Evaluation of an absorbed ELISA and an agar-gel immuno-diffusion test for ovine paratuberculosis in sheep in Australia

The sensitivities and specificities of an absorbed enzyme-linked immunosorbent assay (ELISA) and an agar-gel immuno-diffusion (AGID) test for the detection of Johne's disease in sheep were estimated using data from six known infected and 12 assumed uninfected sheep flocks. Sensitivities were estimated for all histologically positive sheep, as well as by histological lesion score, lesion type (paucibacillary or multibacillary) and sheep body-condition score, with ELISA sensitivities estimated at 95 and 99% specificity. Logistic-regression analysis was used to test for significant effects of lesion score and condition score, with flock included in the model as a random effect. Estimated specificities were 95% (95% Cl: 93.4,95.6%) and 99% (98.4,99.4%) for ELISA cut-point ratios of 2.4 and 3.6, respectively, and 100% (99.7, 100.0%) for the AGID. Estimated sensitivities for all infected sheep were 41.5% (35.0, 48.3%), 21.9% (16.6, 27.9%) and 24.6% (19.1, 30.7%) for ELISA cut-point ratios of 2.4 and 3.6 and for AGID, respectively. Sensitivities of all tests and cut-points varied significantly between flocks and between categories of lesion score and condition score. Sensitivity ranged from 25 to 73, 10 to 47 and 9.2 to 63% between flocks, for the ELISA with cut-points of 2.4 and 3.6, and for the AGID, respectively. Sensitivity was highest in thin sheep and in sheep with multibacillary lesions. The effects of lesion type and condition score on test sensitivity were significant in the logistic regressions for the AGID and ELISA at both cut-points and the flock effect was significant for the AGID but not for the ELISA at either cut-point. (C) 2003 Elsevier B.V. All rights reserved.