急性肠系膜缺血致肠梗死的危险因素及临床预测模型的开发和验证

Risk factors analysis and the development and validation of a clinical prediction model for acute mesenteric ischemia-induced intestinal infarction

目的探讨急性肠系膜缺血(AMI)致肠梗死的危险因素,开发并验证AMI致肠梗死的Nomogram图临床预测模型。方法收集2014年1月至2020年12月云南大学附属医院201例AMI患者作为建模组,昆明医科大学第一附属医院105例AMI患者作为验证组。建模组患者年龄为(62±12)岁,男70例,女131例。验证组患者年龄为(62±10)岁,男69例,女36例。回顾性分析2组患者的临床表现、实验室指标及影像学指标等临床资料,进行单因素、多因素二元logistic回归分析,筛选出AMI进展为肠梗死的危险因素。使用R软件开发AMI致肠梗死的Nomogram图临床预测模型,通过C指数、校准曲线和临床决策分析曲线评价该Nomogram图的区分度、校准度和临床净获益情况。结果休克(OR=13.69,95%CI:2.31~88.87)、体温≥38.0℃(OR=6.39,95%CI:1.85~22.11)、白细胞≥18×10^(9)/L(OR=1.19,95%CI:1.03~1.37)、CT图像肠管改变(OR=0.17,95%CI:0.04~0.66)、腹膜刺激征(OR=0.06,95%CI:0.01~0.26)、pH≤7.34(OR=0.00,95%CI:0.00~0.01)是AMI致肠梗死的危险因素(均以“否”为参照)。体温≥38.0℃(灵敏度:81.5%,特异度:87.0%),白细胞≥18×10^(9)/L(灵敏度:84.0%,特异度:83.0%),pH≤7.34(灵敏度:70.4%,特异度:76.7%)是AMI患者进展为肠梗死的截断值。使用上述参数建立AMI致肠梗死的Nomogram图预测模型并进行内部和外部验证,验证组和建模组的C指数分别为0.96(95%CI:0.921~0.999)和0.98(95%CI:0.962~0.996);验证组和建模组的校准曲线均在参考线附近走形,与参考线的走行偏离程度较低;验证组和建模组的临床决策分析曲线表明该Nomogram图临床预测模型可为AMI患者带来较好的临床净获益。结论休克、体温≥38.0℃、白细胞≥18×10^(9)/L、CT图像肠管改变、腹膜刺激征、pH≤7.34是AMI致肠梗死的危险因素;使用上述参数建立的AMI致肠梗死的Nomogram图预测模型可有效预测AMI患者进展为肠梗死的概率。

Objective To investigate the risk factors of intestinal infarction caused by acute mesenteric ischemia(AMI),and develop and validate a clinical prediction model of Nomogram plot for intestinal infarction caused by AMI.Methods From January 2014 to December 2020,201 patients with AMI in the Affiliated Hospital of Yunnan University were selected as the modeling group,and 105 patients with AMI in the First Affiliated Hospital of Kunming Medical University were collected as the validation group.The age of patients in the modeling group was(62±12)years,and there were 70 males and 131 females.The age of patients in the validation group was(62±10)years old,and there were 69 males and 36 females.The clinical data including clinical manifestations,laboratory indicators and imaging indicators of patients in the two groups were retrospectively analyzed,and univariate and multivariate binary logistic regression analyses were performed to screen out risk factors for the progression of AMI to intestinal infarction.A Nomogram plot clinical prediction model for AMI causing intestinal infarction was developed using R software,and the differentiation,calibration and net clinical benefit of this Nomogram plot were evaluated by C-index,calibration curve and clinical decision analysis curve.Results Shock(OR=13.69,95%CI:2.31-88.87),body temperature≥38.0℃(OR=6.39,95%CI:1.85-22.11),white blood cells(WBC)≥18×10^(9)/L(OR=1.19,95%CI:1.03-1.37),intestinal changes on CT(OR=0.17,95%CI:0.04-0.66),peritoneal irritation(OR=0.06,95%CI:0.01-0.26),and pH≤7.34(OR=0.00,95%CI:0.00-0.01)were risk factors for intestinal infarction caused by AMI.Body temperature≥38.0℃(sensitivity:81.5%,specificity:87.0%),WBC≥18×10^(9)/L(sensitivity:84.0%,specificity:83.0%)and pH≤7.34(sensitivity:70.4%,specificity:76.7%)were cut-off values for progression to intestinal infarction in patients with AMI.The Nomogram plot prediction model of intestinal infarction induced by AMI was established by using the above parameters and validated internally and externally.The C index of validation group and modeling group were 0.96(95%CI:0.921-0.999)and 0.98(95%CI:0.962-0.996),respectively.The calibration curves of the validation group and the modeling group were shaped near the reference line,and the deviation from the reference line was low.The clinical decision analysis curves of validation group and modeling group showed that the clinical prediction model of Nomogram could bring better clinical net benefit for AMI patients.Conclusions Shock,temperature,white blood cells,intestinal changes on CT,peritoneal irritation sign,and PH are independent risk factors for AMI-induced intestinal infarction.The Nomogram plot established by these factors can effectively predict the probability of AMI patients progressing to intestinal infarction.