Engi Algharably1, Jonathan Gross2, Lina Alasfar2, Reinhold Kreutz1, Johannes Schulte2, Ahmed Abulfathi3
1Charité – Universitätsmedizin Berlin, corporate member of Freie Universität Berlin and Humboldt Universität zu Berlin, Institute of Clinical Pharmacology and Toxicology, 2Charité – Universitätsmedizin Berlin, corporate member of Freie Universität Berlin, Humboldt-Universität zu Berlin, Department of Pediatric Oncology and Hematology, 3CDDS, Certara USA, Radnor
Background and objectives: Busulfan is a bifunctional alkylating that is commonly used in conditioning regimens prior to allogeneic hematopoietic stem cell transplantation (HSCT) or as part of high dose chemotherapy regimens with autologous stem cell rescue. Busulfan displays a narrow therapeutic index and wide inter- and intraindividual variability, particularly in pediatric patients, necessitating therapeutic drug monitoring (TDM) to ensure effective treatment while minimizing toxicity. However, there is currently no consensus on TDM protocols leading to variability in sampling times or frequencies that vary among centers according to local practices and resources. Characterizing busulfan PK using a population approach is a necessary first step to optimize TDM-guided dosing practice in patients undergoing HSCT. The objectives of this work were to 1) develop a population PK model to characterize the PK of busulfan; 2) quantify and identify the sources of variability of busulfan PK. Methods: Data from 62 consecutive patients who underwent HSCT between March 2017 and October 2023 at the Department of Pediatric Oncology and Hematology of the Charité University Hospital, Berlin, Germany were included. Busulfan was administered intravenously and samples were withdrawn before start of the infusion and at 0h, 30min, 1h, 2h, 4h, 6h and 9h after the end of infusion once daily over the 3–4 consecutive treatment days. Initial doses ranged from 3.1 to 5.2 mg/kg according to a dosing nomogram using weight cut-off values (Bartelink et al., 2012). Subsequent TDM adjustments were implemented to achieve target cumulative area under the concentration-time curve (cAUC) levels, specifically 50–70 mg·h/L for reduced-intensity conditioning and 80–105 mg·h/L for myeloablative conditioning. Nonlinear mixed-effects modeling was conducted using Phoenix NLME (version 8.4, Certara Inc.), employing the first-order conditional estimation with extended least-squares (FOCE-ELS) method for model development. Interindividual variability (IIV) was explored on all PK parameters and was assumed to be log-normally distributed. Different residual error models including proportional error were tested. Nonlinear elimination was evaluated by Michalis-Menten kinetics. Covariates including total body weight, body surface area, fat free mass, normal fat mass, age, sex, malignancy, baseline serum albumin and interacting comedications were tested. The latter included paracetamol, metronidazole and fludarabine. Results: In this pediatric cohort (39% female), a total of 1,531 busulfan concentration measurements were collected, averaging 6.3 samples per individual (range: 5.9–49). The median age was 4.3 years (range: 0.25–20.25), and the median weight was 17.1 kg (range: 4.8–78). A two-compartment model with zero-order infusion and first-order elimination fits the data best. The final model includes allometric scaling on the disposition parameters with total body weight normalized to 17.1 kg, and fixed exponents of 0.75 for central clearance (Cl), intercompartmental clearance (Q) and 1 for volumes of distribution of central (V1) and peripheral compartments (V2). A maturation component was incorporated into the Cl using a sigmoid Emax model based on postmenstrual age to account for metabolic ontogeny. The PK parameters were estimated with sufficient precision, yielding typical values of 10.14 L for V1, 22.91 L for V2, 3.34 L/h for Cl, and 0.57 L/h for Q for a 17.1 kg-individual. Including covariates such as sex, malignancy, baseline serum albumin and interacting co-medications did not improve the model performance. Conclusion: The model adequately characterizes busulfan PK. This model will be used to enhance TDM-guided dosing, aiming to achieve target cAUC levels throughout the treatment period and reduce amount of required samples for this.
Bartelink IH. Clin Pharmacokinet. 2012; 51(5):331-45
Reference: PAGE 33 (2025) Abstr 11319 [www.page-meeting.org/?abstract=11319]
Poster: Drug/Disease Modelling - Oncology