IntroductionPrevious research demonstrated that medical scent detection dogs have the ability to distinguish SARS-CoV-2 positive from negative samples with high diagnostic accuracy. To deploy these dogs as a reliable screening method, it is mandatory to examine if canines maintain their high diagnostic accuracy in real-life screening settings. We conducted a study to evaluate the performance of medical scent detection dogs under real-life circumstances.MethodsEight dogs were trained to detect SARS-CoV-2 RT-qPCR-positive samples. Four concerts with a total of 2802 participants were held to evaluate canines’ performance in screening individuals for SARS-CoV-2 infection. Sweat samples were taken from all participants and presented in a line-up setting. In addition, every participant had been tested with a SARS-CoV-2 specific rapid antigen test and a RT-qPCR and they provided information regarding age, sex, vaccination status and medical disease history. The participants’ infection status was unknown at the time of canine testing. Safety measures such as mask wearing and distance keeping were ensured.ResultsThe SARS-CoV-2 detection dogs achieved a diagnostic specificity of 99.93% (95% CI 99.74% to 99.99%) and a sensitivity of 81.58% (95% CI 66.58% to 90.78%), respectively. The overall rate of concordant results was 99.68%. The majority of the study population was vaccinated with varying vaccines and vaccination schemes, while several participants had chronic diseases and were under chronic medication. This did not influence dogs’ decisions.ConclusionOur results demonstrate that SARS-CoV-2 scent detection dogs achieved high diagnostic accuracy in a real-life scenario. The vaccination status, previous SARS-CoV-2 infection, chronic disease and medication of the participants did not influence the performance of the dogs in detecting the acute infection. This indicates that dogs provide a fast and reliable screening option for public events in which high-throughput screening is required.

A prevalence study was conducted on German sheep flocks including goats if they cohabitated with sheep. In addition, a novel approach was applied to identify an infection at the herd-level before lambing season with preputial swabs, suspecting venereal transmission and ensuing colonisation of preputial mucosa with Coxiella (C.) burnetii. Blood samples and genital swabs were collected from breeding males and females after the mating season and were analysed by enzyme-linked immunosorbent assay (ELISA) and quantitative polymerase chain reaction (qPCR) respectively. In total, 3367 animals were sampled across 71 flocks. The true herd-level prevalence adjusted for misclassification probabilities of the applied diagnostic tests using the Rogan-Gladen estimator for the prevalence estimate and a formula by Lang and Reiczigel (2014) for the confidence limits, ranged between 31.3% and 33% (95% confidence interval [95% CI] 17.3–45.5) detected by the ELISA and/or qPCR. Overall 26–36.6% (95% CI 13–56.8) were detected by ELISA, 13.9% (95% CI 4.5–23.2) by the qPCR and 7.9–11.2% (95% CI 0.08–22.3) by both tests simultaneously. The range of results is due to data obtained from literature with different specifications for test quality for ELISA. Among eight farms with females shedding C. burnetii, three farms (37.5%) could also be identified by preputial swabs from breeding sires. This indicates less reliability of preputial swabs if used as a single diagnostic tool to detect C. burnetii infection at the herd-level.

Latent class analysis is a well-established method in human and veterinary medicine for evaluating the accuracy of diagnostic tests without a gold standard. An important assumption of this procedure is the conditional independence of the tests. If tests with the same biological principle are used, this assumption is no longer met. Therefore, the model has to be adapted so that the dependencies between the tests can be considered. Our approach extends the traditional latent class model with a term for the conditional dependency of the tests. This extension increases the number of parameters to be estimated and leads to negative degrees of freedom of the model, meaning that not enough information is contained in the existing data to obtain a unique estimate. As a result, there is no clear solution. Hence, an iterative algorithm was developed to keep the number of parameters to be estimated small. Given adequate starting values, our approach first estimates the conditional dependencies and then regards the resulting values as fixed to recalculate the test accuracies and the prevalence with the same method used for independent tests. Subsequently, the new values of the test accuracy and prevalence are used to recalculate the terms for the conditional dependencies. These two steps are repeated until the model converges. We simulated five application scenarios based on diagnostic tests used in veterinary medicine. The results suggest that our method and the Bayesian approach produce similar precise results. However, while the presented approach is able to calculate more accurate results than the Bayesian approach if the test accuracies are initially misjudged, the estimates of the Bayesian method are more precise when incorrect dependencies are assumed. This finding shows that our approach is a useful addition to the existing Bayesian methods, while it has the advantage of allowing simpler and more objective estimations.

Prevalence of Coxiella burnetii in German sheep flocks and evaluation of a novel approach to detect an infection via preputial swabs at herd-level-ERRATUM. Epidemiology and Infection 148,

Veselkoff N and Volk HA (2022) Expert considerations and consensus for using dogs to detect human SARS-CoV-2-infections.