Risk of Intracranial Hemorrhage in US Patients with Hemophilia a: Real-World Retrospective Cohort Study Using the Athndataset

Background Intracranial hemorrhage (ICH), a serious complication of hemophilia A (HA), causes high rates of mortality and morbidity (Zanon. Blood Transfus. 2019). The risk of developing ICH is 10 to 20 times higher in patients with HA (PWHA) than in individuals without (Zanon. 2019). Patient data from 1998 through 2008 identified several risk factors for ICH, including presence of an inhibitor and prior ICH (Witmer. Br J Haematol. 2010). However, standard of care and the treatment landscape have evolved (Srivastava. Haemophilia. 2020). This study leverages clinical information from the ATHNdataset to assess the incidence of ICH in non-neonate (≥2 years of age) PWHA and determine the association between ICH risk and key characteristics.

Methods This retrospective cohort study used data from PWHA (male, aged >2 and <75 years) who had visit information from January 1, 2010, through September 30, 2020, as identified in the ATHNdataset. The American Thrombosis & Hemostasis Network (ATHN) is the steward of the ATHNdataset, a Health Insurance Portability and Accountability Act-compliant, de-identified patient health dataset containing data from individuals with bleeding disorders receiving care through ATHN affiliates who opt in to contribute. The unadjusted incidence rate (IR) and incidence rate ratio (IRR) were evaluated by inhibitor and prophylaxis status. Association between patient characteristics and risk of ICH (HR) was evaluated by time-to-event analysis using univariate and multivariate Cox proportional-hazards (CoxPH) models. Stratified bootstrap analysis and penalty-based ridge regression sensitivity analyses were done to account for the rarity of ICH and small sample bias.

Results In over 10 years and 9 months of collected data, 135 of 7837 PWHA (1.7%) had an ICH, with a median follow-up time of 10.7 patient-years. The incidence of ICH by demographic, clinical, and treatment characteristics is described in Table 1. In the unadjusted analysis, stratification by inhibitor status had an IRR (IR+/IR-) of 1.76 (95% CI: 0.97-3.20; P=0.059). IRR (IR+/IR-) by prophylaxis status in the unadjusted analysis was 0.63 (95% CI: 0.43-0.94; P=0.02). Univariate CoxPH modeling and multivariate CoxPH modeling with interaction, with associated bootstrap and ridge regression analyses, identified factors associated with differential risk of developing ICH (Table 2). Characteristics associated with a greater risk of developing ICH include being between the ages of 2 and 12 years, having ever been on Medicaid, having had HIV, hepatitis C (HCV), or hypertension (HTN) comorbidities, and never having received factor VIII (FVIII) or prophylactic treatment. In the multivariate population analysis, the estimated HR for patients never receiving prophylaxis was 7.67 (95% CI: 0.81-3.27). This HR estimate shrunk to 2.03 (95% CI: 0.26-2.21) in the associated bootstrapping analysis and 3.09 in the associated highest penalty ridge regression analysis but was still found to be statistically significant. Inhibitor status was not found to be statistically associated with ICH. Results related to an interaction between prophylaxis and inhibitor status (Table 2) could not be interpreted because of the small sample size.

Discussion ICH is a serious, life-threatening complication for PWHA, hence the need to identify factors associated with increased risk of ICH. This is the first study in a decade evaluating key demographic, clinical, and HA treatment characteristics for associated risk of PWHA developing ICH. These results demonstrate that risk factors for ICH include age (greatest risk in children between the ages of 2 and 12 years), having ever been on Medicaid, having had HIV, HCV, or HTN comorbidities, and never having received FVIII or prophylactic treatment. It was previously shown that prophylaxis confers a protective effect against ICH in instances of severe disease (Witmer 2010) and in children and adolescents (Andersson. Br J Haematol. 2017). Inhibitor positivity has been shown to increase risk for ICH (Witmer 2010; Zanon. Haemophilia. 2012); however, this study was unable to corroborate those findings. A larger sample size is required to draw substantive conclusions regarding the joint impact of prophylaxis and inhibitor status on risk of ICH. As the treatment landscape for HA evolves with the increased use of non-factor products and gene therapy, continued evaluation of risk factors for ICH will be needed.

Caicedo:Takeda: Current Employment, Current holder of stock options in a privately-held company. Denne:Takeda: Current Employment, Current holder of stock options in a privately-held company. Ewenstein:Takeda: Current Employment, Current holder of stock options in a privately-held company. Mokdad:Takeda: Current Employment, Current holder of stock options in a privately-held company. Xing:Takeda: Current Employment, Current holder of stock options in a privately-held company. Recht:Oregon Health & Science University: Ended employment in the past 24 months; Catalyst Biosciences, CSL Behring, Genentech, Grifols, Hema Biologics, Novo Nordisk, Pfizer, Sanofi, Takeda, uniQure: Consultancy; Bayer, Biomarin, CSL Behring, Genentech, Grifols, Hema Biologics, LFB, Novo Nordisk, Octapharma, Pfizer, Sanofi, Spark Therapeutics, Takeda, uniQure: Research Funding; Foundation for Women and Girls with Blood Disorders; Partners in Bleeding Disorders: Thrombosis and Hemostasis Societies of North America: Membership on an entity's Board of Directors or advisory committees; American Thrombosis and Hemostasis Network; Yale University School of Medicine: Current Employment.