Evaluation of the MilA ELISA for the diagnosis of herd infection with Mycoplasma bovis using bulk tank milk and estimation of the prevalence of M. bovis in Australia

Abstract

Infection with Mycoplasma bovis has been identified as a growing threat in dairy industries worldwide and there is an urgent need for an inexpensive and accurate herd-level screening tool to identify herds that have been exposed to M. bovis. This study aimed to evaluate the use of the MilA ELISA for testing bulk tank milk (BTM) samples for antibodies against M. bovis and estimate a suitable cut-off and diagnostic sensitivity (DSe) and specificity (DSp) for this assay. An optimal cut-off was then applied for investigating the geographical and seasonal distribution of infection with M. bovis in Australia. A total of 5554 BTM samples from 2683 dairy herds were collected during March, August and December 2017. BTM samples were tested in the MilA ELISA and a cut-off of 29 antibody units (AU) was estimated to be optimal using Bayesian latent class analysis which makes no assumption about the true disease status of herds under investigation. At this cut-off, the DSe and DSp were estimated to be 96.6% (95% highest probability density [HPD] interval: 87.0, 99.8) and 94.2% (95% HPD: 89.9, 97.4), respectively. The diagnostic specifications were found to vary markedly with stage of the production cycle, suggesting that targeted sampling was needed to maximize accuracy. We also found distinct differences in the apparent prevalence of M. bovis in different dairying regions, as well as seasonal variation. The highest apparent prevalence of M. bovis was observed in samples collected in March and an overall drop in the proportion of positive herds was seen from March to December. Overall, this study provides insights into the dynamics of BTM antibodies against M. bovis in Australian dairy herds and how the MilA ELISA can be applied for bulk tank milk testing.