Response: Comparing joint arthroplasties in severe hemophilia A with severe hemophilia B

den Uijl and colleagues found that the overall number of joint arthroplasties performed in their patients with hemophilia B (an indirect index of the severity of congenital coagulation disorder) is not different from that in patients with hemophilia A, at variance with our recent report of a significantly less frequent need of these operations in hemophilia B.¹  Their conclusions are based on 78 hemophilia A and 9 hemophilia B patients who underwent arthroplasty in a single institution, whereas our conclusions stem from much larger absolute numbers of patients derived from all the hemophilia centers in Italy (253 with hemophilia A vs 15 with hemophilia B) and also from a systematic review of the literature that was consistent with our findings of fewer arthroplasties in hemophilia B. We suspect that the small sample size on which den Uijl et al based their tests of statistical significance entails a high risk of type 2 error (the error of not finding a difference that does exist), although a smaller risk of type 1 error (the error of finding a difference that does not exist) cannot be ruled out even in our much larger cohort. Furthermore, the Dutch cohort was treated early and regularly with continuous prophylaxis at a significantly higher rate than that for patients included in the Italian cohort. Obviously the efficacy of intensive prophylaxis had an impact on the natural history of the disease, making hemophilia A and B become similar in severity, and thereby explaining, at least in part, the differences between the Italian and Dutch cohorts.

With this as a preamble, we agree with the Dutch investigators that the issue of the different severity of the 2 hemophilias remains unsettled, because the rate of joint arthroplasty is only a very indirect index of disease severity and entails the potential influence of several confounders. We also agree with den Uijl et al that large multinational studies are warranted to reach a firm conclusion on the issue of varied severity, with possible clinical implications on the need for early continuous prophylaxis. In agreement with Makris,²  we believe that the different rate of gene mutation types (null vs non-null) make biologically plausible that hemophilia B is less severe than hemophilia A. Other studies^3-5  are consistent with our suggestions of lesser severity of hemophilia B.