Is VIII worse than IX?

In this issue of Blood, Tagariello and colleagues report that patients with severe hemophilia A are more likely to undergo hip or knee arthroplasty than those with severe hemophilia B and suggest that there is a difference in the severity of the 2 hemophilias.

Traditional teaching has been that the clinical manifestations of hemophilia A and B are identical and cannot be differentiated without measurement of the specific FVIII and IX clotting factor activities. However, the possibility that hemophilia B is a milder disease was proposed 50 years ago by Quick, who based his opinion on his personal experience with 24 hemophilia B patients diagnosed in his department at a time prior to the availability of clotting factor concentrates.¹  Quick's suggestion has been supported by some more recent data. In a preliminary Canadian report, adult severe hemophilia B patients were found to have had 35% fewer bleeds than severe hemophilia A patients,²  while a second Canadian study, on virtually the whole of the Canadian hemophilia population, reported that significantly more severe hemophilia A patients used prophylaxis (69%) compared with severe hemophilia B patients (32%). The effect was most pronounced in those under 2 years of age but was noted at every age group, including those over the age of 50.³  A higher bleed frequency for hemophilia A and subsequently hemophilia severity score was also observed by Schulman and colleagues in Sweden, but the number of severe patients in the derivation cohort was relatively small at 43 individuals.⁴ 

The study by Tagariello et al in this issue describes a 3-fold higher frequency of joint arthroplasty in patients with hemophilia A compared with hemophilia B of the same severity, defined as a coagulation factor level of < 1 U/dL. This difference was not due to confounders such as age, HIV or hepatitis C infections, or the presence of an inhibitor.⁵  The data reported were retrospective but covered the whole of the Italian hemophilic population, a major strength of the study. The authors also performed a systematic review of published hemophilia arthroplasty series and found that in the 7 other series, 147 (91.3%) of the 161 patients with hemophilia undergoing joint arthroplasty had hemophilia A. Although these other series did not report the total number of patients registered in their centers, this proportion is significantly higher than expected from the proportion of hemophilia A patients in the large comprehensive registries (84% in Canada, the United Kingdom, and Italy).⁵ 

A potential criticism of this study is that joint arthroplasty is an end-stage event and a surrogate of severity. The frequency of bleeds would have been a better variable to compare in the 2 hemophilias, but this information was not available. Heterogeneity in the treatment of bleeds or referral for arthroplasty could have influenced the results, but it is not obvious why these would have varied by hemophilia type. Furthermore, it is not known if the lack of FVIII (compared with IX) results in more bleeds, if the bleeds respond less well to treatment, or if the bleeds are more severe or more destructive.

An important issue that could potentially explain the results is that severe hemophilia A and B are currently defined on the basis of a clotting factor level of < 1 U/dL⁶  and not on the phenotype. The accuracy of the commonly used clotting assays at these low levels has been questioned.⁷  It is also well recognized that patients with “severe” disease with clotting factors of < 1 U/dL show heterogeneity in the phenotype in terms of bleed frequency as well as in their thrombin-generating capacity.⁸  It is likely that the group of < 1 U/dL patients contains individuals with truly no VIII/IX activity, as well as some with low but detectable activity. If the proportion of these was different for the 2 hemophilias, it could have influenced the results. Tagariello et al tried to address this issue by comparing the patient groups with genetic defects likely to result in absent clotting factor and those with less disrupting defects but found no difference in the arthroplasty rate.

At present, it is reasonable to conclude that when defining the 2 severe hemophilias in the same way as < 1 U/dL level, there appears to be a difference in the bleeding phenotype. The Tagariello study does not give us the definitive answer but does suggest that the observation is true. This will undoubtedly be the start of a new avenue to firstly reproduce and confirm these findings in other cohorts and to try to explain the pathophysiology of this observation.