Development And Validation Of A Risk Prediction Model For Progressing To Acute On Chronic Liver Failure In Patients With Acute Exacerbation Of Chronic Hepatitis B Patients

Objective:To develop and validate a risk prediction model for progressing to acute on chronic liver failure(ACLF)in patients with acute exacerbation(AE)of chronic hepatitis B(CHB)patients.Methods : 474 patients with AE of CHE patients admitted to the Department of Infectious Diseases of Affiliated Hospital of Zunyi Medical College from January 2011 to February 2017 were retrospectively analyzed.252 patients were assigned to risk-model derivation and the other 222 patients to model validation.Univariate risk factors associated with the development of ACLF were entered into a multivariate logistic regression model for screening independent risk factors.The risk prediction model was established based on independent risk factors.The predictive value of our model was evaluated by the receiver operating haracteristic(ROC)curve and the area under the curve(AUC).The predictive value of MELD(model for end-stage liver disease)and MELD-Na were also evaluated in this study.Results:1.474 patients with AE of CHB patients were included in this study.36 patients developed to ACLF after admission,including 28 males and 8 females,the average age was 43.33 ± 10.29 years old,other 438 patients did not progress to ACLF,including 371 males and 67 females,the average age was 37.68 ± 11.26 years old;2.Univariate risk factors associated with the development of ACLF were high serum levels of HBV DNA,indirect bilirubin(IBIL)and total bilirubin(TBIL),low serum levels of cholinesterase(CHE),prealbumin(PA),gamma-glutamyl transpeptidase(GGT)and uric acid(UA),prolonging of prothrombin time(PT),activated partial prothrombin time(APTT)and prothrombin activity(PTA),and high international normalized ratio(INR)level;3.Logistic regression analysis showed that HBV DNA and INR are the independent risk factors associated with the development of ACLF,the Logistic regression model was Y=-14.049+0.687×lgHBV DNA+4.798×INR.4.The AUC of risk prediction model was 0.764,it is higher than that of MELD(0.663)and MELD-Na(0.665).Our risk prediction model also had higher sensitivity(0.722),specificity(0.791)and positive predictive value(0.289),it also had higher positive likelihood ratio(3.51)and negative likelihood ratio(0.35)than those of MELD and MELD-Na.Conclusions:The risk factors for patients with AE of CHB patients progressing to ACLF include high serum levels of HBV DNA,IBIL and TBIL,low serum levels of CHE,PA,GGT and UA,and coaulopathy;the independent risk factors are HBV DNA and INR.Our risk prediction model show high sensitivity and specificity in predicting the development of ACLF in patients with AE of CHB patients.