Reproducibility of a Commercial Enzyme-Linked Immunosorbent Assay for Bovine Paratuberculosis among Eight Laboratories

Abstract

Interlaboratory reproducibility of an absorbed enzyme-linked immunosorbent assay (ELISA) kit for detection of bovine serum antibodies to Mycobacterium paratuberculosis was evaluated. A panel of 30 bovine sera (15 positives and 15 negatives) was tested in triplicate microtiter wells on each of 2 days at 8 different laboratories. One laboratory had invalid results because of positive or negative serum control optical density (OD) readings beyond the acceptable range specified by the kit. The coefficient of variation (CV) for mean OD values was influenced by low ODs on test negative sera at 2 laboratories, thus the CVs on positive sera were considered a more representative measure of kit reproducibility. Between-well CVs averaged 6.7% ± 2.8% (mean ± standard deviation), and between-day CVs averaged 14.5% ± 9.8% among the 7 laboratories with valid assays on the 15 positive sera. The OD values were converted to positive or negative classifications for each assay well, and the results were compared. Among 1,392 assays in 7 laboratories, 98.6% were in agreement. Eleven of 18 discrepant results were due to a sample that consistently gave OD values near the cutoff for a positive test. Exclusion of that serum from the analysis resulted in a 99.8% rate of agreement among laboratories. Results indicated that the absorbed ELISA kit provided reproducible results within and between laboratories.