Evaluation of P22 ELISA for the Detection of Mycobacterium bovis-Specific Antibody in the Oral Fluid of Goats

The ante-mortem diagnosis of tuberculosis (TB) in ruminants is based mainly on the intradermal tuberculin test and the IFN-γ assay. Antibody (Ab)-based tests have emerged as potential tools for the detection of TB infected animals using serum, plasma, or even milk samples. Oral fluids have also been evaluated as alternative samples with which to detect specific Abs against Mycobacterium bovis in pigs or wild boars, but not in ruminants. The objective of this study was, therefore, to evaluate the performance of an in house-ELISA for TB diagnosis (P22 ELISA) in goats as an experimental model for the diagnosis of TB using oral fluid samples. Oral fluid samples from 64 goats from a TB-infected herd (n = 197) and all the animals from a TB-free herd (n = 113) were analyzed using the P22 ELISA. The estimated sensitivity (Se) and specificity (Sp) were 34.4% (95% CI: 22.4–45.6) and 100% (95% CI: 97.4–100), respectively. The optimal cut-off point was set at 100% according to the ROC analysis. Those animals with a higher level of Abs in their oral fluid attained a higher lesion score (p = 0.018). In fact, when taking into account only the setting of the animals with severe lesions (n = 16), the ELISA showed a Se of 75% (95% CI: 53.7–96.2). Results of the present study suggest that the P22 ELISA is highly specific but has a limited value detecting infected animals in oral fluid samples. Nevertheless, its performance is significantly higher in the presence of severe lesions