BACKGROUND AND AIM: Acute liver failure (ALF) is defined as acute liver injury (ALI) associated with coagulopathy. A follow-up strategy for ALI and characterization of ALI patients with a risk of progressing to ALF have never been established. To establish predictive markers for progression from ALI to ALF, this study compared the clinical characteristics and laboratory data on the day of registration to data from a regional referral system of patients with ALI. METHODS: This prospective, observational study enrolled 365 consecutive patients with ALI/ALF between 2007 and 2016. We evaluated 109 ALI patients, 27 of whom satisfied the ALF criteria during observation and another 82 patients who recovered without progression to ALF. RESULTS: Four patients died; all were in the ALF group. The variables of age, incidence of autoimmune hepatitis, model of end-stage liver disease score, values for total bilirubin and prothrombin time (PT)-international ratio, and Japan Hepatic Encephalopathy Prediction Model (JHEPM) probability at registration were significantly higher in ALF patients than in ALI patients. In multivariate analysis, PT and JHEPM were identified as risk factors for progression to ALF. The cut-off values of 13%, 4.9%, 65%, and 1.32% for the model of end-stage liver disease score, JHEPM probability, PT, and PT-international ratio values, respectively, had high negative predictive values. Furthermore, among patients whose JHEPM was underestimated, none died due to ALF. CONCLUSION: The JHEPM probability is a predictive parameter that can be used to decide a follow-up treatment strategy for ALI patients.
BACKGROUND AND AIM: Acute liver failure (ALF) is defined as acute liver injury (ALI) associated with coagulopathy. A follow-up strategy for ALI and characterization of ALI patients with a risk of progressing to ALF have never been established. To establish predictive markers for progression from ALI to ALF, this study compared the clinical characteristics and laboratory data on the day of registration to data from a regional referral system of patients with ALI. METHODS: This prospective, observational study enrolled 365 consecutive patients with ALI/ALF between 2007 and 2016. We evaluated 109 ALI patients, 27 of whom satisfied the ALF criteria during observation and another 82 patients who recovered without progression to ALF. RESULTS: Four patients died; all were in the ALF group. The variables of age, incidence of autoimmune hepatitis, model of end-stage liver disease score, values for total bilirubin and prothrombin time (PT)-international ratio, and Japan Hepatic Encephalopathy Prediction Model (JHEPM) probability at registration were significantly higher in ALFpatients than in ALI patients. In multivariate analysis, PT and JHEPM were identified as risk factors for progression to ALF. The cut-off values of 13%, 4.9%, 65%, and 1.32% for the model of end-stage liver disease score, JHEPM probability, PT, and PT-international ratio values, respectively, had high negative predictive values. Furthermore, among patients whose JHEPM was underestimated, none died due to ALF. CONCLUSION: The JHEPM probability is a predictive parameter that can be used to decide a follow-up treatment strategy for ALI patients.