Comparison of classical isolation protocols with a 24 h screen to detect viable salmonellas in faeces

A 24 h screen which detects three viable salmonella cells per g of faeces was compared with classical isolation procedures for their ability to identify salmonella-positive samples from a pig rearing unit. The screen involved an overnight enrichment in Muller-Kauffmann tetrathionate (MK) broth, subculture for 4 h in M broth containing 10 micrograms ml-1 novobiocin, followed by detection of the presence of salmonellas by BacTrace and Salmonella-tek ELISAs. The classical protocols were: (1) an overnight and 48 h incubation in MK or selenite cysteine broth; or (2) overnight incubation in buffered peptone water and 24 h subculture in Rappaport-Vassiliadis broth (BPW-RV). Salmonellas were isolated from the broth cultures on xylose lysine deoxycholate and brilliant green agars. Thirty four of 100 samples were positive for salmonellas but no single isolation protocol identified all of them. The best of the classical isolation protocols, 48 h incubation in MK broth, identified 27 (79%) of the 34 positive samples whilst the screen identified 26 (76%) of the 34 positive samples. False-positive results were obtained from all isolation protocols except BPW-RV.