Associations between factor VIII levels as estimated from pharmacokinetic studies and patient-reported treated bleeds in severe hemophilia A

Introduction Hemophilia A is a hereditary bleeding disorder caused by a deficiency in factor (F) VIII, with the severity of the bleeding phenotype broadly correlating with the residual FVIII activity. People with mild hemophilia beyond 10 IU/dL rarely experience hemarthroses, yet joint deteriorations from such bleeds, common in severe hemophilia, further predispose to subsequent bleeds. Here, we used infusions as reported in routinely collected patient diaries with half-life estimates derived from pharmacokinetic (PK) studies to correlate the occurrence of on-demand infusions due to bleeding episodes with FVIII levels. Method We included all adult people with severe hemophilia A (PWSHA) that participated in our hemophilia biobank, had at least half a year of documentation provided by patient diaries, and for whom a PK study was available. Assuming one-compartmental pharmacokinetics and using the superposition principle, we calculated FVIII levels for each subject over their whole observation period. PK parameters were derived from PK studies conducted with WAPPS-Hemo. For estimating the association of FVIII levels with bleedings, we fitted repeated time-to-event Anderson-Gill regressions, modelling FVIII as time-varying covariate, adjusting for presence of hemophilic arthropathy. To account for multiple events per subject, robust variance estimates were calculated. Results As of October 2025, patient diaries from 109 PWSHA had been collected at our center, 35 (32.1%) of which had a PK study available and were included in this study (median age 34.4; IQR, 26.3-43.5), amounting to a total of 19 324 documented infusions (50.7% with extended half-life products) and 45 704 000 IU used over a median observation period of 4.71 years (2.67-8.28, [ Fig. 1 ]). At their first documentation, 27 (77.1%) subjects already had presence of hemophilic arthropathy noted in their medical history. A total of 1 325 bleedings (72.2% joint bleeds) were reported, resulting in a median annualized bleeding rate of 4.6 (1.3-8.1) and a joint bleeding rate of 3.5 (0.8-5.8). Over the observation period, the median average FVIII activity was 16.4 IU/dL (13.6-22.0) and the median time subjects were above 1 and 10 IU/dL was 79.0% (65.5-89.5) and 41.3% (32.9-50.3), respectively. In a repeated time-to-event model on total bleeds, hazard ratios were 0.92 (95% CI, 0.88-0.97) for every 1-unit increase in FVIII activity and 3.32 (1.37-8.08) for presence of hemophilic arthropathy. [ Fig. 2 ] shows the expected number of annual (joint) bleeds when holding FVIII constant at 1 and 10 IU/dL with and without hemophilic arthropathy, respectively. Fig. 1 Study characteristics of included subjects. The number of subjects with hemophilic arthropathy refers to the time point of first documentation. Data are presented as median with interquartile ranges or frequencies with percentages. Annual bleeding rates are mean rates estimated from a negative binomial regression model with time in years as offset variable. Fig. 2 A) Annualized bleeding (ABR) and joint bleeding rates (AJBR) observed over the course of the study, and joint bleeds as a percentage of total bleeds (AJBR/ABR). B) Percentage of time the individual subjects were above the depicted FVIII levels (transparent blue lines). The thick super-imposed blue line and the blue shaded band represent summary estimates over all subjects (median and IQR, respectively). C-D) Expected number of (joint) bleeds estimated from a repeated time-to-events model for a subject with or without hemophilic arthropathy holding FVIII levels constant at 1 or 10 IU/dL. Conclusion In our study on PWSHA, we noted a high number of on-demand infusions due to bleeding at FVIII levels otherwise considered sufficiently protective in mild hemophilia. Our study highlights the impact of pre-existing joint impairment on occurrence of subsequent bleeds, emphasizing the need for preserving joint health and individualizing prophylaxis. Publication History Article published online: 17 February 2026 © 2026. Thieme. All rights reserved. Georg Thieme Verlag KG Oswald-Hesse-Straße 50, 70469 Stuttgart, Germany