Acute liver failure carries a high rate of mortality, but if metabolic support can be maintained for a critical period, liver healing and recovery are possible. Current techniques of temporary hepatic support are cumbersome and inconsistently effective. We studied the ability of dispersed hepatocytes to provide metabolic support when transplanted to rats with liver failure induced by dimethylnitrosamine (DMNA), a rapidly metabolized agent that is selectively toxic to liver cells. DMNA (20 mg/kg) was administered intravenously to 92 Lewis rats. Animals were divided into four groups receiving the following treatments 24 hours after DMNA administration: group I-intraperitoneal transplantation of hepatocytes prepared from 2.0 gm of normal isologous rat liver; group II-infusion into the portal vein of hepatocytes prepared from 1.5 gm of liver; group III-infusion of saline into the portal vein; group IV-no further treatment. The percentages surviving in each group 3 weeks after DMNA administration were 63%, 71%, 17%, and 6%, respectively. Mean serum glutamic oxaloacetic transaminase (SGOT) levels 3 days after DMNA administration were similar in the four groups, indicating that the degree of liver damage was equivalent. A significantly higher proportion of hepatocyte treated rats survived. Liver histology after DMNA administration showed hemorrhagic central lobular necrosis. A return to near-normal architecture occurred by 3 weeks in surviving animals. In group II hepatocytes were seen in portal venules, sinusoids, and central veins. We conclude that dispersed hepatocytes, transplanted either intraperitoneally or via the portal vein, can provide sufficient metabolic support to allow for recovery from drug-induced hepatic necrosis.