Background: Mycoplasma bovis-associated disease can cause tremendous production losses, welfare issues and high antimicrobial use. Therefore, screening cattle for M. bovis antibodies before entering the herd is a popular and possibly cost-efficient way to reduce disease introduction. However, interpretation of results can be challenging due to variable accuracy between tests and populations. This study's objectives were to compare the diagnostic test accuracy of three commercially available M. bovis antibody ELISAs (ID-screen, Bio K302 and Bio K432) and to explore optimal cutoff values for screening purposes.
Methods: A prospective diagnostic test accuracy study was performed on 170 serum samples from youngstock using Bayesian latent class modelling. Samples were categorised using manufacturer and generated cutoff values.
Results: Using the manufacturers' guidelines, ID-screen, Bio K432 and Bio K302 showed 97.6%, 67.4% and 33.6% sensitivity, and 78.8%, 97.6% and 99.1% specificity, respectively. Optimised cutoffs resulted in 94.8%, 82.6% and 78.3% sensitivity, and 94.2%, 92.5% and 79.4% specificity, respectively.
Conclusions: The highest diagnostic accuracy for detecting M. bovis antibodies was obtained by ID-screen (≥110%). However, by adjusting cutoff values, the sensitivity of Bio-X tests could be markedly increased, making these tests also applicable as screening tools.
Limitations: Interpretation needs to be careful as antibodies may be linked to both infectious and non-infectious status.
Keywords: Bayesian latent class analysis; Bio K302; Bio K432; ID-screen; purchase; sensitivity.
© 2022 British Veterinary Association.