Hemophilia A and hemophilia B: focus on arthropathy and variables affecting bleeding severity and prophylaxis
- Manuscript handled by: L. Aledort
- Final decision: F. R. Rosendaal, 6 June 2013
Correspondence: Miguel Escobar, University of Texas Health Science Center, 6655 Travis, Suite 400, Houston, TX 77030, USA.
Tel.: +713 500 8360; fax: 713 500 8364.
Hemophilia A (HA) and hemophilia B (HB) are X-linked, recessive disorders. Although their clinical manifestations are essentially indistinguishable, it has been suggested that bleeding episodes in patients with HA are generally more severe and occur at higher frequency than in patients with HB. Nevertheless, considerable debate remains regarding the relative severity of HA and HB. Based on the relative risk of undergoing joint arthroplasty, it appears that patients with HA have more severe joint deterioration compared with patients with HB. Although it is difficult to speculate on the factors that might modify bleeding severity in patients with hemophilia, recent observations indicate that other coagulation proteins, such as tissue factor pathway inhibitor or polymorphisms in coagulation factor genes and genetic defects associated with hypercoagulability may account for the variability in clinical phenotype among patients with hemophilia. Numerous studies have provided evidence supporting the clinical and social benefits of administration of clotting factor in prophylaxis. However, it is still unclear why this approach is more commonly utilized in patients with HA than in those with HB.
Hemophilia A (HA) and hemophilia B (HB) are X-linked, recessive, inherited blood disorders [1, 2]. As of 2005, it was estimated that approximately 400 000 people worldwide were affected by hemophilia . The prevalence of HA is approximately five times that of HB , with a worldwide frequency estimated at one per 5000–7000 male births . In HA and HB, respectively, the underlying genetic defect results in a deficiency or complete absence of the plasma proteins, factor (F) VIII and FIX, respectively. Severe deficiency of FVIII or FIX leads to recurrent hemarthroses and bleeding episodes in soft tissues and other organs. FVIII circulates in an inactive, pro-factor form in complex with the carrier von Willebrand factor (VWF). Proteolytic activation by thrombin releases FVIII from VWF and results in production of the activated co-factor, FVIIIa. Activation enhances the affinity of FVIIIa for anionic phospholipid surfaces, which facilitates its association with the serine protease FIXa in the tenase complex. As a result of assembling with FIXa on the surface of the tenase complex, FVIIIa enhances the catalytic efficiency of FIXa toward FX by several orders of magnitude, resulting in the conversion of FX to FXa. This reaction is critical for the propagation of coagulation because FXa catalyzes the conversion of prothrombin to thrombin .
In patients with hemophilia, the bleeding phenotype usually correlates with measurable plasma factor activity, and the bleeding manifestations in both HA and HB are identical. The FVIII and FIX Subcommittee of the Scientific and Standardisation Committee of the International Society on Thrombosis and Haemostasis defines disease severity in hemophilia based on the plasma level of FVIII or FIX activity; the severe form is associated with a factor level < 0.01 IU mL−1 (1% of normal), the moderate form with a factor level of 0.01–0.05 IU mL−1 (1–5% of normal), and the mild form with a factor level > 0.05–0.40 IU mL−1 (more than 5–40% of normal) . Patients with moderate disease (FVIII/FIX levels between 1% and 5% of normal) experience less severe symptoms; bleeding into the joints and muscles occurs after minor injuries, but excessive bleeding may occur after surgery or dental procedures. Patients with mild disease (FVIII/FIX levels between 5% and 40% of normal) do not experience spontaneous bleeds; bleeding generally only occurs after surgery, dental procedures or trauma. Variability in the severity of hemorrhagic episodes is known to occur among patients with similar circulating plasma levels of FVIII or FIX . Although the clinical manifestations of HA and HB are essentially indistinguishable, recent evidence suggests that the impact of severe disease on joint status may be more detrimental in patients with HA than those with HB .
Molecular defects in hemophilia
The types of mutations that underlie coagulation factor deficiency differ between HA and HB. Databases that register HA and HB mutations indicate a total of 2107 known mutations for HA and 1094 mutations for HB [6, 7]. Hemophilia most commonly arises from point mutations (90% of patients) including missense mutations, nonsense mutations or RNA splice site mutations . Rearrangements or inversions are rare in hemophilia except for an inversion and translocation of exons 1–22, which results in complete disruption of the FVIII gene. This rearrangement is the most common mutation leading to HA and causes approximately 45% of severe cases . In contrast to HA, a large variety of mutations are responsible for HB; gross gene rearrangements/deletions/inversions account for approximately 3% of mutations overall in HB .
While null mutations that prevent synthesis of FVIII or FIX are usually associated with undetectable factor activity, non-null mutations account for variable factor levels and have been shown to be a main determinant of bleeding severity . For example, small mutations within or near exon 14 of the FVIII gene poly-A runs, a potential source of replication, transcription or translation errors, may lead to partial correction of the FVIII open reading frame, resulting in continued production of some functional FVIII protein .
In approximately 5% of HA patients, FVIII protein levels are relatively high (~30% of normal), but the FVIII protein is not functional and is classified as cross-reacting material (CRM+). The mutations thought to be responsible for CRM+ HA are generally missense mutations found in the A2-domain of FVIII . Approximately one-third of patients with HB are classified as CRM+ . The mutations responsible for this form of HB generally affect post-translational modifications, including γ-carboxylation, FIX activation, ligand binding and substrate recognition . Another form of HB is HB Leyden, which is characterized by severe childhood bleeding with FIX antigen and activity levels < 1% of normal . Upon puberty or with the administration of androgens, FIX protein levels increase to between 30% and 60% of normal, which is sufficient to resolve the spontaneous bleeding phenotype [13, 14].
Mutations are also among the major risk factors for the formation of alloantibodies with inhibitory activity toward FVIII or FIX, estimated to affect approximately 25–30% of treated patients with severe HA and approximately 1.5–5% of treated patients with severe HB [15, 16]. In HA, a higher frequency of inhibitor development is observed among patients with more severe mutations, such as the intron 22 inversion, large deletions and nonsense mutations, compared with patients carrying missense mutations or small deletions (35% vs. 5%) . In HB, gene deletions or rearrangements have been found to be associated with an approximately 50% risk of inhibitor development compared with 20% for frameshift, premature, stop or splice-site mutations .
Bleeding frequency and impact of disease
While recurrent hemorrhage is characteristic of severe hemophilia, it appears that the bleeding frequency and, potentially, the impact of bleeding episodes may differ between HA and HB. Regarding bleeding frequency, in a recently published 3-year study of 88 patients with moderate and severe HA and HB, Nagel et al. concluded that bleeding occurred more frequently in patients with HA (14.4 bleeds per patient year−1) than in patients with HB (8.63 bleeds per patient year−1) . In the same study, consumption of FVIII compared with FIX, either in the prophylactic setting or on demand, was not statistically different between patients with HA and those with HB. This finding may be due to the lower in vivo recovery of infused recombinant FIX compared with that of recombinant FVIII [17, 18] paired with the increased frequency of FVIII administration (three times per week) in a prophylactic setting compared with FIX administration (two times per week) .
Bleeding, particularly intracranial hemorrhage, remains the leading cause of death for patients with severe hemophilia . According to a survey of 6018 patients with hemophilia in the United Kingdom who were not infected with human immunodeficiency virus (HIV), hemophilia increased the all-cause death rate, with standardized mortality ratios of 1.19 (95% confidence interval [CI], 1.09–1.29) for patients with mild/moderate disease and 2.69 (95% CI: 2.37–3.05) for patients with severe disease . However, the relative impact of severe bleeding episodes in patients with HA compared with HB on morbidity and mortality has been a subject of ongoing debate, largely fueled by a lack of conclusive epidemiologic findings and the small population of patients with HB. In the aforementioned UK survey, covering a 23-year period, the all-cause death rates for mild or moderate hemophilia were similar for HA and HB (adjusted death rate ratio of 0.96; 95% CI, 0.77–1.21) . While there was a trend toward a higher all-cause death rate for patients with severe HA compared with severe HB, it was not statistically significant (adjusted death rate ratio of 0.71; 95% CI, 0.49–1.04). Additionally, after adjusting for age, calendar period and inhibitor status, mortality rates due to intracranial hemorrhage and from bleeding of any type were not different for HA compared with HB . However, other studies have shown that the impact of bleeding episodes in patients with HA was generally more serious than the impact of such episodes in patients with HB when evaluated by bleed frequency , proportion of patients who had undergone arthroplasties  and Hemophilia Severity Score (HSS) . The HSS was proposed by Schulman et al. as a composite score based on the annual incidence of joint bleeds, joint score and annual clotting factor consumption in 100 patients with hemophilia evaluated over 10 years . The investigators demonstrated that patients with severe HA had a significantly higher HSS than patients with severe HB (0.50 vs. 0.29; P = 0.031). However, for patients with moderate or mild hemophilia, HA was shown to have a lower HSS than HB, reaching statistical significance for mild disease (0.0008 vs. 0.052; P = 0.027) but not moderate disease (0.073 vs. 0.115; P = 0.98).
Similarly conflicting results were found in a 5-year evaluation of all patients with HA and HB treated at the German Vivantes Klinikum . In this study, a greater proportion of patients with HA (111 of 181) had severe hemophilia compared with those with HB (12 of 34). However, a larger percentage of patients with severe HB had experienced intracranial bleeding (4 of 12) compared with those with severe HA (5 of 111). The authors suggest that patients with HB may have a milder bleeding phenotype than those with HA, but are simultaneously at a higher risk of intracranial hemorrhage .
Recurrent joint bleeds in patients with hemophilia lead to chronic arthropathy, a frequent complication and an indicator of the severity of the disease . In a retrospective analysis of data collected from 29 Italian hemophilia centers, hemophilic arthropathy was shown to be more common among patients with HA than those with HB when patients were assessed based on their relative risk of undergoing joint arthroplasty . In this study, investigators reported that patients with HA had a > 3-fold higher risk of undergoing joint arthroplasty compared with patients with HB (odds ratio [OR], 3.38; 95% CI, 1.97–5.77). These results were not affected by HIV status, hepatitis C virus status or inhibitor status, according to Cox regression analysis. The authors of this study have suggested that the difference in clinical severity between HA and HB (reflected by joint arthroplasty rates) may be due to the fact that the point mutations that frequently lead to HB are often less severe than the mutations that cause HA, such as null mutations.
More recent data suggest that long-term complications are more common in HA compared with HB [17, 23]. In the Nagel et al. 3-year study (N = 88 with moderate and severe HA and HB), a higher percentage of patients with HA (14.7%; 10 of 68 patients with HA) underwent a surgical procedure due to musculoskeletal complications than did patients with HB (4.7%; 1 of 21 patients with HB) . Although less common, hip abnormalities were evaluated using the Universal Data Collection (UDC) database because these abnormalities can significantly impact mobility. Hip abnormalities were more prevalent among patients with HA (18%; 1125 of 6419 patients with HA) compared with those with HB (14%; 247 of 1773 patients with HB; P = 0.0003), with HA also found to be a significant independent predictor of hip abnormality using logistic regression (OR, 1.3; 95% CI, 1.0–1.4) .
Potential mechanisms underlying variations in disease severity
While it is difficult to speculate on the factors that might modify bleeding severity in patients with hemophilia, recent literature has proposed that other coagulation proteins or the patient's genetic background may affect clinical phenotype. Tissue factor pathway inhibitor (TFPI) inhibits activated FVII/tissue factor (FVIIa/TF) complex in an FXa-dependent manner. In one study, free plasma levels of TFPI were found to be lower in HB than in HA for patients with severe disease (HB, 9.3 ng mL−1; HA, 11.8 ng mL−1; P = 0.0250), mild or moderate disease (HB, 8.8 ng mL−1; HA, 12.4 ng mL−1; P = 0.0316), and when disease severity was not taken into account (HB, 9.0 ng mL−1; HA, 12.1 ng mL−1; P = 0.0001) . The authors speculated that this difference may potentially contribute to the less severe bleeding observed in HB patients . In addition, Maroney et al. demonstrated that in a HA FVIII-null (F8−/−) mouse model, selective elimination of TFPI from hematopoietic cells reduced bleeding and increased clot volume following vascular injury, suggesting that a lack of FVIIa/TF inhibition by TFPI may compensate for FIX or FVIII deficiency in patients with hemophilia . Although the role of TFPI in moderating bleeding severity has not been directly investigated in hemophilia patients, a study of low TFPI in FV-deficient patients suggests a possible protective effect of low TFPI levels against severe bleeding . Plasma levels of TFPI were lower in FV-deficient patients compared with controls, and patients with partial FV deficiency had an intermediate level of TFPI. Low TFPI levels decreased the FV requirement for minimal thrombin generation in FV-deficient plasma to < 1% . In the same study, patients with HA were also found to have significantly lower free TFPI levels compared with control patients (9.8 ± 2.4 ng mL−1 vs. 15.3 ± 3.3 ng mL−1; P = 0.001), suggesting that TFPI deficiency may also confer some protective effect in HA.
In a genotype analysis of 114 patients with severe hemophilia with either a ‘mild’ phenotype (≤ 5 bleeds in the preceding year, < 10 WFH clinical score and < 10 Pettersson radiological score) or a severe phenotype (all others), a polymorphism in the FVII gene (Arg353Gln) was associated with a 3.5-fold to 6.6-fold increase in the risk of developing a ‘severe’ phenotype (P = 0.044) . This particular FVII gene polymorphism is associated with decreased levels of circulating FVII . It is certainly intriguing to assume that the extrinsic coagulation pathway plays a significant role in moderating the pathophysiology and subsequently the disease severity in hemophilia. This hypothesis, however, remains to be proven and will require longitudinal studies with sufficient numbers of patients.
Mutations or functional polymorphisms in other genes of the blood coagulation pathway that are associated with thrombophilia, including FV Leiden and prothrombin G20210A, have been shown to influence the clinical severity of HA. For example, in one study of patients with mild compared with severe HA, coinheritance of the prothrombin 20210A allele and of the intron 22 inversion in the FVIII gene was associated with a milder phenotype in patients with severe HA. Patients with this genotype had fewer bleeding episodes, reduced FVIII concentrate consumption and had a tendency toward a lower incidence of hemophilic arthropathies compared with patients without the prothrombin 20210A allele but with intron 22 inversion . The coexistence of resistance to activated protein C conferred by FV Leiden may result in a mild bleeding phenotype in some patients with severe HA [30, 31]. For example, in a study of 137 patients with severe HA, patients (n = 6 [4.4%]) who were heterozygous carriers for FV Leiden consumed less factor concentrate than non-carriers (geometric mean, 310 units kg−1 year−1 vs. 1185 units kg−1 year−1). In addition, a larger proportion of the patients who had FV Leiden experienced ≤ 10 bleeding episodes in their worst year compared with non-carriers (50% vs. 11%; P = 0.03). However, these study findings were from a small subset of the overall study population; most patients enrolled in the study who had severe HA and a milder bleeding phenotype did not carry the FV Leiden mutation . Overall, clinical studies in the hemophilia population (predominantly HA) regarding the benefit derived by carrying FV Leiden have collectively provided mixed findings, with evidence suggestive of protection in some studies but not in others [33-35].
Prophylaxis in HA and HB
Prophylactic therapy, an approach to converting the bleeding phenotype from severe to moderate or mild, appears to be utilized more frequently in patients with HA compared with those with HB [19, 36]. Analysis of recent UDC data from hemophilia treatment centers in the United States found that 56% of patients with severe HA received prophylaxis, compared with 46% of patients with severe HB (M. Escobar, personal communication). Similarly, in a survey of Canadian adult and pediatric patients with hemophilia, patients with HA used clotting factor prophylaxis at a higher rate than did patients with HB. Among patients without inhibitors, 23% of 2087 patients with HA received clotting factor prophylaxis compared with 11% of 498 patients with HB . Similar trends were found for HA compared with HB among patients with severe hemophilia (69% vs. 32%) and moderate hemophilia (18% vs. 5%), while the proportion of patients with mild hemophilia on prophylaxis was the same for HA and HB (1% vs. 1%). In addition, the frequency of prophylaxis utilization was greater in patients with severe HA than in those with severe HB, regardless of age group (0 to > 70 years) .
It is unclear whether the lower frequency of prophylaxis in patients with HB compared with HA reflects differences in the bleeding phenotype, a lack of prospective data, insufficient understanding of the pharmacokinetics of plasma-derived compared with recombinant FIX, or is simply a feature of treatment conventions .
Recent clinical observations point to certain differences between patients with HA and those with HB in terms of bleeding severity, inhibitor development and response to treatment. It is possible that defects in genes encoding other coagulation proteins may modulate the severity of the clinical manifestations in hemophilic patients; however, future studies are warranted. Bleeding episodes in patients with HA appear to occur more frequently and result in more severe joint damage. Although there are no comparative reports of the relative success of prophylaxis with replacement clotting factors in HA vs. HB, prophylaxis is generally utilized more frequently in patients with HA than in those with HB.
Medical writing support was provided by M. Paul, of MedErgy and was funded by Novo Nordisk. This publication was supported by Novo Nordisk.
Disclosure of Conflict of Interest
M. Escobar has received research funding from Pfizer and fees for advisory boards and consultation from Baxter, Bayer, Pfizer, Biogen and Novo Nordisk, Inc. S. Sallah is an employee of Novo Nordisk.