Simulation-Based Evaluation of Busulfan Sampling Strategies in Adults and Adolescents Undergoing Allogeneic...

Busulfan, used in hematopoietic stem cell transplantation, exhibits a narrow therapeutic index and high pharmacokinetic (PK) variability. Therapeutic drug monitoring (TDM) is currently employed to achieve target exposure. However, there is no consensus on an optimal limited sampling strategy (LSS) for busulfan on the once-daily (QD) regimen. A virtual cohort of adult patients (n = 1000) receiving OD intravenous busulfan 3.2 mg/kg as a 3-h infusion was generated by Monte Carlo simulation using three published busulfan population PK models. For each one- and two-sample LSS, individual PK parameters were estimated from Day-1 sample(s) by maximum a posteriori Bayesian estimation and used to estimate the 4-day cumulative AUC. A model-informed precision-dosing (MIPD) analysis quantified the target-attainment benefit of TDM per design. The best-performing designs were validated in an external cohort (n = 37) with adults and adolescents. Across all three models, a two-sample early-plus-late schedule was the most informative, and the recommended two-sample design was 3 and 7 h after the start of infusion. TDM guided by a well-chosen LSS approximately doubled cumulative-target attainment, whereas a single early 3-h sample was markedly weaker. In the external cohort, two-sample designs recovered the cumulative 4-day AUC to ∼5%-8% MAPE versus ∼11%-13% for a single 3-h sample. For OD intravenous busulfan in adults and adolescents, a two-sample schedule at 3 and 7 h after the start of infusion supports the correct dosing decision and, through MIPD-guided dose adjustment, approximately doubles target attainment; a single 6-h sample is a simpler but less robust alternative.