Establishment And Validation Of HM-AD Model In Decompensated Cirrhosis Without Acute-on Chronic Liver Failure:A Large Prospective Cohort Study

Background and aim: Cirrhotic patients with acute decompensation(AD)frequently develop acute-on chronic liver failure(ACLF),which is associated with high mortality rates.Recently,the CLIF Consortium Acute Decompensation score(CLIF-C AD score)has been developed to predict prognosis in AD patients without ACLF using the CANONIC study database.The aims of this study were to develop and validate a CLIF-C AD score based prognostic model by assessing new objective variables.Methods: A total of 1014 AD patients without ACLF were recruited in a prospective observational cohort study including one training and validation cohort.Cox proportional hazard analysis was performed in the training cohort to develop the novel CLIFC AD score(Hwamei acute decompensation score,HM-AD score)model.The performance of the new model was tested in the validation cohort by C index and a receiver operating characteristic curve(ROC).Results: During follow-up,91 deaths were reported,with a 3-month mortality of 9%.On multivariate cox regression analysis,bilirubin,hemoglobin and CLIF-C AD score were found to be significantly associated with 3-month mortality.The new model named HM-AD score had higher C index than CLIF-C AD score,CTP score,MELD score,MELD-Na score in predicting 3-month mortality(0.770 vs 0.721 vs 0.657 vs 0.677 vs 0.692,P<0.01).In the validation cohort of347 patients,HM-AD score remained better than other conventional models(0.782 vs 0.728 vs0.671 vs 0.639 vs 0.641 P<0.01).A risk stratification based on HM-AD model can be used for doctors to manage patients.Conclusions: A modified CLIF-C AD score named HM-AD score by incorporating bilirubin and hemoglobin has better predictive value of 3-month mortality than other conventional models.