The pharmacokinetics of CI-921 were studied after 65 infusions over a 20-fold dose range (13-270 mg/m2 per day) in 16 patients during a phase 1 trial. CI-921 was given by a 15 min infusion on three consecutive days. Plasma samples were collected after the first and third infusions, and urine, at 6 h intervals throughout the 3 days. CI-921 concentrations were measured by an HPLC method. Maximum plasma concentrations ranged from 3-86 mumol/l. The plasma concentration-time disposition curves were mainly biphasic over the 24-h postinfusion period. There was no significant difference by the paired t-test between the Cmax, AUC, CL, Vss, MRT, t1/2 alpha, or t1/2 beta calculated for the first and third infusions. The means (range) of model-independent pharmacokinetic parameters were: CL, 158 (94-290) ml/h per kg; Vss, 319 (219-614) ml/kg; MRT, 2.1 (1.1-3.5) h; t1/2 alpha, 0.5 (0.2-1.1) h; and t1/2 beta, 2.6 (1.1-5.0) h. There was a strong linear correlation between the dose and the AUC and Cmax, suggesting linear kinetics over this dose range. A very small amount (less than 1%) of the total dose was excreted as unchanged CI-921 in the urine, mostly in the 12-h postinfusion period.