Comparison of test performance of a conventional PCR and two field-friendly tests to detect Coxiella burnetii DNA in ticks using Bayesian latent class analysis

( )-infected livestock and wildlife have been epidemiologically linked to human Q fever outbreaks. Despite this growing zoonotic threat, knowledge of coxiellosis in wild animals remains limited, and studies to understand their epidemiologic role are needed. In -endemic areas, ticks have been reported to harbor and spread and may serve as indicators of risk of infection in wild animal habitats. Therefore, the aim of this study was to compare molecular techniques for detecting DNA in ticks. In total, 169 ticks from wild animals and cattle in wildlife conservancies in northern Kenya were screened for DNA using a conventional PCR (cPCR) and two field-friendly techniques: Biomeme's qPCR Go-strips (Biomeme) and a new PCR high-resolution melt (PCR-HRM) analysis assay. Results were evaluated, in the absence of a gold standard test, using Bayesian latent class analysis (BLCA) to characterize the proportion of positive ticks and estimate sensitivity (Se) and specificity (Sp) of the three tests. The final BLCA model included main effects and estimated that PCR-HRM had the highest Se (86%; 95% credible interval: 56-99%), followed by the Biomeme (Se = 57%; 95% credible interval: 34-90%), with the estimated Se of the cPCR being the lowest (24%, 95% credible interval: 10-47%). Specificity estimates for all three assays ranged from 94 to 98%. Based on the model, an estimated 16% of ticks had DNA present. These results reflect the endemicity of in northern Kenya and show the promise of the PCR-HRM assay for surveillance in ticks. Further studies using ticks and wild animal samples will enhance understanding of the epidemiological role of ticks in Q fever.