Selection of patients with acute hepatic failure for liver transplantation remains difficult, and there is no definite proof of a survival effect. We therefore did a retrospective study in 75 consecutive patients referred over a 12-year period. In two-thirds we identified a cause, mostly viruses or drugs. Patients were grouped by the Clichy and King's College criteria. In 20 there was no indication for transplantation. Of the 5 with autoimmune hepatitis, 3 died, significantly differing from the other 15 ( P = 0.009). The remaining 55 met our criteria, except 1. All 9 patients with absolute contraindications died. Of the 46 enlisted, 7 died without transplantation. One-year survival after transplantation was 69%, compared with 58% by "intention to treat." For patients enlisted, transplantation reduced mortality by 78% ( P = 0.069). The Clichy and King's College criteria reliably predict survival without transplantation, except in autoimmune hepatitis. Our study strongly suggests that transplantation improves survival.