The development of neutralizing antibodies against factor VIII is a common and serious complication of replacement therapy, occurring mainly in the early stages of treatment. Meta-analyses of observational studies have suggested a higher risk of inhibitor development with concentrates produced by recombinant technologies (rFVIII) than with those derived from human plasma (pdFVIII) containing von Willebrand factor, which was recently confirmed in a randomized trial. In this trial cumulative incidences of inhibitor development were 44.5% for rFVIII and 26.8% for pdFVIII, for a hazard ratio (HR) of 1.87 (95% confidence interval (CI95) 1.17-2.96). Given the particularly high risk with rFVIII , it has been suggested to restrict the use of rFVIII to low risk patients, and treat high-risk patients with pdFVIII. We investigated such a strategy in a post-hoc analysis of the SIPPET study, in which we used the FVIII genotype (F8 gene mutation) to classify patients by prior risk.


SIPPET is an open label international randomized trial on which 251 previously untreated (n=142) or minimally treated (less than five exposure to blood components other than concentrate or cryoprecipitate, n=109) in 42 centers to be treated exclusively with a concentrate from the class of rFVIII or pdFVIII. Patients were tested for inhibitors before entry and at regular intervals during 50 exposure days, 3 years or the development of an inhibitor of at least 0.4 Bethesda units (BU). The trial ran from 2010 to 2014 and was terminated when the prespecified number of patients was included. Patients who had not reached 50ED by that time were censored. Patients were classified at high risk when they carried a null mutation (inversion, large deletion, frameshift, nonsense mutation) in the F8 gene and as low risk when they carried another causative variant (missense, splice site, polymorphisms, no mutation). We estimated cumulative incidences, hazard ratios and numbers needed to harm (NNH) for rFVIII vs pdFVIII for high- and low risk patients.


Among 251 patients, 76 developed an inhibitor (all > 0.7 BU) of which 50 were high- titer (> 5 BU). Among 197 patients classified as high risk, 65 developed an inhibitor (cumulative incidence 38.2%, CI95 30.8-45.6), whereas among the 38 patients classified as low risk 7 developed an inhibitor (cumulative incidence 23.9%, CI95 8.2-39.6). High and low risk patients were equally distributed over the two arms of the trial, i.e., 96 out of 126 treated with rFVIII were high risk, and 101 out of 125 treated with pdFVIII. Among high risk patients, cumulative incidence was 30.7% when treated with pdFVIII , and 46.5% when treated with rFVIII (risk difference 15.8%). Among low risk patients, no inhibitors developed with pdFVIII, whereas the cumulative incidence was 43.2% with rFVIII (risk difference 43.2%). This implies that the Number Needed to Harm was 5.6 overall, 6.3 for high-risk patients, and 2.3 in low risk patients.


Risk stratification by the type of F8 mutation does not identify previously untreated patients with hemophilia A who have a low inhibitor risk when exposed to rFVIII. Other means need to be found to reduce the occurrence of inhibitors with rFVIII.


Palla:Pfizer: Other: travel support . Mannucci:NovoNordisk: Speakers Bureau; Kedrion: Speakers Bureau; Grifols: Speakers Bureau; Bayer: Speakers Bureau. Peyvandi:Bayer: Speakers Bureau; SOBI: Speakers Bureau; Kedrion Biopharma: Consultancy, Other: research funding paid to Luigi Villa Foundation, Research Funding; Alexion: Other: research funding paid to Luigi Villa Foundation, Research Funding; Octapharma: Consultancy; Ablynx: Membership on an entity's Board of Directors or advisory committees, Other: research funding paid to Luigi Villa Foundation, Research Funding; Biotest: Other: research funding paid to Luigi Villa Foundation, Research Funding, Speakers Bureau; CSL Behring: Speakers Bureau; LFB: Consultancy; Grifols: Speakers Bureau; Novo Nordisk: Other: research funding paid to Luigi Villa Foundation, Research Funding, Speakers Bureau.

Author notes


Asterisk with author names denotes non-ASH members.

This icon denotes a clinically relevant abstract