It is unclear from preliminary laboratory studies whether a high- or a low-dose interferon treatment strategy is optimal. As part of an ongoing study of mechanisms of interferon action, we have evaluated toxicity in a two-arm protocol in which patients were randomly assigned to receive lymphoblastoid interferon by either a low-dose treatment strategy (2 X 10(6) units/m2 daily X 28 days then daily X 5 days every other week by im injection) or a high-dose treatment strategy (5 X 10(6) units/m2 by continuous iv infusion over 24 hours, escalating by 5 X 10(6) units/m2/day as tolerated over 10 days, repeated every 28 days). The main toxic effects in both arms were fever, fatigue, and anorexia. Marked interpatient differences within each dose arm were greater than differences between arms. Additional significant toxic effects included nausea and vomiting, hypotension, leukopenia, thrombocytopenia, and evidence of hepatic toxicity. Minor changes in serum electrolytes were noted. Coagulation studies were normal. The dose-limiting toxic effect for the high-dose arm was myelosuppression. Median maximum tolerated dose among high-dose strategy patients was 18 X 10(6) units/m2, but there was marked interpatient variation. We conclude that both dose schedules were relatively well-tolerated. Because of individual variation in tolerance, high-dose treatment should include a dose escalation strategy.