Diagnostic performance of the Pourquier ELISA for detection of antibodies against Mycobacterium avium subspecies paratuberculosis in individual milk and bulk milk samples of dairy herds

The objective of the study was to determine the diagnostic performance of the Pourquier ELISA for detection of antibodies against Mycobacterium avium subsp. paratuberculosis (Map) in individual milk samples and in bulk milk samples. For individual milk samples the specificity of the Pourquier ELISA was estimated by testing a panel of individual milk samples from certified Map-free herds. The relative sensitivity of the assay in individual milk samples and agreement of the results with those of serum samples was estimated by testing panels of paired serum-milk samples from seropositive cattle, whole-herd investigations, and moderate or heavy shedders. The specificity of the ELISA for individual milk samples was still 99.8% at a cut-off of 20% sample to positive (S/P) value, clearly lower than the cut-off defined by the manufacturer (30% S/P). The relative sensitivity for individual milk samples as compared with positive serum samples was 87% for a cut-off of 20% S/P, and 80% for a cut-off of 30% S/P. The sensitivity of this ELISA for detection of high shedders was >90% both for individual milk and serum samples, also agreement was very good (κ = 0.91 for all paired samples). The specificity of the Pourquier ELISA in bulk milk samples was investigated by testing bulk milk samples from certified Map-free herds. Feasibility of bulk milk testing was investigated by titrating ELISA positive individual milk samples in negative milk. In addition, 383 bulk milk samples from herds with a known within-herd seroprevalence were tested. The specificity of the ELISA for bulk milk samples was 100% at a cut-off of 12.5% S/P. At the cut-off recommended by the manufacturer (30% S/P) performance of the bulk milk ELISA related to herd status (>=2 seropositive cows) was rather poor, corresponding with a sensitivity of 24% and a specificity of 99% relative to serology. However, at the revised cut-off for bulk milk of 12.5% S/P and a within-herd seroprevalence of >=3%, sensitivity and specificity relative to serology were 85% and 96%, respectively. Given the current herd-level seroprevalence in The Netherlands, these test characteristics corresponded with positive and negative predictive values for bulk milk of 67% and 94%, respectively. In conclusion, the diagnostic performance of the Pourquier ELISA for individual milk samples creates opportunities for a cheaper and more feasible testing scheme, while the diagnostic performance for bulk milk samples warrants further consideration.