Comparison of 2 culture methods and PCR assays for Salmonella detection in poultry feces.

The present work compared 2 culture methods and the combinations of pre-enrichment and enrichment culture methods with PCR assays [buffered peptone water-PCR and tetrathionate-PCR or modified semisolid Rappaport-Vassiliadis (MSRV)-PCR] for motile and nonmotile Salmonella strain detection using artificially contaminated poultry feces. The specificity and positive predictive values were equal to one in both culture methods. Specificity and positive predictive values, accuracy, sensitivity, and negative predictive values were higher for motile than nonmotile Salmonella strains in culture methods. Only Salmonella enterica serovar Gallinarum was detected by the MSRV method with low accuracy, sensitivity, and negative predictive value. The detection level of motile strains was 2 ×10(0) to 22 × 10(2) cfu per 25 g for these methods, whereas it was 6.9 × 10(2) cfu per 25 g in culture methods for Salmonella Gallinarum. Extending the incubation time of the enrichment medium to 6 d in the TT method did not improve the isolation rates. In general, all selective plating media did not show any statistical differences in the parameters of performance studied. On the other hand, accuracy and sensitivity values were higher in MSRV-PCR and tetrathionate-PCR methods than in the buffered peptone water-PCR method. Specificity and positive predictive values were equal to one in most of the cases. In terms of detection limits, motile Salmonella strains were recovered from 5 × 10(0) cfu per 25 g in MSRV-PCR and tetrathionate-PCR methods, whereas the detection limit was better for nonmotile Salmonella in MSRV-PCR methods than in the tetrathionate-PCR method. Kappa coefficients showed that there was a very good agreement between tetrathionate and MSRV methods for motile Salmonella strains, whereas these methods did not show any concordance for nonmotile Salmonella strains. When buffered peptone water-PCR was compared with both tetrathionate-PCR and MSRV-PCR, agreement was poor for motile Salmonella strains and slight to fair for nonmotile Salmonella strains. The difference in isolation rate obtained with the methods used for motile and nonmotile Salmonella strains must be taken into account when a poultry fecal sample is considered negative for the presence of Salmonella.