A Phase II clinical trial was conducted to evaluate the efficacy of intravenous fluorouracil (5-FU) and subcutaneous recombinant interferon-alpha-2b (rIFN-alpha-2b) in the treatment of hepatocellular carcinoma (HCC) and to define factors that might be predictive of a response to treatment. Twenty-nine patients were registered on the protocol. 5-FU was administered as a continuous intravenous (i.v.) infusion (dose = 750 mg/m2) for 5 consecutive days. rIFN-alpha-2b was administered subcutaneously (SC) (dose = 5 x 10(6) um/m2) once a day on days 1, 3, and 5 of the 5-FU infusion. The treatment was repeated at 14-day intervals. Responses were assessed at the end of one course of therapy, which was equivalent to four treatments. Of the 28 patients evaluable for response, 5 (18%) had a partial response, and 1 (4%) had a minor response. Responses lasted from more than 2 to more than 24 months (median, 11.5 months). Ten (36%) patients experienced no response, and 12 (43%) had progressive disease. The 6 responders were part of a group of 16 patients who had pretreatment levels of serum alpha-fetoprotein (AFP) of 50 ng/ml or less and a group of 8 whose tumors involved 50% or less of the liver parenchyma. Mucositis, which occurred in 54% of the patients, was the most common toxicity associated with the treatment regimen. Diarrhea and dermatitis were observed in 16% and 17% of the patients, respectively; fatigue, thrombocytopenia, granulocytopenia, neurologic toxicity, and nausea and vomiting were not commonly seen. The regimen of i.v. 5-FU and SC rIFN-alpha-2b was well tolerated and induced durable partial response in 31% (5 of 16) of patients with HCC who had low levels of serum AFP and in those with 50% or less of liver replacement. In contrast, the treatment regimen was ineffective in patients with HCC who had high levels of serum AFP or extensive liver disease.