Jelien Den Hollander 1, Sjoerd Koopman 1, Marjon Cnossen 2, Robert Sidonio Jr. 3, Ron Mathôt 1
1 Hospital Pharmacy-Clinical Pharmacology, Amsterdam UMC (Amsterdam, The Netherlands), 2 Department of Pediatric Hematology and Oncology, Erasmus MC Sophia Children’s Hospital (Rotterdam, The Netherlands), 3 Department of Pediatrics, Emory University School of Medicine (Atlanta, The United States)
Introduction
Von Willebrand Disease (VWD) is the most common inherited bleeding disorder, caused by a quantitative deficiency or qualitative dysfunction of Von Willebrand Factor (VWF) [1]. As VWF protects factor VIII (FVIII) from premature degradation, FVIII levels may also be reduced [1]. Long-term prophylaxis with a VWF concentrate is recommended in patients experiencing severe and frequent bleeding episodes [2]. Heterogeneous breakthrough bleeding patterns, however, suggest considerable interindividual variability (IIV) in pharmacokinetic (PK) exposure [3]. In turn, this advocates for the use of PK-guided dosing to enable individualised prophylactic therapy.
Objectives
To develop a population pharmacokinetic (popPK) model for the plasma-derived VWF/FVIII concentrate with a 1:1 activity ratio (Wilate®) in patients with severe VWD and to explore the relationship between VWF/FVIII activity and bleeding events during prophylaxis.
Methods
A nonlinear mixed-effects PK modelling approach was applied to VWF:ristocetin cofactor activity (VWF:RCo) and FVIII activity (FVIII:C) data from paediatric and adult patients enrolled in one prospective single-dose study (WIL-12) [4] and two one-year prospective phase 3 prophylaxis trials (WIL-31: NCT04052698, and WIL-33: NCT04953884) [3].
During popPK development, one- and two-compartment models were evaluated. The inclusion of IIV on all PK parameters, covariance structures, inter-occasion variability, study-specific residual error model structures, and a potential interaction between VWF:RCo exposure and FVIII:C clearance were assessed. Two model components were included a priori: 1) a turnover component describing endogenous production and elimination of VWF:RCo/FVIII:C, and 2) allometric scaling with fixed exponents. Sex, VWD subtype, blood group and study were evaluated as covariates on all PK parameters using a stepwise covariate modelling approach.
The lower limits of quantification (LLOQ) were 5.0 IU/dL for VWF:RCo and 3.1 IU/dL for FVIII:C. Observations below the LLOQ were handled using the M3 method. Model evaluation was performed using goodness-of-fit (GOF) plots, prediction-corrected visual predictive checks (pcVPCs) and bootstrap analysis (n=1000).
In the pharmacodynamic analysis, individual post-hoc PK profiles were constructed, and VWF/FVIII activity levels were retrospectively assessed at time of bleeding for patients included in WIL-31 and WIL-33.
Results
62 patients were included with a median age of 18.5 years (range: 1–68 years) and a median body weight of 59.5 kg (12.5-138.3 kg). Eleven patients (18%) had VWD type 1, thirteen (21%) type 2A, two (3%) type 2B, two (3%) type 2M, and thirty-four (55%) type 3. A total of 825 VWF:RCo and 855 FVIII:C activity levels were available, of which 19% and 12%, respectively, were below the LLOQ.
A two-compartment turnover model best described the activity over time data of both VWF:RCo and FVIII:C. For a typical 70 kg individual, VWF:RCo clearance was 3.1 dL/h (coefficient of variation (CV): 51%) and FVIII:C clearance was 1.7 dL/h (CV: 47%). Corresponding central volume of distributions were 40.6 dL (CV: 22.4%) for VWF:RCo and 35.6 dL (CV: 21.3%) for FVIII:C. No interaction between VWF:RCo exposure and FVIII:C clearance, nor covariance between their clearance IIV terms, was detected. The GOF plots, pcVPCs, and bootstrap of the final models demonstrated adequate model performance.
A total of 83 spontaneous and 61 traumatic bleeding events occurred in 27 patients. Fifty percent of the spontaneous and traumatic bleedings occurred at VWF:RCo activity levels below 3.6 IU/dL (range: 1.5–53) and 3.9 IU/dL (range: 1.8–45.0), respectively. Corresponding FVIII:C activity levels were 13.6 IU/dL (range: 1.3–72.1) and 13.5 IU/dL (range: 1.3–68.7), respectively.
Conclusion
We developed a popPK model for a plasma-derived VWF/FVIII 1:1 concentrate that characterises the PK of both VWF:RCo and FVIII:C in VWD patients receiving prophylaxis. This model may support clinicians in personalising long-term prophylaxis by enabling assessment of individual PK exposure metrics, including trough activity levels, area under the curve, time above predefined thresholds and VWF:RCo and FVIII:C activity levels at the time of bleeding events.
References:
References
1 Leebeek FWG, Eikenboom JCJ. Von Willebrand’s Disease. N Engl J Med 2016; 375: 2067–80.
2 Connell NT, Flood VH, Brignardello-Petersen R, Abdul-Kadir R, Arapshian A, Couper S, Grow JM, Kouides P, Laffan M, Lavin M, Leebeek FWG, O’Brien SH, Ozelo MC, Tosetto A, Weyand AC, James PD, Kalot MA, Husainat N, Mustafa RA. Ash isth nhf wfh 2021 guidelines on the management of von willebrand disease. Blood Adv 2021; 5.
3 Sidonio RF, Boban A, Dubey L, Inati A, Kiss C, Boda Z, Lissitchkov T, Nemes L, Novik D, Peteva E, Taher AT, Timofeeva MA, Vilchevska K V., Vdovin V, Werner S, Knaub S, Khayat CD. von Willebrand factor/factor VIII concentrate (Wilate) prophylaxis in children and adults with von Willebrand disease. Blood Adv American Society of Hematology; 2024; 8: 1405–14.
4 Kessler CM, Friedman K, Schwartz BA, Gill JC, Powell JS. The pharmacokinetic diversity of two von willebrand factor (VWF)/ factor VIII (FVIII) concentrates in subjects with congenital von willebrand disease: Results from a prospective, randomised crossover study. Thromb Haemost 2011; 106.
Reference: PAGE 34 (2026) Abstr 11973 [www.page-meeting.org/?abstract=11973]
Poster: Drug/Disease Modelling - Other Topics