Arterial ketone body ratio (AKBR [acetoacetate/beta-hydroxybutyrate]) was measured in nineteen patients under medical supportive therapy for fulminant hepatic failure (FHF), in order to evaluate its predictive value relative to liver transplantation. Of the 19 patients 8 (42%) were salvaged and 11 (58%) died. Seven of 8 survivors showed an increased AKBR over 0.6 at 24-hr after admission, and all of them showed AKBR over 0.8 at 48-hr with subsequent maintenance of the value over 1.0. By contrast, all 11 nonsurvivors demonstrated sustained suppression of AKBR below 0.4 from 24 to 72 hr after admission. AKBR values at 24 and 48 hr showed statistically significant differences between survivors and nonsurvivors. Neither the grade of portal systemic encephalopathy (PSE) nor other conventional laboratory parameters--such as AST, bilirubin, ammonia, prothrombin time, hepaplastin test, fibrinogen, and platelet count--could discriminate between survivors and nonsurvivors by univariate analysis. These results indicate that AKBR can accurately predict the prognosis of FHF at the initial 24-48 hr after admission, and that it can play an important role in setting the indication of FHF for liver transplantation.