The ultimate goal in the treatment of haemophilia patients with inhibitors is to eradicate permanently the inhibitor and induce tolerance. Here we summarize our experience at the Malmö centre regarding tolerance induction according to the Malmö Treatment Model. The protocol includes immunoadsorption if needed, neutralization of inhibitor and replacement with factor concentrates, cyclophosphamide intravenously for 2 days (12-15 mg kg-1 bw) and then orally (2-3 mg kg-1 bw) for an additional 8-10 days and intravenous gammaglobulin daily at dosages of 0.4 g kg-1 bw for 5 days. This protocol has been applied in 23 haemophilia patients with inhibitors, 16 haemophilia A patients and seven haemophilia B patients. Altogether 36 attempts have been made to induce tolerance. Ten of the 16 haemophilia A (62.5%) and 6/7 patients with haemophilia B (86%) became tolerant after the treatment. The chances of success or failure are roughly equal, if the series is considered in a historical perspective. The data showed that the chances of success in tolerance induction with the Malmö protocol were best in those patients with low inhibitor titres, with relatively low historical inhibitory peak and with a long interval since the previous replacement therapy. This was especially true where no inflammatory state was present at the start or during tolerance induction. The advantage with this method compared to the high-dose regimen is that in the successful cases tolerance can be achieved within 3-4 weeks.

During treatment of a haemophilia A patient with a high-responding inhibitor against factor VIII coagulant activity (VIII:C), we observed a difference in recovery of VIII:C depending upon which factor concentrate was infused. Inhibitor plasma samples or IgG fraction from seven patients were tested against a panel of seven different commercially available factor VIII concentrates of which five were plasma-derived and two recombinant. In two of the plasma samples, inhibitor titres manifested a wide range of values depending upon which concentrate was used in the test system. Thus, inhibitor neutralization was less and VIII:C recovery greater when factor VIII concentrates containing large amounts of von Willebrand factor were used than when highly purified concentrates containing no von Willebrand factor or only trace amounts were used. In both of these two patients the inhibitor was directed against the light chain of factor VIII, and it is possible that the epitope of the light chain with which the inhibitor reacts is partly blocked by the von Willebrand factor. We conclude that inhibitors may differ in their reactivity with factor VIII molecules contained in clotting factor concentrates, and that there is factor VIII epitope variation between different concentrates. These findings have implications for the selection of concentrates for the treatment of inhibitor patients and the haemostatic effect may be improved if a concentrate giving the lowest inhibitor titre is chosen. Thus, in vitro testing of inhibitor reactivity with a panel of concentrates is recommended when treatment of inhibitor patients with factor VIII concentrates is considered.

In order to describe the haemostatic role of a variation in inhibitor reactivity with different factor VIII (FVIII) concentrates, we have compared inhibitor titres against a panel of FVIII concentrates and correlated titre with the capacity to inhibit thrombin generation. Three plasma-derived concentrates were tested in vitro in mixing experiments with inhibitor plasmas from 11 patients with severe haemophilia A: Fanhdi, which contains von Willebrand factor (VWF) with a final ratio of approximately 1:1 (VWF IU per IU FVIII:C); Haemate-P with a ratio of 2.5:1 and Hemofil-M containing only trace amounts of VWF. In addition, the recombinant FVIII concentrate Kogenate Bayer containing no VWF was included. Inhibitor titres and the capacity to generate thrombin were measured. A statistically significant difference in measured titres was found with the highest titres recorded against Hemofil-M. The inhibitor titres needed to inhibit 50% maximum thrombin generation were the lowest for Kogenate Bayer and the highest and similar for Fanhdi and Haemate-P with intermediate titres needed for inhibition of Hemofil-M. In this study, the thrombin generation assay provides additional indications for the role of VWF in the treatment of patients with inhibitors. The VWF-containing concentrates Fanhdi and Haemate-P, added to FVIII-deficient plasma with the presence of inhibitor, generate more thrombin than do the purified concentrates Hemofil-M and Kogenate Bayer.

Treatment of severe bleeding and the performance of surgery in haemophilia patients with inhibitors creates severe problems. It is generally agreed that treatment is most effective if circulating levels of factor VIII/IX can be achieved long enough for control of haemostasis. Immunoadsorption with protein A for the removal of inhibitor has improved treatment for patients with initial inhibitor titres too high to neutralize by infusion alone. This is a summary of our experience in Malmö regarding immunoadsorption and haemostasis. A total of 19 applications with immunoadsorption in 10 patients were performed. On all occasions it was possible to eliminate totally the inhibitor or reduce it to low levels that could easily be neutralized with factor concentrate. Haemostatic levels of coagulation factors could be maintained for 5-9 days in all but one patient. This period was sufficient to stop ongoing haemorrhage or prevent excessive bleeding at surgical interventions.

Severity assessment of patients with haemophilia A (HA) is traditionally based on FVIII activity (FVIII:C). Clinical phenotype of HA patients often differs between individuals with the same FVIII:C determined with clotting and chromogenic assays. The aim of this study was to assess the influence of the FVIII:C on thrombin generation (TG) assay parameters both in vitro and ex-vivo postinfusion plasma. For in-vitro approach, influence of FVIII:C was evaluated on TG parameters in several dilutions of a normal plasma pool with commercial FVIII-depleted-plasma (FVIIIDP) and in others experiments, adding increasing amounts of different commercial FVIII concentrates (Fanhdi, Haemate-P, Hemofil-M and Kogenate Bayer) to FVIIIDP. In a series of 50 postinfusion samples, from HA patients of different severity, we assayed TG and FVIII:C (chromogenic and clotting). In vitro experiments, the 50% of maximum TG peak (TGMP) was achieved using only 5% FVIII:C and the TGMP was obtained with 40% of normal VIII:C. Impaired response compared with normal plasma was found in FVIIIDP using addition of increasing amounts of different commercial FVIII concentrates. An overall good correlation between the two FVIII assays was observed (y = 0.9115x - 0.273, r = 0.975, P < 0.001); TGMP and the Lag-Phase-Time (LPT) provided some discrepant results when compared with the total range of FVIII:C measurements. In contrast, correlations for TGMP, LPT and endogenous thrombin potential were improved in samples restricted to FVIII:C <5%. We conclude that TG parameters tentatively could be a tool to tailor the global haemostatic capacity in haemophilic patients.