A Short Communication: Comparison of Methods to Estimate Busulfan Exposure in Pediatric Hematopoietic Stem Cell Transplant Recipients

Background: Noncompartmental analysis (NCA) and model-based Bayesian methods (MBMs) are used for cumulative area-under-the-concentration–time curve (AUC cum ) estimates during therapeutic drug monitoring. Understanding their predictive differences can assist with switching between them and result interpretation. This study aims to compare NCA-based and MBM-based busulfan-AUC cum predictions in pediatric hematopoietic stem cell transplant recipients. Methods: Data on busulfan (administered once daily through intermittent infusion) were obtained from 4 hospitals. Busulfan concentrations were measured 3, 3.25, 4, 5, 6, 8, and 24 hours after infusion initiation over 4 treatment days. The busulfan dose, pharmacokinetic profile, and patient covariate factors were supplied to Kinetica and InsightRX Nova using the Shukla model to generate NCA-based and MBM-based busulfan exposure predictions. Differences in AUC cum estimates at the end of the treatment course were compared using a Wilcoxon signed-rank sum test and Bland–Altman plots. The effect of restricting data to specific treatment days on the MBM-based predictions was also examined. Results: Data from 89 pediatric hematopoietic stem cell transplant recipients (2304 busulfan samples) were analyzed. The MBM consistently estimated higher AUC cum values than NCA, with absolute and relative percentage differences (mean ± SD) of 2.8 ± 3.1 mg·L −1 ·h and 3.4 ± 3.7%, respectively ( P < 0.001); 1.1% of individual comparator estimates displayed a relative percentage difference ≥15%. When using the MBM and comparing sampling across all 4 days to day 1 sampling or to day 1 and 3 sampling, 14.6% and 3.4%, respectively, of estimates displayed a relative percentage difference ≥15%. Conclusions: When switching from NCA to MBM during therapeutic drug monitoring, a small increase in target busulfan AUC cum may be required to maintain similar drug exposure. Sampling on only day 1 may not accurately estimate AUC cum in all patients when using MBM.