Bayesian Estimation of Methotrexate Pharmacokinetics in Children with Acute Lymphoblastic Leukaemia and Prediction of Folinic Acid Rescue
I. Grabnar(1), B. Faganel Kotnik(2), P. Jovic(1), M. Pislar(1), V. Dolzan(3), J. Jazbec(2)
(1)Faculty of Pharmacy, University of Ljubljana, Ljubljana, Slovenia; (2)Department of Oncology and Hematology, University Children’s Hospital Ljubljana, Ljubljana, Slovenia; (3)Institute of Biochemistry, Faculty of Medicine, University of Ljubljana, Ljubljana, Slovenia
Objectives: Acute lymphoblastic leukaemia is the most common childhood cancer and high dose-methotrexate (MTX) is still the mainstay in the chemotherapy. Due to high variability in pharmacokinetics and well established relationship between treatment toxicity and MTX exposure, TDM and adaptive folinic acid rescue are essential for clinical management of patients. We have previously studied the association of genetic polymorphism in the folate metabolic pathway with MTX pharmacokinetics and toxicity [1]. The aim of the present study was to evaluate Bayesian predictions of MTX concentrations in the elimination phase based on scarce samples.
Methods: Routine TDM data from 64 patients were used. Four courses of MTX treatment were infused over 24 hours and blood samples were collected at 24, 36, 42 and 48 hours. Additional samples were obtained, if MTX concentration at 48 hours was above 0.5 μmol/L. Pharmacokinetic parameters of MTX were estimated by NONMEM VI.
Additional data from 37 patients were used for evaluation of prediction performance of the Bayesian method. MTX conentrations at 48 h were predicted from measured concentrations at 24 and 42 h and predictive performance of the model was evaluated by comparisson with the actual concentration measurements at 48 h.
Results: Pharmacokinetic data from 252 courses and 919 MTX concentration measurements were available for analysis. MTX concentration profiles were fitted with a two compartment model. In a typical patient CL was estimated at 7.12 L/h (IIV 31%), volumes of the central and peripheral compartment were 9.73 (IIV 6%) and 3.61 (IIV 63%) L, respectively and distribution clearance was 0.134 L/h (IIV 65%). IOV in CL was estimated 14.8%. Residual variability was estimated at 0.0642 μmol/L (additive component) and 68.1% (proportional component).
In the independent group of patients bias of the predicted concentrations at 48 h was 10% with precision of 77%. Dosage adjustment of folinic acid rescue based on predicted MTX concentration was accurate in 81% of courses.
Conclusion: Bayesian estimation is a useful tool for prediction of MTX concentration in the elimination phase and can be used for adjustments of folinic acid dosing.
References:
[1] Faganel Kotnik B, Dolzan V, Grabnar I, Jazbec J. Relationship of the reduced folate carrier gene polymorphism G80A to methotrexate plasma concentration, toxicity, and disease outcome in childhood acute lymphoblastic leukemia. Leuk Lymphoma. 2010;51(4):724-6.
