Mannitol Salt Agar-Cefoxitin Combination as a Screening Medium for Methicillin-Resistant Staphylococcus aureus

In disk diffusion tests, cefoxitin is now considered a betterindicator than oxacillin for the presence of the mecA gene inStaphylococcus aureus. A logical extension of this work is theincorporation of cefoxitin into media selective for methicillin-resistantStaphylococcus aureus (MRSA). This paper describes the developmentand subsequent testing of mannitol salt agar containing 4 mg/litercefoxitin with a unique collection of well-characterized MRSAstrains, including low-level methicillin-resistant strains andan equal number of known mecA-negative S. aureus strains. Theagar supported the growth of 96.6% of the mecA-positive strainsin the collection and inhibited the growth of 100% of the mecA-negativestrains. These results suggest that selective media based oncefoxitin are superior to those based on oxacillin for the detectionof MRSA.

INTRODUCTION

Top

Introduction

Culture media selective for methicillin-resistant Staphylococcusaureus (MRSA) have traditionally been based on blood agar, mannitolsalt agar (MSA), or Baird-Parker agar containing methicillinor oxacillin alone or in combination with other antibiotics.An enrichment step consisting of culture in nutrient broth withup to 7.5% sodium chloride prior to inoculation of one of theselective agars mentioned above has also been advocated. Countlessreports in the literature (2, 3, 5, 7, 10) have described screeningmedia for MRSA, and this possibly reflects the fact that phenotypicmethods based on methicillin or oxacillin have never achievedlevels of sensitivity and specificity which have been totallyacceptable. However, we have continued to base our screeningmedia on these agents simply because we have had nothing better.Recent reports (4, 8, 9) have shown that cefoxitin is a betteragent for prediction of methicillin resistance in S. aureus,and disk susceptibility testing with cefoxitin now replacesdisk susceptibility testing with methicillin and oxacillin inan increasing number of centers. We describe the use of a collectionof well-characterized MRSA strains in the development of a selectivemedium which replaces oxacillin with cefoxitin.

MATERIALS AND METHODS

Top

Materials and Methods

MSA (DM160D) was from Mast Diagnostics (Merseyside, United Kingdom),cefoxitin sodium salt was from Sigma-Aldrich, and Etest wasfrom AB Biodisk (Solna, Sweden).

A total of 89 mecA-negative and 87 mecA-positive Staphylococcusaureus strains, described below, and three S. aureus controlstrains (two mecA-positive strains, ATCC 33591 and ATCC 43300,and one mecA-negative strain, ATCC 29213) were used in the study.Pulsed-field gel electrophoresis and/or multilocus sequencetyping (MLST) had previously been performed with all mecA-positivestrains (9, 10).

The following groups of strains were tested in this study: (i)a 100-strain collection of 75 mecA-positive S. aureus (MRSA)strains and 25 mecA-negative S. aureus (methicillin-susceptibleS. aureus [MSSA]) strains described previously (10), which wereused in the initial stages in order to set an appropriate concentrationof cefoxitin for use in the medium; (ii) 64 S. aureus strainsthat tested negative for the mecA gene (of these, 14 were classifiedas borderline resistant S. aureus (BORSA) strains on accountof their elevated MICs for oxacillin); (iii) 8 mecA-positivestrains which have proved difficult or impossible to classifyas such by the use of cefoxitin disks (9); and (iv) an additionalfour MRSA strains that were initially classified as BORSA strainsbut that were later found to be positive for the mecA gene.

The presence or absence of the mecA gene and the presence ofthe nuc gene were determined for all strains by PCR (9, 10).

The MICs for the 100-strain collection were determined by usingEtest on Iso-Sensitest Agar (Oxoid, Basingstoke, United Kingdom).

MSA plates were prepared with cefoxitin at concentrations of2, 3, or 4 mg/liter. The test strains were applied at concentrationsof 10² and 10⁵ viable organisms in the following way: the suspensionswere prepared by using fresh overnight cultures on blood agar,matched to a McFarland standard, and diluted in phosphate-bufferedsaline to suspensions such that when 1 µl was appliedwith a disposable loop and spread over an area of approximately6 cm by 2 cm, a count of either 100 or 100,000 colonies wasachieved. A control plate of blood agar was used for each strainand dilution. Incubation was carried out in air at 35 to 37°Cfor 48 h, and the result was read at 18 and 48 h. The growthof any visible colonies after incubation was recorded as a positiveresult.

The remaining strains were tested as described above, but onlyon MSA containing 4 mg/liter cefoxitin.

RESULTS

Top

Results

The MIC results for the 100-strain collection were as follows:all 25 mecA-negative strains exhibited cefoxitin MICs of $≤$ 2 mg/liter.Seven mecA-positive strains had MICs of 4 mg/liter, and theremaining 68 mecA-positive strains had MICs >4 mg/liter.

MSA plates with 2 mg/liter cefoxitin allowed the growth of allMRSA and MSSA strains after 48 h incubation. MSA with 3 mg/litercefoxitin allowed the growth of four MSSA strains and all ofthe MRSA strains. MSA containing 4 mg/liter cefoxitin allowedthe growth of none of the 25 MSSA strains and all 75 MRSA strains.We therefore chose to further evaluate a concentration of 4mg/liter cefoxitin.

The growth of all strains at 18 h and at 48 h is shown in Table1. The performance of the medium for the eight strains thatwere the most difficult to grow is shown in Table 2.

View this table:
[in this window]
[in a new window]
TABLE 1. Growth of 89 MRSA and 90 MSSA strains on MSA containing 4 mg/liter cefoxitin after 18 and 48 h incubation and with two inocula

View this table:
[in this window]
[in a new window]
TABLE 2. Performance of the medium with the eight most-difficult-to-grow MRSA strains after 48 h incubation on MSA with cefoxitin

DISCUSSION

Top

Discussion

Methicillin or oxacillin has been the agent of choice in selectivemedia for mecA-positive S. aureus for over 20 years. Mannitoland salt have often been combined as an aid in identificationand a selective agent, respectively. Mannitol salt agar with1 mg/liter oxacillin was shown to have a sensitivity of 90.7%and a specificity of 96.0% (10) for the strains (100-straincollection) used in this study. A similar medium, oxacillinresistant screening agar (1), which is also based on mannitol,salt, and lithium chloride but which contains 2 mg/liter oxacillin,was shown to have a sensitivity of 74% when it was used to testpatient specimens plated directly, although its sensitivityincreased when it was combined with an enrichment broth.

In recent years cephalosporins have been reported to be particularlysuccessful when they are used as an alternative to oxacillin.A phenyl red mannitol broth containing aztreonam and ceftizoxime(11) allowed the growth of all reference MRSA strains tested,although only a small number of strains were tested. Disks containingthe cephamycins cefoxitin and moxalactam on Mueller-Hinton agarwere shown to be 100% sensitive and specific for the differentiationof MRSA and MSSA (4). By the use of Iso-Sensitest agar and a30-µg cefoxitin disk, a similarly high sensitivity andspecificity were reported for a total of 457 S. aureus strains,many of which exhibited low-level resistance to oxacillin (8).A recent report that described the use of a chromogenic mediumand cefoxitin (6) indicated that this medium showed a sensitivityof 89% and a specificity of 99.5% for the detection of MRSAin clinical material.

In our work, the typing of the mecA-positive Staphylococcusaureus strains (by pulsed-field gel electrophoresis and/or MLST)was used only to ensure that the MRSA strain collection wasas heterogeneous as possible. Our results demonstrate that cefoxitinis far superior to oxacillin when it is used in this particularmannitol salt medium. By the use of a 48-h incubation and a10²-CFU/ml inoculum, the sensitivity and specificity of thecefoxitin MSA plate method were 100% (compared to 90.7% withoxacillin) and 100% (compared to 96.0% for oxacillin), respectively,when the comparison was made for the strains used in both studies.Over and above the strains used for the initial calibrationof the medium, we chose a collection of MRSA strains which haveproved difficult or impossible to categorize correctly by usingoxacillin or cefoxitin (9). These strains give inhibition zoneswhich fall within ±3 mm of the suggested zone breakpointfor cefoxitin (9). Three of a total of eight strains would havebeen missed by use of this medium, and five of eight strainswould have been missed on oxacillin-containing medium. Whenwe included these eight difficult-to-identify MRSA strains inthe calculation, the sensitivity of the cefoxitin medium was96.0% (85.0% for oxacillin) and the specificity was 100% (96%for oxacillin).

It should be noted, however, that all MSA media are not alike;their salt contents differ, and this may affect the growth ofsome strains (10). The particular MSA used in the present studycontains 3% salt, whereas most other commercially produced MSAmedia contain 7.5% salt, and the results reported here may notbe achievable with other brands.

Based on the cefoxitin MICs for our 100 reference strains, acefoxitin concentration of 4 mg/liter was chosen. Laboratorystudies with a battery of MSSA strains tested repeatedly haveshown that the plated medium has a shelf life of 30 days ata temperature of 2 to 8°C. In contrast, selective mediacontaining oxacillin deteriorate after 1 week. Our use of twoinocula serves to illustrate the "inoculum effect" on the timeto detection. Small numbers of bacteria take longer to producevisible growth, up to 48 h in the present study, while heavierinocula, such as those generated after overnight incubationin enrichment broth, can produce growth on solid media 1 dayearlier (1). The choice of the two inocula was based on theexpected numbers of bacteria in specimens without enrichment(10²) and after enrichment (10⁵). Without the use of an enrichmentstage, our medium requires 48 h of incubation.

In routine use our medium has proved to be robust and reliable.Patient material is plated directly on the medium, and the platesare examined on the following day. Any yellow colonies are testedby a tube coagulase test and a cefoxitin susceptibility test.Those strains that test positive for coagulase and that areresistant to cefoxitin are further tested by the MRSA latextest, and a presumptive identification of MRSA is made. At thispoint, many laboratories would use a PCR test for the mecA genefor confirmation, and in areas where MRSA is frequently detected,it may be acceptable to trust the results of the cefoxitin disktest. Plates which are negative at 18 h are incubated for afurther 24 h before they are discarded. The number of S. aureusstrains which are mannitol negative is not known, but as faras we have been able to ascertain, the number is low and noneof the strains in the collection (which were not selected byusing MSA medium) were mannitol negative. However, ever morefrequently patient specimens contain methicillin-resistant (andcefoxitin-resistant) coagulase-negative staphylococci (CoNS).These are capable of growth on the medium, but S. epidermidisproduces pink colonies, which are easily distinguishable fromthe yellow (mannitol-positive) colonies of S. aureus. A numberof CoNS other than S. epidermidis are mannitol positive andtherefore produce yellow colonies. In our setting and basedon the results for all samples screened for 1 year, the problemoccurs in 10% of the samples. With the low prevalence of MRSAin our area of Sweden (<1%), most of the secondary work iswith the CoNS. In an area with a higher prevalence (e.g., 10to 40% in many areas of Europe), MRSA will completely dominatethe findings and CoNS will not be perceived as a problem. Irrespectiveof the prevalence of MRSA, since one has to do the cefoxitintest at this point anyway, the addition of a few negative coagulasetube tests which can be discarded before further work is doneis of little consequence.

ACKNOWLEDGMENTS

We thank the Antibiotic Section, Institute for Infectious DiseaseControl (SMI), Stockholm, Sweden, and the Statens Serum Institut,Copenhagen, Denmark, for bacterial strains and Anders Rhod Larsenfor MLST.

FOOTNOTES

* Corresponding author. Mailing address: Department of Clinical Microbiology, Central Hospital, SE 351 85 Växjö, Sweden. Phone: 46 470 58 74 60. Fax: 46 470 58 74 55. E-mail: robert.smyth{at}ltkronoberg.se.

REFERENCES

Top