Currently, there is an urgent need to establish the optimal dosing of TREO in conditioning prior to hematopoietic stem cell transplantation, especially in children. For that purpose, pharmacokinetic analyses are ongoing within clinical phase II and III trials. In this paper, HPLC methods for determination of prodrug treosulfan and/or its biologically active epoxides in human plasma or serum are reviewed for the first time, including the spectrum of analytes being quantified, detection type, and derivatization methodology. The major focus is addressed to the stability of TREO and its monoepoxide related with different strategies of patients’ blood processing, e.g. blood pH lowering to different values, no pH adjustment; centrifugation of blood immediately after collection or within a few hours later. This issue is crucially important for the robust bioanalysis because the epoxytransformation of TREO is a nonenzymatic, highly pH and temperature-dependent reaction. In-depth analysis of the literature results demonstrates that some methodologies of blood treatment could produce the systematic underestimation of TREO concentrations. Consequently, the drug clearance and volume of distribution will be overestimated, which might false the association of the drug exposure with the regimen-related toxicity and clinical outcomes. The paper indicates the deficiencies of the blood processing strategies and offers hints for their refinement. The provided information ought to be important in the current investigations of the personalized TREO pharmacokinetics.