Gemcitabine (GEM) is a novel deoxycytidine analogue which has shown promising antitumor activity in solid tumor models and a broad range of schedule-dependent MTDs (12-4560 mg/m(2)) in preliminary clinical studies. The present phase I trial evaluated escalating doses of weekly GEM using a 30-min infusion at a starting dose-level of 300 mg/m(2)/wk x 3 every 28 days. At least 3 patients entered each dose-level step and 3 more cases were treated when significant toxicity was seen. A total of 39 patients with various advanced solid tumors and prior chemotherapy entered this study. Six escalation steps (102 courses) were tested to define the MTD at 1,370 mg/m(2)/wk. No definite dose-effect relationships were observed for myelosuppression up to 1,095 mg/m(2)/wk. However, increased severity of leucopenia (dose-limiting) and greater non-hematologic toxicity as well as a higher number of toxic treatment delays, requiring subsequent dose attenuation in 6 out of 12 patients, were observed at 1,370 mg/ m(2)/wk. In all, 6 out of 11 patients experiencing WHO grade greater than or equal to 3 toxicity (11/21 events recorded in 11/18 courses) were treated at the MTD. Clinically significant toxicity included (patients with WHO grade 2-3): leucopenia (44%), thrombocytopenia (26%), anemia (23%), fever (69%), emesis (38%) and AST/ALT rise (26%). Mild proteinuria, ankle edema, skin rash, hair loss and mucositis were seen in less than or equal to 5%. The good tolerability and the evidence of antitumor activity of GEM at doses greater than or equal to 875 mg/m(2)/wk (1 CR and 3 PRs in 15 bladder cancer patients) encourage further phase II studies at much higher dose-levels (1,370 mg/m(2)) than previously suggested.