TEXT

Top Abstract Text References

Coagulase-negative staphylococci (CoNS) are important causes of nosocomial bacteremia, especially in patients with intravenous catheters or indwelling medical devices (1, 3, 7, 14). In some institutions, 70 to 75% of CoNS are resistant to oxacillin; consequently, vancomycin is the therapy of choice. If the isolate is oxacillin susceptible, treatment with a penicillinase-resistant penicillin is preferred to minimize selection of vancomycin-resistant organisms. A susceptibility test method that provides accurate oxacillin results is critical.

The National Committee for Clinical Laboratory Standards (NCCLS) changed the breakpoints for oxacillin testing of CoNS in January 1999, because of discrepancies between results of mecA genetic assays, which are considered the reference method for detecting oxacillin resistance in staphylococci, and the former NCCLS breakpoints of 4 µg/ml for resistance and 2 µg/ml for susceptibility. Currently, for microtiter dilution and disk diffusion testing, respectively, CoNS isolates with an MIC of 0.5 µg/ml or a zone diameter of 17 mm are considered resistant, and those with an MIC of 0.25 µg/ml or a zone diameter of 18 mm are considered susceptible (9). To comply with these guidelines, it was necessary for laboratories using the Vitek (bioMerieux Vitek, Inc.) to use an alternate method when testing CoNS. Vitek updated its software and increased the number of oxacillin concentrations tested to accommodate the revised oxacillin breakpoints for CoNS in September 1999. The purpose of the present study was to evaluate the performance of the Vitek card GPS105 with software version VTK-R07.01. The results were compared to disk diffusion and detection of the mecA gene by PCR.

Isolates of CoNS (n = 202) from cultures of clinical specimens of blood and urine were collected prospectively and stored at 70°C. Prior to testing, isolates were subcultured twice on sheep blood agar. Isolates were identified as CoNS using Gram stain, catalase, and latex agglutination tests (Staphaurex; Murex Biotech Ltd., Dartford, England). CoNS were identified to species with the API Staph-Ident system (bioMerieux Vitek, Inc.). Disk diffusion testing was performed according to NCCLS guidelines (10) by using Mueller-Hinton agar without additional NaCl (BD Biosciences, Sparks, Md.) and 1-µg oxacillin disks (BD Biosciences). Zone diameters were measured after plates were incubated in ambient air at 35°C for 24 h. Vitek cards GPS105 were inoculated and incubated according to manufacturer's recommendations. Results were interpreted by Vitek software version VTK-R07.01. If disk diffusion and Vitek results disagreed, both tests were repeated. Multiplex PCR for detection of the mecA gene and the Staphylococcus IS431 sequence was performed as previously described (13); if isolates did not contain the mecA gene or an IS431 sequence, amplification of the Staphylococcus 16S rRNA sequence was performed as described previously (6).

The 202 CoNS isolates tested (141 from urine, 61 from blood) in this study belonged to 11 species (Table 1). The mecA gene was detected in 126 isolates (62.4%), of five species: Staphylococcus epidermidis, S. haemolyticus, S. hominis, S. warneri, and S. xylosus. S. capitis, S. cohnii, S. lugdunensis, S. saprophyticus, S. sciuri, and S. simulans were consistently mecA negative.

Initial Vitek and disk diffusion results are summarized by source in Table 1. The majority of the CoNS from blood were S. epidermidis. The mecA gene was detected in 44 (72%) of all blood isolates; all were determined to be oxacillin resistant by both disk diffusion and Vitek. All 115 mecA-negative isolates were determined to be susceptible by Vitek and disk diffusion (sensitivity and specificity of 100%).

Of the urine isolates (n = 141), 80 were mecA positive and 61 were mecA negative. Both Vitek and disk diffusion methods did not perform as well on isolates of CoNS from urine: 6 of the mecA positive isolates were determined to be susceptible by Vitek, and 11 of the mecA negative isolates were determined to be resistant (sensitivity, 92.5%; specificity, 82%). The 11 isolates that were mecA negative but oxacillin resistant as determined by Vitek and disk diffusion included six S. saprophyticus isolates, two S. lugdunensis isolates, and one isolate each of S. capitis, S. cohnii, and S. sciuri. Three species (S. epidermidis, S. hominis, and S. haemolyticus) comprised the majority of the isolates (122 of 141; 86.5%). For these three species, the sensitivity and specificity of Vitek were high (92.4 and 97.6%, respectively). After repeat testing, Vitek correctly identified 76 of the 79 mecA-positive isolates as resistant and 100% of the mecA-negative isolates as susceptible (sensitivity, 96.2; specificity, 100%).

Discrepancies between results of mecA PCR assays when CoNS are tested and the NCCLS microtiter dilution and disk diffusion breakpoints for oxacillin against S. aureus have been recognized for several years. For this reason, various investigators recommended lowering the oxacillin MIC breakpoint for CoNS: York et al. (15) suggested 1 µg/ml for susceptibility, Cormican et al. (2) and McDonald et al. (8) suggested 0.5 µg/ml, and Marshall et al. (7) suggested 0.25 µg/ml.

Tenover et al. (12) concurred with the breakpoints of 0.25 µg/ml for susceptibility and 0.5 µg/ml for resistance because these values allowed maximum sensitivity for detecting mecA-positive isolates of S. epidermidis without drastically compromising specificity. These authors acknowledged, however, that these breakpoints were considerably less specific for mecA in species of CoNS other than S. epidermidis. Tenover et al. (12) also suggested modified breakpoints for disk diffusion: 17 mm for resistance and 18 mm for susceptibility. The NCCLS considered data from all studies and adopted the oxacillin MIC and disk diffusion breakpoints for CoNS advocated by Tenover et al. Our results show that the new Vitek revised cards and software system accurately predicts the presence or absence of mecA in the three most commonly encountered species of CoNS.

For species of CoNS other than S. epidermidis, S. hominis, and S. haemolyticus, we found poor correlation between mecA PCR and oxacillin testing by Vitek or disk diffusion. In particular, isolates of S. capitis, S. cohnii, S. lugdunensis, S. saprophyticus, and S. sciuri that had an oxacillin MIC of 0.5 µg/ml as determined by Vitek (and were resistant by disk diffusion) were consistently mecA negative. Similar findings have been reported by other investigators (4, 5, 12). Hussain et al. (4) tested 493 isolates of 11 species of CoNS (no S. scuiri) and found mecA-negative isolates of S. capitis, S. cohnii, S. lugdunensis, and S. saprophyticus with oxacillin MICs of 0.5 to 2.0 µg/ml by agar dilution. In a second study, Hussain et al. (5) tested 463 isolates of 13 species of CoNS (no S. scuiri) and found mecA-negative S. capitis, S. cohnii, S. lugdunensis, S. saprophyticus, S. caprae, S. warneri, and S. xylosus with oxacillin MICs of 0.5 µg/ml. In the latter study, isolates were tested by a latex agglutination test that detects penicillin-binding protein (PBP) 2a. Compared to mecA PCR, the sensitivity of that latex test was 100% and the specificity was 99.5%, whereas the specificity of the new NCCLS breakpoint was 60.8%. The mechanism(s) of decreased susceptibility to oxacillin in mecA-negative CoNS with an oxacillin MIC of 0.5 to 2.0 µg/ml is unknown, but alterations in PBPs other than PBP2a have been documented in some strains of S. haemolyticus and S. saprophyticus (11) and, therefore, suggested as one possibility (12). It is not known whether penicillinase-resistant penicillins can eradicate clinical isolates of mecA-negative CoNS with an oxacillin MIC of 0.5 to 2.0 µg/ml. Clinical trials designed to answer this question are needed.