The realization in the 1980s that liver transplantation was feasible in acute liver failure created a need for prognostic indicators to determine the subset of patients most likely to benefit from that intervention. Although many clinical and investigational parameters have been shown to reflect prognosis in acute liver failure, only a few come close to fulfilling the fundamental requirements of a prognostic model in this context-i.e., ease of evaluation, early applicability, and accuracy. The King's College criteria were described in 1989 and created separate models for patients who had and who had not received acetaminophen. 1 The parameters tested were basic clinical and laboratory data, resulting in simple-to-use and universally applicable models. The 5 variables included in the group that did not receive acetaminophen were etiology, age, time from jaundice to encephalopathy, serum bilirubin, and coagulation parameters (prothrombin time or international normalized ratio [INR]). More recently, Model for End-Stage Liver Disease (MELD) scores, which have been introduced and validated for use in prognostication in chronic liver disease, were used to assess prognosis in non-acetaminophen-related acute liver failure. 2,3 The 3 variables that contribute to MELD score are serum bilirubin, serum creatinine, and INR.
In this issue of Liver Transplantation, Dhiman and colleagues describe another prognostic model and compare its performance to the King's College criteria and MELD scores in a population of patients with proven (mainly hepatitis B) or presumed acute viral hepatitis. 4 They identified 6 parameters: age ΟΎ50 years, jaundice to encephalopathy time greater than 7 days, grade 3-4 encephalopathy, cerebral edema, prothrombin time ΟΎ35 seconds, and serum creatinine ΟΎ1.5mg/dL. Any 3 of these factors indicates a poor prognosis. A comparative evaluation suggested that this model was better than both the MELD