POPULATION PHARMACOKINETICS OF HIGH-DOSE ZIDOVUDINE IN A PHASE-I CANCER STUDY

High doses of zidovudine, fluorouracil and calcium folinate have been simultaneously administered as a weekly 24-hour infusion in a chemothe rapeutic protocol for cancer patients with solid tumours. The establis hed therapeutic regimen, fluorouracil and calcium folinate, acts by in hibiting thymidylate synthase in tumour tissues. The theoretical basis for incorporation of zidovudine into this chemotherapeutic regimen is to interfere with the thymidine salvage pathway lending to more effec tive depletion of thymidine nucleotide levels and enhanced cytotoxicit y to rapidly proliferating cells. Patient plasma sampling was performe d prior to termination of the 24-hour infusion to assess steady-state levels and during a 4-hour postinfusion period to observe the decay ki netics. Samples were assayed for zidovudine by high-performance liquid chromatography. Pharmacokinetic profiles were evaluated for 11 patien ts administered doses at escalating levels in the range of 7 to 15 g/m (2) of zidovudine (6 patients received 3 dose levels, 1 patient receiv ed 2 levels, and 4 patients received only 1 level). The data was analy sed using NONMEM, a computer program for population pharmacokinetic an alysis. A 2-compartment pharmacokinetic model with parallel first-orde r and Michaelis-Menten elimination was used as the structural model, a nd a constant variance model was used to model intraindividual error. First-order-clearance was modelled as being proportional to a covariat e (serum creatinine clearance). Population estimates for volumes of th e central and peripheral compartments were 24.3 and 62.5L, respectivel y. Estimates of first-order clearance and intercompartmental clearance in Wh were (0.436 x serum creatinine clearance) [in ml/min] and 50.2, respectively. The Michaelis-Menten constants, V-max and K-m, were 22. 0 mg/L/h and 9.28 mg/L, respectively, where V-max is the maximum veloc ity of the capacity-limited process and K-m is the concentration at ha lf the maximum velocity. The 2-compartment parallel elimination model described provided a good fit for the nonlinear kinetics observed for zidovudine at high doses. NONMEM I was found to be useful for pharmaco kinetic analysis of a small population with dense sampling.