Population pharmacokinetic study of methotrexate in children with acute lymphoblastic leukemia.

A population pharmacokinetic model was developed to describe the factors that may affect the pharmacokinetics of methotrexate (MTX) in Chinese child patients with acute lymphoid leukemia (ALL) and to predict the individual pharmacokinetic parameters in these patients. One hundred and eighteen children with ALL who received MTX at the dose of 2 - 3.5 g/m(2) were enrolled in this study. 96 children were enrolled in the index group and 22 children in the validation group. The data were analyzed using nonlinear mixed effect model (NONMEM) software. A linear two-compartment model with linear elimination best described the data. The forward inclusion-backward elimination method was used to investigate the different covariates, including age, body weight, gender, etc. The Bayesian method was used to predict the individual pharmacokinetic parameters. Validation was applied using an internal and external approach. The population pharmacokinetic parameters and 95% confidence interval (CI) were obtained as follows: The clearance of central compartment (CL1), apparent volume of distribution of central compartment (V1), the clearance between central and peripheral compartment (CL2), and apparent volume of distribution of peripheral compartment (V2) were 5.04 (3.93 - 6.15) l/min, 16.1 (12.5 - 19.7) l, 0.203 (0.102 - 0.304) l/min and 7.05 (3.86 - 10.20) l, respectively. The inter-individual variability of CL1, V1, CL2, and V2 were 49.60%, 29.36%, 137.64%, and 107.70%, respectively. Gender, body surface area and the amount of alkalinization agent during 24 hours before MTX administration had significant effects on CL1. A strong relationship was found in this study between CL2 and age, as well as between V2 and age. A good correlation was further proved through the validation model. Moreover, some secondary pharmacokinetic parameters were estimated: the elimination half-life t1/2 was 2.34 h (CV = 36.7%), elimination constant k(e) was 0.33 h(-1) (CV = 33.2%), and the area under plasma concentration versus time curve AUC was 582.92 mg x h x l(-1) (CV = 55.9%). Our model combine Bayesian approach enabled a satisfactory estimation of MTX concentration in individual patients. The results of this study allowed clinicians to assess the MTX pharmacokinetic parameters based on the specific demographic characteristics of patients.