Use of real-time PCR on faecal samples for detection of sub-clinical Salmonella infection in cattle did not improve the detection sensitivity compared to conventional bacteriology

There is a need for more sensitive detection methods to improve effectiveness of control programmes of Salmonella enterica subsp. enterica serotypes (Salmonella) in cattle. We assessed the performance of a rapid, molecular-based, real-time PCR (rt-PCR) method against the conventional bacteriological culture-reference method (BCRM) on cattle faecal samples for detection of sub-clinical Salmonella infections in cattle. Thirty faecal samples were artificially contaminated with either 10 or 50 CFU of one of five strains of S. Dublin (SD) and S. Typhimurium (ST). The overall detection sensitivity of both rt-PCR and BCRM was 100% for ST and 78% for SD. Furthermore, 163 faecal samples from cattle herds with suspected Salmonella infection were tested to compare the relative performance of rt-PCR to BCRM on samples from naturally infected herds. The relative sensitivity of rt-PCR was 20% (3/15 BCRM positive samples) while the relative specificity and accuracy was 99% and 92%, respectively. Both methods had limitations for detecting low levels of SD (<1 CFU/g). Hence, the evaluated rt-PCR method did not provide a sensitive alternative to the BCRM for detection of bacteria in faecal samples of sub-clinically, Salmonella-infectious cattle.