Disk with High Oxacillin Content Discriminates between Methicillin-Resistant and Borderline Methicillin-Susceptible Staphylococcus aureus Strains in Disk Diffusion Assays Using a Low Salt Concentration

Abstract

A separation between mecA⁺ strains of Staphylococcus aureus and strains lacking mecA was achieved by the disk diffusion assay and the agar dilution method, utilizing disks containing 5 μg of oxacillin and inocula of approximately 5 × 10⁵ CFU/spot, respectively, provided that agar with 0 to 0.5% NaCl and incubation at 30°C were employed. The 5-μg oxacillin disks clearly discriminated between borderline methicillin-susceptible and mecA⁺ strains. The oxacillin MICs were more affected by the inoculum density and salt concentration than were the methicillin MICs, and oxacillin MICs of 4 to 16 μg/ml were obtained for strains lacking mecA. Significantly higher levels of β-lactamase production and reduced oxacillin susceptibilities were recorded for strains lacking mecA, in particular strains of phage group V, when agar with ≥2% NaCl was used than when agar with 0 to 0.5% NaCl was employed. The results indicate that the borderline methicillin-susceptible phenotype is a salt-dependent in vitro phenomenon of questionable clinical relevance.

Strains of Staphylococcus aureus with reduced susceptibility to penicillinase-resistant penicillins are categorized as follows: (i) methicillin-resistant S. aureus (MRSA), which produce the low-affinity penicillin binding protein (PBP) 2a, encoded by the mecA gene (25); (ii) strains with modified PBPs (29) due to altered penicillin binding capacity as a result of point mutations or due to hyperproduction of PBPs (8, 10); and (iii) borderline methicillin-susceptible S. aureus (BSSA), also referred to as borderline oxacillin-resistant S. aureus, generally considered to be due to hyperproduction of type A β-lactamase by strains of phage group V, the 94-96 complex, that harbor the pBW15 plasmid (17, 18).

A routine method for the detection of strains with reduced susceptibility should be quick and easy to perform with good precision and accuracy. In order to improve identification of MRSA strains, several parameters, such as the preparation and size of the inocula, incubation time and temperature, and medium type, brand, and composition have been evaluated (2, 6, 9, 16, 28). The disk diffusion method is often recommended, but an international consensus has not been reached. The use of disks containing 1 μg of oxacillin for identification of oxacillin-resistant strains and of amoxicillin-clavulanic acid disks for discrimination between MRSA and BSSA strains is recommended by the National Committee for Clinical Laboratory Standards and the Swedish Reference Group for Antibiotics (20, 26), while the use of disks with 5 μg of oxacillin is recommended in France (5). Additionally, a consensus has not been reached with respect to the incubation temperature and the composition of the media. Reports have shown that neither the disk diffusion method nor MIC determination is able to identify heterogeneous or salt-intolerant mecA⁺ strains as MRSA or to discriminate between BSSA and MRSA strains (7, 9, 11, 14, 19).

The aim of the present study was to evaluate the impact of NaCl on (i) oxacillin and methicillin MICs; (ii) the ability of disks with 5 μg of oxacillin, versus those with amoxicillin-clavulanic acid, to discriminate between MRSA and BSSA strains; and (iii) the production and release of β-lactamase in the disk diffusion method.

(This study was presented in part at the 8th International Symposium on Staphylococci and Staphylococcal Infections, Aix-les-Bains, France, 23 to 26 June 1996.)

Organisms and test conditions.

In the present study, 107 S. aureus strains identified by established methods (13) were used. The presence of the mecA gene was determined by PCR (22) with a Gene Amp PCR System 2400 (Perkin-Elmer). S. aureus ATCC 29213 and ATCC 25923 and one clinical mecA⁺ strain of S. haemolyticus (D38) were used as reference strains. Fifty-six strains were mecA⁺ and categorized as MRSA. Twenty of the strains lacking mecA were categorized as BSSA (methicillin MIC ≥ 4 μg/ml [19]) by using agar with 2% NaCl (17) and incubating at 30°C (23). β-Lactamase was produced by 49 mecA⁺ strains and by 49 strains lacking mecA (21). The strains were characterized by phage typing (3). Inocula from overnight cultures on blood agar plates were suspended in 50 mM phosphate buffer (pH 7.0) to match the turbidity of a McFarland 0.5 standard. Cation-supplemented Mueller-Hinton agar (BBL Microbiology System, Cockeysville, Md.) with 0, 0.5, or 2% (wt/vol) NaCl was used throughout the study. Agar with 3 or 5% NaCl was included in a semiquantitative β-lactamase test, and agar with 5% NaCl was used for MIC determinations. Statistical analyses included Fisher’s exact test, Student’s t test, and McNemar’s test (1).

Susceptibility of and discrimination between MRSA, BSSA, and methicillin-sensitive S. aureus (MSSA).

Oxacillin (Sigma Chemical Co., St. Louis, Mo.) and methicillin (Astra, Södertälje, Sweden) MICs were determined for 36 mecA⁺ strains and 51 strains lacking mecA by the agar dilution method. Incubation at 30°C for 24 h was employed (23). Significantly higher MICs (P < 0.001) were obtained for the strains lacking mecA, in particular the BSSA strains, and for the highly heterogeneous MRSA strains when agar with 2 or 5% NaCl was used than when agar with 0 or 0.5% NaCl was employed (Table 1). However, a separation between mecA⁺ strains and those lacking mecA was not achieved, and the oxacillin MICs obtained for MRSA strains that showed high MICs at 0 to 0.5% NaCl were one to four twofold dilution steps lower when agar with 2 or 5% NaCl was used (P < 0.001). Similar results have been reported elsewhere (9, 19), and it has been proposed that NaCl concentrations above 2.5% not be used due to the salt intolerance of some clones of MRSA (11). Our results support this view, but we would advocate the use of NaCl at a concentration no higher than 0.5% for MIC determinations.

TABLE 1.

Ranges of oxacillin and methicillin MICs obtained for MSSA, BSSA, and MRSA

Antimicrobial agent	% NaCl	MIC range (MIC50) at a concentration ofa:
		5 × 104 CFU/spot			5 × 105 CFU/spot
		MSSA	BSSA	MRSA	MSSA	BSSA	MRSA
Oxacillin	0	0.25–0.5 (0.25)	0.25–2 (0.5)	1–1,024 (16)	0.25–1 (0.5)	0.5–4 (1)	16–1,024 (64)
	0.5	0.25–0.5 (0.25)	0.25–4 (0.5)	2–1,024 (32)	0.25–1 (0.5)	1–8 (1)	16–1,024 (64)
	2	0.25–1 (0.25)	0.5–8 (0.5)	8–512 (64)	0.25–2 (0.5)	0.5–16 (2)	16–1,024 (128)
	5	0.25–1 (0.25)	0.5–8 (2)	4–256 (32)	0.25–2 (0.5)	0.5–16 (4)	8–512 (64)
Methicillin	0	1–2 (2)	2–8 (2)	8–2,048 (16)	2–4 (2)	2–8 (4)	16–2,048 (64)
	0.5	2–4 (2)	2–8 (4)	16–1,024 (16)	2–4 (2)	4–8 (4)	32–1,024 (32)
	2	2b (2)	4–16 (4)	16–2,048 (32)	2–8 (2)	4–16 (4)	32–2,048 (64)
	5	2–4 (2)	2–16 (4)	16–1,024 (64)	2–16 (2)	4–16 (8)	32–1,024 (128)

^aMIC₅₀, MIC at which 50% of the isolates were inhibited. 

^bStrains for which MICs were >2 μg/ml were categorized as BSSA. 

A separation between mecA⁺ strains and those lacking mecA was achieved when a larger inoculum (approximately 5 × 10⁵ CFU/spot) was used instead of the recommended 5 × 10⁴ CFU/spot (20) (Table 1). The oxacillin MICs were more affected by the inoculum size than were the methicillin MICs, in particular for the heterogeneous MRSA. A few BSSA strains showed oxacillin MICs of 4 to 16 μg/ml. Inoculum dependence of oxacillin MICs and reduced specificity upon agar screening have also been reported by others (7). These results may be explained by the lower degree of stability of oxacillin to β-lactamases than that of methicillin (17, 24). In the present study, the methicillin MICs classified all mecA⁺ strains as resistant (≥16 μg/ml) and all strains lacking mecA as susceptible (≤8 μg/ml) (20, 26) when the larger inoculum was used. Classification of the strains lacking mecA as BSSA or MSSA varied under different growth conditions.

Incubation at 30 and 35°C for 24 h was employed in the disk diffusion method. Transmitted light was used for inspection of zones for minute or single colonies and for recording of zone diameters. The strains lacking mecA (BSSA and MSSA) could not be clearly separated from mecA⁺ strains when disks with 1 μg of oxacillin or with 20 μg of amoxicillin plus 10 μg of clavulanic acid (AB Biodisk, Solna, Sweden) were used (Table 2). After incubation at 35°C, 33 and 61% of the strains lacking mecA showed no zone of inhibition on agar with 0 and 2% NaCl, respectively (data not shown). Poor performances by these disks have been reported (9, 14), and we earlier recommended the use of disks containing 5 μg of oxacillin in combination with PDM agar (Biodisk) for S. aureus and novobiocin-resistant coagulase-negative staphylococci (22). However, the disks containing 1 μg of oxacillin proved to be the most suitable for identification of methicillin-resistant, novobiocin-sensitive, coagulase-negative staphylococci (22). Disks with a high content of oxacillin have been recommended previously (5, 16, 27), although evaluations have produced less favorable results (2). The poor performance in these studies could be due to the fact that species-specific breakpoints (22) were not applied; i.e., S. aureus and coagulase-negative strains were not separated (2, 16), and detection of the mecA gene was not feasible at that time. In the present study, a clear separation between the mecA⁺ strains and the strains lacking mecA was obtained with the disks containing 5 μg of oxacillin when agar with 0 to 0.5% NaCl and incubation at 30°C were employed (Table 2). The zone histograms for the strains lacking mecA showed a shift toward smaller zones, indicating enhanced growth at 35°C compared to 30°C. The opposite pattern was observed for the mecA⁺ strains, yielding a low degree of discrimination between strains with and without the mecA gene. These results were in accordance with those reported by French and coworkers, who advocated incubation at 30°C (6).

TABLE 2.

Distribution of zone diameters obtained with strains lacking mecA and with mecA⁺ strains under different conditions, using disks containing 5 μg of oxacillin or 20 μg of amoxicillin with 10 μg of clavulanic acid^a

^aStrains were grown at 30 or 35°C in medium containing 0, 0.5, or 2% NaCl. MS, strains lacking mecA gene; MR, mecA⁺ strains. Interpretive breakpoints, indicated by the horizontal lines, are proposed for the 5-μg oxacillin disk and adopted from those given by Liu et al. for the coamoxicillin disk (14). 

The influence of salt concentration on β-lactamase production.

The BSSA phenotype has been correlated with extensive production of type A β-lactamase among strains of phage group V, the 94-96 complex (17, 18). Based on the 1974 results of Kim and Chipley, who showed that the release of staphylococcal β-lactamase increases in an environment containing 5 to 10% NaCl (12), a semiquantitative method designed to reflect the milieu in the disk diffusion method was applied to ascertain whether the salt concentration could affect the production and release of β-lactamase during susceptibility testing. Duplicate 1-μl drops (approximately 10⁵ CFU/spot) were inoculated onto agar supplemented with 0.5 μg of methicillin/ml as a β-lactamase inducer (1.0 μg/ml was used for strains with methicillin MICs of ≥4 μg/ml). After 24 h at 30°C, the agar plates were flooded with nitrocefin (BBL Microbiology Systems; 200 μg/ml). After 40 min ± 15 s at room temperature, the zone diameters of hydrolyzed nitrocefin surrounding the colony spots were recorded with an accuracy of 0.1 mm. The total intra- and interassay variation, 0.6532 mm, was calculated from results obtained by testing six strains on five different occasions with two or four tests of each strain per agar plate (1).

At 0.5% NaCl, all strains produced similar amounts of β-lactamase; there was no deviating strain or group of strains. These results were used as reference values. At 2 to 5% NaCl, a correlation between the salt concentration and the β-lactamase production was evident (Fig. 1). Four patterns were identified, corresponding to zones obtained for four reference strains: type A β-lactamase, NCTC 9789; type B, 22260; type C, Sal77; and type D, FAR10 (D. Kernodle, Nashville, Tenn.). The type A pattern, produced by 27 MRSA, 13 BSSA, and 14 MSSA strains, was characterized by a large zone (≥12.0 mm) when agar with ≥2% NaCl was used. Strains with the type B pattern (11 MRSA, 4 BSSA, and 6 MSSA) showed unchanged or reduced zone sizes at the higher salt concentrations compared to 0.5% NaCl. Strains with the type C pattern (seven MRSA, two BSSA, and five MSSA) showed zone sizes intermediate between those of the type A and type B patterns. The type D pattern (four MRSA and five MSSA) differed from the other three patterns in that the zone size of the hydrolyzed nitrocefin showed small changes between 0.5, 2, and 3% NaCl but increased substantially at 5%. The type A, B, and C patterns differed significantly (P < 0.001) at all salt concentrations, whereas the type D occasionally differed from that of type A or B. The increased β-lactamase production by the type A strains was reflected in significantly higher MICs and smaller zones than those of other strains at high salt concentrations but not at 0 to 0.5% NaCl. The type A β-lactamase was mainly produced by strains of phage group V, which was in accordance with results published by McMurray and coworkers (18).

FIG. 1.

Ranges of changes in β-lactamase production by 31 strains of S. aureus lacking mecA grown in the presence of 2, 3, and 5% NaCl versus production in medium with 0.5% NaCl. Based on changes in β-lactamase production, the strains were classified as type A (——), type B (–...–), type C (.....), and type D (–––) β-lactamase producers.

In conclusion, the agar dilution method performed according to recommendations provides only an approximation of the degree of susceptibility and does not reveal MRSA strains. An inoculum of ≥5 × 10⁵ CFU/spot has to be used for screening for MRSA, and a method of verifying the presence or absence of the mecA gene in strains that grow at oxacillin concentrations of ≥4 μg/ml is necessary. The amoxicillin-clavulanic acid disk should be replaced by the disk containing 5 μg of oxacillin, and agar with a low salt concentration and incubation at 30°C should be employed for the identification of MRSA by the disk diffusion method. Since the BSSA phenotype is related to the salt concentration, we propose that it does not constitute a true therapeutic problem but rather is a laboratory phenomenon; this is supported by the fact that BSSA infections can be treated with the same drugs as are used for treatment of MSSA infections (4, 15).