Cut-off evaluation of ID Screen Mycoplasma bovis ELISA for use on bulk tank milk in New Zealand

Repeated testing of bulk tank milk (BTM) samples and testing of serum samples has played a major role in New Zealand's Mycoplasma bovis (M. bovis) eradication programme. We evaluated the performance of the ID Screen Mycoplasma bovis indirect ELISA on BTM samples and identified the maximal diagnostic sensitivity (DSe) and diagnostic specificity (DSp) for testing for antibodies against M. bovis in New Zealand dairy herds. Also, we investigated factors influencing DSe over the milking season and associated with the presence of antibodies against M. bovis in a sample. Data from 63 dairy herds were analysed using Bayesian latent class analysis for two conditionally dependent tests. A BTM ELISA sample-to-positive ratio (SP%) cut-off of 24 was the estimated optimal threshold yielding the optimal combination of DSe and DSp - a DSe of 78.6 % (95 % highest posterior density [HPD] interval, 50.8-93.5) and a DSp of 98.5 % (95 % HPD interval, 97.2-99.4). The presence of antibodies against M. bovis in a BTM sample was associated with herd size and the somatic cell count concentration. DSe varied markedly throughout the milking season, and was dependent on the total BTM sample volume at the time of sampling. The DSe of the ID Screen ELISA was highest for samples tested during the early and mid-stages of the milking season, and when milk vat volumes were low relative to the number of cows in milk.