Sensitivity and specificity of a competitive ELISA using frozen-thawed milk or serum for the diagnosis of small ruminant lentivirus infection in goats using a Bayesian latent class model

Abstract

Small ruminant lentivirus (SRLV) infection in goats causes progressive inflammatory infiltration of multiple organs including the mammary gland, and has been hypothesized to impact udder health and milk quality on dairy goat operations. The objective of this study was to evaluate frozen-thawed milk as an alternative substrate to serum for detection of antibodies to SRLV by a commercial competitive ELISA. Paired milk and serum samples collected on the same day from 116 goats were compared using a commercial SRLV antibody test kit, cELISA (VMRD, Pullman, WA, USA). Agreement between sample types and sensitivity, specificity, positive and negative predictive values of the cELISA using frozen-thawed milk and serum were calculated using a Bayesian latent class model. Agreement was excellent (κ = 0.97), with only 2/116 pairs yielding discordant results. The sensitivity and specificity of cELISA on milk were 97.0% and 95.9%, respectively. Individual frozen-thawed milk samples were a suitable alternative to frozen-thawed serum samples for the detection of SRLV antibodies in goats using a commercial competitive ELISA.