postoperative adverse events in ERCP, and can provide a theoretical basis for preventing the occurrence of postoperative adverse events after ERCP.
万方数据