High Percentage of Methicillin-Resistant Staphylococcus aureus Isolates with Reduced Susceptibility to Glycopeptides in The Netherlands

Abstract

While testing the in vitro activities of 14 antimicrobial agents against 107 methicillin-susceptible Staphylococcus aureus (MSSA) and 250 methicillin-resistant S. aureus (MRSA) isolates collected in The Netherlands, we found to our surprise that 19 (7.6%) MRSA isolates were suspected of having reduced susceptibilities to the glycopeptides when the Etest system (AB Biodisk, Solna, Sweden) was used with a large inoculum (no. 2 McFarland standard) and an extended incubation time (48 h) on brain heart infusion agar for MIC testing. Eventually, 15 of these isolates were classified as heterogeneously resistant to glycopeptides (heterogeneously glycopeptide-intermediate S. aureus [hGISA] isolates) according to the population analysis profile-area under the curve analysis. The MICs at which 50 and 90% of isolates are inhibited obtained with the Etest system with the large inoculum were as follows: for MSSA isolates, 3.0 and 4.0 μg/ml, respectively, for both teicoplanin and vancomycin; for MRSA isolates, 3.0 and 8.0 μg/ml, respectively, for teicoplanin, and 3.0 and 4.0 μg/ml, respectively, for vancomycin. This is the first report of hGISA isolates in The Netherlands.

The prevalence of resistance to methicillin among Staphylococcus aureus isolates in The Netherlands is low (≤2%) (2, 4). Therefore, most infections with S. aureus can be treated with a semisynthetic penicillin. In many parts of the world, however, the occurrence of methicillin-resistant S. aureus (MRSA) is widespread, and vancomycin is used for empirical therapy for staphylococcal infections (21). However, MRSA isolates with reduced susceptibilities to glycopeptides have now been reported from many countries, including the neighboring countries of Germany and Belgium (3, 8). Until now, no MRSA isolates with reduced susceptibilities to glycopeptides have been reported in The Netherlands.

The emergence of MRSA isolates with reduced susceptibilities to glycopeptides has urged the need to create new agents for the treatment of MRSA infections (20). The ability of clinicians to choose among several antimicrobial agents with different mechanisms of action for the treatment of infections caused by these microorganisms could reduce the increasing selection pressure for resistance to glycopeptides by gram-positive microorganisms in hospitals (26). New antimicrobial agents, like linezolid, have become more and more important as alternative treatments for MRSA infections.

In this study we compared the in vitro activities of linezolid and 13 other antimicrobial agents against 107 methicillin-susceptible S. aureus (MSSA) and 250 MRSA isolates from The Netherlands. Furthermore, we report on the first MRSA isolates with reduced susceptibilities to glycopeptides isolated in The Netherlands.

(This study was presented at the 41st Interscience Conference on Antimicrobial Agents and Chemotherapy, Chicago, Ill., December 2001.).

MATERIALS AND METHODS

Bacterial strains.

The 250 MRSA isolates used in this study were collected in The Netherlands between 1989 and 1998 and are part of the MRSA strain collection of the National Institute of Public Health and the Environment, Bilthoven, The Netherlands. Identification of the isolates as S. aureus and methicillin resistance were determined by multiplex PCR for the mecA gene and the coagulase gene, as described previously (25). Strains were selected on the basis of their different phage types. The 250 MRSA isolates included in the study comprised 247 different phage types. Three isolates were not typeable.

The 107 MSSA isolates were from cultures of blood collected between May 1998 and June 1999 from consecutive patients at the following six hospitals: St. Elisabeth Hospital and Tweesteden Hospital, Tilburg, The Netherlands; Pasteur Hospital, Oosterhout, The Netherlands; Tweesteden Hospital, Waalwijk, The Netherlands; and St. Ignatius Hospital and Hospital de Baronie, Breda, The Netherlands. Only one isolate was included from each patient per admission period. Isolates were identified by a latex agglutination test (Staphaurex Plus; Murex Diagnostics Ltd., Dartford, England), by the detection of free coagulase by the tube coagulase test with rabbit plasma (14), and by the detection of DNase (DNase agar; Oxoid Ltd., Basingstoke, England). If the results of these tests were discordant, an AccuProbe culture identification test (Gen-Probe; San Diego, Calif.) was performed according to the manufacturer's instructions. The result of the AccuProbe test was considered the “gold standard” for the identification of S. aureus. At the time of collection, the blood culture isolates were classified as methicillin susceptible (oxacillin MIC ≤ 2 μg/ml) by broth microdilution susceptibility testing, performed as described by the National Committee for Clinical Laboratory Standards (NCCLS). Furthermore, no growth was observed by the oxacillin agar screen test, performed as described by the NCCLS (16).

Antimicrobial agents and MIC testing.

The MICs of the following 14 antimicrobial agents were determined: linezolid, oxacillin, vancomycin, teicoplanin, gentamicin, tobramycin, co-trimoxazole, quinupristin-dalfopristin, ciprofloxacin, erythromycin, clindamycin, rifampin, fusidic acid, and mupirocin.

All MICs were determined with the Etest system (AB Biodisk, Solna, Sweden) according to the instructions of the manufacturer. The oxacillin Etest strip was placed onto a Mueller-Hinton agar plate supplemented with 2% NaCl, and the plate was incubated at 35°C for 24 h. For vancomycin and teicoplanin, the Etest strips were placed on brain heart infusion (BHI) agar, and a large inoculum (no. 2 McFarland standard) and an extended incubation time (48 h) were used to be able to detect heterogeneously glycopeptide-resistant S. aureus (heterogeneously glycopeptide-intermediate S. aureus [hGISA]) isolates. All other Etest strips were placed onto Mueller-Hinton agar plates, and the plates were incubated for 18 to 24 h at 35°C.

The MIC of vancomycin was also determined by broth microdilution susceptibility testing, as described by the NCCLS (16).

Isolates were categorized as susceptible or resistant to an antimicrobial agent according to the breakpoints published by the NCCLS (16). The following breakpoints were applied to the antimicrobial agents for which the NCCLS does not provide breakpoints: an isolate was classified as susceptible to linezolid if the MIC was ≤4 μg/ml (6), as susceptible to mupirocin if the MIC was ≤4 μg/ml and as high-level resistant to mupirocin if the MIC was >256 μg/ml (5), and as susceptible to fusidic acid if the MIC was ≤0.5 μg/ml and as resistant to fusidic acid if the MIC was ≥2 μg/ml (19). For determination of reduced susceptibilities to vancomycin and teicoplanin by the Etest system with the large inoculum, the interpretive criteria proposed by the manufacturer and recently published by Walsh et al. (27), not the NCCLS breakpoints, were used. This meant that the MICs of both vancomycin and teicoplanin were ≥8 μg/ml or that the MIC of teicoplanin alone was ≥12 μg/ml.

PAP-AUC ratio.

Modified population analysis profiles (PAPs) were determined as described previously by Walsh et al. (27) and Wootton et al. (28). After 24 h of incubation in Tryptone soy broth (Oxoid), an undiluted culture and dilutions of 1/10⁸ and 1/10⁵ were spiral plated (Don Whitley spiral platers) onto BHI agar (Oxoid) plates containing 0.5, 1, 2, 2.5, 4, and 8 μg of vancomycin per ml. The colonies were counted after 48 h of incubation in air at 35°C. The numbers of CFU per milliliter were plotted against the vancomycin concentration by using Prism software (GraphPad, San Diego, Calif.). The modified PAPs were inspected visually, and the data were analyzed differently from the way in which they were analyzed previously (11). The area under the curve (AUC) was calculated for each strain by using a logarithmic graph of the viable count versus the vancomycin concentration. Every test sample run was accompanied by strains Mu3 (ATCC 700698) and Mu50 (ATCC 700699) and a glycopeptide-sensitive S. aureus strain (ATCC 29213) as controls. A ratio was then calculated by dividing the AUC for the test strain by the AUC for Mu3 (the archetype Japanese hGISA strain).

The PAP-AUC ratio criteria used to determine vancomycin resistance, glycopeptide-intermediate resistance, and heterogeneous glycopeptide-intermediate resistance were determined as described previously and are based on multiple (n = 30) tests with both Mu3 and Mu50: ≤0.90 for glycopeptide sensitivity, 0.90 to 1.3 for heterogeneous glycopeptide-intermediate resistance, and ≥1.3 for glycopeptide-intermediate resistance.

RESULTS

The MIC range, the MIC at which 50% of isolates are inhibited (MIC₅₀), the MIC₉₀, and the percentage of isolates susceptible to the antimicrobial agents tested are presented in Tables 1 and 2 for the MSSA and MRSA isolates, respectively.

TABLE 1.

In vitro activities of linezolid and other antimicrobial agents against 107 MSSA strains isolated in The Netherlands^a

Antibiotic	MIC (mg/liter)			Susceptibility
	Range	50%	90%	% S	% R
Oxacillin	0.125-1.0	0.5	0.5	100	0
Teicoplanin (Etest)b	1.0-4.0	3.0	4.0	100d	0
Vancomycin (Etest)b	1.5-6.0	3.0	4.0
Vancomycin (broth)c	≤0.12-4.0	0.5	1.0	100e	0
Gentamicin	≤0.064-0.38	0.125	0.25	100	0
Tobramycin	0.064-0.38	0.19	0.25	100	0
Co-trimoxazole	0.032-0.064	0.047	0.064	100	0
Q-D	0.038-1.0	0.38	0.5	100	0
Ciprofloxacin	0.125-≥32	0.25	0.5	96.3	0.9
Erythromycin	0.047-≥256	0.25	0.38	96.3	0.9
Clindamycin	0.064-1.5	0.125	0.125	99.1	0
Rifampin	≤0.016-0.064	≤0.016	≤0.016	100	0
Fusidic acid	0.032-12	0.125	0.19	94.4	5.6
Mupirocin	0.094-0.38	0.19	0.38	100	0
Linezolid	0.075-2.0	1.0	1.5	100	0

^aAbbreviations: Q-D, quinupristin-dalfopristin; S, susceptible; R, resistant.

^bEtest system with large inoculum and 48 h of incubation.

^cBroth microdilution susceptibility testing.

^dAccording to Etest criteria.

^eAccording to the NCCLS breakpoint criteria.

TABLE 2.

In vitro activities of linezolid and other antimicrobial agents against 250 MRSA strains isolated in The Netherlands^a

Antibiotic	MIC (mg/liter)			Susceptibility
	Range	50%	90%	% S	% R
Oxacillin	0.019-≥256	128	≥256	0	100b
Teicoplanin (Etest)c	0.25-96	3.0	8.0	92.4	7.6e
Vancomycin (Etest)c	1.0-16	3.0	4.0
Vancomycin (broth)d	≤0.12-4	0.5	1.0	100f	0
Gentamicin	0.023-≥256	16	≥256	43.6	48.0
Tobramycin	0.032-≥256	16	≥256	36.4	48.8
Co-trimoxazole	0.032-≥32	0.125	12	83.6	16.4
Q-D	0.019-1.0	0.38	0.75	100	0
Ciprofloxacin	0.125-≥32	16	≥32	42	54
Erythromycin	0.023-≥256	≥256	≥256	32.4	57.6
Clindamycin	0.016-≥256	0.19	≥256	64	32.8
Rifampin	0.094-≥256	0.016	≥256	67.2	23.2
Fusidic acid	≤0.016-≥256	0.125	0.5	90.8	7.2
Mupirocin	≤0.016-≥1024	0.19	0.25	98.8	1.2
Linezolid	0.19-2.0	1.0	1.5	100	0

^aAbbreviations: Q-D, quinupristin-dalfopristin; S, susceptible; R, resistant.

^bAll isolates were known to possess the mecA gene.

^cEtest system with a large high inoculum and 48 h of incubation.

^dBroth microdilution susceptibility testing.

^eAccording to Etest criteria.

^fAccording to the NCCLS breakpoint criteria.

All S. aureus isolates showed susceptibility to linezolid and quinupristin-dalfopristin in vitro.

Only eight of the MSSA isolates showed resistance to one of the antimicrobial agents tested. One isolate was resistant to ciprofloxacin (MIC = 32 μg/ml), one isolate was resistant to erythromycin (MIC = 256 μg/ml), and six isolates were resistant to fusidic acid (Table 1).

Only three (1.2%) MRSA isolates showed resistance to the topical agent mupirocin, and only one of these isolates had high-level resistance (MIC > 256 μg/ml). A low percentage (7.2%) of resistance to fusidic acid was found.

According to the Etest criteria, 19 (7.6%) of the MRSA isolates, for which vancomycin and teicoplanin MICs were ≥8 μg/ml or teicoplanin MICs were ≥12 μg/ml, were suspected of having reduced susceptibilities to the glycopeptides (Table 3). For 15 of these isolates the PAP-AUC ratio was between 0.9 and 1.3, which is the criterion for the heteroresistant phenotype (hGISA). For four isolates, the PAP-AUC ratio did not confirm the reduced susceptibilities to the glycopeptides found by the Etest. For one isolate (isolate 15 in Table 3), the PAP-AUC ratio was close to 0.90, namely, 0.88.

TABLE 3.

Results for 19 MRSA isolates found to have reduced susceptibilities to glycopeptides by Etest criteria

Isolate	Yr of isolation	Country of origin	Isolate source	MIC (μg/ml) with Etest system		Vancomycin MIC (μg/ml) by broth microdilution	PAP-AUC ratioa	Interpretation for hGISAb
				Vancomycin	Teicoplanin
1	1990	Turkey	—c	6	12	1	1.09	+
2	1991	France	—	8	12	1	1.12	+
3	1992	Turkey	—	8	8	1	0.99	+
4	1992	Turkey	—	4	24	1	1.03	+
5	1994	India	Perineum	6	12	0.5	0.82	−
6	1994	France	Burn wound	8	16	1	1.16	+
7	1994	Italy	Pus	4	12	2	1.05	+
8	1994	—	Nares	8	16	2	1.07	+
9	1994	South Africa	Nares	4	24	2	0.84	−
10	1994	Greece	Skin	3	12	1	1.13	+
11	1994	Italy	Wound	4	12	1	1.04	+
12	1994	France	Pus	3	12	0.5	0.78	−
13	1995	Germany	Perineum	8	12	2	1.28	+
14	1995	Greece	Pus	6	12	1	1.06	+
15	1998	Argentina	Wound	2	12	1	0.88	−
16	1998	Ivory Coast	Perineum	12	16	4	1.23	+
17	1998	—	Pus	6	12	0.5	1.00	+
18	1998	Italy	Nares	16	96	2	0.97	+
19	1998	—	—	12	16	1	0.93	+

^aPAP-AUC ratio criteria for vancomycin resistance are defined in the text.

^bInterpretation of PAP-AUC ratio: +, hGISA; −, glycopeptide-sensitive S. aureus.

^c—, unknown.

None of the hGISA isolates would have been detected directly by the broth microdilution method. The vancomycin MIC was 4 μg/ml for only one isolate (isolate 16) by broth microdilution testing. This isolate should have been further investigated according to NCCLS and Centers for Disease Control and Prevention guidelines, according to which strains for which the vancomycin MIC is ≥4 μg/ml should be assessed more closely (27).

While the MIC₅₀ and MIC₉₀ of vancomycin were 3.0 and 4.0 μg/ml, respectively, for both MSSA and MRSA isolates by the Etest, the MIC₅₀ and MIC₉₀ of vancomycin were 0.5 and 1.0 μg/ml, respectively, for both MSSA and MRSA isolates by broth microdilution (Tables 1 and 2).

The MIC₅₀s and MIC₉₀s of linezolid for MSSA and MRSA isolates were equal. However, the MIC₉₀s of both quinupristin-dalfopristin and teicoplanin were higher for MRSA isolates than for MSSA isolates.

DISCUSSION

The purpose of the present study was to assess the in vitro activities of linezolid and 13 other antimicrobial agents against MSSA and MRSA isolates collected in The Netherlands. To our surprise we found that 15 MRSA isolates were heterogeneously resistant to glycopeptides. Staphylococci displaying this phenotype have been reported from various locations over the world, including Japan, the United States, and European countries (27), but this is the first report of S. aureus isolates with reduced susceptibilities to glycopeptides from The Netherlands. The collection of MRSA isolates used in this study is composed of isolates collected in The Netherlands between 1989 and 1998. During that period, as at present, the prevalence of MRSA isolates in The Netherlands was very low, and the MRSA strains isolated were often recovered from patients who had been hospitalized in other countries. Most isolates in the collection are of European origin (25).

The country of origin was recorded for 12 of the 15 isolates confirmed to be hGISA, and none of these isolates originated from The Netherlands: 3 isolates were from Turkey, 3 were from Italy, 2 were from France, 2 were from Greece, 1 was from Germany, and 1 was from Ivory Coast (Table 3). hGISA isolates have been reported from Greece, France, Germany, and Italy before (8, 12, 15, 17). To our knowledge, there are no reports of hGISA isolates from Turkey or Ivory Coast. The prevalence of MRSA isolates in a Turkish hospital, collected as part of the European SENTRY Antimicrobial Surveillance Program between January 1997 and December 1999, was 37.5% (4). In 1996 and 1997 the prevalence of MRSA in Ivory Coast was 16.8% (13). With this high prevalence of MRSA isolates, one can predict that the rate of glycopeptide usage is high and there is a strong chance that one would eventually find hGISA isolates.

In earlier surveillance susceptibility testing studies with part of the MRSA isolate collection used in the present study, namely, those isolates collected between 1989 and 1995, no resistance to vancomycin was described (2). In that study, however, susceptibility testing was performed by an agar dilution method according to the guidelines of the Dutch Committee on Antibiotic Susceptibility Guidelines with an inoculum of 5 × 10³ CFU/spot incubated on Iso-Sensitest agar at 37°C for 24 h.

When the present study was designed, there was very little literature about the value of the Etest system for the detection of hGISA. The manufacturer, however, recommended the use of BHI agar, a larger inoculum, and a longer incubation period. Since we did not have any experience with this method and broth microdilution was the routine method for susceptibility testing in our laboratory, we tested all isolates for vancomycin resistance by the broth microdilution method as well. None of the isolates would have been correctly classified as (h)GISA by the broth microdilution method. For one isolate (isolate 16 in Table 3), the vancomycin MIC was 4 μg/ml by broth microdilution, and the isolate should have been assessed more closely according to NCCLS and Centers for Disease Control and Prevention guidelines (27). The PAP-AUC ratio for this isolate (1.23) was the second highest detected among the isolates tested.

Since this study was designed, more and more has been published about screening methods for the detection of S. aureus isolates heterogeneously resistant to vancomycin. Although well-standardized microdilution susceptibility testing methods are able to detect S. aureus isolates with reduced susceptibilities to vancomycin, they cannot detect heteroresistance (22). Wootton et al. (28) found that the Etest with a large inoculum (no. 2 McFarland standard), a longer incubation time (48 h), and rich BHI medium was the only method that correctly identified Mu3, the heteroresistant isolate that was described by Hiramatsu et al. (11). In a study comparing different methods for the detection of staphylococci with reduced susceptibilities to glycopeptides, Walsh et al. (27) found the Etest method with a large inoculum to be a reliable and sensitive screening method for the detection of glycopeptide resistance, including heteroresistance. In the same study, the PAP-AUC ratio proved to be a reliable method for confirmation of the Etest results. In the present study, 4 of the 250 MRSA isolates (1.6%) were incorrectly identified as having reduced susceptibilities to the glycopeptides when the Etest method with the large inoculum was used. This is comparable to the 2.1% (7 of 329 MRSA isolates) rate of false-positive results described by Walsh et al. (27).

No S. aureus isolates resistant to linezolid were detected in the present study, and as in earlier studies (1), no differences in the in vitro activities of linezolid against MSSA and MRSA isolates were found. The development of resistance to linezolid during treatment has been described in vancomycin-resistant Enterococcus faecium isolates (10). Recently, the first MRSA isolate resistant to linezolid was described (23). This isolate was recovered from a patient after 4 weeks of treatment with linezolid for dialysis-associated peritonitis; the linezolid-resistant MRSA isolates were, however, not related to the linezolid-susceptible isolates initially isolated.

While linezolid is bacteriostatic for staphylococci (6), quinupristin-dalfopristin is bactericidal. However, the bactericidal activity of quinupristin-dalfopristin is variable, and S. aureus strains that have cross-resistance to macrolides, lincosamides, and streptogramin B antibiotics were not killed by quinupristin-dalfopristin in vitro (7). Clindamycin susceptibility was found to be a good surrogate indicator of the bactericidal activity of quinupristin-dalfopristin in vitro (7). Clindamycin resistance, however, is not uncommon among MRSA isolates. In the present study, 32.8% of the MRSA isolates showed resistance to clindamycin, but even that rate is low, considering the rate of resistance to clindamycin among MRSA isolates recovered in European centers participating in the SENTRY program between 1997 and 1999, among which the rate of resistance was as high as 73.5% (4).

Interpretation of the activities of antimicrobial agents in vitro remains a difficult task. The choice of a susceptibility testing method can influence the MIC. The linezolid MICs determined by the Etest tended to be lower than those determined by broth microdilution (6). Tubau et al. (24) found that the linezolid MICs were usually 1 to 2 dilutions lower by the Etest method than by the microdilution method. In a European multicenter study, Gemmell et al. (9) found that the mean MIC of linezolid by the Etest was approximately twofold lower than the mean MIC by the broth microdilution method. It was hypothesized that this effect is due to the bacteriostatic activity rather than the bactericidal activity of linezolid. This difference between the MICs obtained by the Etest and those obtained by the broth microdilution method was not found for vancomycin. In that study, however, the tests with vancomycin were conducted with Mueller-Hinton agar, while we used rich BHI medium. Another study (18) showed that use of the Etest method with a large inoculum resulted in only a minimal difference in the vancomycin and teicoplanin MICs for susceptible American Type Culture Collection control strains compared to the MICs obtained by the agar dilution method, while the vancomycin MICs for the hGISA control strains were 1 dilution lower by the agar dilution test than by the Etest. In our study, however, the vancomycin MICs for both MSSA and MRSA isolates obtained by the Etest method with a large inoculum were increased compared to those obtained by the broth microdilution method. We did not find any MSSA isolate suspect for reduced susceptibility to glycopeptides by the macrodilution Etest method. Until now, only one MSSA isolate with reduced susceptibility to glycopeptides has been reported (18). We believe that the rich BHI medium, the large inoculum, and the longer incubation period used for the Etest method account for the higher vancomycin MIC₅₀s and MIC₉₀s that we found with the Etest system compared to those obtained by broth microdilution testing. This could also explain the higher vancomycin and teicoplanin MIC₅₀s and MIC₉₀s that we found compared to those detected in other studies, which used other susceptibility testing methods. Although use of broth microdilution techniques would have made it easier to compare our results with those of other studies, this would have meant that we would have missed the isolates with reduced susceptibilities to the glycopeptides.

Modification of the medium on which susceptibility testing is performed in order to increase the level of expression of a resistance mechanism, which makes resistant isolates easier to detect, is well accepted. For instance, for the detection of MRSA, Mueller-Hinton agar supplemented with NaCl is used to increase the chance of detecting oxacillin resistance (25).

Low levels of resistance to the agents mupirocin and fusidic acid, which are used for the eradication of MRSA carriage, were found. Only one isolate with high-level resistance to mupirocin was detected. Low-level mupirocin resistance, which was found for two other isolates, has little influence on the efficacy of treatment with topical mupirocin, while in the case of high-level resistance, it is unlikely that topical mupirocin treatment will eradicate the organism (5).

The relatively small proportion of MRSA isolates resistant to co-trimoxazole (16.4%) found in this study is comparable to the proportion of co-trimoxazole-resistant MRSA isolates reported in hospitals in Europe (23%), Canada (16%), and the United States (26%) that participated in the SENTRY program from 1997 to 1999 (4). The rate of co-trimoxazole resistance among MRSA isolates from the Latin American SENTRY program participants during the same period was much higher (65.4%).

The in vitro data presented in this and other studies, together with promising clinical results, show that linezolid is a valuable new addition to the arsenal of antimicrobial agents for the treatment of infections caused by gram-positive microorganisms. Even infections caused by resistant S. aureus isolates, like MRSA or (h)GISA, could possibly be treated successfully with linezolid.