ARTICLE IN PRESS

Evaluation of separate vs pooled swab cultures,different media, broth enrichment andanatomical sites of screening for the detectionof methicillin-resistant Staphylococcus aureusfrom clinical specimens

I. Grmek-Kosnika,*, A. Ihanb, U. Dermotaa, M. Remsc, M. Kosnikd,H. Jorn Kolmose

aRegional Institute of Public Health Kranj, Gosposvetska 12, SI-4000 Kranj, SloveniabInstitute of Microbiology and Immunology, Medical Faculty, University Ljubljana, Zaloska 4,SI-1000 Ljubljana, SloveniacRegional Hospital of Jesenice, Cesta Marsala Tita 112, SI-4000 Jesenice, SloveniadUniversity Clinic of Respiratory and Allergic Diseases, 4204 Golnik, SloveniaeDepartment of Clinical Microbiology, Odense University Hospital, DK-5000 Odense, Denmark

Received 17 July 2003; accepted 24 January 2005

01do

KEYWORDSMRSA; Pooling;Detection ofcolonization;Surveillance cultures

95-6701/$ - see front matter Q 200i:10.1016/j.jhin.2005.01.032

* Corresponding author. Tel.: C386E-mail address: irena.grmek-kosnik

Summary Early identification of methicillin-resistant Staphylococcusaureus (MRSA) carriers is a major component of an MRSA control programme.The cost and laboratory workload could be markedly reduced by processingmultiple swabs from one person in one culture broth (specimen pooling). Weevaluated the sensitivity for MRSA detection and the growth rate of pooledswabs compared with individual processing. In total, 1254 swabs from 423subjects (two to five swabs per subject) were submitted for detection ofMRSA. Swabs were suspended in 2-mL volumes of sterile Todd–Hewitt Brothand divided into two 1-mL aliquots. One aliquot of the suspension wasprocessed as a single specimen, and the other aliquot was mixed (pooled)with other suspensions in which swabs from the same patient weresuspended. Forty-four (10%) pooled samples were positive for MRSA.Specimens from seven additional patients that were negative when pooledwere positive when processed separately. There was no case where thepooled specimen was positive but the separate specimens were negative.The diagnostic sensitivity of pooled surveillance cultures compared with

Journal of Hospital Infection (xxxx) xx, 1–7

www.elsevierhealth.com/journals/jhin

5 The Hospital Infection Society. Published by Elsevier Ltd. All rights reserved.

4 2017112; fax: C386 4 2017113.@zzv-kr.si

I. Grmek-Kosnik et al.2

ARTICLE IN PRESS

single cultures, when only subjects colonized by MRSA were considered, was86% and the false-negative rate was 14%. Eighty percent of the pooledpositive cultures were detected by the third day and all were detected bythe fourth day. Fifty-four percent of the specimens processed separatelywere detected by the second day and all were detected by the fourth day.Pooling of specimens decreases the sensitivity of MRSA detection comparedwith processing each swab separately, particularly in swabs with a lownumber of colony-forming units. In all subjects whose pooled samples werenegative but whose swabs examined separately were positive, the swabsexamined separately were negative on primary plates and positive only afterculturing in enrichment broth.Q 2005 The Hospital Infection Society. Published by Elsevier Ltd. All rightsreserved.

Introduction

The spread of methicillin-resistant Staphylococcusaureus (MRSA) has become amajor problem in manyhospitals worldwide.1 In the healthcare setting,early identification of MRSA carriers is a majorcomponent of an MRSA control programme. Manyhospitals practise on-admission screening for MRSAof high-risk patients to prevent nosocomial andinterhospital spread, and to isolate and decolonizeMRSA carriers appropriately.2–6

Over the past decade, laboratory procedureshave dramatically improved the detection ofresistant micro-organisms in terms of speed, sensi-tivity and accuracy. New ways of optimizing thesensitivity of isolation have been described, involv-ing the addition of methicillin or oxacillin tomannitol salt agar and the use of enrichment insalt broth.7–9 Selective media have been developedto achieve isolation and presumptive identificationin a single step.

The increased prevalence of MRSA has had agreat impact on the workload of microbiologylaboratories. Also, as laboratory budgets havebeen reduced, the detection and identificationprotocols for MRSA screening need to be based notonly on sensitivity, but also on speed of detection,workload and cost per patient.

The cost and laboratory workload could bemarkedly reduced by processing multiple swabsfrom one person in one culture broth (specimenpooling). This study was performed to evaluate thesensitivity for MRSA detection and the growth rateof pooled swabs, compared with individual process-ing as in usual laboratory practice. In addition, wecompared the performance of broth enrichment,non-selective media and selective media. We alsoevaluated the prevalence of MRSA according toanatomical site in three groups: patients who had

been MRSA positive in the past, patients withunknown MRSA status, and healthcare workers.

Materials and methods

Screening policy for MRSA

The regional hospital in Jesenice routinely screensany patient with a history of MRSA colonization, allpatients admitted from nursing homes and allinterhospital transfers. It is usual practice to takethree swabs: one from both nostrils, one from thethroat and one from the axilla and groin regions. Inpatients in intensive care units, a swab of theperineal area is also taken. In addition, if thepatient has wounds, leg ulcers or sacral sores,swabs from these areas are taken. A catheterspecimen of urine is collected in patients with anindwelling urethral catheter, and a tracheal aspi-rate is examined in intubated patients.

Bacterial isolates

The study was conducted over a six-month periodbeginning in January 2002 at the Institute of PublicHealth Kranj, Slovenia. In total, 1254 specimenswere obtained for MRSA culture, consisting ofnostrils (415 specimens), perineum (105 speci-mens), throat (290 specimens), axilla and groin(390 specimens), and skin and soft tissue (54specimens).

Cotton-tipped swabs were transported in Stuarttransport medium Code 111C (Copan, Italy). In thelaboratory, the swabs were placed in 2 mL of sterileTodd–Hewitt Broth (THB) (Becton Dickinson andCompany, Le Pont de Claix, France) with STAPH/STREP selective supplement containing nalidixicacid and colistin sulphate (CNA Supplement, Oxoid,

Separate vs pooled swab cultures for detection of MRSA 3

ARTICLE IN PRESS

Basingstoke, UK), and vortexed for 5 s. Swabs werediscarded and the suspensions of THB were dividedin two; 1 mL of the suspension was processed as asingle specimen, and the remaining 1 mL was mixed(pooled) with other suspensions from the samepatient.

Separate processing of swab cultures

This procedure is used routinely in our laboratory.One millilitre of suspension was vortexed, and 50 mLof suspension was inoculated on to a blood agarplate (BA) with an oxacillin disk and on to aselective medium, ORSAB (oxacillin resistancescreening agar base) with supplementary oxacillin(2 mg/L) and polymixin B and aniline blue indicator(Oxoid, Basingstoke, UK). An oxacillin disk (1 mg)(Becton Dickinson and Company, Le Pont de Claix,France) was put on the most dense part of thebacterial suspension on the blood agar. The plateswere incubated at 37 8C. Plates were examined at24 and 48 h for the presence of mannitol-ferment-ing colonies, which were blue on ORSAB.

The remaining 900 mL of suspension was incu-bated at 37 8C. After overnight incubation, 50 mL ofthe suspension was inoculated on to ORSAB (THB-ORSAB) and BA (THB-BA). THB-ORSAB and THB-BAplates were incubated at 37 8C. If the plates werenegative after 24 h, 50 mL of the remaining suspen-sion was inoculated on to BA and ORSAB again. Theremaining THB suspensions were stored at K30 8C.All plates were incubated aerobically at 37 8C andexamined at 24 and 48 h for typical colonies ofS. aureus. All suspicious colonies around theoxacillin disk from BA and dark blue colonies fromORSAB were subcultured on to blood agar andidentified by conventional tests.

Pooling of swab specimens

One millilitre THB suspensions of all single swabsfrom the same subject were transferred to theempty sterile tube. Pooled suspensions weremanaged separately from the single suspensions.They were coded to remain blind until the end ofthe examination. Pooled suspensions’ enrichmentbroths were incubated at 37 8C for 24 h. Afterovernight incubation, 50 mL of pooled suspensionswere inoculated into BA and ORSAB. In pilotexperiments, we found that growth of bacteria inimmediate subculturing of the pooled suspensionwas inadequate for analysis. Therefore, we did notperform immediate subculturing of the pooledsuspensions.

Identification of S. aureus and confirmationof MRSA

Isolates of MRSA were identified by Gram stain,catalase activity, DNase activity (DNase agar)(Oxoid, Basingstoke, UK), mannitol salt fermenta-tion and a latex agglutination test (Slidex StaphPlus) (bioMerieux, Marcy l’Etoile, France). TheMRSA screen test (Innogenetics, Belgium) wasused for the detection of PBP 2a associated withoxacillin-resistant staphylococci. Tube coagulaseactivity (rabbit plasma) (bioMerieux, France) andPYR test (Oxoid, Basingstoke, UK) were determinedwhen necessary.

Susceptibility to oxacillin was determined by thedisk diffusion method, according to the NationalCommittee for Clinical Laboratory Standards guide-lines, with Mueller–Hinton agar. All isolates weretested for oxacillin resistance by screening onMueller–Hinton agar containing 6 mg/L oxacillin.10

S. aureus ATCC 25923 and ATCC 43300 were used tomonitor antibiotic potency for quality control. Thefirst MRSA isolate from each subject was frozen atK70 8C.

Statistical analysis

We considered the classical bacteriological tech-nique (examination of single surveillance speci-mens) to be the reference method. Comparison ofisolation rates of pooled surveillance swabs withsingle swabs was done by the two paired proportionmethod (Chi-square test), in which the numbers ofspecimens positive by one test were compared withthe numbers positive by the other test. Statisticalsignificance was accepted for P!0.05.

Results

Prevalence of MRSA

In total, 1254 specimens from 423 subjects (264patients and 159 healthcare workers) were includedin the study. Ninety-four subjects (22%) were knownto have been MRSA carriers in the past. Threehealthcare workers and two patients had undergonedecolonization before collection of specimens. Of1254 specimens, 415 swabs (33%) were taken fromthe nose, 390 (31%) from the axilla and groin, 290(23%) from the throat, 105 (8%) from the perineumand 54 (4%) from elsewhere on the skin. Overall,MRSA was detected from at least one specimenfrom the same patient in 11.9% of specimens(prevalence of MRSA was 11.9%).

I. Grmek-Kosnik et al.4

ARTICLE IN PRESS

Single vs pooled processing of specimens

We examined 423 pooled specimens. Poolings weremade by mixing two to five (mean 3.0) swabs fromthe same patient. Forty-four (10%) single andpooled specimens were positive. Specimens fromseven patients were positive when analysed separ-ately but negative when pooled. There was no casewhere pooled specimens were positive but separatespecimens were negative (Table I). In five patients,pooled specimens were negative and separatespecimens were positive; these five patients hadbeen previously colonized by MRSA (three noses,one throat, one groin). In four cases, MRSA wereisolated only on ORSAB after broth enrichment inTHB. Also, in two specimens from healthcareworkers (one nose, one throat), MRSA were isolatedonly on ORSAB after broth enrichment in THB.Examination of separate specimens was signifi-cantly more efficient than examination of pooledsamples (Chi-square 5.5, P!0.05). The diagnosticsensitivity of pooled surveillance cultures com-pared with single cultures, when only subjectscolonized by MRSA were considered, was 86% with afalse-negative rate of 14%. Overall, the negativepredictive value of pooled specimens was 98%. Allpatients with specimens that were negative whenpooled and positive when analysed separately hadMRSA isolated from only one anatomical site,whereas 23 of 44 (52%) patients with pooled positiveresults had MRSA isolated from two or moreanatomical sites (P!0.05).

Time to detection

The mean time to MRSA detection (standarddeviation) was 53 (9.8) h for the pooled specimenscompared with 38 (17) h for the specimensprocessed separately. Eighty percent of the pooledpositive cultures were detected on the third day(48 h after receiving specimens in the laboratory)and all had been detected by the fourth day. Fifty-four percent of the specimens processed separately

Table I Comparison of diagnostic yield of pooled vsseparate surveillance swabs in 423 subjects

Separate

Positive Negative Total

Pooled Positive 44 0 44Negative 7 372 379Total 51 372 423

Only one positive single surveillance swab per subject wasconsidered. The difference in diagnostic yield between testswas statistically significant (Chi-square test, P!0.05).

were detected on the second day (24 h afterreceiving specimens in the laboratory), another34% were detected by the third day and all had beendetected by the fourth day.

In the seven specimens that were positive whenanalysed separately but negative when pooled, themean time to MRSA detection (standard deviation)was 62 (13) h. Forty-three percent of the processedspecimens were detected by the third day and 57%were detected by the fourth day.

Influence of culture media

When culturing specimens separately, BA gaveresults in 24 h (Table II). This was rapid but had asensitivity of only 54.6%. ORSAB was not signifi-cantly more sensitive (66.3%) than BA, even after72 h. A significant improvement in sensitivity wasachieved by THB enrichment (P!0.01). Brothenrichment improved detection compared withdirect culture on BA by 26.8%, but could not replacedirect culture because of a delay in results. Also, wewould have missed 3.5% of MRSA carriers if directculture on ORSAB had not been performed. Afterbroth enrichment, subculture on ORSAB was notsignificantly more sensitive than subculture on BA.Non-selective BA was the only medium that gavepreliminary results in 54.6% of the positive speci-mens after 24 h.

Anatomical site of MRSA detection

In patients in whom MRSA had been found in thepast (NZ85), MRSA was found in 19% of all screeningcultures and in 27 (31.8%) subjects. Prevalence washighest in the throat (24%), perineum (22%) andnose (21%), and less prevalent on the skin (20%)and axilla and groin (13%) (Figure 1). If the nose andthroat alone had been screened, six of 27 caseswould have been missed (five were positive in axillaand groin alone, one was positive on skin alone). Inpatients with unknown MRSA status (NZ179), MRSAwas found in 7.5% of all screening culturesperformed and in 21 (11.7%) subjects. Prevalencewas highest in the nose (9.6%) and less prevalent onaxilla and groin (7.4%), skin (6.2%), perineum (6.1%)and throat (5.1%) (Figure 2). If nose and throatalone had been screened, three of 21 cases wouldhave been missed (two were positive on perineumalone, one was positive on axilla and groin alone). Inhealthcare workers from endemic wards (NZ159),MRSA was found in 0.6% of all screening culturesand in 1.9% of subjects (NZ3). Two were positivein the nose (1.2%) and one in the throat (0.6%), but

Table II Positive cultures of methicillin-resistant Staphylococcus aureus (MRSA) from separate specimens ondifferent culture media

Media Separate specimenspositive cultures

(NZ86)

Detectiontime (h)

Statistical significancebetween two durations

of incubation

Number (%) of MRSA that wouldhave been missed when a

particular medium was not used

BA24 47 (54.6%)a 24 NS 0 (0)BA48 47 (54.6%) 48ORSAB24 41 (47.7%)b 24 P!0.05 3 (3.5)ORSAB48 57 (66.3%) 48THB-BA24 70 (81.4%)a 48 NS 22 (25)THB-BA48 70 (81.4%) 72THB-ORSAB24 72 (83.7%)b 72 NSTHB-ORSAB48 72 (83.7%) 96

BA24(48), direct culture on blood agar with oxacillin disk examined at 24 (48) h for typical colonies of MRSA; ORSAB24(48), directculture on selective medium ORSAB examined at 24 (48) h for typical colonies of MRSA; THB-BA24(48), broth subculture on blood agarwith oxacillin disk examined at 24 (48) h for typical colonies of MRSA; THB-ORSAB24(48), broth subculture on selective medium ORSABexamined at 24 (48) h for typical colonies of MRSA; NS, not significant.a Difference between results, P!0.01.b Difference between results, P!0.05.

Separate vs pooled swab cultures for detection of MRSA 5

ARTICLE IN PRESS

no subjects were positive in axilla and groin or skin(Figure 3).

Discussion

In attempting to control MRSA, it is important toidentify the major reservoir, i.e. both colonized andinfected patients. Surveillance methods, specificallythose including screening cultures to identify pre-viously unrecognized carriers, are essential for thesuccess of infection control programmes. An idealscreening method would quickly identify all patientswith MRSA, give no false-negative results and haveboth sensitivity and negative predictive values of

Figure 1 Prevalence of methicillin-resistant Staphylo-coccus aureus (MRSA) by screening swab site among 85previously positive patients. N, number of specimens.

100%. Two factors are important in achieving thisgoal: the use of optimal culture methods (identifi-cation and susceptibility tests), and the optimalselection of anatomical sites for culture.11

The main advantage of processing several swabsfrom one person in one broth (i.e. pooling speci-mens) is the reduced cost (two to five times,depending on the number of specimens per person).However, we have shown important disadvantagesof pooling, namely lower sensitivity for detectingMRSA carriers and a delay in obtaining a result. Onereason for reduced sensitivity may be the higherdilution of MRSA in the pooled specimen, as allcases that were negative when the specimens werepooled but positive when the specimens were

Figure 2 Prevalence of methicillin-resistant Staphylo-coccus aureus (MRSA) by screening swab site among 179patients previously negative or with unknown status. N,number of specimens.

Figure 3 Prevalence of methicillin-resistant Staphylo-coccus aureus (MRSA) by screening swab site among 159health workers from endemic wards. N, number ofspecimens.

I. Grmek-Kosnik et al.6

ARTICLE IN PRESS

analysed separately had only one positive swab.Another reason might be overgrowth of MRSA byother bacteria and the inhibitory influence of otherbacteria on its growth.12–14 Furthermore, in thecase of positive pooled results, we could notdetermine the site of MRSA colonization. In theliterature, we found no data on the effect of poolingspecimens on sensitivity. Our experience indicatesthat approximately 14% of MRSA-colonized subjectsmay be missed if pooling alone is used.

Although the reading of the pooled specimenswas blind, the reading of single samples was notblind to the patients’ previous carrier status, andcould have biased our results. Out of the sevenpoolings that were falsely negative, five werepreviously positive patients and two were pre-viously negative healthcare workers. Techniciansreading the plates need to be alert to the possibilityof finding MRSA on previously negative individuals aswell as on previously positive subjects.

Broth enrichment and direct culture on aselective solid medium enhances the isolation ofMRSA. Different authors got 2–65% more positiveresults when using such methods.15–17 Enrichment isparticularly necessary in previously positivepatients, patients after decolonization, patientsat high risk for colonization, and in epidemiologicalsurveys. The disadvantage of broth enrichment is adelay in detecting the growth of MRSA.

We also looked at the roles of an enrichmentbroth, non-selective media, selective media andduration of incubation. Non-selective BA with anoxacillin disk had the advantage of allowing apresumptive positive report at 24 h, but lacked

specificity. Colonies had to be further identified.However, on the basis of our studies, we recom-mend THB enrichment for 24 h and subculturing onto a selective medium such as ORSAB. This methodallows detection of low levels of MRSA. We use achromogenic medium (ORSAB-Oxoid) containingoxacillin (2 mg/L), polymixin B and aniline blueindicator. This medium detects MRSA in 24 h asintense blue pigmented colonies, easy to identifyeven if present in very low numbers. The disadvan-tage is that ORSAB is not completely selective forMRSA and S. haemolyticus also gives blue-pigmentedcolonies on ORSAB. These colonies may be confusedwith MRSA until they are tested for coagulase andPYR. Other investigators have studied the perform-ance of selective media. Simor et al. comparedORSAB with a conventional mannitol salt agar platesupplemented with oxacillin (2 mg/L) for thedetection of MRSA in clinical specimens.18 Thesensitivities were comparable but the specificity ofORSAB was much higher. A comparison with non-selective media was not made. Gaillot et al.compared CHROMagar Staph aureus (CSA) foridentification of S. aureus, with conventionaldetection of S. aureus on 5% horse blood agarplates on 2000 consecutive clinical samples.19 Thesensitivity of CSA was significantly higher (95.5% vs81.9%), whereas the specificities were comparable(approximately 99%).18,20–22

Our study examined the relative value ofindividual culture sites in the detection of MRSAcarriage more rigorously among three differentgroups. In patients known to be MRSA positive inthe past, cultures from the throat were more oftenpositive than the other four sites studied, but otherbody sites were often positive when the throat swabwas negative. In patients with unknown MRSA statusand in healthcare workers from endemic wards,cultures of the anterior nares were more oftenpositive than the other four sites studied. If noseand throat cultures alone were performed, 17.6% ofMRSA carriers would have been missed.

The increasing prevalence of MRSA has forcedclinical microbiology laboratories to institute accu-rate and rapid detection methods. Our experienceindicates that approximately 14% of MRSA-colon-ized subjects may bemissed if pooling alone is used.Pooling is not an appropriate method for routinedetection of MRSA carriers.

References

1. Fluckiger U, Widmer AF. Epidemiology of methicillin-resistant Staphylococcus aureus. Chemotherapy 1999;45:121—134.

Separate vs pooled swab cultures for detection of MRSA 7

ARTICLE IN PRESS

2. Lucet JC, Chevret S, Durand-Zaleski I, Chastang C, Regnier B.Prevalence and risk factors for carriage of methicillin-resistant Staphylococcus aureus at admission to the inten-sive care unit. Arch Intern Med 2003;163:181—188.

3. Solberg CO. Spread of Staphylococcus aureus in hospitals:causes and prevention. Scand J Infect Dis 2000;32:587—595.

4. Girou E, Pujade G, Legrand P, Cizeau F, Brun-Buisson C.Selective screening of carriers for control of methicillin-resistant Staphylococcus aureus (MRSA) in high-risk hospitalareas with a high level of endemic MRSA. Clin Infect Dis 1998;27:543—550.

5. Mishal J, Sherer Y, Levin Y, Katz D, Embon E. Two-stageevaluation and intervention program for control of methi-cillin-resistant oxacillin-resistant Staphylococcus aureus inthe hospital setting. Scand J Infect Dis 2001;33:498—501.

6. Kotilainen P, Routamaa M, Peltonen R, et al. Elimination ofepidemic methicillin-resistant Staphylococcus aureus from auniversity hospital and district institutions, Finland. EmergInfect Dis 2003;9:169—175.

7. Kampf G, Weist K, Swidsinski S, Kegel M, Ruden H.Comparison of screening methods to identify methicillin-resistant Staphylococcus aureus. Eur J Clin Microbiol InfectDis 1997;16:301—307.

8. Van Enk RA, Thompson KD. Use of a primary isolationmedium for recovery of methicillin-resistant Staphylococcusaureus. J Clin Microbiol 1992;30:504—505.

9. Davies S, Zadik PM. Comparison of methods for the isolationof methicillin-resistant Staphylococcus aureus. J Clin Pathol1997;50:257—258.

10. National Committee for Clinical Laboratory Standards.Methods for dilution antimicrobial susceptibility tests forbacteria that grow aerobically. 5th ed. Wayne, PA: Com-mittee for Clinical Laboratory Standards; 2002 [approvedstandard M7-A5].

11. Sanford MD, Widmer AF, Bale MJ, Jones RN, Wenzel RP.Efficient detection and long-term persistence of the carriageof methicillin-resistant Staphylococcus aureus. Clin InfectDis 1994;19:1123—1128.

12. Ji G, Beavis R, Novick RP. Bacterial interference causedby autoinducing peptide variance. Science 1997;276:2027—2030.

13. Heikkila MP, Saris PE. Inhibition of Staphylococcus aureus bythe commensal bacteria of human milk. J Appl Microbiol2003;95:471—478.

14. Rao GG, Wong J. Interaction between methicillin-resistantStaphylococcus aureus (MRSA) and methicillin-sensitiveStaphylococcus aureus (MSSA). J Hosp Infect 2003;55:116—118.

15. Brayshaw PD. Methicillin-resistant Staphylococcus aureus:evaluation of detection techniques on laboratory-passagedorganisms. Br J Biomed Sci 1999;56:170—176.

16. Perry PL, Coombs GW, Boehm JD, Pearman JW. A rapid(20 h) solid screening medium for detecting methicillin-resistant Staphylococcus aureus. J Hosp Infect 1998;40:67—72.

17. Gardam M, Brunton J, Willey B, McGeer A, Low D, Conly J. Ablinded comparison of three laboratory protocols for theidentification of patients colonized with methicillin-resist-ant Staphylococcus aureus. Infect Control Hosp Epidemiol2001;22:152—156.

18. Simor AE, Goodfellow J, Louie L, Louie M. Evaluation of anew medium, oxacillin resistance screening agar base, forthe detection of methicillin-resistant Staphylococcus aureusfrom clinical specimens. J Clin Microbiol 2001;39:3422.

19. Gaillot O, Wetsch M, Fortineau N, Berche P. Evaluation ofCHROMagar Staph. aureus, a new chromogenic medium, forisolation and presumptive identification of Staphylococcusaureus from human clinical specimens. J Med Microbiol2000;38:1587—1591.

20. Kluytmans J, Van Griethuysen A, Willemse P, Van Keulen P.Performance of CHROMagar selective medium and oxacillinresistance screening agar base for identifying Staphylo-coccus aureus and detecting methicillin resistance. J ClinMicrobiol 2002;40:2480—2482.

21. Zadik PM, Davies S, Whittaker S, Mason C. Evaluation of anew selective medium for methicillin-resistant Staphylococ-cus aureus. J Med Microbiol 2001;50:476—479.

22. Merlino J, Leroi M, Bradbury R, Veal D, Harbour C. Newchromogenic identification and detection of Staphylococcusaureus and methicillin-resistant S. aureus. J Clin Microbiol2000;38:2378—2380.