Recently, a study of docetaxel in combination with the new orally administered P-glycoprotein (P-gp) inhibitor R101933 showed that this combination was feasible. However, due to the low oral bioavailability of R101933 and high interpatient variability, no further attempts to increase the level of P-gp inhibition were made. Here, we assessed the feasibility of combining docetaxel with intravenously (i.v.) administered R101933, and determined the disposition of docetaxel with and without the P-gp inhibitor. Patients received i.v. R101933 alone at a dose escalated from 250 to 500 mg on day 1 (cycle 0), docetaxel 100 mg/m2 as a 1-h infusion on day 8 (cycle 1) and the combination every 3 weeks thereafter (cycle 2 and further cycles). 12 patients were entered into the study, of whom 9 received the combination treatment. Single treatment with i.v. R101933 was associated with minimal toxicity consisting of temporary drowsiness and somnolence. Dose-limiting toxicity consisting of neutropenic fever was seen in cycles 1 and 2 or in further cycles at both dose levels. The plasma pharmacokinetics of docetaxel were not changed by the R101933 regimen at any dose level tested, as indicated by plasma clearance values of 22.5±6.2 l/h/m2 and 24.2±7.4 l/h/m2 (P=0.38) in cycles 1 and 2, respectively. However, the faecal excretion of unchanged docetaxel decreased significantly after the combination treatment from 2.5±2.1% to less than 1% of the administered dose of docetaxel, most likely due to inhibition of the intestinal P-gp by R101933. Plasma concentrations of R101933 were not different in cycles 0 or 2 and the concentrations achieved in the first 12-h period after i.v. infusion were capable of inhibiting P-gp in an ex vivo assay. We conclude that the combination of 100 mg/m2 i.v. docetaxel and 500 mg i.v. R101933 is feasible, lacks pharmacokinetic interaction in plasma, and shows evidence of P-gp inhibition both in an ex vivo assay and in vivo as indicated by the inhibition of intestinal P-gp.