One is the indiscriminate use of assays for BAb that are known to give false positive results or report irrelevant Ab with low affinity as if they were high-affinity BAb; see Assay for binding antibodies to cytokines. In such cases, the clinician assumes that a patient without specific Ab response to IFN-b, and therefore without NAb, has developed 'non-neutralising' BAb to the interferon.

Another important reason is the use of antiviral neutralisation assays whose clinical relevance is uncertain; see Clinical relevance of NAb assays?. In these cases, NAb developed by patients will often go unnoticed, and BAb in such patients, whether or not they are tested correctly, will again be characterized as 'non-neutralising'.

We believe that appropriately measured specific BAb induced during IFN-beta therapies are probably always neutralising. Most of these BAb are of IgG class (mw: 150,000), and there is - unpublished - evidence that there are at least two major epitopes on IFN-beta which may bind to these BAb. Thus, high-affinity binding of one or two IgG molecules to a small ligand like IFN-beta (mw: 20,000) is likely to interfere with assembly of the two components of the IFN-a/b receptor, even in cases where the true BAb bind to epitopes not directly associated with receptor binding; see figure.

This is supported by the results of testing BAb and NAb in MS patients treated for up to 18 months with subcutaneous injections of IFN-beta. As shown in the figure below, the frequency of BAb-positives, tested by RIA, rose from 0 to about 80% after 3 months, and this frequency increased to almost 100% during the 18 months of testing. When testing the same patients for NAb, the numbers of positives closely matched those of the BAb-positives, but only when NAb were measured by a high-sensitivity assay. If a low sensitivity NAb assay had been used alone, the false conclusion would be that less than 40% of the patients developed NAb while > 50% developed 'non-neutralising' BAb!