Abstract
Introduction and objective
It has previously been reported that people with hemophilia are at higher risk for reduced bone mineral density compared to the general population. In addition, a retrospective, single-center study of people with Hemophilia A (HA) and Hemophilia B (HB) found a higher rate of fractures in people with HA or HB (24.8 per 1000 patient-years; n=382) compared to a large prospective cohort study of the general US population which found 9.6 fractures per 1000 patient-years (n=343,574; Gay et. al. BJH 2015). This study investigated the relative risk of adverse bone health outcomes between people with HA, HB or von Willebrand Disease (VWD), and the general population.
Methods
Males with ≥ 1 inpatient diagnosis or ≥ 2 outpatient diagnoses (at least 30 days apart) of HA, HB or VWD were identified in the Truven Health MarketScan Commercial and Medicare Supplemental database (2002-2015). Controls were matched males of similar age without these diagnoses, and had no factor replacement use. All eligible subjects were required to have no evidence of bypassing agent use, continuously enrolled in a health plan for ≥ 2 years, and no fractures in the baseline period (first 12 months after database entry). Patients were then followed from the end of the baseline period to the end of continuous eligibility in the database.
Multivariate regression models were used to examine the impact of each bleeding disorder on bone health outcomes. Poisson regression models were used to evaluate the number of fractures and joint replacements; Cox proportional hazards models were used to assess time to the first diagnosis of osteoporosis or osteopenia. All models controlled for demographics, year of entry in the MarketScan database, health insurance plan, replacement clotting factor use (yes or no), concurrent medications, and comorbid conditions (e.g. HIV/HCV status) measured in the baseline period.
Results
The analyses included 4,546 people with HA, 794 with HB, 3,555 with VWD and 69,900 controls. Mean age (standard deviation; SD) was 35.9 (23.6) for people with HA, 37.0 (23.5) for HB, 30.1 (23.5) for VWD, and 34.5 (24.2) for controls.
During an overall median follow-up of 2.3 years (2.8, 2.8, 3.0, and 2.2 years for HA, HB, VWD and controls, respectively), unadjusted incidence rates per 1000 person-years were higher for each adverse bone health outcome in people with each bleeding disorder (HA, HB, VWD) than controls (Fractures: 9.8, 11.1, 5.6 vs. 4.4; Joint Replacements: 24.8, 18.0, 13.5 vs. 7.0; Osteoporosis: 7.5, 8.8, 4.9 vs. 2.8; Osteopenia: 11.4, 13.6, 8.9 vs. 4.7).
After controlling for between-group differences, significant positive associations were found for each bleeding disorder with fractures, joint replacements, osteoporosis, and osteopenia (Table 1). All results remained significant after adjusting for multiple comparisons, with the exception of fractures in people with VWD vs. controls.
No significant differences were observed in the risk of any of the adverse bone health outcomes studied between bleeding disorders.
Conclusion
Patients with HA, HB and VWD appear to be at higher risk of adverse bone health outcomes compared to the general population. While the increased risk of fractures, osteoporosis and osteopenia in this study suggest a potential role for endogenous clotting factors in maintaining bone health, the effect of disease severity or replacement factor regimen (on-demand or prophylaxis) on bone health could not be examined with these data. Future studies should investigate the possible pleiotropic effect of replacement clotting factor to understand how bone health in patients with bleeding disorders is affected by factor replacement.
Disclosures
Booth: Shire: Employment. Lu:Shire: Employment. Gallo:Shire: Employment. Ito:Shire: Employment. Valentino:Shire: Employment.
Into the future with novel emerging therapies for haemophilia