急性胰腺炎并发胰腺感染的危险因素及其预测模型的构建

Risk factors for pancreatic infection in patients with acute pancreatitis and establishment of prediction model

目的分析急性胰腺炎(AP)并发胰腺感染的危险因素,并构建列线图预测模型。方法选取2020年1月-2023年1月天津医科大学朱宪彝纪念医院收治的54例AP并发胰腺感染患者为感染组,随机选取133例AP未并发胰腺感染患者为非感染组,单因素和多因素分析AP患者并发胰腺感染的危险因素,构建列线图预测模型,并分析预测模型的预测价值。结果并发低氧血症、禁食时间>15 d、胃肠功能障碍时间>2 d、中心静脉导管留置时间>6 d均为AP患者并发胰腺感染的独立危险因素(P<0.05);Hosmer-Lemeshow检验结果显示模型拟合度较好(χ^(2)=9.722,P=0.285);按照诊断概率Logit(P)绘制AP患者并发胰腺感染的受试者工作特征(ROC)曲线,当Logit(P)>0.234时,曲线下面积(AUC)值为0.823,95%CI为0.764~0.882,敏感度为83.33%,特异度为71.43%。结论AP患者并发胰腺感染与并发低氧血症、禁食时间、胃肠功能障碍时间、中心静脉导管留置时间有关,相关列线图预测模型的预测价值较高。

OBJECTIVE To analyze the risk factors for pancreatic infection in the patients with acute pancreatitis(AP)and establish the nomogram prediction model.METHODS A total of 54 AP patients complicated with pancreatic infection who were treated in Tianjin Medical University Chu Hsien-I Memorial Hospital from Jan 2020 to Jan 2023 were assigned as the infection group,and 133 AP patients who were not complicated with pancreatic infection were randomly chosen as the non-infection group.Univariate analysis and multivariate analysis were performed for the risk factors for the pancreatic infection in the AP patients,the nomogram prediction model was established,and the predictive value of the prediction model was analyzed.RESULTS Complication with hyoxemia,fasting time more than 15 days,gastrointestinal dysfunction more than 2 days and central venous catheter indwelling more than 6 days were the independent risk factors for the pancreatic infection in the AP patients(P<0.05).The result of Hosmer-Lemeshow test showed that the model had high fitting degree(χ^(2)=9.722,P=0.285).The receiver operating characteristic(ROC)curve for the pancreatic infection was drawn based on diagnostic probability Logit(P);when Logit(P)was more than 0.234,the area under curve(AUC)was 0.823,with the 95%CI 0.764-0.882,the sensitivity 83.33%,the specificity 71.43%.CONCLUSION The pancreatic infection in the AP patients is associated with the complication of hyoxemia,fasting time,gastrointestinal dysfunction time and central venous catheter indwelling time.The nomogram prediction model has high predictive value.