Management of haemophiliacs with alloimmune antibodies has been a difficult problem since plasma-derived factor (F) VIII and FIX became readily available, and they continue to be a cause of significant morbidity and mortality, as well as being a considerable financial burden on healthcare systems. The priority in these patients is to treat bleeding problems with high doses of FVIII or FIX or with bypassing agents such as FEIBA (factor eight inhibitor bypassing fraction human; Baxter AG, Vienna, Austria), prothrombin complex concentrate or recombinant-activated FVII (rFVIIa) and secondly to attempt immune tolerance. Although a very costly procedure (Knight et al, 2003) this will, if successful, have the best outcome for the patient's quality of life and, in terms of lifetime treatment, there may be overall cost benefits in immune tolerance. Haemophiliacs in whom immune tolerance fails have a relatively bleak outlook and are likely to suffer from chronic arthropathy as well as being at increased risk of life-threatening bleeds. The anti-CD20 monoclonal antibody, rituximab, eliminates circulating B cells and, as it has been used successfully to manage autoimmune diseases (Berentsen et al, 2001; Giagounidis et al, 2002; Hongeng et al, 2002; Leandro et al, 2002; Ahn et al, 2003; Krause et al, 2003), it may offer a therapeutic option for treatment of inhibitors.
We report the use of rituximab (Genentech, San Francisco, CA, USA) in two children with severe haemophilia with inhibitors to factors VIII and IX, which failed to respond to conventional immune tolerance therapy. The treatment was well tolerated by both children. The child with haemophilia B had no clinical improvement or fall in CD19 and he is currently being treated with recombinant activated factor VII (NovoSeven, Novo Nordisk, Denmark) for bleeding episodes. The child with haemophilia A had a good clinical response with a negative inhibitor assay at 11 months follow-up.
A 2-year-old boy with severe haemophilia A (FVIII <1 IU/dl) developed an inhibitor but continued to respond well to on-demand FVIII injections. By 1996, at the age of 5 years, he had developed a target joint in his right knee and had started to have bleeds two to three times per week on regular home treatment with a high purity plasma-derived FVIII. He was therefore switched to on-demand FEIBA. In 1998, after referral to our unit, a decision was made to attempt immune tolerance. A port-a-cath was inserted uneventfully with rFVIIa cover and tolerance was started with 100 IU/kg of recombinant FVIII (Kogenate; Bayer AG, Leverkusen, Germany) twice daily. His inhibitor level fluctuated somewhat over the course of the next 2 years but by the end of 1999 his inhibitor screen was negative, although he continued to have abnormal recovery with 48 h levels of <1 IU/dl. Six months later his inhibitor had reappeared at a low level (0·6 BU, see Fig 1) and a further attempt was made to eliminate it with increased doses of FVIII. He continued to have some clinical response to alternate daily treatment with 100 IU/kg/dose of Kogenate but although he ceased to exhibit an anamnestic response to FVIII, his inhibitor had not been eliminated. In 2002, the use of rituximab was discussed with the family, as case reports were available describing its successful use in treating acquired FVIII inhibitors (Wiestner et al, 2002).
In February 2003, aged 11 years, he was given four, weekly, doses of rituximab (375 mg/m2), which were well tolerated and he continued on 100 IU/kg of Kogenate on alternate days. He and his family noticed an immediate clinical improvement in terms of frequency of bleeds and, incidentally, in his atopic symptoms. His inhibitor titre, which was 12 BU prior to treatment, fell to 0·7 BU at 8 months and it is currently undectable at 11 months, with a FVIII recovery of 198 IU/dl 30 min post-dose (75 IU/kg). His CD19 count (which, together with immunoglobulins, is a marker for B-cell response to rituximab in oncology patients) fell as expected to an undetectable nadir, with an isolated dip in his immunoglobulin (Ig) M to 0·43 g/l (normal range: 0·5–1·9g/l). He has had no problems with opportunistic infection.
A 2·5-year-old boy with severe haemophilia B (FIX <1 IU/dl) developed an inhibitor to plasma-derived FIX. After seeing several haemophilia specialists his parents decided that, as he was not experiencing any spontaneous joint bleeds and still had a response to escalated doses of FIX therapy without any signs of anaphylaxis, the potential risks of immune tolerance (nephrotic syndrome and uncontrolled bleeding with port-a-cath insertion) outweighed the potential benefits.
Until the age of 9 years, he was managed at his local haemophilia centre with on-demand high-dose plasma-derived FIX but he was referred to us when he developed a target joint. His inhibitor titre at this stage was 70 BU. All the involved parties at this time felt that immune tolerance should be attempted despite the longevity of the inhibitor (6 years).
He had a port-a-cath inserted under rFVIIa cover and immune tolerance was started with a regimen of 100 IU/kg of recombinant FIX (Benefix; Genetics Institute Inc., Andover, MA, USA) daily. Initially, the results were encouraging in that his inhibitor titre fell (see Fig 2) but then it started to rise steadily and immune tolerance was stopped after 9 months. He continued with on-demand rFVIIa. Urinalysis remained negative for protein throughout the attempted immune tolerance (Ancliff et al, 2000).
Over the following 2 years, he required hospital admissions for intensive hydro- and physiotherapy, intra-articular corticosteroid injections and management of bleeds. At this point the possibility of treatment with rituximab was explored. In April 2003, he was given four, weekly, rituximab infusions (375 mg/m2) which were completed without significant problems.
After 3 months his inhibitor titre had fallen to <0·5 BU and Benefix, 3000 IU twice weekly, was restarted. However, 2 weeks later his inhibitor had risen again to 28·9 BU and FIX therapy was abandoned. His CD19 count at this time was 0·22 × 109 g/l compared with 0·24 × 109 g/l prerituximab. IgG and IgA levels remained normal throughout and the IgM only fell a little (0·77 g/l pre and 0·34 g/l nadir; normal range 0·5–1·9 g/l).
Current options for the management of relapsed or refractory inhibitors in children with haemophilia are limited. There is reluctance amongst many haemophilia carers to use immunosuppressive regimens (such as the Malmo regimen: immunoadsorption, cyclophosphamide and intravenous immunoglobulin) (Freiburghaus et al, 1999), which are the basis of treatment for acquired FVIII antibodies in adults, because of their potential side-effects, most notably secondary malignancy and infertility, although the total dose of cyclophosphamide is low, and because there is no evidence to suggest they are superior in this setting (Wight et al, 2003). The use of mycophenolate mofotil has been reported (Martinez-Saguer et al, 2003), but the regimen continued for over a year and involved monthly pulses of high-dose dexamethasone with all the attendant concerns about long-term corticosteroid exposure in children. We report the first use of rituximab to treat alloimmune FVIII/IX inhibitors in children with severe haemophilia, one of whom has shown an apparent response at 11 months follow-up, with a reported reduction in bleeds, improved response to treatment, negative inhibitor screen and good FVIII recovery 30 min post-dose. The second child demonstrated no fall in CD19 count after rituximab treatment and had an immediate recrudescence of his inhibitor when re-exposed to FIX. We are unable to explain the unchanged CD19 count in this child that, intuitively, seems likely to relate to his failure to respond. This failure is not a reflection of greater longevity of his inhibitor, as they had been present in the first and second child for 10 and 11 years, respectively, nor can lack of response be blamed on a higher inhibitor titre prerituximab, as the titre was higher in case 1. One difference between the management of these cases was that the first boy remained on FVIII throughout the period of rituximab treatment whereas the second boy was not re-exposed to FIX until 3 months after the rituximab because of concerns about potential anaphylaxis; it may be that tolerance induction requires exposure to the clotting factor at the same time as rituximab.
Of two boys treated with rituximab for refractory FVIII and FIX inhibitors, one appears to have had a response with 11 months follow-up and the other has failed to respond. We suggest that rituximab may have a place in the second-line management of alloimmune inhibitors and that, if other groups can confirm this, it may be appropriate to assess its earlier use in the treatment process, in conjunction with conventional immune tolerance.
Dr A. Worsley and Dr C. Hay were closely involved in the management of both of these children. We would like to thank Dr K. Peggs for his comments and suggestions.