Data reported by Bernuau et al. have strongly supported the measurement of coagulation factor V as the best prognostic indicator in fulminant hepatic failure (FHF) and as the test on which selection for urgent liver transplantation should be made. In this study, we have measured plasma factor V in 110 patients with FHF, in grades I-IV coma, in 88 of whom the etiology was acetaminophen overdose. On admission, patients who did not survive had significantly lower factor V levels (median, 5%; range, 1-27; n = 49), compared with those who did (median, 10%; range, 2-70; P < .001). In the 81 patients with acetaminophen-induced FHF who did not receive a transplant, there was no cutoff level of factor V that clearly separated the patients. On statistical analysis, a positive predictive value (the mortality in patients predicted to have a poor prognosis) of 0.49 was calculated for factor V <20% and 0.57 for factor V < 10%. If the prognostic criteria included deep coma (grades III and IV) as well as factor V <20%, a positive predictive value of 0.73 was calculated. This compared with a value of 0.92 for the well-established King's prognostic criteria based on pH, and a combination of international normalized ratio (INR), renal failure, and coma. In the 17 mixed, nonacetaminophen group of patients who did not receive a liver graft, the positive predictive value was 0.85 for a factor V level <20% and 1.00 for factor V <10%, compared with 0.93 for the King's criteria for that etiologic group. This study demonstrates that the predictive accuracy of plasma factor V level is much less effective than the well-validated King's criteria in the selection of patients with acetaminophen-induced FHF needing liver grafting, although it may be useful in patients with FHF due to other causes.