Population pharmacokinetics (PK) of the vascular-disrupting agent 5,6-dimethylxanthenone-4-acetic acid (DMXAA) in cancer patients

Abstract

13076 Background: DMXAA (AS1404) is a small-molecule vascular-disrupting agent that selectively disrupts established tumor blood vessels. It is being evaluated in phase II trials in combination with cytotoxics. DMXAA is extensively metabolized, predominantly by UDP-glucuronosyltransferase (UGT) 2B7 and 1A9, and exhibits concentration-dependent protein binding. The study aim was to develop a population PK model for DMXAA and to examine covariates influencing drug disposition. Methods: 3,050 DMXAA total plasma concentrations were measured in 124 cancer patients receiving DMXAA monotherapy as a 20-min iv infusion weekly or every three weeks at doses from 6 to 4900 mg/m2 in 3 phase I trials. PK models were fitted to the data from all patients simultaneously using the program NONMEM. Potential covariates including age, body weight, height, body surface area (BSA), sex, and liver and kidney function tests were screened with general additive model analysis and tested in the PK model. Results: DMXAA concentration-time profiles after multiple-dose administration were well described by a 3-compartment model with saturable elimination (Michaelis-Menten kinetic) from the central compartment. BSA and sex were significant covariates on the volume of distribution of the central compartment (V1) and the maximum elimination capacity (Vm), respectively, accounting for 10 and 12% of interindividual variation (IIV) in these parameters. Population estimates for Vm, Km (concentration at which half-maximal elimination rate is achieved), and V1 were 122 × (1 + 0.502 × (2-SEX)) (μM/h) [females had lower Vm than males; SEX = 1 for males, 2 for females], 103 μM, and 8.15 × (BSA/1.8)0.69 (L), respectively. In the final model, IIV in Vm was 52%. Conclusions: DMXAA plasma disposition is characterized by a saturable elimination process. BSA-guided dosing for DMXAA is important. Evaluation of other covariates that reflect drug glucuronidation (e.g., UGT polymorphisms) may account for a greater fraction of the substantial unexplained PK variability. The population PK model will aid development of limited PK sampling strategies and investigation of PK-pharmacodynamic relationships in further trials. [Table: see text]