NCB de Jager (1), JM Heijdra (2), CJ Fijnvandraat (3), FWG Leebeek (4), MH Cnossen (2), RAA Mathôt (1) for the “OPTI-CLOT” and "WiN" study groups.
(1) Hospital Pharmacy-Clinical Pharmacology, Academic Medical Center Amsterdam, the Netherlands. (2) Department of Pediatric Hematology, Erasmus University Medical Center - Sophia Children’s Hospital Rotterdam, the Netherlands. (3) Department of Pediatric Hematology, Academic Medical Center Amsterdam, Amsterdam, the Netherlands. (4) Department of Hematology, Erasmus University Medical Center Rotterdam, the Netherlands.
Objectives: Von Willebrand factor/factor VIII (VWF/FVIII) concentrate Haemate® P (Humate P) is widely used in treatment of patients diagnosed with Von Willebrand’s Disease (VWD). These patients are characterised by a qualitative or quantitative defect of VWF. VWF is essential in primary and secondary hemostasis as it mediates adhesion and aggregation of platelets at sites of vascular injury, and acts as a chaperone protein for FVIII by protecting it from premature clearance. Treatment consists of replacement therapy using clotting factor concentrates in cases of acute or perioperative bleedings aiming to correct the VWF (and FVIII when present) deficiency. The objective of this study is the development of population pharmacokinetic (PK) algorithms for the perioperative situation, which could allow more accurate dosing of clotting factors, reducing consumption without increasing the bleeding risk.
Methods: VWD patients undergoing minor or major surgery in the Academic Medical Centre Amsterdam, Erasmus Medical Centre, Leiden University Medical Centre, Radboud University Medical Centre or University Medical Centre Groningen between 2000-2015 who received Humate® P were included in the dataset for this study. The retrospective data, consisting of FVIII levels, were analysed using nonlinear mixed-effects modelling (NONMEM v7.4) in combination with the First-Order Conditional Estimation with Interaction (FOCE+I) method. [1] With respect to endogenous FVIII baseline levels, historical baseline measurements were added to each observation in order to produce more accurate individual predictions. As pre-operative FVIII levels were occasionally higher then FVIII baseline levels, a ‘virtual’ dose was introduced in the data. A bioavailability parameter is used in each occasion to correct for this difference, without affecting the PK estimates. Covariate relationships, containing demographics, surgery characteristics and clinical features, were evaluated using a forward inclusion- followed by a backwards elimination method to explain inter-individual variability (IIV). A bootstrap method was used to check the robustness of the PK parameter estimates. Model performance evaluations were based on goodness-of-fit plots and visual predictive checks (VPC).
Results: PK parameter estimates were based on 96 adult and 8 pediatric VWD patients, undergoing 139 and 8 surgeries respectively. Median age, body weight and surgery duration (range) were 51 years (0.5-82), 77 kg (8.8-118) and 81 minutes (7-470), respectively. Furthermore, median FVIII baseline values and pre-operative FVIII levels (range) were 0.41 (0.01-0.97) and 0.76 (0.01-3.11) IU, respectively. The included patients were diagnosed with type I (n=55), type IIA (n=24), type IIB (n=7), type IIM (n=9), type IIN (n=3) or type III (n=6) VWD. PK profiles, containing 734 FVIII level observations were best described using a one-compartment model. A proportional error model was used to describe residual variability. PK parameters were allometrically scales using the ¾ power-model for Vd an CL. Typical values of the model for the volume of distribution (Vd) and clearance (CL) were 4.6 L/70 kg and 0.032 L/h/70 kg, respectively. Corresponding inter-individual variability of Vd and CL were 34% and 69%. The residual error was 0.23%. Covariate analysis identified an association between CL and the duration of the surgery: CL decreases as surgery duration increases. Furthermore, patients in ASA class III or IV exhibited a 41% decrease of CL. Bootstrap results confirmed the robustness of the model. Furthermore, the VPC of the final model, using 1000 replicate simulations of 104 patients, resulted in predictions of the simulated data that were well-matched with the observed level-time profiles.
Conclusions: Time-courses of obtained perioperative FVIII levels after administration of Humate® P were described adequately by the developed PK model. This model facilitates PK-guided dosing of Humate® P in VWD patients undergoing a minor or major surgical procedure, which potentially likely result in improvement of quality and cost-effectiveness of care.
References:
[1] Beal S, Sheiner LB, Boeckmann AJ, Bauer RJ. NONMEM User Guides. 1989-2011; Icon Development Solutions, Ellicott City, Maryland, USA.
Reference: PAGE 27 (2018) Abstr 8490 [www.page-meeting.org/?abstract=8490]
Poster: Drug/Disease Modelling - Other Topics