Abstract
Background
Hemophilic arthropathy (HA) is a major cause of disability in patients with severe hemophilia A. However, long-term data describing progression to advanced radiographic HA across major joint sites remain limited. We investigated risk factors for progression to advanced HA over extended follow-up.
Methods
We conducted a retrospective cohort study of inhibitor-negative patients with severe hemophilia A followed at our institute between 1984 and 2020. Bilateral elbows, knees, and ankles were evaluated (six joints per patient). HA progression was defined as reaching Arnold–Hilgartner stage ≥ IV. Time to progression was analyzed using Cox proportional hazards models with cluster-robust standard errors, with participant as the clustering variable, to account for within-participant correlation across joints. Covariates included joint site, Arnold–Hilgartner stage at initiation of regular prophylaxis, age at first visit to our institute, prophylaxis status at the first visit, and laterality.
Results
The cohort included 109 patients (mean age at first visit to our institute, 14.3 years). At the first visit, Arnold–Hilgartner stage ≥ I changes were present in 15% of elbows (29/200), 8% of knees (16/200), and 27% of ankles (51/186). Across 583 joints and 8,634 joint-years of follow-up, 94 joints progressed to stage ≥ IV (10.9 per 1,000 joint-years of follow-up). In multivariable analysis, baseline radiographic status at the start of regular prophylaxis was strongly associated with progression (vs stage 0: stage I HR 14.95; stage II HR 11.84; stage III HR 34.47). Patients already receiving regular prophylaxis at the first visit had a lower risk of progression (HR 0.54). After adjustment, joint-site differences were not evident, and the independent association of age at first visit to our institute with progression was not estimated with sufficient precision.
Conclusions
Baseline radiographic status at initiation of regular prophylaxis was strongly associated with long-term progression to advanced radiographic HA. These findings support initiating regular prophylaxis before radiographic joint changes become established.
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1186/s12891-026-10343-9.
Keywords: Hemophilia A, Hemophilic arthropathy, Prophylaxis, Arnold–Hilgartner classification, Retrospective cohort study
Background
Hemophilia A is a congenital bleeding disorder caused by deficiency or absence of coagulation factor VIII (FVIII) [1]. Severe hemophilia is defined as FVIII activity < 1 IU/dL (< 1% of normal). Patients with severe hemophilia often develop spontaneous joint bleeding, which triggers chronic synovitis and progressive osteochondral damage. Recurrent hemarthrosis leads to iron deposition in the synovium and a persistent inflammatory response, which impairs cartilage homeostasis and contributes to chondrocyte loss and structural deterioration [2–6]. As a result, hemophilic arthropathy (HA) causes pain, restricted range of motion, and functional disability, substantially affecting quality of life [7, 8]. HA typically involves the elbow, knee, and ankle joints; among these, the ankle is frequently affected and can limit daily activities even in young patients [9, 10].
Prophylactic administration of clotting factor concentrates (CFCs) is a key strategy to prevent joint bleeding and reduce HA progression. Prophylaxis is commonly categorized as primary prophylaxis initiated in the absence of documented joint disease, before the second clinically evident joint bleed, and before three years of age; secondary prophylaxis initiated after two or more joint bleeds but before the onset of joint disease; and tertiary prophylaxis initiated after documented joint disease is present [11]. In clinical practice and research, radiography-based scoring remains widely used to stage HA. The Arnold–Hilgartner classification provides a standardized radiographic assessment, and stage IV or higher indicates cartilage destruction accompanied by joint space narrowing, which reflects clinically important structural progression [12].
Despite major treatment advances, access to effective hemophilia care remains markedly unequal worldwide. In many low-income and lower-middle-income countries, a substantial proportion of people with hemophilia have limited or no access to diagnosis, specialist care, and CFCs [13]. Severe hemophilia also imposes substantial direct and indirect costs on patients, families, healthcare systems, and society [7, 14]. At the same time, recent therapeutic developments include extended half-life FVIII products, non-factor prophylaxis with emicizumab, and gene therapy [11, 15–17]. This contrast increases the need for long-term data that clarify the association between joint status at prophylaxis initiation and subsequent structural progression, particularly for patients who begin regular treatment after structural joint changes have developed.
Evidence supports the benefit of early prophylaxis, particularly primary prophylaxis, in reducing joint damage. However, much of the available evidence derives from follow-up periods of approximately 6–10 years [18, 19]. Data describing progression from radiographic stage at prophylaxis initiation to advanced HA across different joints over several decades remain limited. This question is relevant both in settings where prophylaxis may be delayed and in contemporary care for patients who already have joint damage when regular prophylaxis is started.
Therefore, this retrospective cohort study investigated the risk of developing advanced HA, defined as Arnold–Hilgartner stage ≥ IV, in inhibitor-negative patients with severe hemophilia A followed at our institute. We evaluated whether the risk differed by joint (elbow, knee, and ankle) and how baseline radiographic stage at initiation of regular prophylaxis, age at first visit to our institute, and treatment status at first visit related to subsequent progression. We hypothesized that Arnold–Hilgartner stage at initiation of regular prophylaxis would be associated with the risk of reaching stage ≥ IV and that the ankle would show the highest event rate among the studied joints.
Methods
Study design and setting
This retrospective cohort study evaluated the development of hemophilic arthropathy (HA) in patients with severe hemophilia A who attended the outpatient clinic of the University of Occupational and Environmental Health, Japan, between 1984 and 2020. This was a single-center referral cohort rather than a population-based sample. Patients were primarily referred from the clinical catchment area of our institution, and the cohort was not assembled through a national registry. Therefore, the proportion of all Japanese patients with severe hemophilia A represented by this cohort could not be estimated from the available dataset.
Participants
We enrolled patients with severe hemophilia A who visited our institute during the study period and had no inhibitors (alloantibodies against factor VIII [FVIII]). We evaluated six joints per patient (right/left elbow, knee, and ankle). We excluded observation time after suspension of regular prophylaxis; thus, each joint contributed follow-up time only while the patient met the study definition of treatment status and follow-up availability.
Because a complete screening log (e.g., the number of patients screened and excluded) was not available in the current dataset, we did not present a cohort flow diagram. This limitation is addressed in the Discussion.
Radiographic assessment and definition of hemophilic arthropathy
Plain radiographs of the elbow, knee, and ankle joints were obtained in routine clinical practice using anteroposterior and lateral views. The timing of radiography was determined at the physician’s discretion (e.g., based on symptoms or clinical necessity) rather than at fixed intervals. Because radiography was not performed according to a prespecified schedule, protocol-defined radiographic intervals were not available. Lack of radiography during follow-up reflected ongoing clinical observation without symptoms or physician-assessed need for imaging, rather than withdrawal from follow-up.
Arthropathy was graded using the Arnold–Hilgartner classification [12]: stage 0, normal joint; stage I, no skeletal abnormalities on radiographs with soft tissue swelling; stage II, osteoporosis and epiphyseal overgrowth; stage III, subchondral cysts with preserved joint space; stage IV, joint space narrowing with cartilage destruction; and stage V, fibrous contracture with disorganization of joint structures. We defined the HA progression event as progression to Arnold–Hilgartner stage IV or higher, because stage IV indicates radiographic joint space narrowing and clinically relevant structural progression.
Exposure definition: regular prophylaxis
We recorded the age at initiation of regular FVIII prophylaxis. Patients receiving standard half-life FVIII products at a rate of ≥ 2.5 infusions per week for > 45 weeks per year were considered to have initiated regular prophylaxis [20, 21]. Prophylaxis dose in IU/kg was not consistently available across the 36-year study period; therefore, prophylaxis was classified by infusion frequency and annual treatment duration rather than dose intensity. Extended half-life FVIII products and non-factor therapies such as emicizumab were not evaluated as exposures in this analysis.
Follow-up and joint-years of follow-up
For each joint, the observation period began at the patient’s first visit to our institute and ended at the earliest of: (1) the date when the joint reached Arnold–Hilgartner stage ≥ IV or (2) the date of the last visit or last available radiographic assessment (censoring). Observation time was calculated at the joint level and expressed as joint-years of follow-up for incidence calculations.
When the treatment status changed from on-demand replacement to regular prophylaxis, a new observation interval was created from the start date of regular prophylaxis. Joint arthroplasty or arthrodesis before progression to Arnold–Hilgartner stage ≥ IV was considered a potential competing terminal event. In the available analytic records, no joint arthroplasty or arthrodesis was identified before the endpoint; therefore, a Fine–Gray competing-risk model was not applied. The denominators differed across analyses according to the analytic purpose. Baseline radiographic summaries were based on joints with available radiographs at the first visit, whereas incidence analyses were based on joints that contributed evaluable joint-level follow-up time. Analyses involving prophylaxis status used observation intervals, because each joint could contribute separate intervals before and after initiation of regular prophylaxis. The starting point of follow-up was defined as the first visit to our institute.
Statistical analysis
We estimated hazard ratios (HRs) with 95% confidence intervals (CIs) for progression to Arnold–Hilgartner stage ≥ IV using Cox proportional hazards models. Incidence rates were expressed per 1,000 joint-years of follow-up. Joint bleeding rate was not included in the statistical models because joint bleeding episodes were not systematically recorded in a standardized format across the entire 36-year study period.
The multivariable model included joint site (elbow, knee, or ankle), Arnold–Hilgartner stage at the start of regular prophylaxis, age at the first visit to our institute, regular prophylaxis status at the first visit to our institute, and laterality (left vs right). Because multiple joints from the same participant were included in the analyses, observations within participants were not assumed to be independent. Therefore, all models were estimated using cluster-robust standard errors with participant as the clustering variable to account for within-participant correlation.
Because the association between regular prophylaxis status at the first visit to our institute and progression appeared to differ by baseline arthropathy status, we additionally estimated HRs stratified by Arnold–Hilgartner stage at the start of prophylaxis (stage 0 vs stage ≥ I).
As sensitivity analyses, we first refitted the main model by treating age at first visit to our institute as a continuous variable. We then fitted a second reduced model that also excluded laterality. These analyses were used to evaluate whether the main estimates were sensitive to reduced model complexity.
Ethics approval
The Institutional Review Board of the University of Occupational and Environmental Health, Japan, approved this study (approval number: UOEHCRB21-045). Details regarding ethics approval and consent to participate are provided in the Declarations section.
Results
Study population and baseline radiographic findings
The analysis included 109 patients. The mean age at the first visit to our institute was 14.3 years (SD 9.1), and the median age was 12.4 years (interquartile range, 7.9–18.4). At the first visit, Arnold–Hilgartner stage ≥ I arthropathy was present in 29/200 elbows (15%), 16/200 knees (8%), and 51/186 ankles (27%) (Table 1). Regular prophylaxis was initiated in 87 patients (80%): 20 (18%) at < 3 years, 25 (23%) at 3– < 6 years, 20 (18%) at 6– < 12 years, and 22 (20%) at ≥ 12 years of age. Available participant and treatment characteristics, including the definition of regular prophylaxis and the availability of prophylaxis dose and annualized joint bleeding rate data, are summarized in Supplementary Table S1.
Table 1.
Arnold–Hilgartner stage for each joint on radiographs at the first visit to our institute
| Right | Left | ||
|---|---|---|---|
| Elbow | 0 | 87 | 84 |
| I | 1 | 0 | |
| II | 5 | 4 | |
| III | 6 | 5 | |
| IV | 2 | 5 | |
| V | 1 | 0 | |
| Knee | 0 | 93 | 91 |
| I | 1 | 1 | |
| II | 3 | 2 | |
| III | 2 | 3 | |
| IV | 0 | 1 | |
| V | 2 | 1 | |
| Ankle | 0 | 66 | 69 |
| I | 5 | 6 | |
| II | 5 | 7 | |
| III | 12 | 11 | |
| IV | 3 | 0 | |
| V | 1 | 1 |
Details of the Arnold–Hilgartner stage of each joint on available radiographs at the first visit to our institute in 109 patients analyzed. Arnold–Hilgartner stage I or higher was observed at the first visit in 15% of elbows (29 of 200), 8% of knees (16 of 200), and 27% of ankles (51 of 186)
Radiographs were obtained as part of routine clinical care rather than at prespecified intervals. Therefore, protocol-defined radiographic intervals could not be summarized, and radiographic follow-up was interpreted in the context of ongoing clinical observation.
Incidence of advanced hemophilic arthropathy
Across 583 joints, the total observation time was 8,634 joint-years of follow-up. Hemophilic arthropathy reaching Arnold–Hilgartner stage ≥ IV occurred in 94 joints, corresponding to an incidence of 10.9 per 1,000 joint-years of follow-up. By joint site, stage ≥ IV arthropathy developed in 31/201 elbows (10.5 per 1,000 joint-years of follow-up), 14/202 knees (4.4 per 1,000 joint-years), and 49/180 ankles (19.7 per 1,000 joint-years) (Table 2).
Table 2.
Incidence of progression to hemophilic arthropathy by joint site
| Number of joints observed | Joint-years of follow-up | Number of HA progression events | Incidence rate per 1,000 joint-years of follow-up | |
|---|---|---|---|---|
| Elbow | 201 | 2,944 | 31 | 10.5 |
| Knee | 202 | 3,198 | 14 | 4.4 |
| Ankle | 180 | 2,492 | 49 | 19.7 |
| Total | 583 | 8,634 | 94 | 10.9 |
Overall and joint-specific incidence of progression to hemophilic arthropathy, defined as reaching Arnold–Hilgartner stage ≥ IV. Incidence rates are shown per 1,000 joint-years of follow-up
HA hemophilic arthropathy
The incidence of progression to stage ≥ IV increased with higher Arnold–Hilgartner stage at the start of the analytic interval. Among intervals classified as stage 0 at interval start (517 intervals), 22 progressed to stage ≥ IV (2.9 per 1,000 joint-years). In contrast, 6/21 intervals at stage I progressed (36.6 per 1,000 joint-years), 23/41 intervals at stage II progressed (42.3 per 1,000 joint-years), and 43/47 intervals at stage III progressed (121.9 per 1,000 joint-years).
Among 412 intervals not receiving regular prophylaxis at the first visit to our institute, 54 progressed to stage ≥ IV (14.2 per 1,000 joint-years). Among 456 intervals receiving regular prophylaxis, 40 progressed (8.3 per 1,000 joint-years). The incidence did not differ materially by laterality (right: 52/295 joints, 12.1 per 1,000 joint-years; left: 42/288 joints, 9.6 per 1,000 joint-years) (Table 3).
Table 3.
Incidence of hemophilic arthropathy by interval-level characteristics
| Number of analytic units | Joint-years of follow-up | Number of HA progression events | Incidence rate per 1,000 joint-years of follow-up | |||
|---|---|---|---|---|---|---|
| Interval-start Arnold–Hilgartner stage | Arnold–Hilgartner stage | 0 | 517 | 7,574 | 22 | 2.9 |
| I | 21 | 164 | 6 | 36.6 | ||
| II | 41 | 544 | 23 | 42.3 | ||
| III | 47 | 353 | 43 | 121.9 | ||
| Regular prophylaxis at first visit to our institute | No | 412 | 3,797 | 54 | 14.2 | |
| Yes | 456 | 4,837 | 40 | 8.3 | ||
| Laterality | Right | 295 | 4,280 | 52 | 12.1 | |
| Left | 288 | 4,354 | 42 | 9.6 | ||
Incidence of progression to hemophilic arthropathy, defined as reaching Arnold–Hilgartner stage ≥ IV. Values are shown by analytic characteristics. Observation intervals were created when treatment status or radiographic status changed; therefore, counts for interval-start Arnold–Hilgartner stage and prophylaxis status represent analytic intervals and do not equal the number of unique joints shown in Table 2. Laterality is shown at the joint level. Incidence rates are expressed per 1,000 joint-years of follow-up. HA, hemophilic arthropathy
Risk factors for progression to Arnold–Hilgartner stage ≥ IV
In univariate Cox analysis, compared with the elbow, the knee showed a lower risk of progression (HR 0.40, 95% CI 0.22–0.74; p = 0.003), whereas the ankle showed a higher risk (HR 1.86, 95% CI 1.16–2.98; p = 0.009). In multivariable analysis, the differences by joint site were not clearly evident (knee: HR 0.66, 95% CI 0.38–1.17; p = 0.156; ankle: HR 0.91, 95% CI 0.56–1.50; p = 0.720) (Table 4).
Table 4.
Hazard ratios for progression to Arnold–Hilgartner stage ≥ IV
| Univariate | Multivariable | |||||||
|---|---|---|---|---|---|---|---|---|
| HR | 95% CI | p | HR | 95% CI | p | |||
| Joint | ||||||||
| Elbow | Reference | Reference | ||||||
| Knee | 0.40 | 0.22 | 0.74 | 0.003 | 0.66 | 0.38 | 1.17 | 0.156 |
| Ankle | 1.86 | 1.16 | 2.98 | 0.009 | 0.91 | 0.56 | 1.50 | 0.720 |
| Arnold–Hilgartner stage at initiation of regular prophylaxis | ||||||||
| 0 | Reference | Reference | ||||||
| I | 13.12 | 6.14 | 28.04 | < 0.001 | 14.95 | 6.60 | 33.82 | < 0.001 |
| II | 13.72 | 6.99 | 26.94 | < 0.001 | 11.84 | 5.48 | 25.56 | < 0.001 |
| III | 42.59 | 25.26 | 71.80 | < 0.001 | 34.47 | 18.12 | 65.60 | < 0.001 |
| Age at first visit to our institute | ||||||||
| < 6 | Reference | Reference | ||||||
| 6– < 12 | 2.80 | 0.55 | 14.34 | 0.216 | 2.45 | 0.60 | 9.91 | 0.210 |
| 12– < 18 | 9.42 | 1.95 | 45.55 | 0.005 | 2.63 | 0.68 | 10.23 | 0.162 |
| ≥ 18 | 16.28 | 3.23 | 82.18 | 0.001 | 2.03 | 0.43 | 9.44 | 0.369 |
| Regular prophylaxis at first visit to our institute | ||||||||
| No | Reference | Reference | ||||||
| Yes | 0.44 | 0.28 | 0.68 | < 0.001 | 0.54 | 0.32 | 0.90 | 0.019 |
| Laterality | ||||||||
| Right | Reference | Reference | ||||||
| Left | 0.80 | 0.58 | 1.11 | 0.187 | 0.91 | 0.58 | 1.42 | 0.675 |
Univariate and multivariable Cox analyses of factors associated with progression to hemophilic arthropathy, defined as reaching Arnold–Hilgartner stage ≥ IV. The multivariable model included joint site, Arnold–Hilgartner stage at initiation of regular prophylaxis, age at first visit to our institute, regular prophylaxis status at first visit to our institute, and laterality. Cox models were estimated using cluster-robust standard errors with participant as the clustering variable
HA hemophilic arthropathy, HR hazard ratio, CI confidence interval
Arnold–Hilgartner stage at the initiation of regular prophylaxis was strongly associated with progression. Compared with stage 0, the multivariable HRs were 14.95 (95% CI 6.60–33.82) for stage I, 11.84 (95% CI 5.48–25.56) for stage II, and 34.47 (95% CI 18.12–65.60) for stage III (all p < 0.001) (Table 4).
In univariate analysis of age at first visit to our institute (reference: < 6 years), the HRs were 2.80 (95% CI 0.55–14.34; p = 0.216) for 6– < 12 years, 9.42 (95% CI 1.95–45.55; p = 0.005) for 12– < 18 years, and 16.28 (95% CI 3.23–82.18; p = 0.001) for ≥ 18 years. In multivariable analysis, the independent association between age at first visit to our institute and progression was not estimated with sufficient precision, as reflected by wide 95% CIs (6– < 12 years: HR 2.45, 95% CI 0.60–9.91; p = 0.210; 12– < 18 years: HR 2.63, 95% CI 0.68–10.23; p = 0.162; ≥ 18 years: HR 2.03, 95% CI 0.43–9.44; p = 0.369) (Table 4).
Regular prophylaxis at the first visit to our institute was associated with a lower risk of progression compared with no regular prophylaxis at the first visit (univariate HR 0.44, 95% CI 0.28–0.68; p < 0.001; multivariable HR 0.54, 95% CI 0.32–0.90; p = 0.019) (Table 4). Laterality was not clearly associated with progression (left vs right: multivariable HR 0.91, 95% CI 0.58–1.42; p = 0.675) (Table 4).
Stratified analysis by baseline arthropathy status at prophylaxis initiation
Because the association between prophylaxis status at the first visit to our institute and progression appeared to differ by baseline arthropathy status, we performed analyses stratified by Arnold–Hilgartner stage at prophylaxis initiation (stage 0 vs stage ≥ I). Among joints with stage 0 at prophylaxis initiation, regular prophylaxis at the first visit to our institute was associated with a lower risk of progression (HR 0.21, 95% CI 0.08–0.53; p = 0.001). In contrast, among joints with stage ≥ I at prophylaxis initiation, prophylaxis status at the first visit to our institute was not clearly associated with progression (HR 0.95, 95% CI 0.56–1.62; p = 0.843) (Table 5).
Table 5.
Stratified analysis according to arthropathy status at initiation of regular prophylaxis
| Arnold–Hilgartner stage at initiation of regular prophylaxis | Stage 0 | Stage ≥ I | |||||||
|---|---|---|---|---|---|---|---|---|---|
| HR | 95% CI | p | HR | 95% CI | p | ||||
| Regular prophylaxis at first visit to our institute | No | Reference | Reference | ||||||
| Yes | 0.21 | 0.08–0.53 | 0.001 | 0.95 | 0.56–1.62 | 0.843 | |||
Analysis of the association between prophylaxis status at the first visit to our institute and subsequent progression to hemophilic arthropathy, stratified by Arnold–Hilgartner stage at initiation of regular prophylaxis (stage 0 vs stage ≥ I)
Among joints classified as stage 0 at prophylaxis initiation, regular prophylaxis at the first visit to our institute was associated with a lower risk of progression to hemophilic arthropathy (HR 0.21, 95% CI 0.08–0.53; p = 0.001). In contrast, among joints classified as stage ≥ I at prophylaxis initiation, prophylaxis status at the first visit to our institute was not clearly associated with subsequent progression (HR 0.95, 95% CI 0.56–1.62; p = 0.843)
CI confidence interval, HA hemophilic arthropathy, HR hazard ratio
The Kaplan–Meier curves showed lower progression-free survival with higher Arnold–Hilgartner stage at prophylaxis initiation in the overall joint-level analysis and within each joint site (Fig. 1). This pattern was most evident for joints classified as stage III at prophylaxis initiation.
Fig. 1.
Kaplan–Meier curves for progression-free survival according to Arnold–Hilgartner stage at initiation of regular prophylaxis. Kaplan–Meier curves show time to progression to Arnold–Hilgartner stage ≥ IV according to Arnold–Hilgartner stage at initiation of regular prophylaxis. Curves are shown for all joints and separately for elbows, knees, and ankles. AH, Arnold–Hilgartner
Sensitivity analyses supported the stability of the main results. When age at first visit to our institute was treated as a continuous variable, the HRs for Arnold–Hilgartner stages I, II, and III remained similar to the main model (stage I HR 13.99, 95% CI 6.04–32.40; stage II HR 11.26, 95% CI 5.21–24.30; stage III HR 34.09, 95% CI 17.98–64.62). Similar results were obtained in the second sensitivity model, which treated age as a continuous variable and excluded laterality from the model (stage I HR 13.54, 95% CI 5.91–31.03; stage II HR 11.58, 95% CI 5.42–24.71; stage III HR 34.33, 95% CI 18.13–65.01) (Supplementary Table S2). In a descriptive supplementary analysis limited to regular prophylaxis-related analytic intervals, later age at prophylaxis initiation showed a descriptive relationship with a higher proportion of advanced Arnold–Hilgartner stages at the start of prophylaxis (Supplementary Table S3).
Discussion
Principal findings and original contribution
The main finding of this study was that Arnold–Hilgartner stage at initiation of regular prophylaxis was strongly associated with subsequent progression to advanced radiographic HA. The 36-year joint-level design allowed the elbows, knees, and ankles to be evaluated separately across 8,634 joint-years of follow-up using a clinically meaningful structural endpoint. The original contribution of this study is its joint-level, stage-specific evaluation of the association between radiographic status at prophylaxis initiation and long-term progression. Adjusted joint-site differences were not evident. The association between age at first visit to our institute and progression was estimated with limited precision and should not be interpreted as evidence that age is unimportant.
Interpretation in relation to prophylaxis timing
Primary prophylaxis initiated before recurrent joint bleeding or established arthropathy is recommended in international guidelines and reduces long-term joint damage [11, 18, 19, 22–24]. Secondary prophylaxis initiated after two or more joint bleeds but before established arthropathy can also reduce bleeding and preserve joint outcomes [18, 24]. Tertiary prophylaxis remains clinically beneficial after arthropathy has developed. In the SPINART trial, late prophylaxis reduced bleeding, pain, and healthcare resource use and improved joint health and quality of life, although structural progression on magnetic resonance imaging did not differ clearly from on-demand treatment over three years [25]. Because joint bleeding histories were not systematically available, the present cohort could not be classified directly according to standard definitions of primary, secondary, and tertiary prophylaxis. These previous studies provide clinical context for our finding that joints classified as stage ≥ I at prophylaxis initiation had a higher long-term risk of progression than stage 0 joints. However, this result does not indicate that prophylaxis initiated after radiographic changes is ineffective or should be withheld.
The interpretation of timing nevertheless requires caution. Joint bleeding episodes were not systematically recorded across the full study period, and the age variable in the Cox model was age at first visit to our institute rather than age at prophylaxis initiation. The study therefore cannot define an optimal chronological age or number of joint bleeds at which prophylaxis should begin. Instead, the results indicate that radiographic status at prophylaxis initiation identifies joints with different long-term risks of advanced structural progression.
Clinical implications for long-term joint management
These findings have three practical implications. First, the present findings, together with international recommendations, support initiation of regular prophylaxis before radiographic HA is established whenever feasible [11]. Second, patients who begin prophylaxis after radiographic changes have been identified should not be considered treatment failures. They may still benefit from bleeding prevention, symptom control, and improved function, but their existing joint changes identify a group that requires continued joint assessment and individualized management [11, 25]. Third, long-term assessment should include the elbows, knees, and ankles because the adjusted risk of progression did not differ clearly by joint site and the distribution of HA varies among these joints [9, 26].
The present study did not compare imaging modalities or radiographic schedules and therefore cannot support a fixed interval for plain radiography. However, documenting joint status when regular prophylaxis is initiated may help clinicians identify patients who remain at increased risk despite treatment. In these patients, clinical joint examination and appropriate imaging can be incorporated into longitudinal care according to symptoms, age, treatment history, and local resources [11].
Global and contemporary relevance
The findings also have relevance beyond the historical treatment setting of this cohort. Access to diagnosis, specialist care, and continuous prophylaxis remains limited in many low-resource settings [13]. In such settings, patients may begin prophylaxis only after structural joint changes have developed, making long-term data on radiographic progression clinically relevant. In well-resourced settings, extended half-life FVIII products and emicizumab have expanded prophylactic options, while gene therapy trials have shown reduced bleeding and FVIII use [15–17]. Most follow-up in the present study preceded the widespread use of these therapies, and they were not evaluated as study exposures. Accordingly, the incidence rates observed in this historical cohort should not be directly extrapolated to patients receiving contemporary prophylactic regimens.
Nevertheless, the central observation remains relevant to patients with pre-existing arthropathy when contemporary therapy is started. The study shows that established radiographic changes at prophylaxis initiation are associated with later progression over extended follow-up. Prospective registries should examine whether this association persists under extended half-life FVIII, emicizumab, and gene therapy while collecting standardized information on bleeding, adherence, dose intensity, physical joint findings, and imaging.
Strengths, generalizability and limitations
The strengths of this study include the extended observation period, joint-level evaluation of 583 joints, use of an advanced radiographic endpoint, adjustment for within-participant correlation, and sensitivity analyses of model specification. However, this was a single-center referral cohort rather than a population-based sample. The proportion of all Japanese patients with severe hemophilia A represented by the cohort could not be estimated. Referral patterns may have selected patients with more severe disease, better access to specialist care, or more complete follow-up. In addition, patients entered the cohort at different ages and disease stages across the 36-year period, which may limit generalizability and influence the estimated incidence.
Several measurement limitations should also be considered. A complete screening log was unavailable, which may have introduced selection bias if unrecorded or excluded patients differed from those analyzed. Radiographs were obtained during routine care rather than at fixed intervals, which may have delayed detection of asymptomatic progression and introduced interval censoring. Lack of imaging reflected ongoing clinical observation without symptoms or physician-assessed need for radiography rather than withdrawal from follow-up, but this practice may still have affected event timing. Prophylaxis dose, FVIII trough levels, adherence, treatment interruptions, and annualized joint bleeding rates were not consistently available. These factors may have influenced both radiographic status at prophylaxis initiation and later progression. Previous studies have shown that prophylaxis intensity and treatment interruption can affect long-term joint outcomes [27–32].
Finally, joint arthroplasty or arthrodesis before the radiographic endpoint could act as a competing terminal event. No such procedure was identified before progression to stage ≥ IV in the available analytic records, so a Fine–Gray competing-risk model was not applied. If procedures were not captured, progression risk may have been underestimated. Future studies should use standardized imaging intervals and prospectively record surgical endpoints, treatment exposure, adherence, and bleeding outcomes.
Conclusions
Baseline radiographic status at initiation of regular prophylaxis was strongly associated with long-term progression to advanced radiographic HA in patients with severe hemophilia A. These findings support initiation of regular prophylaxis before radiographic HA is identified and indicate that joints with established changes at prophylaxis initiation require continued assessment and individualized management. The independent association between age at first visit to our institute and progression was estimated with limited precision.
Supplementary Information
Supplementary file 1: Supplementary Table S1. Available participant and treatment characteristics. Supplementary Table S2. Sensitivity analyses for model specification. Supplementary Table S3. Distribution of Arnold–Hilgartner stage at initiation of regular prophylaxis by age at prophylaxis initiation.
Acknowledgements
Not applicable.
Abbreviations
- CFCs
Clotting factor concentrates
- CI
Confidence interval
- FVIII
Coagulation factor VIII
- HA
Hemophilic arthropathy
- HR
Hazard ratio
- UMIN-CTR
UMIN Clinical Trials Registry
Authors’ contributions
HS, RS, TT, YY, KT, MT, and TA contributed to the study design and implementation and radiograph analysis. HS and YF contributed to the analysis of the results, formal analysis, and manuscript writing. RF and AS conceived the original draft and supervised the project. All authors had full access to all the data in the study and accept responsibility for submission for publication. All authors read and approved the final manuscript.
Funding
This work was supported by a Grant-in-Aid for Scientific Research (C) from the Japan Society for the Promotion of Science (JSPS; Tokyo, Japan) to HS (grant number 23K08644).
Data availability
The datasets generated and/or analysed during the current study are not publicly available due to protection of personal information but are available from the corresponding author on reasonable request. The other contact points are the Office of the Center for Clinical Research Promotion, Division of Clinical Research, University of Occupational and Environmental Health (E-mail: rinshokenkyu@mbox.clnc.uoeh-u.ac.jp; Tel: + 81 93 691 7127) and the Ethics Committee, Graduate School of Medicine, University of Occupational and Environmental Health.
Declarations
Ethics approval and consent to participate
The authors conducted the present study in accordance with the Declaration of Helsinki and the Ethical Guidelines for Life Science and Medical Research Involving Human Subjects. The Institutional Review Board of the University of Occupational and Environmental Health, Japan, approved this study (approval number: UOEHCRB21-045) and waived the requirement for informed consent according to institutional policy for retrospective studies. An opt-out approach was used in accordance with institutional policy. This study was registered in the UMIN Clinical Trials Registry as UMIN000057959.
Consent for publication
Not applicable.
Competing interests
R. Shirayama reports grant support from Chugai Pharmaceutical Co., Ltd., paid to her institution; honoraria for lectures from Bayer Yakuhin, Ltd., Chugai Pharmaceutical Co., Ltd., CSL Behring K.K., Fujimoto Pharmaceutical Corporation, Japan Blood Products Organization, KM Biologics Co., Ltd., Novo Nordisk Pharma Ltd., Pfizer Japan Inc., Sanofi K.K., and Takeda Pharmaceutical Co., Ltd.; and advisory board fees from Sanofi K.K. All other authors declare no competing interests.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Bolton-Maggs PH, Pasi KJ. Haemophilias A and B. Lancet. 2003;361(9371):1801–9. [DOI] [PubMed] [Google Scholar]
- 2.Melchiorre D, Milia AF, Linari S, Romano E, Benelli G, Manetti M, et al. RANK-RANKL-OPG in hemophilic arthropathy: from clinical and imaging diagnosis to histopathology. J Rheumatol. 2012;39(8):1678–86. [DOI] [PubMed] [Google Scholar]
- 3.Den Uijl IE, Mauser Bunschoten EP, Roosendaal G, Schutgens RE, Biesma DH, Grobbee DE, et al. Clinical severity of haemophilia A: does the classification of the 1950s still stand? Haemophilia. 2011;17(6):849–53. [DOI] [PubMed] [Google Scholar]
- 4.White GC 2nd, Rosendaal F, Aledort LM, Lusher JM, Rothschild C, Ingerslev J, et al. Definitions in hemophilia. Recommendation of the scientific subcommittee on factor VIII and factor IX of the scientific and standardization committee of the International Society on Thrombosis and Haemostasis. Thromb Haemost. 2001;85(3):560. [PubMed]
- 5.Pollmann H, Richter H, Ringkamp H, Jurgens H. When are children diagnosed as having severe haemophilia and when do they start to bleed? A 10-year single-centre PUP study. Eur J Pediatr. 1999;158(Suppl 3):S166–70. [DOI] [PubMed] [Google Scholar]
- 6.Roosendaal G, Mauser-Bunschoten EP, De Kleijn P, Heijnen L, van den Berg HM, Van Rinsum AC, et al. Synovium in haemophilic arthropathy. Haemophilia. 1998;4(4):502–5. [DOI] [PubMed] [Google Scholar]
- 7.Fischer K, van der Bom JG, Mauser-Bunschoten EP, Roosendaal G, van den Berg HM. Effects of haemophilic arthropathy on health-related quality of life and socio-economic parameters. Haemophilia. 2005;11(1):43–8. [DOI] [PubMed] [Google Scholar]
- 8.Hooiveld M, Roosendaal G, Vianen M, van den Berg M, Bijlsma J, Lafeber F. Blood-induced joint damage: long-term effects in vitro and in vivo. J Rheumatol. 2003;30(2):339–44. [PubMed] [Google Scholar]
- 9.Chang CY, Li TY, Cheng SN, Pan RY, Wang HJ, Lin SY, et al. Prevalence and severity by age and other clinical correlates of haemophilic arthropathy of the elbow, knee and ankle among Taiwanese patients with haemophilia. Haemophilia. 2017;23(2):284–91. [DOI] [PubMed] [Google Scholar]
- 10.Rodriguez-Merchan EC. Musculoskeletal complications of hemophilia. HSS J. 2010;6(1):37–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Srivastava A, Santagostino E, Dougall A, Kitchen S, Sutherland M, Pipe SW, et al. WFH guidelines for the management of hemophilia, 3rd edition. Haemophilia. 2020;26(Suppl 6):1–158. [DOI] [PubMed] [Google Scholar]
- 12.Arnold WD, Hilgartner MW. Hemophilic arthropathy. Current concepts of pathogenesis and management. J Bone Joint Surg Am. 1977;59(3):287–305. [PubMed] [Google Scholar]
- 13.Pierce GF, Adediran M, Diop S, Dunn AL, El Ekiaby M, Kaczmarek R, et al. Achieving access to haemophilia care in low-income and lower-middle-income countries: expanded Humanitarian Aid Program of the World Federation of Hemophilia after 5 years. Lancet Haematol. 2022;9(9):e689–97. [DOI] [PubMed] [Google Scholar]
- 14.O’Hara J, Hughes D, Camp C, Burke T, Carroll L, Garcia Diego DA. The cost of severe haemophilia in Europe: the CHESS study. Orphanet J Rare Dis. 2017;12:106. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Mahlangu J, Powell JS, Ragni MV, Chowdary P, Josephson NC, Pabinger I, et al. Phase 3 study of recombinant factor VIII Fc fusion protein in severe hemophilia A. Blood. 2014;123(3):317–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Mahlangu J, Oldenburg J, Paz-Priel I, Negrier C, Niggli M, Mancuso ME, et al. Emicizumab prophylaxis in patients who have hemophilia A without inhibitors. N Engl J Med. 2018;379(9):811–22. [DOI] [PubMed] [Google Scholar]
- 17.Ozelo MC, Mahlangu J, Pasi KJ, Giermasz A, Leavitt AD, Laffan M, et al. Valoctocogene roxaparvovec gene therapy for hemophilia A. N Engl J Med. 2022;386(11):1013–25. [DOI] [PubMed] [Google Scholar]
- 18.Schobess R, Kurnik K, Friedrichs F, Halimeh S, Krumpel A, Bidlingmaier C, et al. Effects of primary and secondary prophylaxis on the clinical expression of joint damage in children with severe haemophilia A. Results of a multicenter non-concurrent cohort study. Thromb Haemost. 2008;99(1):71–6. [DOI] [PubMed]
- 19.Manco-Johnson MJ, Abshire TC, Shapiro AD, Riske B, Hacker MR, Kilcoyne R, et al. Prophylaxis versus episodic treatment to prevent joint disease in boys with severe hemophilia. N Engl J Med. 2007;357(6):535–44. [DOI] [PubMed] [Google Scholar]
- 20.Scott MJ, Xiang H, Hart DP, Palmer B, Collins PW, Stephensen D, et al. Treatment regimens and outcomes in severe and moderate haemophilia A in the UK: the THUNDER study. Haemophilia. 2019;25(2):205–12. [DOI] [PubMed] [Google Scholar]
- 21.Blanchette VS, Key NS, Ljung LR, Manco-Johnson MJ, van den Berg HM, Srivastava A. Definitions in hemophilia: communication from the SSC of the ISTH. J Thromb Haemost. 2014;12(11):1935–9. [DOI] [PubMed] [Google Scholar]
- 22.Manco-Johnson MJ, Soucie JM, Gill JC; Joint Outcomes Committee of the Universal Data Collection, US Hemophilia Treatment Center Network. Prophylaxis usage, bleeding rates, and joint outcomes of hemophilia, 1999 to 2010: a surveillance project. Blood. 2017;129(17):2368–74. [DOI] [PMC free article] [PubMed]
- 23.Astermark J, Petrini P, Tengborn L, Schulman S, Ljung R, Berntorp E. Primary prophylaxis in severe haemophilia should be started at an early age but can be individualized. Br J Haematol. 1999;105(4):1109–13. [DOI] [PubMed] [Google Scholar]
- 24.Fischer K, van der Bom JG, Mauser-Bunschoten EP, Roosendaal G, Prejs R, de Kleijn P, et al. The effects of postponing prophylactic treatment on long-term outcome in patients with severe hemophilia. Blood. 2002;99(7):2337–41. [DOI] [PubMed] [Google Scholar]
- 25.Manco-Johnson MJ, Lundin B, Funk S, Peterfy C, Raunig D, Werk M, et al. Effect of late prophylaxis in hemophilia on joint status: a randomized trial. J Thromb Haemost. 2017;15(11):2115–24. [DOI] [PubMed] [Google Scholar]
- 26.Liu S, Zhou RF, Jin ZB, Wu M, Zhang PY. Age-related severity and distribution of haemophilic arthropathy of the knee, ankle and elbow among Chinese patients with haemophilia. Haemophilia. 2020;26(1):129–35. [DOI] [PubMed] [Google Scholar]
- 27.Petrini P. What factors should influence the dosage and interval of prophylactic treatment in patients with severe haemophilia A and B? Haemophilia. 2001;7(1):99–102. [DOI] [PubMed] [Google Scholar]
- 28.Fischer K, Astermark J, van der Bom JG, Ljung R, Berntorp E, Grobbee DE, et al. Prophylactic treatment for severe haemophilia: comparison of an intermediate-dose to a high-dose regimen. Haemophilia. 2002;8(6):753–60. [DOI] [PubMed] [Google Scholar]
- 29.Fischer K, Steen Carlsson K, Petrini P, Holmstrom M, Ljung R, van den Berg HM, et al. Intermediate-dose versus high-dose prophylaxis for severe hemophilia: comparing outcome and costs since the 1970s. Blood. 1970;122(7):1129–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Fischer K, van den Berg M. Prophylaxis for severe haemophilia: clinical and economical issues. Haemophilia. 2003;9(4):376–81. [DOI] [PubMed] [Google Scholar]
- 31.Nijdam A, Foppen W, De Kleijn P, Mauser-Bunschoten EP, Roosendaal G, van Galen KP, et al. Discontinuing early prophylaxis in severe haemophilia leads to deterioration of joint status despite low bleeding rates. Thromb Haemost. 2016;115(5):931–8. [DOI] [PubMed] [Google Scholar]
- 32.van Dijk K, Fischer K, van der Bom JG, Scheibel E, Ingerslev J, van den Berg HM. Can long-term prophylaxis for severe haemophilia be stopped in adulthood? Results from Denmark and the Netherlands. Br J Haematol. 2005;130(1):107–12. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary file 1: Supplementary Table S1. Available participant and treatment characteristics. Supplementary Table S2. Sensitivity analyses for model specification. Supplementary Table S3. Distribution of Arnold–Hilgartner stage at initiation of regular prophylaxis by age at prophylaxis initiation.
Data Availability Statement
The datasets generated and/or analysed during the current study are not publicly available due to protection of personal information but are available from the corresponding author on reasonable request. The other contact points are the Office of the Center for Clinical Research Promotion, Division of Clinical Research, University of Occupational and Environmental Health (E-mail: rinshokenkyu@mbox.clnc.uoeh-u.ac.jp; Tel: + 81 93 691 7127) and the Ethics Committee, Graduate School of Medicine, University of Occupational and Environmental Health.