Sensitivity and specificity of bacterial culture, qPCR, and somatic cell count for detection of goats with Staphylococcus aureus intramammary infection using Bayesian latent class models

Abstract

Staphylococcus aureus (S. aureus) is an important udder pathogen affecting goat milk production. The ability to detect goats with subclinical mastitis caused by S. aureus is essential in udder health control programs. In Norway, the industry recommends using somatic cell count (SCC) as a screening tool, and conventional bacterial culture (BC) as a confirmatory test for goat milk samples, but a commercial qPCR, Mastitis 4 qPCR (DNA Diagnostics, Risskov, Denmark) is also available. However, few studies have validated the use of these methods for the detection of goats with S. aureus intramammary infection (IMI). Therefore, the objective of this retrospective study was to estimate the diagnostic sensitivity and specificity of BC, qPCR, and SCC for the detection of goats with IMI caused by S. aureus using Bayesian latent class analysis. We analyzed the BC and qPCR results of aseptically collected milk samples and SCC results from milk recording samples from 319 goats from three herds using different SCC cut-offs. At a SCC cut-off of 2000,000 cells/mL, the estimated median prevalence in each herd was 12.7% (95% highest posterior density credible interval [CI] 6.5–19.8), 15.7% (95% CI 9.3–23.0), and 1.5% (95% CI 0.0–4.3). The median sensitivity was 93.0% (95% CI 80.2–100), 93.6% (95% CI 82.3–100) and 78.2% (95% CI 62.3–91.2) for BC, qPCR, and SCC, respectively. The estimated median specificity of BC was 99.5% (99% CI 98.4–100), for qPCR, 98.9% (95% CI 97.5–100), and for SCC 61.5% (95% CI 56.0–67.1). The results show that BC, which is today’s standard method for diagnosing IMI, has a high accuracy for detection of goats with S. aureus IMI, but qPCR had a sensitivity and specificity similar to BC, and may act as an alternative.