Establishment And Evaluation Of Short-term Prognosis Model Of Acute-on-chronic Liver Failure

Objective:To explore the factors affecting the short-term(90 d)prognosis of patients with Acute-on-chronic liver failure(ACLF),and to establish a new predictive model.Methods:A total of 159 patients with ACLF were divided into two groups:death group(n=87)and survival group(n=72).The clinical data that may affect the prognosis were collected and compared.The independent risk factors affecting clinical outcome were screened by multivariate Logistic analysis,and a new predictive model was established.The ROC curve is used to evaluate and predict the model,and the DeLong method is used to compare the area under the ROC curve(AUC)。Results:Multivariate logistic regression analysis showed that TBil,Alb,Na~+and HE were independent prognostic factors of ACLF(OR=1.007,P=0.001;OR=0.884,P=0.010;OR=0.906,P=0.021;OR=21.879,P=0.000);The HTNA model was established jointly by TBil,Alb,Na~+and HE.The AUC(95%CI)of HTNA model was 0.835(0.768-0.889),and the AUC(95%CI)of MELD model was 0.638(0.558-0.713).The difference between them was statistically significant(Z=4.282,P<0.0001),Compared with MELD,the HTNA had significantly accuracy(94.44%VS69.44%)and Yoden index(0.53 VS 0.28).Conclusion:TBil,Alb,Na~+and HE are independent risk factors for short-term prognosis in patients with ACLF.The HTNA model established by the four factors can predict the clinical outcome more accurately and has better predictive value.