Abstract
Two Ag-ELISAs, an IgG-specific antibody detection ELISA (IgG ELISA) and a card agglutination test (CATT) for the detection of Trypanasoma evansi infections in buffaloes in Indonesia, were compared. Diagnostic sensitivity estimates were obtained by testing sera from 139 Indonesian buffaloes which had been found to be infected by parasitological tests. Diagnostic specificity was estimated by testing sera from 263 buffaloes living in Australia. Response-operating characteristic curves were constructed, and optimal ELISA cut-off values, which minimized the number of false-negative and false-positive results, were chosen. The IgG ELISA had the highest sensitivity (89%) and the CATT had the highest specificity (100%). There was a significant difference between the sensitivities (71 and 81%), but not between the specificities (75 and 78%), of the two Ag-ELISAs. The four tests were further compared by calculation of post-test probabilities of infection for positive and negative test results using a range of prevalence values, and likelihood ratios. The results suggested that the CATT was the best test to 'rule-in' infection (i.e. the highest probability of infection in test-positive animals) and the IgG ELISA was the best test to 'rule-out' infection (i.e. the lowest probability of infection in test-negative animals).