Validation of an indirect immunofluorescence assay (IFA) for the detection of IgG antibodies against Coxiella burnetii in bovine serum

Abstract

There is limited knowledge of the true prevalence and distribution of coxiellosis in dairy and beef cattle populations in Australia. For this to occur, apparent prevalence estimates need to be reliably adjusted, accounting for diagnostic sensitivity (DSe) and diagnostic specificity (DSp) of the test used. However, there are few tests available with known diagnostic specifications suitable to inform screening and surveillance activities in the Australian context. We initially modified and optimised a human indirect immunofluorescence assay (IFA) test for the detection of IgG antibodies against phase I and/or phase II Coxiella burnetii in bovine sera and determined an optimal screening dilution cut-off to be 1:160. Direct comparison of the modified IFA with the commercial IDEXX enzyme-linked immunosorbent assay (ELISA) kit (Q Fever Ab Test IDEXX Laboratories, United States of America) was performed by testing 458 serum samples from four distinct cattle populations across the east coast of Australia and New Zealand. Cross classified test results were then analysed using Bayesian latent class modelling, to validate the tests in the absence of a gold standard reference test. Results from this analysis indicate that the IFA, at a 1:160 serum dilution, has an estimated DSe of 73.6% (95% Credible Interval (CrI) 61.1, 85.9) and DSp of 98.2% (95% CrI 95.1, 99.7). The commercial IDEXX ELISA kit was found to have a higher DSe of 87.9% (95% CrI 73.9, 96.4) and similar DSp of 97.7% (95% CrI 93.2, 99.7). Evaluation of the diagnostic performance of the IFA and ELISA methods, specifically for use in cattle will enable more accurate interpretation of prevalence estimates of C. burnetii exposure to be reported for cattle in Australia and other countries.

Keywords: Bayesian latent class model; Cattle; Coxiella burnetii; Coxiellosis; ELISA; IFA; Validation.