Analysis of Borderline Oxacillin-Resistant Staphylococcus aureus (BORSA) Strains Isolated in Tunisia

Abstract

Twenty-three strains of Staphylococcus aureus with borderline resistance to oxacillin were studied. These strains were not detected by the cefoxitin test, tests for penicillin-binding protein 2a (PBP2a), mecA, and mecA_LGA251 were negative, and the strains were genetically unrelated. To detect all strains resistant to oxacillin, laboratories should routinely test for both cefoxitin and oxacillin.

TEXT

Resistance to methicillin in Staphylococcus aureus is commonly mediated by penicillin-binding protein 2a (PBP2a), a low-affinity penicillin-binding protein encoded by the chromosomal mecA gene (4). In 2011, Garcia-Alvarez et al. described a novel mecA gene homologue, called mecA_LGA251, associated with resistance to β-lactam antibiotics. This gene has been present in clinical strains of methicillin-resistant S. aureus (MRSA) which have been isolated in the United Kingdom and in Denmark (11). The intrinsic resistance, mecA mediated, may be either homogenous, which is easily detectable, or heterogeneous (4). Oxacillin disc diffusion has been the traditional method for methicillin resistance screening. However, this test often fails to detect heterogeneous MRSA populations. Since 2001, the 30-μg cefoxitin disc test has proven to be more efficient in predicting methicillin resistance (10, 24, 25). This test has been recommended by different committees, such as the Clinical and Laboratory Standards Institute (CLSI) (6) and the Comité de l'Antibiogramme de la Société Française de Microbiologie (CA-SFM) (7), for the purpose of predicting mecA-mediated resistance in Staphylococcus spp. Detection of the PBP2a or the mecA gene are the reference methods for methicillin susceptibility testing, but these assays are not feasible in most clinical laboratories (4). Resistance to methicillin may be extrinsic, non-mecA mediated, in S. aureus strains with low-level resistance to oxacillin, known as borderline oxacillin-resistant Staphylococcus aureus (BORSA) (16–19). Typically, this borderline phenotype results from excess production of β-lactamase. It was described initially by McDougal and Thornsberry in 1986 (17). According to these authors, these strains were neither heteroresistant nor multiresistant, and they produced large amounts of normal staphylococcal β-lactamase which partially hydrolyze oxacillin and became fully susceptible to oxacillin in the presence of β-lactamase inhibitors (17). However, the borderline phenotype has been attributed to other mechanisms, i.e., the production of an inducible, plasmid-mediated methicillinase or different modifications in the PBP genes due to spontaneous amino acid substitutions in the transpeptidase domain (19, 21).

For the present report, we analyzed 23 strains of S. aureus with reduced susceptibility to oxacillin that were isolated in Sfax University Hospital (Tunisia).

From 2006 to 2011, 1,895 clinical strains of S. aureus were recovered in Sfax University Hospital. Among these strains, 415 (21.9%) were MRSA and 23 (1.2%) had reduced susceptibility to oxacillin. These 23 strains were obtained from various clinical specimens: 14 were from pus, 4 from blood cultures, 3 from tracheal aspirates, 1 from a urine sample, and 1 from an ear. Fourteen strains were isolated from patients in the dermatological unit.

Antibiotic susceptibility testing was performed using the disc diffusion method on Mueller-Hinton agar (Bio-Rad). MICs were determined by the broth microdilution method in unsupplemented Mueller-Hinton medium. Antimicrobial susceptibility results were interpreted according to the CLSI guidelines (6). β-Lactamase production was determined by using nitrocephin discs (Cefinase; bioMérieux). All of the strains were β-lactamase producers and were borderline resistant to oxacillin: their inhibition zone diameters ranged from 10 mm to 13 mm for 1 μg of oxacillin in Mueller-Hinton agar after incubation at 35°C and at 30°C, and their MICs varied from 2 to 4 μg/ml. These strains were susceptible not only to cefoxitin (inhibition zone, ≥28 mm) but also to amoxicillin-clavulanic acid (inhibition zone, ≥25 mm; MICs, <0.125 to 0.25 μg/ml), cefotaxime (inhibition zone, ≥ 25mm; MICs, 0.5 to 2 μg/ml), and imipenem (inhibition zone, ≥42 mm; MICs, <0.125 to 0.125 μg/ml).

PBP2a production detected using the PBP2a latex agglutination test as recommended by the manufacturer (bioMérieux) and detection of the mecA and the mecA_LGA251 genes by PCR (9, 20) were negative for all strains. The lack of the mecA gene confirmed that these strains were not MRSA. The β-lactamase-hyperproducing BORSA phenotype was suggested. To confirm this hypothesis, the activity of clavulanic acid in combination with oxacillin was tested, as previously demonstrated by others (5, 22). We found that these strains became fully susceptible to oxacillin in the presence of β-lactamase inhibitor. In other words, for all of the strains, significant increases (≥5 mm) of the inhibition zone diameter for oxacillin were shown after the addition of 4 μg of clavulanic acid, and >2-fold dilution decreases in oxacillin MICs were obtained in the presence of 4 μg/ml of clavulanic acid (Table 1). The oxacillin zone diameter and MICs were unaffected by the clavulanic acid in quality control strains (methicillin-susceptible S. aureus ATCC 25923 and 29213 and MRSA ATCC 43300).

Table 1.

Characteristics of borderline oxacillin-resistant S. aureus strains

Straina	Specimen	Date of isolation (day/mo/yr)	Ward	Inhibition zone diameter (mm)b						MIC (μg/ml)b					Other resistance markersc	erm gened	PFGE typee
				OXA	OXA-CLA	FOX	AMC	CTX	IPM	OXA	OXA-CLA	AMC	CTX	IPM
V288	Eye	01/07/2006	Dermatology	10	17	32	29	28	52	4	0.25	0.125	1	<0.125	KAN, TOB, GEN, ERY, L, PT, TET, OFX, RIF, FUC	ermC	A
V451	Pus	03/10/2006	Dermatology	11	20	32	30	30	44	4	0.25	0.125	1	<0.125	KAN, TOB, GEN, ERY, L, PT, TET, OFX, RIF, FUC	ermC	A
R156	Tracheal aspirate	10/01/2007	Intensive care unit	12	20	32	28	27	46	4	0.25	<0.125	0.5	0.125	ERY, TET	ermC	B
V41	Pus	23/01/2007	Dermatology	13	22	32	26	30	48	2	0.125	<0.125	1	<0.125	ERY, TET	ermC	C
V168	Burn	12/04/2007	Dermatology	12	20	30	28	31	48	2	0.25	0.125	1	<0.125	ERY, TET	ermC	D
T306	Pus	23/06/2007	Thoracic surgery	11	17	32	28	26	50	2	0.25	0.125	1	<0.125	TET	Neg	E
V117	Pus	28/03/2008	Dermatology	13	20	32	28	27	48	2	0.25	0.25	0.5	<0.125	KAN	Neg	F
V277	Blood	09/07/2008	Dermatology	12	22	33	30	28	48	2	0.25	0.25	1	<0.125	ERY, L, TET, OFX, RIF, FUC	ermC	G
L170	Ear	21/08/2008	Maxillofacial	12	21	30	30	30	48	4	0.25	0.25	2	<0.125	KAN, TET	Neg	F
V332	Pus	05/09/2008	Dermatology	12	22	30	36	34	52	2	0.125	0.125	1	<0.125	ERY, L, TET, RIF, FUC	ermC	G
V365	Pus	17/10/2008	Dermatology	12	21	32	32	30	48	4	0.125	<0.125	1	<0.125	ERY, L, TET, OFX, RIF, FUC	ermC	G
C4587	Pus	01/12/2008	Infectious disease	11	19	31	29	29	46	4	0.125	<0.125	1	<0.125	KAN, L, TET	Neg	H
PL1480	Urine	19/12/2008	Internal Medicine	12	20	30	28	29	46	4	0.5	0.125	2	<0.125	KAN, TET	Neg	I
V12	Pus	10/01/2009	Dermatology	12	23	30	40	38	56	4	0.25	0.125	2	0.125	ERY, L, TET, OFX, RIF, FUC	ermC	J
UB15	Blood	25/01/2009	Burn unit	12	21	28	29	30	49	4	0.125	0.125	1	0.125	KAN, TET	Neg	H
UB86	Tracheal aspirate	13/04/2009	Burn unit	13	21	30	28	30	46	2	0.125	0.125	1	0.125	KAN, ERY, TET	ermC	K
V92	Burn	27/04/2009	Dermatology	12	21	30	32	34	50	4	0.125	0.125	1	0.125	ERY, L, TET, OFX, RIF, FUC	ermC	L
UB101	Tracheal aspirate	27/05/2009	Burn unit	11	20	30	26	32	46	4	0.25	0.25	1	<0.125	TET	ermC	M
V20	Pus	04/11/2009	Dermatology	11	20	30	26	25	46	2	0.25	0.125	1	0.125	TET, OFX	Neg	N
V136	Pus	16/01/2010	Dermatology	13	21	30	30	28	46	2	0.25	0.125	2	<0.125	KAN, TOB, GEN, ERY, L, TET, OFX, RIF, FUC, SXT	ermC	O
V206	Pus	04/05/2010	Dermatology	13	20	30	29	26	48	2	<0.125	0.125	2	<0.125	L, TET, OFX, RIF, FUC	Neg	P
V18	Blood	30/06/2010	Dermatology	12	20	32	25	28	50	2	<0.125	0.125	2	<0.125	KAN, TOB, TET	Neg	Q
N721	Blood	28/01/2011	Neonatology	12	20	30	25	25	50	2	<0.125	0.125	2	<0.125	FUC	Neg	R
ATCC 25923				22	23	28	36	31	40
ATCC 29213										0.25	0.25	0.25	1	<0.125
ATCC 43300				12	13					16	8

^aS. aureus ATCC 25923 (MSSA) was used as quality control for disc diffusion testing, S. aureus ATCC 29213 (MSSA) as quality control for MIC testing, and S. aureus ATCC 43300 (MRSA) as quality control for oxacillin disc diffusion and MIC testing.

^bOXA, oxacillin; CLA, clavulanic acid; FOX, cefoxitin; AMC, amoxicillin-clavulanic acid; CTX, cefotaxime; IPM, imipenem.

^cKAN, kanamycin; TOB, tobramycin; GEN, gentamicin; ERY, erythromycin; L, lincomycin; PT, pristinamycin; TET, tetracycline; CHL, chloramphenicol; OFX, ofloxacin; RIF, rifampin; FUC, fucidic acid; SXT, trimethoprim-sulfamethoxazole.

^dNeg, negative.

^ePFGE, pulsed-field gel electrophoresis.

These 23 strains showed different antibiotypes (Table 1). All of these strains were susceptible to fosfomycin, to glycopeptides, and to chloramphenicol. The ermABC genes were tested for by PCR as described previously (1). ermC was amplified from 13 strains which were resistant to erythromycin.

The typing of the strains was performed by pulsed-field gel electrophoresis (PFGE) as described previously (3). Genomic DNA was digested with the restriction enzyme SmaI, and the fragments were separated in agarose gels by electrophoresis according to the manufacturer's recommendations. Image normalization and construction of similarity matrices were carried out using Fingerprinting II software (Bio-Rad). PFGE revealed that BORSA strains were genetically unrelated in our hospital (Fig. 1). Eighteen genotypically different strains were identified, suggesting that the BORSA strains had originated from different ancestors.

Fig 1.

Dendrogram of PFGE patterns of 23 BORSA strains.

Borderline-resistant strains of S. aureus have been reported to be associated with both nosocomial and community-acquired infections in some institutions and have been isolated from various infection sites, including skin, surgical wounds, respiratory samples, abscesses, and blood (2, 12–15, 23). Outbreaks of BORSA infections have been reported in two different dermatological units in Denmark (2, 14). These two outbreaks were caused by two different clones, based on the same typing methods (PFGE and spa typing), whereas, in our study, strains that were isolated in the dermatological unit were not closely related.

The incidence of BORSA strains is uncommon (19). It is certainly underestimated given that many clinical microbiology laboratories use only the cefoxitin test for detection of oxacillin resistance in Staphylococcus spp. Indeed, the cefoxitin test is a marker of resistance to oxacillin by acquisition of the mecA gene, and it is unable to detect BORSA strains (10, 24, 25).

The detection of BORSA strains may influence the choice of antibiotics in the treatment. It has been proposed by some authors that there is no apparent reason for BORSA strains to be considered resistant to all other β-lactams (5, 18, 22). Therefore, infections caused by these strains can probably be safely and effectively treated with β-lactam antibiotics, including cloxacillin, or with the use of a β-lactam–β-lactamase inhibitor combination (16, 26, 27). However, other authors have demonstrated that oxacillin has not been effective against BORSA strains (8, 23). As the treatment of infections caused by BORSA with β-lactams is full of risk, therapy will be conducted according to the MIC of oxacillin and to the severity of the infection. In our study, most of the patients were effectively treated with pristinamycin. The five patients who were treated with β-lactams (cefotaxime or imipenem) also recovered.

In conclusion, clinical laboratories should be in a position to recognize BORSA strains by routine susceptibility testing and, especially, should be able to differentiate them from truly methicillin-resistant or -susceptible S. aureus. Hence, for the detection of BORSA strains, we recommend the use of both the cefoxitin disc test as a marker of resistance to oxacillin by acquisition of the mecA gene and the oxacillin disc test on Mueller-Hinton agar incubated at 30°C or on Mueller-Hinton agar with NaCl.