Short communication: comparing real-time PCR and bacteriological cultures for Streptococcus agalactiae and Staphylococcus aureus in bulk-tank milk samples

For more than 30 yr, a control plan for Streptococcus agalactiae and Staphylococcus aureus has been carried out in more than 1,500 dairy herds of the province of Brescia (northern Italy). From 2010 to 2011, the apparent prevalence of Strep. agalactiae has been relatively stable around 10%, but the apparent prevalence of Staph. aureus has been greater than 40% with an increasing trend. The aim of this paper was to estimate and compare the diagnostic accuracy of 3 assays for the detection of Strep. agalactiae and Staph. aureus in bulk-tank milk samples (BTMS) in field conditions. The assays were a qualitative and a quantitative bacteriological culture (BC) for each pathogen and a homemade multiplex real-time PCR (rt-PCR). Because a gold standard was not available, the sensitivities (Se) and specificities (Sp) were evaluated using a Bayesian latent class approach. In 2012 we collected one BTMS from 165 dairy herds that were found positive for Strep. agalactiae in the previous 2-yr campaigns of eradication plan. In most cases, BTMS collected in these herds were positive for Staph. aureus as well, confirming the wide spread of this pathogen. At the same time we also collected composite milk samples from all the 8,624 lactating cows to evaluate the within-herd prevalence of Strep. agalactiae. Streptococcus agalactiae samples were cultured using a selective medium Tallium Kristalviolette Tossin, whereas for Staph. aureus, we used Baird Parker modified medium with added Rabbit Plasma Fibrinogen ISO-Formulation. In parallel, BTMS were tested using the rt-PCR. Regarding Strep. agalactiae, the posterior median of Se and Sp of the 2 BC was similar [qualitative BC: Se=98%, posterior credible interval (95%PCI): 94-100%, and Sp=99%, 95%PCI: 96-100%; quantitative BC: Se=99%, 95%PCI: 96-100%, and Sp=99%, 95%PCI: 95-100%] and higher than those of the rt-PCR (at 40 cycle threshold, Se=92%, 95%PCI: 85-97%; Sp=94%, 95%PCI: 88-98%). Also in case of Staph. aureus, the posterior medians of BC were generally higher than those of rt-PCR. In fact, although the Se of BC was slightly lower (rt-PCR at 40 cycle threshold, median Se=99%, 95%PCI: 97-100%, and qualitative BC, median Se=94%, 95%PCI: 87-99%), the Sp was much higher (rt-PCR at 40 cycle threshold, median Sp=67%, 95%PCI: 38-97%; qualitative BC, median Sp=95%; 95%PCI: 76-100%). Our study confirms that BC and rt-PCR are reliable diagnostic tools to detect Strep. agalactiae and Staph. aureus, and rt-PCR results should be confirmed by BC carried out on BTMS and possibly on composite milk samples.