[en] Performances of four detection methods were evaluated for recovery of Salmonella spp. in naturally contaminated fecal specimens of porcine origin. The NMKL 71 method consisted of enrichment in Rappaport-Vassiliadis broth and plating on xylose-lysine-desoxycholate medium, whereas the SP-VG-M002 method relied on a Diasalm enrichment followed by streaking on xylose-lysine-tergitol 4 agar (XLT-4). The VIDAS SLM method was composed of double enrichment in Muller-Kauffmann tetrathionate broth and in M broths before processing in a VIDAS device. lf the results were positive, the VIDAS ICS immunoenrichment was performed and the result transferred onto three different selective media. The VIDAS ICS protocol is an immunoconcentration step followed by plating on XLT-4. Seventy-eight samples were tested with all four methods simultaneously, leading to 34 positive samples with at least one method. For this assay, VIDAS SLM revealed 31 positive samples (91.2%), whereas the average positive percentage of the three other methods was 37.3% (P < 0.001). Two-paired comparisons with the VIDAS SLM method were also performed. McNemar values were systematically highly significant (P < 0.001). The proportion of agreement was significantly inferior (P < 0.05) for the comparison of VIDAS ICS and VIDAS SLM (68.7%) compared with the two other paired comparisons (average percentage, 81.5%). The conclusion reached by this trial is that VIDAS SLM significantly improves the recovery of Salmonella in naturally contaminated fecal specimens. For the paired-comparisons, NMKL 71 and SP-VG-M002 were comparable in terms of efficiency, whereas the VIDAS ICS protocol, as established by the manufacturer for food samples only, seemed less efficient than the other two.