The Prevalence of Non-Neutralizing Anti-FVIII Antibodies in the Canadian Hemophilia Population.

In approximately 25% of hemophilia A patients, inhibitory anti-FVIII antibodies develop following treatment with FVIII protein infusion therapy. This is currently the most serious treatment-related complication in this population. While the presence of these antibodies has been evaluated by a functional clotting assay for the past 30 years (the Bethesda Assay), there is a growing appreciation that the anti-FVIII immune response may also include the appearance of non-neutralizing antibodies in some patients. The prevalence, natural history and clinical relevance of these antibodies are the subject of this study.

Plasma samples from 602 hemophilia A patients formed the study material for this project. These samples had been stored at –80C since the conversion from plasma-derived to recombinant FVIII (rFVIII) concentrates in Canada. The results of prior Bethesda assay testing were available for 392 of these samples. The ELISA-based test for non-neutralizing anti-FVIII antibodies utilized three different rFVIII products as antigens: two full-length rFVIII products and one B domain-deleted rFVIII. Each microtiter plate run included six negative and one positive control sample. All samples were tested in duplicate and samples with positive results (absorbance >mean + 3SD) were subjected to serial dilution testing. 93 of the samples were also tested by the same assay in a second laboratory.

Anti-FVIII antibodies were documented using this ELISA-based assay in 11.5% of this population. 11.5% and 8.8% of the patients demonstrated non-neutralizing antibodies to the two full-length rFVIII concentrates Advate® and Kogenate®, respectively. In contrast, antibodies to the B domain-deleted concentrate, Xyntha®, were only found in 4.6% of patients. 2.8% of patients had antibodies to all three concentrates while 3.1% of patients had antibodies to only Advate (2.8%) or Kogenate (0.3%). Twenty four of the 392 patients (6.1%) tested with the Bethesda assay had inhibitory anti-FVIII antibodies (range 0.6 – 267 BUs). A third of these Bethesda +ve cases demonstrated positivity in this assay to all three rFVIII products, while another 30% had antibodies against one or two of the rFVIII concentrates. 48 of the 368 cases that tested negative in the Bethesda assay were positive for non-neutralizing anti-FVIII antibodies (13%). To date, 20 of the samples testing positive for non-neutralizing antibodies have been studied by serial dilution analysis and 9 of these samples were only positive at the initial dilution of 1:50. Seven cases had positive dilution titers of 1:200 (3 with inhibitory antibodies ranging from 0.6 to 98 BUs) and two samples were positive at a dilution of 1:400, both of which had positive Bethesda results of 30 and 115 BUs, respectively. To assess the inter-laboratory consistency of this testing protocol, 93 selected samples were assayed with the identical methodology in a second laboratory. These studies showed positive results for the presence of antibodies to the two full-length rFVIIIs in 32-45% of cases. In contrast, the range of positive results for the two laboratories with the B domain-deleted rFVIII was between 19 and 37%.

This study of a large Canadian hemophilia A population has demonstrated the presence of non-neutralizing anti-FVIII antibodies in 11% of the population. Antibodies to the B domain-deleted rFVIII were significantly less frequent. When evaluated, inhibitory antibodies were present in 6% of this population and in the Bethesda negative cases, non-neutralizing antibodies were found in 13% of patients. Finally, many of these non-neutralizing antibodies appear to be of low titer, and their clinical significance must await further study with clotting factor recovery and half-life analysis.

No relevant conflicts of interest to declare.