Seventeen patients with advanced, previously treated malignancies were entered into a phase I trial utilizing recombinant DNA produced alpha 2 leukocyte interferon (rIFN-alpha 2). Sixteen patients were evaluable. Patients were to receive rIFN-alpha 2 by either the I.V. or I.M. route for 35 consecutive days. The dosage was identical by both routes, and patients were escalated from 3 X 10(6) to 10 X 10(6) to 30 X 10(6) to 50 X 10(6) and to 100 X 10(6) I.U. every 7 days. No patient was able to tolerate the consecutive treatment protocol as planned. Dose-limiting toxicity was a flu-like syndrome in 10 patients and was usually associated with a fall in performance status. Confusion resulted in study withdrawal for five patients, four receiving rIFN-alpha 2 by the I.M. route. Hematologic and liver function abnormalities were common, usually transient, and not associated with clinical sequelae. One patient with non-Hodgkin's lymphoma showed substantial improvement; otherwise, all had stable or progressive disease. Pharmacologic studies indicated substantial serum levels at doses greater than or equal to 10 X 10(6) I.U. regardless of route. No consistent changes in NK activity, lymphocyte subpopulations, or immunoglobulin levels were noted, and no patient developed antibodies to rIFN-alpha 2. The dose and schedule used here indicate that high levels of serum rIFN-alpha 2 activity are obtainable by either the I.M. or I.V. route. Intermittent rather than daily dosage is more likely to be better tolerated and should be considered for phase II trials.