Validation of a new pharmacokinetic population model for perioperative dosing of a highly purified plasma-derived von Willebrand factor (pdVWF): the WIL-PKPOP-CHIR study

Introduction Given the substantial variability in the pharmacokinetic (PK) of von Willebrand factor (VWF) and factor VIII (FVIII) following infusion of VWF concentrate, there is a need for developing a tool to optimize perioperative dosing in patients with all von Willebrand disease (VWD) types and maintain VWF: Antigen, VWF: Activity, and FVIII:C levels within the target range. Thus, a PK population model describing both VWF and FVIII kinetics was developed for a highly purified plasma-derived VWF concentrate (pdVWF), using monocentric retrospective data obtained in real-life perioperative settings in France [ 1 ]. Several covariates were identified as significant predictors of interindividual VWF PK variability, including body weight, VWD type, blood group, VWF propeptide, and age. Method To validate the predictive performance of this model, a real-life, multicentre, non-interventional, prospective study will be conducted. This external validation will be performed by comparing individual observed plasma levels of VWF:Ag, VWF:Act, and FVIII:C with concentrations predicted by the PK population model. Results This study will include approximately 50 patients with VWD requiring perioperative treatment with the pdVWF. Dosing regimens and timing of VWF:Ag, VWF:Act, and FVIII:C measurements will follow each centre’s routine clinical practice. The maximal post-operative follow-up will be 10 days. No additional procedures or monitoring beyond standard care will be performed. Validation will follow an adaptive strategy based on VWF:Act data. A first analysis will be performed after 25 evaluable patients (approximately 100 VWF:Act measurements) with an expected root mean square error (RMSE)≤0.30 between observed and predicted values. If this criterion is met, the model will be validated on the full dataset of 50 patients (approximately 200 measurements). If RMSE>0.30, the model will be re-estimated using data of the first 25 patients before being re-evaluated in subsequent patients. The model will be considered externally validated if the final RMSE is