急性胰腺炎并发门静脉系统血栓形成的危险因素及预测模型构建

Risk factors and prediction model built for portal vein system thrombosis in patients with acute pancreatitis

目的探讨急性胰腺炎患者并发门静脉系统血栓形成的危险因素并构建预测模型。方法回顾性分析2020年1月至2022年12月在新疆医科大学第一附属医院接受治疗的188例急性胰腺炎患者临床资料,分为血栓组(46例)和非血栓组(142例)。分析急性胰腺炎合并门静脉系统血栓的危险因素并构建预测模型,运用受试者工作特征(ROC)曲线、校准曲线、Hosmer-Lemeshow检验以及临床决策曲线进行评价。结果门静脉系统血栓发生率为24.47%(46/188)。Logistic回归分析结果显示,腹水(OR=2.28,95%CI:1.04~5.00,P=0.040)、白细胞计数≥10×109/L(OR=3.69,95%CI:1.55~8.78,P=0.003)、三酰甘油-葡萄糖指数≥9.21(OR=4.38,95%CI:1.90~10.11,P=0.001)是急性胰腺炎患者并发门静脉系统血栓的独立危险因素。将上述危险因素绘制ROC曲线,曲线下面积为0.790。Hosmer-Lemeshow检验χ2=4.293,P=0.891,表明预测值和观测值之间没有完美拟合偏差。校准曲线的Brier值为0.144,表示该模型在预测中的准确性较高,预测结果与实际观测值的差异相对较小。制作急性胰腺炎并发门静脉系统血栓的预测模型,根据得分实现初步预测。结论腹水、白细胞计数≥10×109/L以及三酰甘油-葡萄糖指数≥9.21是急性胰腺炎并发门静脉系统血栓形成的独立危险因素,基于以上危险因素构建的列线图具有良好的区分度和准确性,可为治疗策略提供参考。

Objective To investigate the risk factors for patients with acute pancreatitis(AP)combined with portal vein system thrombosis(PVST)and develop a predictive model.Methods A retrospective analysis was conducted in clinical data of patients undergoing treatment for pancreatitis at the First Affiliated Hospital of Xinjiang Medical University from January 2020 to December 2022.188 cases of AP were finally included based on the inclusion and exclusion criteria,with 46 cases in the thrombosis group and 142 cases in the non-thrombosis group.Clinical data of the two groups were compared to analyze the risk factors for PVST in patients with AP and develop a predictive model.Receiver operating characteristic curve(ROC),calibration curve,Hosmer-Lemeshow test,and clinical decision curve were used for evaluation.Results PVST occurred in 46(24.47%)of the 188 patients with AP.Multivariate Logistic regression analysis showed that ascites(OR=2.28,95%CI:1.04-5.00,P=0.040),white blood cell count 10×109/L(OR=3.69,95%CI:1.55-8.78,P=0.003),and TYG≥9.21(OR=4.38,95%CI:1.90-10.11,P=0.001)were independent risk factors for PVST in patients with AP.ROC curve analysis of these risk factors yielded an area under the curve of 0.790.The Hosmer-Lemeshow test showedχ2 value of 4.293 and P value of 0.891,indicating that there was no statistically significant lack of fit between the predicted values and observed values.The Brier score of the calibration curve was 0.144,indicating high accuracy of the model in prediction with small differences between the predicted and observed values.The created predictive model for PVST in patients with AP allowed for initial risk assessment based on the calculated scores.Conclusions Ascites,white blood cell count10×109/L,and TYG≥9.21 are independent risk factors for the development of PVST in patients with AP.The constructed column chart based on these risk factors demonstrates good discriminatory power and accuracy,providing references for the treatment strategies for PVST with AP.