HBV相关慢加急性肝衰竭预后评估模型的建立与验证

The establishment and verification of prognosis evaluation model for HBV related chronic and acute liver failure

目的探讨影响乙型肝炎病毒相关慢加急性肝衰竭(ACHBLF)预后的危险因素并建立预后模型。方法纳入277例ACHBLF患者,包括建模组和验证组。收集2组患者一般临床特征资料及肝功能、凝血功能和血常规等实验室检查结果,以及患者的生存情况。在建模组中采用单因素及多因素Cox回归分析ACHBLF患者预后的影响因素,并建立预后模型。应用受试者工作特征(ROC)曲线评估模型的预测价值,Hosmer-Lemeshow检验评价模型的拟合优度,并采用Kaplan-Meier生存曲线评估患者的生存情况。结果验证组血清总胆红素(TBIL)、直接胆红素和碱性磷酸酶低于建模组,年龄、男性比例、纤维蛋白原、淋巴细胞(LYMPH)、单核细胞、白细胞(WBC)高于建模组(P<0.05)。TBIL、WBC升高,LYMPH、凝血酶原活动度(PTA)降低是ACHBLF患者死亡的独立危险因素。构建新模型TPWL=1.059×TBIL-1.272×PTA+1.090×WBC-0.602×LYMPH,根据建模组ROC曲线确定TPWL模型的最佳临界值为-17.52,模型在建模组和验证组的符合率分别为91.75%和81.93%,ROC曲线下面积分别为0.961和0.914,Hosmer-Lemeshow检验结果P分别为0.210和0.062。Kaplan-Meier生存曲线分析显示TPWL评分≥-17.52患者生存率明显低于TPWL评分<-17.52患者。结论本研究建立的预测模型对评估ACHBLF患者预后有较好的价值。

Objective To investigate the risk factors affecting the prognosis of hepatitis B virus associated chronic acute liver failure(ACHBLF)and to establish a prognostic model.Methods A total of 277 ACHBLF patients were included in this study.The general clinical characteristics,liver function,coagulation function,routine blood test and the survival of the patients were collected.In the training cohort,univariate and multivariate Cox regression analysis was performed to evaluate the independent factors,and a new prognosis model was established.The prognostic prediction ability of the new model was evaluated in training cohort and validation cohort.Area under the receiver operating characteristic(ROC)curve was used to evaluate the predictive value of the models.Hosmer-Lemeshow test was used to evaluate the goodness of fit of the model.Kaplan-Meier survival curve was used to predict the prognosis of patients.Results The total bilirubin(TBIL),direct bilirubin and alkaline phosphatase were lower in the validation group than those in the model group.Data of age,proportion of male,fibrinogen,lymphocytes(LYMPH),monocytes and white blood cells(WBC)were higher in the validation group than those in the model group(P<0.05).The increased TBIL and WBC and the decreased LYMPH and prothrombin activity(PTA)were independent risk factors for ACHBLF death in the model group.The new model TPWL=1.059×TBIL-1.272×PTA+1.090×WBC-0.602×LYMPH and the optimal critical value of TPWL model was-17.52 according to the ROC curve of the model group.The coincidence rates of the model in the modeling group and the validation group were 91.75%and 81.93%,respectively.The areas under the ROC curve were 0.961 and 0.914,respectively.The Hosmer-Lemeshow test results showed that P values were 0.210 and 0.062,respectively.Kaplan-Meier survival curve analysis showed that the survival rate of patients with TPWL score≥-17.52 was significantly lower than that of patients with TPWL score<-17.52.Conclusion The predictive model established in this study has a high value in evaluating the prognosis of ACHBLF patients.