Identification of Methicillin-Resistant or Methicillin-Susceptible Staphylococcus aureus in Blood Cultures and Wound Swabs by GeneXpert

Until a decade ago, clinicians could use epidemiological clues to select empirical therapy for methicillin-susceptible Staphylococcus aureus (MSSA) or methicillin-resistant S. aureus (MRSA) (5). The emergence of MRSA as a community pathogen and the documentation of the inferiority of non-beta-lactam antibiotics in treating MSSA bacteremia greatly complicate initial antibiotic choice (2-4, 7, 8, 9). Early identification and determination of antibiotic susceptibility might help focus initial antibiotic therapy. We compared a multiplex PCR that identifies MSSA and MRSA to standard microbiologic techniques for evaluating the results of blood cultures (BCs) and wound swab (WS) cultures.

Blood was cultured in BacT/Alert, and drug susceptibility was determined with a Vitek 2 system (both from BioMerieux, Durham, NC). For BCs judged to contain gram-positive cocci in clusters (GPCCl), 1-ml aliquots were centrifuged (2 min at 3,000 rpm) to remove charcoal, and the supernatant was studied in a GeneXpert system (Cepheid, Sunnyvale, CA). WS samples were streaked to standard media (blood, chocolate, McConkey, and colistin-nalidixic acid) and then studied in the GeneXpert system within 48 h of collection. GeneXpert real-time PCR detects proprietary sequences of the S. aureus protein A gene, the staphylococcal cassette chromosome, and the methicillin resistance element (1).

Of 223 blood samples, 68 yielded S. aureus by culture, 47 with MRSA and 21 with MSSA. PCR correctly identified 67/68 (98.5%) S. aureus isolates (Tables 1 and 2), including 46/47 (97.9%) MRSA and 21/21 (100%) MSSA isolates. No BC (155/155; 100%) that contained GPCCl without S. aureus contained S. aureus by PCR.

TABLE 1.

Detection of organisms by culture or PCR

Sample	Organism identified by standard culture (no. of samples)	No. of samples in which indicated organism detected by PCR
		MRSA	MSSA	No S. aureus
BC	MRSA (47)	46	1	0
	MSSA (21)	0	21	0
	No S. aureus (155)	0	0	155
WS	MRSA (106)	104	2	0
	MSSA (51)	3	47	1
	No S. aureus (164)	13	18	133
	No S. aureus after adjustment for prior antibiotics and no growth (141)a	4	4	133

^aSee text.

TABLE 2.

PCR sensitivity, specificity, PVs, and effect of concomitant antibiotic use

Sample, organism detected	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)
BC
MRSA	97.9	100	100	99.4
MSSA	100	99.5	95.5	100
No S. aureus	100	100	100	100
WS before adjustment for prior antibiotics and no growth by culturea
MRSA	98.1	92.6	86.7	99.0
MSSA	92.2	92.6	70.1	98.4
Wound swab after adjustment for prior antibiotics and no growth by culturea
MRSA	98.1	96.4	93.7	98.9
MSSA	92.2	97.6	88.7	98.4

^aSee text.

Of 321 WS samples, 106 yielded MRSA and 51 MSSA by culture. PCR identified 104/106 (98.1%) MRSA isolates correctly but misidentified 2 MRSA isolates as MSSA (Table 1). Of 51 MSSA isolates, 47 (92.2%) were identified correctly, 3 incorrectly as MRSA, and 1 incorrectly as no S. aureus by PCR. PCR was positive for S. aureus in 31 of 164 samples that did not contain S. aureus by culture; however, 20 of these 31 were from patients who had received antistaphylococcal antibiotics, and 3 yielded no growth, suggesting inadequate sampling (Tables 1 and 2).

The results of this study show that, once GPCCl have been identified in BCs, PCR technology has high sensitivity, specificity, positive predictive value (PPV), and negative PV (NPV) for correctly identifying MRSA and MSSA. In some cases, appropriate therapy for S. aureus bacteremia could be selected 48 h sooner than with conventional reporting.

For WS samples, however, the results show greater variability. Sensitivity exceeded 98% for MRSA but was lower (92%) for MSSA. The NPV for these two organisms was still high (≥98.4%). Early identification of MRSA would indicate a need for appropriate therapy, and the finding of no S. aureus by PCR strongly suggests the true absence of this pathogen. If culture is regarded as definitive (“gold standard”), PCR provides numerous false positives. Recent studies, however, suggest that S. aureus may go undetected in 15 to 20% of WS samples unless selective media are used (6, 10). We found that many patients with negative culture and positive PCR results were taking antistaphylococcal antibiotics. Thus, it is unclear from the present study whether our finding of S. aureus by PCR and its absence by conventional culture represents a false-positive or a true-positive finding.