This study's objectives were to determine the maximum tolerated dose (MTD) of 9-aminocamptothecin (9-AC), given as a prolonged continuous infusion (CI) for 7-21 days, when formulated in dimethylacetamide/polyethylene glycol 400 (DMA) and then later as a colloidal dispersion (CD), and to determine the steady-state pharmacokinetics of 9-AC. Patients with solid tumors refractory to standard therapy were enrolled on this study. Total dose/cycle of 9-AC/DMA was initially escalated by duration (7-21 days), while keeping the dose rate constant at 6.2 microg/m/h (1.04-3.12 mg/m/4-week cycle). Then, the dose rate was escalated from 6.2 to 21.1 microg/m/h (3.12-10.6 mg/m/4-week cycle) while keeping the infusion duration constant at 21 days. CD formulation was escalated from 14.1 to 25 microg/m/h (7.11-12.60 mg/m/4-week cycle) while keeping the infusion duration constant at 21 days and then escalated from 28.1 to 37.5 microg/m/h (9.44-12.60 mg/m/3-week cycle) while keeping the infusion duration constant at 14 days. Sixty-two patients were evaluable for toxicity; 61 received prior chemotherapy (median 3 regimens/patient). No consistent dose-limiting toxicity (DLT) was encountered with the DMA formulation until dose level 10.60 mg/m/cycle, when two patients experienced DLTs. With the 21-day CD formulation, the MTD was 12.60 mg/m/cycle with three DLTs out of five patients. When 9-AC was given on the 14-day schedule, DLT was seen at 9.44, 11.20 and 12.60 mg/m/cycle, with consistent DLT at the two highest dose levels. All DLTs for both formulations were grade 4 hematologic toxicities (neutropenia and/or thrombocytopenia), while non-hematologic toxicities were relatively mild (including gastrointestinal toxicities and fatigue). One patient with ovarian cancer had a complete response and three had partial responses (PRs). One patient each with non-Hodgkin's lymphoma and cancer of unknown primary had a PR. Pharmacokinetic studies of both formulations of 9-AC revealed a linear relationship between increasing plasma 9-AC lactone concentration and dose. The median plasma 9-AC lactone concentration for 9-AC/CD was approximately twice that achieved by 9-AC/DMA for the same dose level. Both 9-AC formulations, given as a 21-day CI, were well tolerated with dose-limiting myelosupression at the MTD. This dose intensity exceeds that of other 9-AC phase I/II schedules. The recommended phase II dose (RPTD) is 9.42 mg/m/4-week cycle, given as a 21-day infusion. The 14-day schedule of 9-AC/CD was equally myelosuppressive with the RPTD of 9.44 mg/m/3-week cycle, although two heavily pre-treated patients (one with pelvic radiotherapy) could not tolerate this dose. Objective responses were observed in six out of 57 heavily pre-treated patients, most of which had ovarian cancer.