Purpose: This study explored factors affecting the pharmacokinetic variability of imatinib and CGP 74588, and the pharmacokinetic-pharmacodynamic correlations in patients with advanced gastrointestinal stromal tumors.
Experimental design: Thirty-five patients with advanced gastrointestinal stromal tumors received 400 mg of imatinib daily. Six blood samples were drawn: before intake, during 1- to 3- and 6- to 9-hour intervals after intake on day 1, and before intake on days 2, 30, and 60. Plasma imatinib and CGP 74588 concentrations were quantified by reverse-phase high-performance liquid chromatography coupled with tandem mass spectrometry, and analyzed by the population pharmacokinetic method (NONMEM program). The influence of 17 covariates on imatinib clearance (CL) and CGP 74588 clearance (CLM/fm) was studied. These covariates included clinical and biological variables and occasion (OCC = 0 for pharmacokinetic data corresponding to the first administration, or OCC = 1 for the day 30 or 60 administrations).
Results: The best regression formulas were: CL (L/h) = 7.97 (AAG/1.15)(-0.52), and CLM/fm (L/h) = 58.6 (AAG/1.15)(-0.60) x 0.55(OCC), with the plasma alpha1-acid glycoprotein (AAG) levels indicating that both clearance values decreased at a higher AAG level. A significant time-dependent decrease in CLM/fm was evidenced with a mean (+SD) CGP 74588/imatinib area under the curve (AUC) ratio of 0.25 (+/-0.07) at steady state, compared with 0.14 (+/-0.03) on day 1. Hematologic toxicity was correlated with pharmacokinetic variables: the correlation observed with the estimated unbound imatinib AUC at steady-state (r = 0.56, P < 0.001) was larger than that of the total imatinib AUC (r = 0.32, NS).
Conclusions: The plasma AAG levels influenced imatinib pharmacokinetics. A protein-binding phenomenon needs to be considered when exploring the correlations between pharmacokinetics and pharmacodynamics.