The estimation of the mean prevalence in the sample (24%) confirmed the value assessed previously in French areas in which B. ovis
infection is enzootic [20
The estimation of the diagnostic sensitivity and specificity of a test should ideally be derived from testing a statistically relevant panel of animals. The history of these animals and their infection status should be known. The panel should be representative of the region where the test is to be used [36
]. As mentioned above, there is currently no regulatory program in the EU for eradicating the disease or surveying the absence of B. ovis
. with an appropriate definition of a B. ovis
-free status), except in AI units. In these units, animals are kept in specific conditions that could lead to an overestimation of tests specificity. Designing an adequate sampling of a sufficient number of free animals (from free flocks) is therefore very difficult. Moreover, the selection of an appropriate gold-standard-infected population (i.e.
culture-positive animals) is extremely cumbersome and requires a very important budget Furthermore, Brucella
-culture should be performed in conditions providing optimised sensitivity, with a corresponding increase in costs, in order to prevent any consecutive overestimation of the diagnostic sensitivity of the serological tests under study. The technical and financial constraints highlighted above justify the use of the Bayesian approach to estimate the characteristics of serological tests for the diagnosis of B. ovis
infection when the infection status of each animal of the studied population is unknown.
The outcome of our study is consistent with previous results available in the literature. The CFT has a good diagnostic sensitivity and an acceptable specificity but has some technical disadvantages. The EU/OIE/FAO Reference laboratory estimated that this I-ELISA commercial kit was cheaper than the CFT (in-house produced antigen) when the cost was calculated for a daily analysis of 90 serum samples, mainly due to lower manpower costs (data not shown). Moreover, the evaluated I-ELISA is standardised and easier and quicker to perform. Altogether, these advantages allow the analysis of more samples by a single technician in a working day than the CFT. Moreover, the I-ELISA can be automated to speed up the process. Furthermore, haemolysed and anti-complementary sera do not affect I-ELISA performance ([15
]). These advantages may further reduce the cost of this I-ELISA compared to the CFT.
In a screening situation, one important issue is the interpretation of doubtful results obtained with the I-ELISA. We established a cut-off different from that recommended by the manufacturer to avoid doubtful results without decreasing the concordance between the results of the two tests. Nevertheless, it is possible to adjust this cut-off according to the epidemiological context. When the test is used for international or intra-community trade, false positive results due to a lack of specificity could hinder trade and raise economic issues in brucellosis free contexts. Exported rams should come exclusively from uninfected farms (i.e. showing only negative results to regular tests), so the positive predictive value of the test results is low. For this reason, a high cut-off value could be used (for instance 60% OD). In contrast, when the test is used to diagnose B. ovis infection in infected areas or to confirm the infection in rams with clinical lesions with the aim of eradicating the disease, false negative results might slow down eradication and, in particular conditions, it could be acceptable to cull seropositive but uninfected rams. Accordingly, in a context of high prevalence, a low cut-off value could therefore be used (for example 30% OD).
The imperfect diagnostic performance of both tests could justify their use in association [35
]. With the serial association, the overall result is positive when the results are positive with both tests. This interpretation scheme increases the diagnostic specificity of screening and could therefore be recommended for pre-movement testing in a brucellosis free context, in order to avoid false positive results due to the relative lack of diagnostic specificity of this I-ELISA. With the parallel association, the overall result is positive when at least one test gives a positive result.
The parallel association increases the sensitivity of the screening. Accordingly, it could be recommended in a context of suspected infection with the aim of accelerating the eradication of the disease. In this context, many sera may have to be tested in a flock and the rapidity of the process influences the success eradication operations.
The following situations would be suitable for application of screening strategies using the I-ELISA kit:
Rams for export must come from brucellosis free flocks. In this case, the animals could be tested with I-ELISA alone. If all results were negative, rams could be qualified for export. Rams from flocks showing positive results to I-ELISA should not be exported and should be subjected to further investigations (clinical examination, CFT, semen culture or PCR for instance) to confirm or rule out the infection.
In infected flocks in which B. ovis has already been identified, the testing regime should be focused on a quickly eradication of the disease. In this case, all animals (including females) should be submitted to I-ELISA and CFT in parallel. Animals with positive results to at least one of the two tests should be culled and replaced by virgin animals from flocks free of B. ovis infection. The remaining animals resulting negative to both tests could be kept in the flock but should neither be sold nor exported. Further tests should be carried out at 45- to 60-days intervals until all infected animals have been eliminated (i.e. after the entire animal population has been tested negative twice). This strategy raises financial issues in infected flocks with a high prevalence of the disease, because many animals would be culled.
In all flocks, animals should be submitted to a clinical examination two months before the breeding season. Rams with clinical signs of epididymitis should be banned from reproduction and subjected to an I-ELISA test. Rams with positive results to I-ELISA should undergo further investigations (semen culture) to confirm the presence of Brucella ovis.