Identification of Staphylococcus species with the VITEK 2 system: the case of Staphylococcus hominis.

We read with interest the paper by Layer et al. regarding the identification of staphylococci with automated systems including VITEK-2 (bioMérieux, Marcy l’Etoile, France) (6). Of 86 strains analyzed in their study, 4 were Staphylococcus hominis and all were correctly identified by the colorimetric ID-GP VITEK-2 cards. Identical results have been reported in recent studies evaluating 13 and 14 S. hominis isolates (3, 7). Two S. hominis subspecies exist—subspecies hominis and subspecies novobiosepticus—which may be differentiated according to novobiocin susceptibility (5). S. hominis subsp. hominis is more common in clinical samples, but S. hominis subsp. novobiosepticus has recently been shown to cause bacteremia among neonates (1). VITEK-2 actually identifies S. hominis with “low discrimination,” and performance of manual novobiocin susceptibility testing is advised by the manufacturer in order to differentiate S. hominis subsp. hominis and S. hominis subsp. novobiosepticus. Thus, it would be interesting to know what subspecies were used by Layer et al. and whether the system performed similarly with both subspecies. Interestingly, novobiocin resistance testing is one of 43 reactions carried out by the ID-GP card. While a negative result may represent susceptibility but also early termination of the reaction, it is tempting to regard a positive reaction as suggestive of S. hominis subsp. novobiosepticus. We studied nonduplicate consecutive clinical isolates of S. hominis and S. epidermidis (15 each) identified by VITEK-2. S. epidermidis was chosen as the control since it is almost always novobiocin susceptible (4). Isolates were processed by using VITEK-2 according to the manufacturer’s instructions. Isolates were inoculated onto cation-adjusted Mueller-Hinton agar (Hy-Labs, Rehovot, Israel) according to Clinical Laboratory Standards Institute guidelines (2) and tested with a 5g novobiocin disk (BBL, BD Diagnostic Systems, Cockeysville, MD). A zone-ofinhibition diameter of 15 mm was considered to indicate resistance (5). Quality control was monitored using S. aureus ATCC 29213. According to VITEK-2, of 15 S. hominis isolates, 5 (33.3%) were resistant to novobiocin, 8 were susceptible, and 2 yielded an indeterminate result. Of 15 S. epidermidis isolates, 3 (20%) were novobiocin resistant and 12 (80%) were susceptible. However, by disk diffusion, all isolates of both species were novobiocin susceptible. The mean zone-of-inhibition diameter was 34.3 4 mm overall (range, 26 to 43 mm). That for S. hominis was 31.8 3.1 mm (range, 26 to 36 mm), and that for S. epidermidis was 36.8 3.2 mm (range, 33 to 43 mm; P, 0.001 by the Kruskal-Wallis test). The mean zone diameter for the eight isolates that were novobiocin resistant by VITEK-2 was 34.5 3.4 mm, compared to 34.9 3.8 mm for the 20 susceptible strains (P 0.98). Results were similar when S. hominis and S. epidermidis were analyzed separately. Novobiocin susceptibility results from the VITEK-2 ID-GP cards do not correlate well with disk diffusion results, and manual testing is indeed warranted for differentiating S. hominis subsp. hominis and S. hominis subsp. novobiosepticus. Resolving the low-discrimination result obtained with S. hominis is of even greater value given that S. hominis subsp. novobiosepticus is not included in the VITEK-2 advanced expert system knowledge base and, therefore, therapeutic corrections are not suggested for S. hominis subsp. novobiosepticus. Given all of the above, we believe that in future studies evaluating the performance of automated systems with coagulase-negative staphylococci, the issue of S. hominis subsp. hominis and S. hominis subsp. novobiosepticus differentiation should receive more emphasis.