The ELISA is recognized as an efficient diagnostic tool for antibody detection, but there is no standard ELISA assay for detection of antibodies against hemorrhagic septicemia (HS) in cattle. The present study reports on an indirect ELISA assay for antibody detection of HS in dairy cows, and evaluates the sensitivity (Se) and specificity (Sp) of the method using a Bayesian approach. An indirect ELISA was developed with two types of heat extract antigens, Pasteurella multocida strains P-1256 and M-1404, as coating antigens. A checkerboard titration was employed using dairy cow sera immunized with P. multocida bacterin and colostrum-deprived calf sera. The concentrations of heat extract antigen (160μg/mL), sample serum (1:100) and goat anti-bovine immunoglobulin G labeled with horseradish peroxidase (1:2000) were optimal for the assay. The cut-off values were 0.147 and 0.128 for P-1256 and M-1404 coating antigens, and there were no differences in the results of tests with positive and negative sera (p<0.05). The characteristics of three diagnostic tests were evaluated using a one-population Bayesian model, assuming conditional dependence between two types of coating antigen-based ELISAs and indirect hemagglutination assay (IHA). A total of 415 sera samples from dairy cows without HS vaccination and no history of disease were tested. The Se and Sp of the P-1256 and M-1404 ELISAs were higher than those of the IHA. The Se and Sp of the P-1256 ELISA were 90.3% and 90.1%, while the Se and Sp of the M-1404 ELISA were 92.1% and 71.9%. The median values of Se and Sp from the IHA were 36.0% and 58.2%.
Keywords: Bayesian latent class analysis; Heat extract antigen; Hemorrhagic septicemia; Indirect ELISA; Pasteurella multocida.
Copyright © 2017 Elsevier B.V. All rights reserved.