Treosulfan (TREO), a structural analog of busulfan, is currently studied as a myeloablative agent in conditioning regimens before hematopoietic stem cell transplantation in pediatric patients. High exposure to TREO (>1650 mg*h/mL) might be related to early toxicity, especially skin toxicity and mucositis. The aim of the present study was to investigate a potential relationship between exposure to TREO and its monoepoxytransformer (S,S-EBDM), as well as variability of the pharmacokinetics of these entities by means of a population pharmacokinetic approach with a non-linear mixed-effects analysis.
The study included data from 14 children with malignant and non-malignant diseases treated with TREO in daily doses 10–14 g/m2. The parent-metabolite population pharmacokinetic model was developed in NONMEM 7.3 software. Upon the constructed model, an extensive simulation was performed to assess the correlation between exposure to TREO and S,S-EBDM.
It was found that TREO and S,S-EBDM pharmacokinetics was best described with 2-compartmental and 1-compartmental linear models, respectively. The vast majority (>65%) of TREO was transformed to S,S-EBDM. Overall, a considerable interpatient variability of pharmacokinetic parameters was observed, especially the clearance of S,S-EBDM. A weak correlation was found between the exposure to TREO and S,S-EBDM (r = 0.1681, p < 0.0001). Also, patients with an exposure to TREO above 1650 mg*h/mL were most likely to have also a high exposure to S,S-EBDM (35.38 μM*h vs. 43.14 μM*h, p < 0.0001).
In summary, a parent-metabolite population pharmacokinetic model for TREO and S,S-EBDM was developed for the first time. It was shown that there is a weak correlation between exposure to TREO and S,S-EBDM. Therefore therapeutic drug monitoring of not only prodrug but also its active epoxide might be needed.