Comparison of Western immunobloting to an enzyme-linked immunosorbent assay for the determination of anti-Bordetella pertussis antibodies.

During Bordetella pertussis infection, it has been established that an increase of anti-pertussis toxin (PT) and anti-filamentous hemagglutinin (FHA) antibodies occurs. Immunoblots from two manufacturers using FHA and PT antigens were compared with an enzyme-linked immunosorbent assay (ELISA) that used both FHA and PT. One manufacturer used two concentrations of PT bands for the IgG immunoblot, calibrated to the World Health Organization standard for PT in international units (IU/ml), 100 IU/ml (PT-100) and 8 IU/ml (PT). The second immunoblot kit measured antibodies to a single calibrated PT band. Both kits measured IgA antibodies, and one additionally measured IgM antibodies. Two of 41 (5%) ELISA IgM positives were confirmed positive by IgM immunoblotting, suggesting poor specificity of the IgM ELISA. The agreements of the IgG and IgA immunoblots with the ELISA ranged from 72.5% to 85.3%, with only 38 to 51% of IgA positives confirmed by immunoblotting and only 61 to 68% of IgG positives confirmed by immunoblotting. The two immunoblots correlated well with each other, with 91.7% and 94.3% agreement for IgG and IgA, respectively. When the FHA band was used with the PT band as the criterion for positivity, significant differences existed in specificity compared to the ELISA (IgG, 84.1% versus 33.3%; IgA, 82.4% versus 71.0%). When the positive IgA immunoblots (evidence of natural recent infection) were compared to the positive PT-100 IgG immunoblots (evidence of recent infection or vaccination), the PT-100 blot showed a 71% sensitivity in detecting natural recent infection. B. pertussis immunoblots, alone or in combination with ELISAs, can aid in the diagnosis of B. pertussis infection.