Model-informed precision dosing of high-dose IV busulfan in Thai pediatric patients
Félicien Le Louedec (1), Apichaya Puangpetch (2), Usanarat Anurathapan (3), Samart Pakakasama (3), Suradej Hongeng (3), Etienne Chatelut (1), Fabienne Thomas (1), Chonlaphat Sukasem (2)
(1) Laboratoire de Pharmacologie, Institut Claudius Régaud, Institut Universitaire du Cancer de Toulouse Oncopole, Centre de Recherche en Cancérologie de Toulouse, INSERM U1037, Université Paul Sabatier, Toulouse (France) ; (2) Division of Pharmacogenomics and Personalized Medicine, Department of Pathology, Faculty of Medicine Ramathibodi Hospital, Mahidol University, Bangkok 10400 (Thailand) ; (3) Division of Hematology-Oncology, Department of Pediatrics, Faculty of Medicine Ramathibodi Hospital, Mahidol University, Bangkok 10400 (Thailand)
Objectives: Intravenous busulfan is a corner stone in conditioning regimens for pediatric hematopoietic stem cell transplantation. It is administered daily with the objective to target a given cumulative AUC over 4-5 days. Currently, therapeutic drug monitoring (TDM) based on seven blood samples is performed at day 1 and 3 to control the AUC. The objectives of the current work are to identify (1) a covariate-based formula to predict busulfan clearance (CL) for a priori dose individualization, (2) an optimal model-based TDM strategy for a posteriori dose adjustments, both tailored for pediatric Thai patients.
Methods: Busulfan concentration data collected during TDM of patients treated in Ramathibodi Hospital (Bangkok, Thailand) were modeled with a population approach (NONMEM 7.4, Icon PLC). Several methods for modeling the influence of the size of the patients were investigated: power model with the actual body weight (BW), power model with a varying exponent [1], or calculation of the normal fat mass [2]. The influence of the following variables was screened with a stepwise covariate modeling procedure: actual age, age transformed with a maturation function [1,2], sex, malignant disease (MALIGN), fludarabin co-administration, genetic polymorphism of Glutathione S-transferase Alpha-1 (GSTA1, rs3957357 & rs3957356), of GST Pi-1 and copy number variation of GST Mu-1. A limited sampling strategy was explored using CL estimated from all concentration-time points as a reference. Finally, the days when TDM should be performed was assessed through simulations with the R packages mrgsolve and mapbayr [3]. The day-by-day busulfan courses of 1,000 virtual patients for several TDM strategy were simulated, with all the sources of variability taken into account: covariate distribution, inter-individual (IIV), inter-occasion (IOV) and residual variabilities.
Results: Data (135 patients, 323 courses, 2018 plasma concentrations) were well described with a monocompartmental model with IIV and IOV on CL (26.0% and 14.1%, respectively) and on volume of distribution (16.3% and 13.8%, respectively). Patient size was incorporated with a standard power model with actual body weight and a constant exponent. As pharmacogenetic data were missing for 21 patients, the covariate screening was performed on 114 patients and revealed that CL at day 1 was best predicted a priori with the following formula: CL = (BW/25)^0.786 * 0.896^MALIGN * 0.894 ^ GSTA1. Three concentrations (0.25, 2 and 5 hours after the end of the 3h-infusion) were sufficient for a satisfactory Bayesian estimation of CL (relative root-mean-square error: 3.4%). Simulations showed that dose calculations based only on the covariate formula performed poorly: the 90% prediction interval of the cumulative AUC relative to target AUC was of 69-151%. In comparison, TDM-based calculations yielded more precise cumulative AUCs, whichever days the TDM was performed at: day 1 only (85-124%), day 1-3 (90-116%) and day 1-2-3 (93-112%).
Conclusions: This comprehensive approach quantified the benefit of TDM to control busulfan exposure in Thai pediatric patients and suggest to decrease the number of samples from 7 to 3 per day of TDM. A user-friendly shiny web-application is now available for routine dose calculation.
References:
[1] Poinsignon V, Faivre L, Nguyen L, Neven B, Broutin S, Moshous D, et al. New dosing nomogram and population pharmacokinetic model for young and very young children receiving busulfan for hematopoietic stem cell transplantation conditioning. Pediatr Blood Cancer [Internet]. 2020 [cited 2023 Feb 27];67. Available from: https://onlinelibrary.wiley.com/doi/10.1002/pbc.28603
[2] Lawson R, Staatz CE, Fraser CJ, Ramachandran S, Teague L, Mitchell R, et al. Population pharmacokinetic model for once‐daily intravenous busulfan in pediatric subjects describing time‐associated clearance. CPT Pharmacom & Syst Pharma. 2022;11:1002–17.
[3] Le Louedec F, Puisset F, Thomas F, Chatelut É, White-Koning M. Easy and reliable maximum a posteriori Bayesian estimation of pharmacokinetic parameters with the open-source R package mapbayr. CPT: Pharmacometrics & Systems Pharmacology. 2021;10:1208–20.
