Background: Binding antibodies (BAB) against interferon-beta (IFNbeta) are often determined as screening assays before performing an expensive and elaborate neutralizing antibody (NAB) test.
Methods: In this study, we compared four BAB tests, a western blot (WB), a direct binding enzyme-linked immunosorbent assay (ELISA) (dELISA), a capture ELISA (cELISA), and a commercial enzyme immuno-assay (EIA) in 325 multiple sclerosis patients with and without neutralizing antibodies to evaluate the sensitivity and specificity to detect NAB by receiver operating characteristics analysis.
Results: The area under the curve (AUC) values were 0.907 for the dELISA, 0.925 for the cELISA, and 0.776 for the EIA (P < 0.0001 for all). At a sensitivity of 95%, the specificity was approximately 30% in the dELISA, 55% in the cELISA, and 13% in the EIA. The WB as a qualitative BAB detection method had a given sensitivity of 97% and a specificity of 55%. There was a strong and significant correlation between high NAB titers (>500 neutralizing units [NU]) and titers obtained by all quantitative BAB assays. However, low to medium NAB titers (20-500 NU) did not significantly correlate with BAB titers.
Conclusion: We conclude that the cELISA seems to be most suitable for NAB screening, but BAB titers cannot reliably predict NAB titers.