Evaluation of the Vitek 2 System with a Variety of Staphylococcus Species

Staphylococci are widespread in nature and can be isolated from humans and animals and from various food products (4, 5, 11, 13, 15). Some staphylococcal strains are associated with diseases in humans or animals, and others are used for technological purposes (4, 14, 15). Staphylococcus aureus is a well documented pathogen. Of the non-S. aureus staphylococci, S. epidermidis, S. haemolyticus, and S. saprophyticus are those most frequently implicated in disease in humans (4, 9, 15). Other staphylococcal species, such as S. hominis and S. warneri, can be found as contaminants of blood cultures but can also be associated with a variety of infections. In food, particularly in fermented meat products, staphylococci such as S. xylosus, S. carnosus, S. simulans, S. warneri, S. epidermidis, and S. saprophyticus may be found (1). S. xylosus and S. carnosus strains are used as starter cultures in fermented meat products because they contribute to their color and flavor (14). However, the presence of S. aureus in food is a potential public health hazard, since many strains of S. aureus produce enterotoxins (7, 10). Accurate identification of the Staphylococcus species is therefore of great importance in microbiological laboratories.

The Vitek 2 system used with the gram-positive (GP) identification card (bioMérieux, Marcy l'Etoile, France) is an automated machine designed to provide rapid and accurate phenotypic identifications for most clinical staphylococci (2, 6, 8, 16). The colorimetric GP card contains 43 tests. The database of the GP card particularly includes the environmental species S. caprae, S. carnosus, S. equorum, S. gallinarum, and S. vitulinus. However, there has been no documented assessment of this identification card for any collection of Staphylococcus strains, including environmental isolates. The purpose of this study was to assess the ability of the Vitek 2 GP card to identify Staphylococcus species of clinical and environmental origins.

A total of 190 Staphylococcus strains, including 38 type strains representing 35 species (Table 1) and 152 strains from our collection, were studied (Table 2). The latter 152 strains were isolated from clinical samples (n = 60) and from food and plant samples (n = 92). The GP card was used in accordance with the recommendations of the manufacturer. The results were compared with those of the oligonucleotide “Staph array,” a method, which is based on the hybridization of the internal part of the sodA gene (3). The strain was retested with the GP card if the identification results differed between the two methods. In case of a persistent discrepancy, the Staph array identifications were confirmed by DNA sequencing of the sodA gene (12). Results from the GP card were separated into four groups: (i) the “one-choice identification” group corresponded to identical identifications for the GP card and Staph array; (ii) the “low level of discrimination” group corresponded to results for which the GP card indicated the possibility of two or three species, including the result of Staph array and proposed supplementary tests (pigmentation, hemolysis, or novobiocin resistance) to determine the correct identification; (iii) the “misidentification” group corresponded to different identifications for the GP card and Staph array; and (iv) the “no identification” group corresponded to strains for which the GP card gave no result. Correct identification was defined as the association of the first two groups.

Twenty-four reference strains out of 38 (63%) were correctly identified by the GP card (Table 1). Of the 14 misidentified reference strains, 10 were species not included in the database of the GP card. The four other strains were S. capitis subsp. capitis, identified as S. equorum; S. equorum subsp. equorum, identified as S. xylosus; S. haemolyticus, identified as S. warneri/S. pasteuri; and S. vitulinus, identified as S. sciuri. In a study by Layer et al. (6), 70% (20 out of 27) of reference strains were correctly identified. However, the misidentified strains belonged to different species.

Fifty-six of the 60 clinical strains (93.3%), belonging to 14 different species, were identified from the collection. Complementary tests were necessary for three strains. These results were in agreement with those reported by Layer et al. (6). Two other studies found higher identification accuracy (98.2 to 99.3%) (2, 16). In those studies, the numbers of taxa tested and their proportions were different; 66 and 74% of clinical strains belonged to the three species S. aureus, S. epidermidis, and S. hominis, versus only 37% in our study. Three strains belonging to the species S. saprophyticus, S. simulans, and S. delphini were misidentified, and one strain of S. capitis was not identified.

The GP card also correctly identified 67 of the 92 environmental strains (73%). Complementary tests were necessary for 10 strains. Twenty-three strains (25%), belonging to five species, were misidentified (Table 2). Nine misidentified strains belonged to S. fleurettii (n = 1) and S. succinus (n = 8), which were absent in the database of the GP card. Eleven of the 14 remaining misidentified strains were S. equorum strains, which were frequently reported as S. xylosus (n = 10). The last three misidentified strains belonged to S. carnosus and S. vitulinus. One strain each of S. fleurettii and S. succinus were not identified as expected. Except for S. equorum (n = 13) and S. succinus (n = 9), 65 of the 70 environmental strains were correctly identified (92.9%). The misidentifications of S. equorum as S. xylosus were suspected with the vibriostatic O129 test of the GP card. Indeed, all the strains identified as S. xylosus which had a negative test with the vibriostatic compound were misidentified strains. According to our data, the identifications of S. xylosus, S. sciuri, and S. auricularis provided by the GP card required complementary tests for confirmation (Table 3).

In conclusion, the Vitek 2 GP card allowed the identification of 123 (93.2%; n = 132) Staphylococcus strains belonging to the species which are included in the database. The results obtained in this study highlight the interesting performance of the colorimetric Vitek 2 GP card, which can be used in clinical, agrochemical, and food laboratories. However, the GP card demonstrated low accuracy in identification of the species S. equorum and misidentified S. succinus, which is not included in the current Vitek 2 GP database.

• Copyright © 2008 American Society for Microbiology