Comparison of PCR detection of mecA with agar dilution and Etest for oxacillin susceptibility testing in clinical isolates of coagulase-negative staphylococci

Coagulase-negative staphylococci (CoNS) are a significant cause of bacteraemia and other hospital-acquired infections, especially in immunocompromised individuals and patients with prosthetic implants [1–3]. Many CoNS are resistant to penicillin and oxacillin because of production of β-lactamase and penicillin-binding protein 2a (PBP2a), respectively [4]. Production of PBP2a in staphylococci is encoded by the mecA gene. PBP2a has been found in a variety of CoNS, but expression of mecA is different in CoNS and Staphylococcus aureus. Several groups of investigators have described DNA probes [5–7] and PCR assays [8–16] for detection of oxacillin resistance in staphylococci. Based on the results of some of these studies, the National Committee for Clinical Laboratory Standards (NCCLS) has redefined the breakpoints for oxacillin susceptibility in CoNS [17]. In the present study, PCR detection of the mecA gene was compared with Etest and agar dilution based on the new NCCLS breakpoint for identification of oxacillin resistance in clinical isolates of CoNS.

Clinically significant isolates (n = 180) were collected from 13 Norwegian clinical microbiological laboratories over a period of 22 months. Preliminary identification in the local laboratory was on the basis of colony morphology, a positive catalase test, and a negative agglutination test or a negative DNase test. CoNS blood culture isolates that had been collected and stored at Haukeland University Hospital for the previous 2 years were also included in the study (n = 52). The 232 isolates were from the following sources: blood culture (160 isolates), venous and arterial catheters (32), surgical wounds (30), tracheal tubes (four), spinal fluid (two), peritoneal fluid (two), pleural fluid (one) and eye secretion (one). All isolates were identified to the species level by API Staph 32 ID (bioMérieux, Marcy l'Etoile, France) with an automated reader at Haukeland University Hospital.

Etests (AB Biodisk, Solna, Sweden) were performed following the recommendations of the Norwegian Working Group on Antibiotics [18]. Agar dilution was performed with Mueller–Hinton agar (Mast, Bootle, UK) supplemented with NaCl 2% w/v, an inoculum of 0.5× McFarland standard, and incubation for 24 h at 34 ± 1 °C. MecA PCR was performed as described by Predari et al.[9] at the Telelab laboratory, with the following modification in the lysis procedure. The pellet was resuspended in 100 µL TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8.0) with 10 µL lysostaphin (15 U/mL). Following incubation at 37 °C for 15 min, 10 µL sodium dodecyl sulphate 1% w/v and 10 µL proteinase K (25 mg/mL in 50 mM Tris-HCl, pH 8.0) were added, followed by incubation for a further 15 min at room temperature.

In total, 125 (53.8%) of the 232 CoNS isolates were positive for the mecA gene by PCR. Table 1 shows the correlation between agar dilution MICs and the presence of the mecA gene within the different CoNS species identified. The current breakpoint for oxacillin resistance separated the resistant (mecA-positive) and susceptible (mecA-negative) populations with satisfactory precision. Only three mecA-positive isolates had MICs below the actual breakpoint, with values of 0.25 mg/L. Table 2 shows the correlation between mecA PCR and susceptibility tests performed by agar dilution and Etest with an oxacillin breakpoint of ≥ 0.5 mg/L. The sensitivity for agar dilution was 97.6%, and that for Etest was 100%. The specificity for agar dilution was 100%, and that for Etest was 95.4%. If the oxacillin resistance breakpoint had been lowered to ≥ 0.25 mg/L, the sensitivity for agar dilution would have reached 100%, but the specificity would have decreased to 92.6%. The values that decreased the specificity for the oxacillin Etest were in the range 0.5–6.0 mg/L.

CoNS are the pathogens isolated most commonly from bloodstream infections in intensive care unit patients [3]. Of 32 CoNS species recognised, only half are seen in clinical specimens. Among CoNS, S. epidermidis is generally the most common species [1]. In total, 54% of the isolates studied were resistant to oxacillin, as determined by PCR detection of the mecA gene. This value is lower than in other published reports, which have detected resistance rates of 70–80%[2]. In the present study, the sensitivity and specificity of agar dilution were satisfactory when the current oxacillin MIC breakpoint of ≥ 0.5 mg/L was used. Two S. epidermidis and one S. hominis isolate would have been reported as falsely oxacillin-susceptible, but full correlation was achieved among all the other species.

The findings confirm that, among the species encountered commonly, the NCCLS oxacillin breakpoint of ≥ 0.5 mg/L results in correct classification of CoNS as oxacillin-resistant or -susceptible in most cases.

Etests correlated well with PCR for detection of oxacillin resistance in the CoNS isolates, but six of 102 mecA-negative isolates were classified falsely as resistant. Huang et al.[19] found discrepancies between Etest and agar dilution when testing CoNS. Etest results were skewed towards MICs of oxacillin higher than these given by agar dilution. Both Weller et al.[10] and Frebourg et al.[20] found major discrepancies when comparing Etest with PCR, but the results would have been improved if the resistance breakpoint had been lowered from 2 to 0.5 mg/L.

Six reports have either proposed revised criteria for oxacillin susceptibility [2,8] or confirmed [11,14–16] that the new NCCLS oxacillin breakpoint [17] classifies most of the commonly encountered species of CoNS correctly as oxacillin-susceptible or -resistant, but among the more uncommon CoNS species, the oxacillin agar dilution test can lead to the incorrect classification of mecA-negative isolates as oxacillin-resistant. There are, however, small discrepancies among these studies, which may be caused by differences between the isolates analysed or problems with the difficult CoNS speciations [14]. Overall, it seems that the current NCCLS oxacillin breakpoint recommendation will categorise accurately the CoNS species encountered commonly. However, it is recommended that isolates of CoNS should be identified phenotypically and oxacillin resistance should be confirmed with PCR for isolates from patients with serious infections when the choice of antibiotic is crucial [14,15].