OBJECTIVE An accurate and early assessment of the individual patient is critical in deciding whether liver transplantation is indicated in the treatment of fulminant hepatic failure. Based on analysis of patients treated between 1973 and 1985, the Liver Unit at King's College Hospital, London, developed a prognostic model to identify patients with a poor prognosis. The present study was done to determine the applicability of this model in fulminant hepatic failure patients seen at our center in the 1990s. METHODS The records of 145 patients with fulminant hepatic failure, treated conservatively at Queen Elizabeth Hospital, Birmingham between 1990 and 1994, were analyzed. An additional 81 patients, who were transplanted for fulminant hepatic failure during the same period were excluded from the study. RESULTS Application of King's College Hospital criteria at the time of admission to this hospital in the acetaminophen group, had a positive predictive value of 88%, negative predictive value of 65% and predictive accuracy of 71%. The positive predictive value, negative predictive value and predictive accuracy of these criteria for non-acetaminophen-induced fulminant hepatic failure, were 79%, 50% and 68%. Multivariate analysis identified prothrombin time, serum creatinine, white cell count and abnormal potassium levels as independent predictors of mortality in acetaminophen-induced fulminant hepatic failure; and prothrombin time alone in fulminant hepatic failure induced by other etiologies. CONCLUSIONS The King's College Hospital criteria for predicting outcome of fulminant hepatic failure were found to have a slightly lower predictive accuracy than shown in the original study.