Performance of a New MicroScan WalkAway PC30 Panel and Disk Diffusion Method for Detection of Oxacillin Resistance in Staphylococcus spp.

Abstract

The performance of the MicroScan WalkAway PC30 panel for detection of oxacillin resistance was evaluated by use of a collection of 420 staphylococcus isolates. The addition of a cefoxitin test (4 mg/liter) to the oxacillin MIC determination increased its raw performance for Staphylococcus aureus; additional data were required for coagulase-negative staphylococci.

TEXT

Detection of oxacillin (OXA) resistance (OR) in staphylococci is a daily challenge for clinical laboratories. Erroneous susceptibility results could lead to therapeutic failure or inadequate antimicrobial selection pressure (5, 18). Cefoxitin (FOX) testing is currently recommended and used for this purpose (2). We evaluated the performance of a new MicroScan WalkAway panel (Siemens, Sacramento, CA) that includes this compound and the disk diffusion method (DDM) for the detection of OR.

(This work was presented in part at the 20th European Congress of Clinical Microbiology and Infectious Diseases [ECCMID], Vienna, Austria, 10 to 13 April 2010.)

A set of 420 nonduplicate isolates of Staphylococcus aureus (n = 370) and coagulase-negative staphylococci (CNS; n = 50) was collected from two nationwide studies through the French College de Bacteriologie-Virologie-Hygiene network, comprising isolates from cases of community-acquired skin and soft tissue infections (203 S. aureus isolates in 2006) and clinically significant bacteremia (167 S. aureus and 50 CNS isolates in 2007) (9, 10). Identification was performed using (i) specific gyrB PCR for S. aureus, (ii) mass spectrometry (Axima Assurance; Shimadzu) for CNS, and (iii) 16S RNA sequencing when the identification differed from that obtained at the participating-laboratory level (3, 4, 11). The detection of the mecA gene was performed as previously described (10). OR was detected by (i) the FOX DDM according to CLSI guidelines, (ii) the moxalactam (MOX) DDM using the French official criteria (OR is shown by a diameter of <23 mm), and (iii) the MicroScan WalkAway PC30 panel (Siemens, Sacramento, CA) (13, 14). This panel contained oxacillin in doubling dilutions from 0.25 to 2 μg/ml and an additional FOX test (4 μg/ml); each strain was categorized as OR if its oxacillin MIC was greater than 2 μg/ml for S. aureus and 0.5 μg/ml for CNS or if the FOX test was positive (i.e., yielded a growth). The panels were run automatically according to the manufacturer's instructions. Quality control strains (oxacillin-susceptible S. aureus ATCC 25923 and OR S. aureus [ORSA] ATCC 43300) were included in each sample set. All results were within the acceptable range provided by the 2008 CLSI guidelines (2). The contribution of the FOX test was statistically evaluated with the Fisher's exact test. A P value of ≤0.05 was considered to reflect significance. All computations were done with R Project software (http://www.r-project.org).

OR was detected by PCR in 22.7% of S. aureus (84/370) and 66% of CNS (33/50) isolates; the latter group included isolates of S. epidermidis (34/50), S. haemolyticus (8/50), S. hominis (5/50), S. capitis (2/50), and S. schleiferi (1/50). Considering PCR as the gold standard, the sensitivity, specificity, positive and negative predictive values, and numbers of very major errors (VME; false susceptible) and major errors (ME; false resistant) for the different tests and combinations are reported in Tables 1 (S. aureus) and 2 (CNS). For S. aureus isolates, the contribution of the FOX test to the PC30 panel performance was statistically nonsignificant (P = 0.367). Like 9 other ORSA isolates, the only member of the sequence type 80 (ST80) Panton-Valentine leucocidin-producing clone isolated from skin infections was not identified by the FOX-plus-MOX DDM combination (B. Lamy and J. W. Decousser, unpublished data).

Table 1.

Comparison of sensitivity, specificity, and positive and negative predictive values, including the 95% confidence intervals, as well as numbers of very major and major errors of the different tests for the 370 S. aureus strains^a

Test and drug(s)	% Sensitivityb (95% CI)f	% Specificityc (95% CI)	% PPV (95% CI)	% NPV (95% CI)	No. of:
					VMEsd	MEse
Disk diffusion
FOX	75 (64.4; 83.8)	100 (98.7; 100)	100 (94.5; 100)	93.1 (89.7; 95.7)	21	0
MOX	88.1 (79.2; 94.1)	99.6 (98.1; 99.9)	98.6 (92.8; 99.9)	96.6 (93.8; 98.4)	10	1
MOX + FOX	88.1 (79.2; 94.1)	99.6 (98.1; 99.9)	98.6 (92.8; 99.9)	96.6 (93.8; 98.4)	10	1
PC30 panel
OXA MIC	95.2 (88.3; 98.7)	100 (98.7; 100)	100 (95.4; 100)	98.6 (96.5; 99.6)	4	0
FOX test	97.6 (91.7; 99.7)	100 (98.7; 100)	100 (95.6; 100)	99.3 (97.5; 99.9)	2	0
OXA MIC + FOX test	98.8 (93.5; 99.9)	100 (98.7; 100)	100 (98.7; 100)	99.6 (98.1; 99.9)	1	0

^aThe different tests for the 370 S. aureus strains included cefoxitin (FOX) and moxalactam (MOX) alone or in combination in the disk diffusion method and the determination of oxacillin MIC (OXA MIC) and the FOX test alone or in combination in the PC30 panel. PPV, positive predictive value; NPV, negative predictive value.

^bPercentage of the 84 mecA-positive isolates for which positive test results were obtained.

^cPercentage of the 286 mecA-negative isolates for which negative test results were obtained.

^dFalse susceptibility.

^eFalse resistance.

^f95% exact binomial confidence intervals are shown.

Table 2.

Comparison of sensitivity, specificity, and positive predictive value and negative predictive value, including the 95% confidence intervals, as well as numbers of very major and major errors of the different tests for the 50 CNS strains^a

Method and drug(s)	% Sensitivityb (95% CI)f	% Specificityc (95% CI)	% PPV (95% CI)	% NPV (95% CI)	No. of:
					VMEsd	MEse
Disk diffusion
FOX	90.9 (71.8; 96.6)	100 (80.5; 100)	100 (88.1; 100)	85 (58.1; 94.6)	3	0
MOX	100 (89.4; 100)	100 (80.5; 100)	100 (88.1; 100)	100 (80.5; 100)	0	0
MOX + FOX	100 (89.4; 100)	100 (80.5; 100)	100 (88.1; 100)	100 (80.5; 100)	0	0
PC30 panel
OXA MIC	100 (89.4; 100)	94.1 (71.3; 99.8)	97.1 (84.7; 99.9)	100 (79.4; 100)	0	1
FOX test	90.9 (75.7; 98.1)	94.1 (71.3; 99.8)	96.8 (83.3; 99.9)	84.2 (60.4; 96.6)	3	1
OXA MIC + FOX test	100 (89.4; 100)	88.2 (63.5; 98.5)	94.2 (80.2; 99.3)	100 (78.2; 100)	0	2

^aThe different tests for the 50 CNS strains included cefoxitin (FOX) and moxalactam (MOX) alone or in combination in the disk diffusion method and the determination of oxacillin MIC (OXA MIC) and the FOX test alone or in combination in the PC30 panel. PPV, positive predictive value; NPV, negative predictive value.

^bPercentage of the 33 mecA-positive isolates for which positive test results were obtained.

^cPercentage of the 17 mecA-negative isolates for which negative test results were obtained.

^dFalse susceptibility.

^eFalse resistance.

^f95% exact binomial confidence intervals are shown.

OR is sometimes difficult to detect, especially when the mecA gene is heterogeneously expressed (5, 17). Oxacillin susceptibility should be tested in a culture medium containing 2% NaCl or incubated at 30°C; this last solution is not adapted to automated antimicrobial-susceptibility testing systems. In 2002, Felten et al. demonstrated that FOX testing using DDM was the best-performing test for routine detection of OR in S. aureus (7). Additional reports confirmed the contribution of the FOX DDM, which was included in the French and U.S. guidelines (1, 2, 13, 22). The breakpoint varied according to the country; in France, the lower breakpoint for the FOX DDM was 24 mm, while it was 21 mm for the CLSI (2, 13). Additionally, an intermediate category of strains exhibiting a zone diameter between 25 and 26 mm was established, requiring a confirmatory test as mecA or PBP2a detection (13). These discrepancies may explain the gap between the sensitivity results in the study of Felten et al. and in this present work (100% versus 75%) (7). In our study, 14 ORSA isolates exhibited a zone diameter between 22 and 24 mm and 6 isolates a zone diameter between 25 and 26 mm. Concerning the CNS, the lack of performance of the FOX DDM was debated early on and has recently been reported as species dependent (8, 12). Therefore, performance reports should be based on accurate species identification and conclusions limited to the tested species. MOX has been shown to have a higher accuracy than FOX for the detection of OR in CNS (13, 19). Indeed, the combination of MOX with FOX increased the performance of DDM to detect OR in CNS compared to the use of FOX alone (100% sensitivity versus 90.9%) (Table 2).

Recently, a MIC value of 4 μg/ml was established for FOX to detect the presence of the mecA gene (21). In the study of Swenson et al., only 0.3% (1/312) of mecA-positive S. aureus isolates showed a FOX MIC inferior to this breakpoint (21). This variability in the level of the expressed resistance was previously reported for methicillin and attributed to a heterogeneous expression of low-affinity penicillin-binding protein (5). This breakpoint was added to the standard guidelines in the United States and Europe and allowed its inclusion in the panels or cards of automated systems (2, 6). As intended, their performances were significantly improved (15, 20, 23). Until recently, FOX was lacking from the Microscan panel, and misclassification of ORSA strains as susceptible was reported (16). The putative contribution of a FOX test was confirmed in our study: the number of VME decreased (4 to 1), and there was a trend of increasing sensitivity (98.8% versus 95.2%) for the combination compared to the sensitivity obtained with the OXA MIC alone, even though statistical significance was not reached (P = 0.367). Interestingly, the only representative strain of the ST80 ORSA clone was detected by the FOX test, although it was miscategorized by the FOX DDM, as previously described (16). The ability of the FOX test to detect this kind of low-level OR should be confirmed in a larger study. Concerning the CNS strains, the contribution of the FOX tests (DDM and panel PC30) was poor, as previously reported (21). However, the number of strains was insufficient to draw any conclusion, the limited power rendering irrelevant the meaning of the not-significant P value. Nevertheless, according to the confidence intervals of the calculated performances, a putative negative impact of this test could not be excluded. These preliminary results should be strengthened by a more powerful study that includes the species associated with the lower levels of performance of OR detection in the study of John et al. (i.e., S. simulans, S. cohnii, and S. saprophyticus) (12). To date, the published levels of performance of FOX MIC determination in CNS agree with our findings (12, 19).

In conclusion, the new PC30 panel is an accurate automated test for OR detection in clinically relevant S. aureus strains but additional data are required for CNS.