高甘油三酯血症性急性胰腺炎重症化的早期预测模型构建  

Establishment of early predictive model for hypertriglyceridemic acute pancreatitis

作　　者：甘筠益 刘桠名 王倩[1] 汪奕潼 夏梦诗 王晓龙[1] Gan Jun-yi;Liu Ya-ming;Wang Qian;Wang Yi-tong;Xia Meng-shi;Wang Xiao-long(Emergency Department,the 2nd Affiliated Hospital of Chongqing Medical University,Chongqing 400010,China)

机构地区：[1]重庆医科大学附属第二医院急救部,重庆400010

出　　处：《中国急救医学》2022年第11期975-981,共7页Chinese Journal of Critical Care Medicine

摘　　要：目的筛选高甘油三酯血症性急性胰腺炎(HTG-AP)重症化的早期危险因素,构建Nomogram预测模型,并评估其临床价值。方法纳入重庆医科大学附属第二医院2017年1月至2020年12月收治的350例HTG-AP患者为研究对象,按照病情严重程度分为SAP组(n=100)、非SAP组(n=250),收集患者一般资料、实验室指标和各项评分[急性生理学与慢性健康状况评分Ⅱ(APACHEⅡ)、Ranson评分、CT严重指数(CTSI)评分和急性胰腺炎严重程度床边指数(BISAP)评分]等。通过Lasso回归分析筛选上述有统计学意义的变量,纳入多因素Logistic回归分析,以确定HTG-AP患者发生SAP风险的影响因素,构建预测HTG-AP重症化的Nomogram模型。通过分析其ROC曲线、校准曲线和DCA曲线,以评价模型的区分度、一致性和临床适用性。计算模型和各项评分的曲线下面积(AUC),运用Z检验比较各曲线间差异。结果年龄、甘油三酯-葡萄糖指数(TyG-index)、血清淀粉酶(AMY)、C-反应蛋白(CRP)、发热、急性胰周液体积聚(APFC)和急性坏死物积聚(ANC)为HTG-AP重症化的独立影响因素(P<0.05)。构建模型的回归方程为Logit(P)=-5.265-0.042X 1+0.003X 2+0.714X 3+0.009X 4+2.016X 5+1.771 X 6+2.239 X 7。模型经HL拟合优度检验(χ^(2)=6.445,P=0.598),Brier评分值为0.091。经Bootstrap重抽样方法进行内部验证,绝对误差为0.017。模型AUC为0.922(95%CI 0.888~0.956),大于APACHEⅡ评分、CTSI评分、BISAP评分,小于Ranson评分,差异均有统计学意义(P<0.05)。结论年龄、TyG-index、血清AMY、CRP、发热、APFC、ANC为HTG-AP重症化的独立影响因素;建立的Nomogram模型较传统评分系统有一定改善,具有良好的区分度、校准度和临床适用性。Objective To screen risk factors of severe hypertriglyceridemic acute pancreatitis(HTG-AP),to construct a Nomogram prediction model,and to evaluate its clinical value.Methods A total of 350 patients with HTG-AP admitted to the Second Affiliated Hospital of Chongqing Medical University from January 2017 to December 2020 were included in the study and divided into SAP group(n=100)and non-SAP group(n=250)according to the severity of the disease,and the general data,laboratory parameters,and various scores[APACHEⅡ,Ranson,CT severity index(CTSI)and bedside index for severity in acute pancreatitis(BISAP)score]were collected.The above statistically significant variables were screened by Lasso regression analysis and included in multivariate Logistic regression analysis to identify the influencing factors of the risk of developing SAP in HTG-AP patients for constructing a Nomogram model to predict the severity of HTG-AP.The discrimination,consistency,and clinical applicability of the model was evaluated by analyzing its ROC curve,calibration curve,and DCA curve.Finally,the areas under the ROC curve(AUC)of the model and each score were calculated,and the differences between each curve were compared by using the Z test.Results Age,triglyceride and glucose index(TyG-index),serum amylase(AMY),C-reactive protein(CRP),fever,acute peripancreatic fluid collection(APFC)and acute necrotic collection(ANC)were objective factors of severe HTG-AP(P<0.05).The regression equation for constructing the model was Logit(P)=-5.265-0.042X 1+0.003X 2+0.714X 3+0.009X 4+2.016X 5+1.771X 6+2.239X 7.The model was tested for goodness of fit by HL(χ^(2)=6.445,P=0.598),and the Brier score was 0.091.Internal validation was performed by Bootstrap resampling method with absolute error of 0.017.The model AUC was 0.922(95%CI 0.888-0.956),which was greater than the APACHEⅡscore,CTSI score,BISAP score,and less than the Ranson score,and the differences were statistically significant(P<0.05).Conclusions Age,TyG-index,serum AMY,CRP,fever,APFC and ANC are o

关 键 词：急性胰腺炎(AP) 高甘油三酯血症(HTG) 重症化 危险因素 预测模型 

分 类 号：R576[医药卫生—消化系统]