We have previously shown that the combination of 5-fluorouracil (5-FUra) and recombinant alpha-2a-interferon (rIFN-alpha-2a) produced objective responses in 23 of 32 (63%) previously untreated patients with advanced colorectal carcinoma. Because in vitro data suggest that rIFN-alpha-2a modulates the cytotoxic effects of 5FUra in a concentration-dependent manner, a phase I clinical trial was initiated to determine the maximum tolerated dose of rIFN-alpha 2a when administered in combination with 5FUra. A total of 27 patients with advanced colorectal carcinoma were enrolled. The median age was 64 years, and the median performance status was 1. A total of 18 patients had no prior chemotherapy and 19 no prior 5FUra. 5FUra was administered at 750 mg/m2/day by continuous i.v. infusion for 5 days, followed by weekly bolus therapy. rIFN-alpha 2a was administered at 6, 9, 12, 15, or 18 x 10(6) units s.c. beginning on day 1. The dose-limiting toxicity of this regimen was fatigue, resulting in a decrease in performance status, and this was the only toxicity that correlated with increasing dose of rIFN-alpha 2a. Eastern Cooperative Oncology Group grade 3-4 toxicities included leukopenia (6), thrombocytopenia (2), anemia (4), stomatitis (4), diarrhea (4), neurological (2), infection (2), and allergy (2). Three quarters of the patients required interruption of therapy or dose reductions of either 5FUra or rIFN-alpha 2a for toxicity. Among the patients with measurable disease who were previously untreated with 5FUra, 5 of 9 at the lowest dose levels achieved an objective response, including one pathological complete responder, whereas 0 of 9 at the three highest dose levels responded. Among patients previously treated with 5FUra, only 1 achieved an objective response. We conclude that the maximum tolerated dose of rIFN-alpha 2a, when administered with 5FUra as above, is 15-18 x 10(6) units; however, the efficacy of this regimen does not appear to be related to the dose intensity of rIFN-alpha 2a, and future regimens should employ a lower dose, intermittent schedule of rIFN-alpha 2a, which may be better tolerated.