OR03
Reference Values for the Haemophilia Joint Health Score in Patients with Severe Haemophilia derived from the Canadian Bleeding Disorder Registry
K. T. Nguyen1,*, A. K. Chan1, D. Matino2, E. Iserman3, A. Keepanasseril3, K. Strike4,5
1Pediatrics, 2Medicine, 3Department of Health Research Methods, Evidence, and Impact, 4Rehabilitation Science, McMaster University, 5Hamilton Niagara Regional Hemophilia Program, Hamilton Health Sciences, Hamilton, Canada
Introduction: People with haemophilia are prone to recurrent joint bleeding, leading to haemophilic arthropathy. The Haemophilia Joint Health Score (HJHS) is a standardized tool for assessing joint health but lacks age-based reference values, limiting interpretability. Without these, clinicians cannot determine whether a score aligns with expected age-related patterns or indicates early pathology. This study aimed to develop national reference values for the HJHS in severe haemophilia using data from the Canadian Bleeding Disorder Registry (CBDR).
Methods: This cross-sectional study analyzed CBDR data from January 1, 2018, to December 31, 2024. Participants aged ≥4 years with severe haemophilia A or B and at least one HJHS record were included. Total HJHS scores (0–124) were modelled using Generalized Additive Models for Location, Scale, and Shape (GAMLSS). Model selection was guided by the Generalized Akaike Information Criterion (GAIC), and performance was evaluated through 100-fold cross-validation using the expected-to-observed (E:O) ratio and calibration plot.
Results: A total of 545 participants were included (mean age 22±17 years; range 4–75): 475 (87%) had haemophilia A and 70 (13%) haemophilia B. The cohort comprised 296 (54%) pediatric and 249 (46%) adult participants. Most were on prophylaxis (453; 94%), and 355 (79%) had no inhibitor history. Median HJHS was 8 (IQR 2–20), increasing with age. The final model, using a log-normal distribution with cubic spline smoothing, showed strong calibration (E:O 0.93; 95% CI 0.91–0.95) and minimal overfitting. Age-specific percentiles (5th–95th) for ages 4–70 captured expected variability in joint health and the high prevalence of zero scores (141 participants; 26%), most of which occurred among pediatric patients (124; 88%).
Discussion/Conclusion: This study provides the first national age-based HJHS reference values for severe haemophilia A and B in Canada. These percentiles enable interpretation against expected age-specific distributions, improving detection of worse-than-expected joint health. Integrating the curves into clinical and digital tools may enhance standardization, support earlier detection, and enable longitudinal monitoring. Future work should adapt this modelling approach to other national registries to create country-specific reference values and strengthen the HJHS’s global applicability.
Image:
Disclosure of Interest: K. Nguyen: None declared, A. Chan Grant/Research support from: Hamilton Health Sciences and the McMaster Children's Hospital/Hamilton Health Sciences Foundation Chair in Pediatric Thrombosis and Hemostasis., D. Matino Grant/Research support from: Bayer, Pfizer, Novo Nordisk, Sanofi, Octapharma, and Roche paid directly to McMaster University , Consultant for: Sanofi, Sobi, Novo Nordisk, Bayer, Pfizer, Octapharma, Roche for participation in advisory boards, lectures and preparation of educational material, E. Iserman Consultant for: Sanofi, A. Keepanasseril Grant/Research support from: Novo Nordisk, Roche, and the Canadian Drug Agency paid directly to McMaster University, K. Strike Grant/Research support from: Hamilton Health Sciences Health Professional Clinical Research Award, Health Professional Investigator Award, McMaster Children’s Hospital Foundation, Pfizer Canada, and Bayer. Travel support from Pfizer, Bayer, and Novo Nordisk., Consultant for: Sanofi, Bayer, Hemalytic, Takeda, Pfizer, Biogen, Novo Nordisk, Roche, Baxalta, and Octapharma.