Interleukin 6 (IL-6) has antitumor activity comparable to IL-2 in murine models with less toxicity. Because the biological effects of intermittent and continuous infusions may differ, we conducted two concurrent Phase I trials of daily x5, 1-h, and continuous 120-h i.v. infusions to determine the toxicity, biological effects, and maximum tolerated dose of i.v. IL-6. Cohorts of six patients with advanced cancer received escalating doses (1, 3, 10, 30, 100, and 150 microgram/kg/day) of recombinant human IL-6 on days 1-5 and 8-12 of each 28-day course (1-h trial) or on days 1-5 of each 21-day course (120-h trial). Treatment was administered in regular inpatient wards and in outpatient clinics and was withheld in the event of grade 3 toxicity. Sixty-nine patients (1-h trial, n = 40; 120-h trial, n = 29) were enrolled, including 27 with renal cancer and 16 with melanoma. All were ambulatory, and 40 were asymptomatic. Fever (97%), anemia (78%), fatigue (56%), nausea or vomiting (49%), and elevated serum transaminase levels (42%) were the most frequent toxicities. Transient hypotension developed in 23 patients (33%). There were three deaths during the study due to progressive disease and/or infection. There were no objective responses. Dose-related increases in platelet counts and C-reactive protein levels were detected in most patients. Principal dose-limiting toxicities included atrial fibrillation (1 episode in the 1-h trial and 4 episodes in the 120-h trial) and neurological toxicities (3 episodes in the 1-h trial and 4 episodes in the 120-h trial). The neurological toxicities included confusion, slurred speech, blurred vision, proximal leg weakness, paraparesis, and ataxia. These effects were transient and reversed when IL-6 was discontinued. IL-6 can be given by i.v. infusion at biologically active doses with acceptable toxicity. Dose-limiting toxicities consisted mainly of a spectrum of severe but transient neurological toxicities and occasional episodes of atrial fibrillation. The maximum tolerated doses recommended for use with these i.v. schedules in Phase II trials are 100 microgram/kg/day by daily x5 1-h infusion and 30 microgram/kg/day by 120-h infusion. Phase II trials will be performed to determine the antitumor activity of IL-6 and better define its toxicity. Patients in these and other IL-6 studies should be monitored closely for neurological and cardiac effects.