原发性胆汁性肝硬化发生失代偿的预测

目的建立原发性胆汁性肝硬化(PBC)发生失代偿的预测模型,验证并判断其预测价值。方法回顾性分析113例确诊时处于代偿期的PBC患者的人口统计学、实验室检查、临床表现及其他预后模型(Child-Pugh、MELD、Mayo模型)积分,研究终点为发生腹水、肝性脑病、食管胃底静脉曲张出血等失代偿。应用SPSS16.0统计软件,采用多因素(Cox回归、Kaplan-Meier(K-M)等方法建立发生失代偿的预测模型,采用接受者工作特征(ROC)曲线下面积比较所建模型与以往其他模型对PBC发生失代偿的预测价值。结果随访中位数时间31.2个月(3.37～122.43个月)期间,有21例(18.58%)患者达研究终点。所建立的PBC发生失代偿的预测模型(即D-PBC模型)指标包括AST/ALT比值、碱性磷酸酶(ALP)、胆碱酯酶(CHE)和血小板(PLT),PI=0.862×AST/ALT+0.003×ALP(U/L)-0.293×CHE(kU/L)-0.011×PLT(×10~9/L)。与其他模型相比,该预测模型的ROC曲线下面积较大,采用PI>-1.41预测PBC发生失代偿的敏感性高达0.91。结论 D-PBC模型能准确预测代偿期PBC患者临床失代偿的发生。

Prediction of decompensation in patients with primary biliary cirrhosis

Objective To construct the prediction model of decompensated primary biliary cirrhosis(PBC).Methods The clinical characteristics of 113 patients with compensated PBC were retrospectively studied,including demographics, laboratory tests,clinical manifestations,as well as Child-Pugh,model for end-stage liver disease(MELD) and Mayo risk scores.The end-point was the development of clinical decompensation,including ascites,hepatic encephalopathy and/or variceal bleeding.Cox regression and Kaplan-Meier(K-M) method were used to construct a model to predict the decompensation of PBC(D-PBC model).The areas under the receiver operating characteristic(ROC)curves were calculated to compare the predictive accuracy between D-PBC model and three other prognostic models.Results During the follow-up with a median time of 31.2 months(3.37-122.43 months),there were 21 patients(18.58%) reaching the endpoint.The prediction model was consisted of several routine biomarkers including aspartate aminotransferase/alanine aminotransferase ratio(AST/ALT),alkaline phosphatase(ALP),cholinesterase(CHE) and platelet(PLT):PI = 0.862×AST/ALT + 0.003 x ALP(U/L) - 0.293×CHE(kU/L) - 0.011×PLT(×10~9/L).The area under ROC curve of D-PBC model was significantly larger than those of other prognostic models,with a sensibility of 0.91 to predict the incidence of clinical decompensation when PI was > - 1.41.Conclusion D-PBC model could accurately predict clinical decompensation in patients with PBC.