Web Accessible Population Pharmacokinetics Service - Hemophilia: Sources of Variability

Innovation and expected results: As the classical PK approach typically requires 11 samples, which can be a burden to the patient, specific subgroups of individuals are often excluded from analysis-including children, elderly and critical care patients. As WAPPS-Hemo is the largest database of individual PK data, a large subgroup of pediatric patients and patients with inhibitors are available for investigation, allowing for analysis of variability in previously unavailable groups. Determining predictors of variability within patient outcome will allow us to improve the current models used to estimate individual PK, resulting in better individual care and use of resources.

Eligibility criteria: All valid data in the WAPPS-Hemo database will be used. All factor concentrates will be included. All available covariates will be analyzed.

Patients: Individuals with hemophilia A or B, any severity, who are registered on WAPPS-Hemo and for whom infusion and/or PK data are available. Both adult and pediatric patients will be included. Patients with history of inhibitors will be included as a separate subgroup.

Available covariates: age, sex, weight, height, dose and type of factor administered (as total and IU/kg), blood group, baseline factor level, positive history of inhibitors, hematocrit, hemoglobin, serum creatinine and laboratory methods used to measure factor VIII and factor IX.

Design: retrospective data analysis

Study database: The WAPPS-Hemo database will be used for the study. WAPPS-Hemo is a web accessible platform developed and run by the Health Information Research Unit (HiRU) at McMaster University. Data have been provided from participating hemophilia treatment centers worldwide to the scope of obtaining individual PK estimates. Data collected are completely anonymized, and re-use of the data for modelling and validation purposes was agreed upon by the inputting physicians, who committed to inform their patients about use of their data for system improvement.

Data extraction procedure: Data will be extracted from the WAPPS-Hemo database after transforming the centre name into a numeric code. The data extracted will include: patient age, sex, weight, height, dose and type of factor administered (as total and IU/kg), blood group, baseline factor level, positive history of inhibitors, hematocrit, hemoglobin, serum creatinine, laboratory methods used to measure factor VIII and factor IX, infused dose, post-infusion measurement of plasma factor activity level, estimated terminal half-life and time to 0.01, 0.02, and 0.05 IU/mL, with their credibility intervals, area under the curve (AUC), central volume, and clearance. Once extracted, the data will be stored in a secure server, located within the local network of the HiRU, protected by a proxy server and firewall.

Data cleaning: Once extracted, the data will be cleaned. Duplicate submissions will be removed, and all original data that has been duplicated from merged infusions will be deleted. Any data not valid for modelling, such as user input errors, insufficient data, or conditions that exclude use of patient data (such as use of inhibitors) will be removed and excluded from analysis. The dataset will be analyzed to search for outliers. Input errors missed at the source will be corrected where the incorrect measurement has been used (i.e. weight, height).

Data Analysis: Multivariate hierarchical analysis will be performed on various factors (age, sex, weight, height, dose and type of factor administered (as total and IU/kg), blood group, baseline factor level, positive history of inhibitors, hematocrit, hemoglobin, and serum creatinine, laboratory methods used to measure factor VIII and factor IX) to explore sources of variability in patient outcome. Significant predictors will be used as covariates in improvement of the model.