|Home | About | Journals | Submit | Contact Us | Français|
Mupirocin susceptibility testing of Staphylococcus aureus has become more important as mupirocin is used more widely to suppress or eliminate S. aureus colonization and prevent subsequent health care- and community-associated infections. The present multicenter study evaluated two susceptibility testing screening methods to detect mupirocin high-level resistance (HLR), broth microdilution (BMD) MICs of ≥512 μg/ml, and a 6-mm zone diameter for a disk diffusion (DD) test with a 200-μg disk. Initial testing indicated that with Clinical and Laboratory Standards Institute methods for BMD and DD testing, the optimal conditions for the detection of mupirocin HLR were 24 h of incubation and reading of the DD zone diameters with transmitted light. Using the presence or absence of mupA as the “gold standard” for HLR, the sensitivity and specificity of a single-well 256 μg/ml BMD test were 97 and 99%, respectively, and those for the 200-μg disk test were 98 and 99%, respectively. Testing with two disks, 200 μg and 5 μg, was evaluated for its ability to distinguish HLR isolates (MICs ≥ 512 μg/ml), low-level-resistant (LLR) isolates (MICs = 8 to 256 μg/ml), and susceptible isolates (MICs ≤ 4 μg/ml). Using no zone with both disks as an indication of HLR and no zone with the 5-μg disk plus any zone with the 200-μg disk as LLR, only 3 of the 340 isolates were misclassified, with 3 susceptible isolates being classified as LLR. Use of standardized MIC or disk tests could enable the detection of emerging high- and low-level mupirocin resistance in S. aureus.
Mupirocin is a topical antibacterial agent that is used both for the treatment of skin infections and for the suppression or elimination of nasal carriage of Staphylococcus aureus, including methicillin-resistant S. aureus (MRSA) (8). The recommendations of the Healthcare Infection Control Practices Advisory Committee suggest the use of a tiered approach to the prevention and control of infections with multidrug-resistant organisms, including MRSA, in acute-care settings (20). In their recommendations, decolonization is presented as one intervention that may be considered when intensified MRSA control measures are needed; if decolonization is used, susceptibility testing and monitoring for the emergence of resistance to the decolonization agent are recommended in one study (21).
There are two levels of resistance to mupirocin: low-level resistance (LLR), for which the MICs are 8 to 256 μg/ml, and high-level resistance (HLR), for which the MICs are ≥512 μg/ml (11). The mupirocin MICs of strains susceptible to mupirocin are MICs ≤4 μg/ml. HLR is associated with the presence of the plasmid-mediated mupA gene, which encodes a mupirocin-resistant isoleucyl-tRNA synthetase, although S. aureus strains with HLR that lack mupA have occurred (this study) and can also be created in the laboratory (23). LLR results from mutation of the native, chromosomal isoleucyl-tRNA synthetase ileS gene (1). Studies suggest that S. aureus strains with HLR to mupirocin cannot be successfully eliminated with mupirocin and that the occurrence of HLR is increasing (22). It has been suggested that S. aureus strains demonstrating LLR could be eliminated by topical application of mupirocin because of the high concentrations achieved locally, but this has not been demonstrated definitively (11, 21).
Until recently, methods for testing topical agents have not been included in susceptibility testing documents published by the Clinical and Laboratory Standards Institute (CLSI; formerly NCCLS), although guidelines for testing by various methods have been suggested by others (9, 10, 12, 13, 16, 17). The British Society for Antimicrobial Chemotherapy has formal recommendations for the testing of mupirocin (www.bsac.org.uk) that include testing of a 5-μg and a 20-μg mupirocin disk. Their recommendations require MIC testing to determine the level of resistance if a 5-μg disk is used alone. An initial investigation at the Centers for Disease Control and Prevention (CDC), Atlanta, GA, showed that a 200-μg mupirocin disk was able to differentiate isolates with LLR from those with HLR (15). We undertook the study described here to determine the MIC and disk diffusion criteria for the detection of S. aureus strains with high- or low-level mupirocin resistance and to validate quality control tests. Using data from this study, a screen test for prediction of high-level mupirocin resistance is now included in CLSI susceptibility testing documents (3, 6, 7).
The study was conducted in three phases using CLSI broth microdilution and disk diffusion methods (4, 5). Phase 1 was an initial study undertaken at the CDC to study the effects of the time of incubation, the medium used (MIC only), and the method of reading the zone diameters (i.e., by the use of reflected versus transmitted light) and to determine preliminary quality control criteria. Following phase 1, two multilaboratory studies were done simultaneously: a three-laboratory study to validate the MIC and zone diameter criteria for prediction of high- and low-level resistance (phase 2) and an eight-laboratory study to establish final quality control ranges (phase 3).
The laboratories participating in phase 2 were CDC; Evanston Hospital, Evanston, IL; and Sunnybrook Health Sciences Centre, Toronto, Ontario, Canada. For phase 3, the laboratories were CDC; the Clinical Microbiology Institute (CMI), Wilsonville, OR; Duke University Medical Center, Durham, NC; Massachusetts General Hospital, Boston, MA; Evanston Hospital; University of Rochester Medical Center, Rochester, NY; UCLA Medical Center, Los Angeles, CA; and the University of Texas Health Sciences Center at San Antonio, San Antonio, TX.
Broth microdilution plates were prepared using the CLSI reference method (4) with cation-adjusted Mueller-Hinton broth (CAMHB) from three manufacturers, Difco (BD Diagnostic Systems, Sparks, MD), BBL (BD), and Oxoid (Oxoid, Ltd., Basingstoke, Hampshire, England), and mupirocin powder (U.S. Pharmacopoeia; www.usp.org). The concentrations tested ranged from 0.015 to 512 μg/ml. The MICs were read after 16 to 18 h and 24 h of incubation by two readers. For disk diffusion, agar from one manufacturer of Mueller-Hinton agar (BD/BBL) and 200-μg disks from two manufacturers (BD/BBL and Oxoid) were used. The zone diameters were read using both reflected and transmitted light after 16 to 18 h and 24 h of incubation by two readers. Staphylococcus aureus ATCC 29213 and Enterococcus faecalis ATCC 29212 were used for quality control for MIC testing, and S. aureus ATCC 25923 was used for quality control for disk diffusion testing. Preliminary quality control ranges were established using the results from this testing and from previous studies (CMI, unpublished data).
Twenty-six S. aureus strains were tested in phase 1: 17 mupA-negative strains (7 mupirocin susceptible, 9 strains with low-level resistance, and 1 strain with laboratory-derived high-level resistance) and 9 mupA-positive strains. All strains in phase 1 except one mupirocin-susceptible strain were MRSA.
Broth microdilution plates were prepared at CDC using CLSI reference methods with CAMHB from three different manufacturers of CAMHB (BD/Difco, BD/BBL, and Oxoid), frozen, and then shipped on blue ice to the participating laboratories, along with both 5- and 200-μg disks (from both BD/BBL and Oxoid). For MIC testing, the mupirocin concentrations tested ranged from 0.12 to 512 μg/ml. Each laboratory used its in-house Mueller-Hinton agar (MHA; CDC used MHA from BD/BBL; Evanston used MHA from Remel, Lenexa KS; and Sunnybrook used MHA from Oxoid) for disk diffusion testing. All readings were done at 24 h only. Zone diameters for mupirocin were read using transmitted light only.
Each laboratory tested at least 200 S. aureus strains from their collections, which were selected to include large percentages of both strains with high-level mupirocin resistance and strains with low-level mupirocin resistance. The 200 isolates tested at the CDC were from the Active Bacterial Core Surveillance (ABC) MRSA surveillance project. All strains were tested for the presence of the mupA gene using PCR. After initial testing, repeat testing of all methods or further molecular characterization was done on any mupA-positive strain with mupirocin zone diameters of >6 mm with a 200-μg disk or mupirocin MICs of <512 μg/ml or any mupA-negative strain with mupirocin zone diameters of 6 mm (i.e., no zone) with a 200-μg disk or mupirocin MICs of ≥512 μg/ml. The final number and the characteristics of the strains tested in phase 2 are shown in Table Table1.1. All mupirocin-LLR strains were MRSA.
For quality control, S. aureus ATCC 29213 and E. faecalis ATCC 29212 were used for MIC testing and S. aureus ATCC 25923 was used for disk diffusion testing. In addition, a previously characterized mupA-positive strain, S. aureus ATCC BAA-1708, was also tested by disk diffusion only.
Quality control testing was performed and the results were analyzed according to CLSI M23-A3 requirements (2). Broth microdilution plates were prepared at CDC using CLSI reference methods and CAMHB from three different manufacturers (BD/Difco, BD/BBL, and Oxoid), frozen, and then shipped to the participating laboratories, along with both 5- and 200-μg mupirocin disks (from both BD/BBL and Oxoid). For MIC testing, the mupirocin concentrations tested ranged from 0.03 to 512 μg/ml. MHA plates from three different manufacturers (BD/BBL; Remel; and Hardy, Santa Maria, CA) were shipped to all laboratories for disk diffusion testing. All readings were done at 24 h only. Zone diameters were read using transmitted light only.
Each laboratory tested S. aureus ATCC 29213 and E. faecalis ATCC 29212 by the MIC method and S. aureus ATCC 25923 by disk diffusion. mupA-positive S. aureus strain ATCC BAA-1708 was tested by disk diffusion only. Each organism was tested 10 times with a separate inoculum prepared for each replicate, with a maximum of four replicates being tested per day.
PCRs for mupA detection were performed using a forward primer with the sequence GGG CCT TAA TTT CGG ATA GTG CTC and a reverse primer with the sequence TAA TCT GGC TGC GGA AGT GAA ATC. The conditions used for amplification were 95°C for 15 min; 35 cycles of denaturation at 94°C for 30 s, annealing at 68°C for 90 s, and extension at 72°C for 90 s; and a final extension at 72°C for 10 min.
For MIC testing, there was no significant effect of the source of CAMHB or the reader (data not shown); therefore, the data analysis results for the three sources of media (MICs only) and the two readers were combined. Table Table22 shows the results as percent correct for MIC and disk diffusion testing using both reflected and transmitted light. Percent correct is defined as MICs of ≥512 μg/ml or a zone diameter of 6 mm for mupA-positive strains and MICs of ≤256 μg/ml or zone diameters of >6 mm for mupA-negative strains. The combination of time and reading conditions that gave the best combination of sensitivity (percent correct for mupA-positive strains) and specificity (percent correct for mupA-negative strains) for the tests was 24 h of incubation and the use of transmitted light for reading of disk diffusion results. The errors in the detection of mupA-positive strains seen at 24 h by MIC testing were for two different strains for which the results of the testing done with Oxoid CAMHB were read by the same reader. Errors in the prediction of mupA-negative strains were with the one strain with laboratory-derived resistance, which was read as having MICs of ≥512 μg/ml in all three CAMHBs tested.
The phase 2 results of screening for HLR are summarized in Table Table33 for both MIC and disk diffusion testing (using the 200-μg disk) in comparison with the results of screening for HLR according to the presence or absence of mupA by PCR. The sensitivity and specificity of MIC testing for the detection of HLR in media from all three manufacturers were 97 and 99%, respectively, by initial testing and 98 and 99%, respectively, after repeat testing. For disk diffusion testing, the overall sensitivity and specificity were 98 and 99%, respectively, for both initial and repeat testing. For the three CAMHB media used, when the MICs in Difco CAMHB were compared to those in BBL and Oxoid CAMHBs (data not shown), 93.2% of the MICs were the same. MIC results of ±1 dilution (essential agreement) were 100% for Difco CAMHB versus BBL CAMHB and 99.8% for Difco CAMHB versus Oxoid CAMHB. Comparison of the results obtained with BBL versus Oxoid disks showed that 99.7% of the results were ±6 mm for the 200-μg disks and 100% were ±6 mm for the 5-μg disks. Therefore, for further analysis of the data from phase 2 testing, only the results for testing in Difco CAMHB and with BBL disks were used and are shown in Fig. Fig.11 and 2 as scatter plots of the MICs obtained using Difco CAMHB versus the zone diameters obtained using BBL disks. Results for repeat testing with the 200-μg disk are shown in Fig. Fig.1,1, and those for repeat testing with the 5-μg disk are shown in Fig. Fig.22.
On initial MIC testing using the correlation with mupA, six mupA-positive isolates had mupirocin MICs of ≤256 μg/ml (sensitivity errors) and three mupA-negative isolates had MICs of ≥512 μg/ml (specificity errors). By disk diffusion with the 200-μg disk, four mupA-positive isolates had zone diameters of >6 mm (sensitivity errors) and three mupA-negative isolates had no zone (i.e., 6 mm) (specificity errors). Following repeat testing, the errors for most strains resolved; the exceptions were for three phenotypically resistant strains that lacked mupA and three phenotypically susceptible strains that contained mupA.
Using the presence or absence of mupA as the “gold standard,” for MIC testing, the sensitivity and specificity ranged from 97 to 99% (Table (Table3).3). However, on the basis of the mupirocin MIC only, for the repeat testing with the 200-μg disk (Fig. (Fig.1),1), 100% of isolates with mupirocin MICs of ≥512 μg/ml had zone diameters of 6 mm and 100% of isolates with mupirocin MICs of ≤256 μg/ml had zone diameters of >18 mm.
The results for the 5-μg disk are shown in Fig. Fig.2.2. When it was used alone, the 5-μg disk was unable to differentiate mupirocin LLR (MICs, 8 to 256 μg/ml) from HLR (MICs, ≥512 μg/ml). All mupirocin-LLR or -HLR strains and three mupirocin-susceptible strains had zone diameters of ≤11 mm with the 5-μg disk. Similarly, the 200-μg disk used alone was unable to differentiate LLR strains from those that were susceptible (Fig. (Fig.1);1); all LLR strains had zone diameters of >18 mm. However, if the results obtained with the 200-μg disk are used along with the results obtained with the 5-μg disk (Table (Table4),4), it is possible to accurately classify HLR, LLR, and susceptible strains. Of 186 strains with no zone with both the 5- and 200-μg disks, 100% were HLR by MIC testing and 98.4% were mupA positive. Of 93 strains with no zone with the 5-μg disk and any zone with the 200-μg disk, 100% were LLR and mupA negative. Of 342 strains with any zone with either disk, 99.1% were susceptible, <1% were LLR, and 99.1% were mupA negative.
The majority of mupA-positive strains tested were MRSA (174 of 186, or 89%) (Table (Table1).1). Of the mupA-negative strains tested, 104 were methicillin-susceptible S. aureus (MSSA) strains. All of these were mupirocin susceptible; i.e., the MICs were ≤4 μg/ml (Table (Table1).1). Therefore, HLR was detected among both MRSA and MSSA strains. Since no MSSA strains with LLR were included, we are unable to confirm if the test would work for this group.
Summary results for quality control testing and the quality control ranges recommended by the CLSI (7) are shown in Table Table5.5. For S. aureus ATCC 29213, all results were within a narrow 2-dilution range (0.12 to 0.25 μg/ml), and the mode was 0.25 μg/ml. However, because the number of results at 0.12 μg/ml was >66% of the number of results at the mode, this is considered a bimodal distribution, requiring that a 4-dilution range be selected (2). This was also true for E. faecalis ATCC 29212, with all results being 32 to 64 μg/ml and the mode being 64 μg/ml but the number of results at 32 μg/ml being >66% of the number at the mode. For S. aureus ATCC 25923 (Table (Table5),5), the disk diffusion parameters recommend that 97% of the values be in the tested range for the 200-μg disk and 98% of the values be in the tested range for the 5-μg disk. For disk diffusion testing of a mupA-positive strain, S. aureus BAA-1708 (data not shown), all results were 6 mm (i.e., no zone).
Standardized mupirocin susceptibility testing is needed because of increasing mupirocin use for nasal decolonization (19), reports that expanding the use of mupirocin results in greater rates of resistance (24), and the understanding that both mupirocin-low-level-resistant (MICs = 8 to 256 μg/ml) and high-level-resistant (MICs ≥ 512 μg/ml) S. aureus strains must be differentiated when clinical outcome studies are performed (18). We have shown with a three-phase collaborative study that both mupirocin disk diffusion testing and broth dilution testing using CLSI testing methods (3, 6) accurately predict in vitro susceptibility and both low-level and high-level resistance. As demonstrated in the present study, accurate disk readings are achieved with a full 24-h incubation and by the use of transmitted light to read the zone diameters. In addition, no significant broth medium effects were noted with MIC testing when CAMHBs from three different sources were used. Finally, quality control ranges are recommended for disk and broth testing based on CLSI M23-A3 methods (2).
The 200-μg mupirocin disk test differentiates S. aureus strains with high-level mupirocin resistance (no zone of inhibition) from susceptible and low-level-resistant strains (any zone of inhibition) compared to tests that base such differentiations on detection of the presence or absence of mupA by PCR (Table (Table3).3). The sensitivity and specificity of this approach are 98 and 99%, respectively. When the results from the 200-μg disk are compared to the broth dilution MIC results, there is a 100% correlation between MICs of ≥512 μg/ml and no zone of inhibition and MICs of ≤512 μg/ml and a zone of inhibition of >18 mm (Fig. (Fig.1).1). In addition, as suggested by Cookson, the use of a 5-μg mupirocin disk can be used to detect S. aureus stains with low- and high-level resistance, i.e., nonsusceptible isolates (Fig. (Fig.2)2) (8). Our three-laboratory study validated the ability of a 5-μg disk to differentiate mupirocin-susceptible strains (≥12-mm zone diameter and ≤4-μg/ml MIC) from mupirocin-resistant strains (≤11-mm zone diameter and ≥8-μg/ml MIC) with 100% sensitivity and specificity compared to the broth dilution MIC results.
In our study, three strains with mupirocin MICs of >512 μg/ml and no zone of inhibition around the 200-μg disk were mupA negative by PCR. The same three strains also had no zone of inhibition with the 5-μg disk. High-level-mupirocin-resistant, mupA-negative strains have been detected previously (1, 18). These strains could carry a novel mechanism of resistance or, one could postulate, could have multiple base changes in the native, chromosomal isoleucyl-tRNA synthetase gene, elevating MIC values from the low-level to high-level resistance ranges. It has been shown that single base changes in this chromosomal gene result in mupirocin MICs of 8 μg/ml, while two mutations result in MICs of >32 μg/ml (1). Since repeated passage of S. aureus in the presence of subinhibitory levels of mupirocin has resulted in mupA-negative, high-level-resistant mutants, it is tempting to favor the latter mechanism as an explanation for the nonagreement of the results obtained with these three strains (23).
In addition, three strains of S. aureus were mupA positive by PCR and had MICs of <4 μg/ml and inhibitory zone diameters of 30, 32, and 33 mm with the 200-μg disk and 23, 24, and 26 mm with the 5-μg disk. As noted by others (18), mupA-positive strains with low-level-resistant or susceptible MIC values occur because (i) the mupA gene was located on the chromosome rather than a plasmid (low-level resistance) or (ii) a frameshift mutation in the mupA gene is present and resulted in an inactive gene product.
We expect that the strains tested in this multilaboratory study are representative of current S. aureus community and hospital strains. With 622 strains tested in phase 2 and only 3 (0.5%) appearing to be falsely resistant by PCR, we predict that this finding is rare today. False resistance might lead to the use of an alternative topical drug that has in vitro activity but no FDA approval for decolonization or the use of agents for which clinical trial data document in vivo efficacy for nasal decolonization, such as neomycin sulfate-polymyxin B sulfate (Polysporin)-bacitracin (Neosporin), bacitracin zinc-polymyxin B sulfate, or retapamulin. A false-susceptible finding would subject the patient to decolonization with mupirocin where the isolate is resistant and unlikely to be eradicated. Fortunately, false-susceptible results in our population of strains are equally rare (<0.5% of strains tested).
As advocated by de Oliveira et al. (10), we also validated the use of two mupirocin disk concentrations to identify three phenotypic categories: mupirocin susceptible, mupirocin low-level resistant, and mupirocin high-level resistant (Table (Table4).4). Using zone/no zone criteria, no zone around both the 5-μg and 200-μg disks indicates mupirocin high-level resistance. No zone around the 5-μg disk and any zone around the 200-μg disk indicate low-level resistance. Any zone of inhibition around both disks indicates mupirocin susceptibility. The accuracies of the two-disk method for the prediction of high-level resistance, low-level resistance, and susceptibility are 100%, 100%, and 99%, respectively, compared to the results of broth dilution MIC testing.
The results from phase 3, using an eight-laboratory study, recommend acceptable limits for quality control of broth microdilution and disk diffusion (Table (Table5).5). S. aureus ATCC 29213 and E. faecalis ATCC 29212, with ranges at low and intermediate MIC levels, respectively, are recommended for use for broth dilution quality control testing. S. aureus ATCC 25923 is used for quality control testing with the 5- and 200-μg disks. A mupA-positive strain, S. aureus ATCC BAA-1708, should grow in a well or tube containing 256 μg/ml of mupirocin and produce no zone of inhibition with the 200-μg disk. This strain is recommended for quality control testing of the CLSI broth dilution and disk tests for the detection of high-level mupirocin resistance (7).
Reasons for performing mupirocin susceptibility testing may be to identify individual patients who are at risk for mupirocin decolonization failure or to perform surveillance studies for emerging mupirocin resistance. High-level mupirocin resistance has been associated with mupirocin decolonization failures. In a prospective evaluation of mupirocin decolonization, at 3 days posttreatment, the decolonization rates were lower for patients colonized with high-level-mupirocin-resistant MRSA (27.7%) than for patients colonized with either low-level-mupirocin-resistant MRSA (80%) or mupirocin-susceptible MRSA (78.5%) (25). In a randomized controlled trial where a combination of nasal mupirocin use, oral doxycycline therapy, and chlorohexidine washes was compared to no decolonization intervention for preventing MRSA infections in a health care setting, nasal colonization with high-level mupirocin-resistant S. aureus isolates was identified as an independent predictor of decolonization failure in the treatment arm (21). Studies correlating mupirocin HLR and decolonization treatment failure have been done only with MRSA (14, 21), although we have no reason to believe that it would be different with MSSA.
The clinical significance of low-level mupirocin resistance is unclear. One reason for this is that isolates with this phenotype are relatively uncommon. In the randomized controlled trial, there were not enough isolates with low-level resistance to evaluate how this affected the mupirocin decolonization outcome. In the study of mupirocin decolonization by Walker et al. described above, patients colonized with low-level-mupirocin-resistant isolates were initially decolonized (80% at the 3-day follow-up), but at the 1- to 4-week follow-up, the decolonization rate for patients colonized with low-level-resistant strains was similar to the rate for patients colonized with high-level-resistant strains, and these rates were lower than those for patients colonized with susceptible strains (25). However, only 5 of the 40 patients were colonized with low-level-resistant S. aureus isolates; therefore, these results are difficult to interpret. The availability of standardized and validated mupirocin broth dilution and disk diffusion testing methods should facilitate data collection to understand the clinical significance of low-level mupirocin resistance.
Currently, there are no FDA breakpoints for mupirocin, and this has limited the availability of commercial products for mupirocin susceptibility testing in the United States. However, similar broth-based tests are included on panels from some manufacturers of automated susceptibility testing devices sold outside the United States. Commercial disks are currently available in the United States for research use only. Laboratories in the United States that want to perform mupirocin susceptibility testing could prepare their own reagents and validate the test in-house. If mupirocin use increases, it is likely that mupirocin resistance will increase as well. In this case, it will be important for microbiology laboratories to have access to mupirocin susceptibility testing methods and reagents.
We thank all those who performed the technical work or provided expert assistance for this study: David Lonsway (CDC); Lisa Louie and Christine Watt (Sunnybrook Health Sciences Center); Jean Spargo (Massachusetts General Hospital); Bard Kostecki (Evanston Hospital); David Vicino (University of Rochester Medical Center); Farzaneh Sooudipour and Viet Nam Nguyen (UCLA); M. L. McElmeel and L. C. Fulcher (University of Texas Health Science Center); and Stanley Mirrett, Dolores H. Calley, and Hina S. Patel (Duke University Medical Center). We also thank the ABC MRSA project investigators for their contribution of isolates for this study.
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.
Published ahead of print on 5 May 2010.