Abstract
Three serological methods, the Rose-Bengal test (RBT), the complement-fixation test (CFT) and an indirect enzyme-linked immunosorbent assay (I-ELISA) were compared for the detection of
Brucella-infected animals in unvaccinated cattle herds in Eritrea. In this study, 71 herds first were classified as positive or negative for
Brucella infection on the basis of at least one animal being seropositive by RBT and CFT. All the 159 RBT-positive samples from the 26 seropositive herds and 214 RBT-negative samples randomly selected from the seropositive herds and from the 45 negative herds were tested further by CFT and I-ELISA. Using the ELISA titer as main predictor, and incorporating the RBT results, a logistic model was built to predict the CFT-negative or -positive status of individual sera and to estimate sensitivity and specificity. Whilst the ELISA titers (≤20) accurately predicted all the negative sera in herds that were also negative by the CFT, the number of seropositive animals was higher by ELISA in herds that had positive animals. Serum samples which give higher degrees of agglutination with the RBT need not be re-tested with CFT; consideration of the seropositive status of a herd should be taken into consideration on defining the cut-off optical density readings for ELISA.