Received 31 January 2007/Returned for modification 29 March 2007/Accepted 12 June 2007

Transcription

1 JOURNAL OF CLINICAL MICROBIOLOGY, Aug. 2007, p Vol. 45, No /07/$ doi: /jcm Copyright 2007, American Society for Microbiology. All Rights Reserved. The Emergence and Importation of Diverse Genotypes of Methicillin-Resistant Staphylococcus aureus (MRSA) Harboring the Panton-Valentine Leukocidin Gene (pvl) Reveal that pvl Is a Poor Marker for Community-Acquired MRSA Strains in Ireland Angela S. Rossney, 1,2 Anna C. Shore, 3 Pamela M. Morgan, 1 Margaret M. Fitzgibbon, 1,2 Brian O Connell, 1,2 and David C. Coleman 3 * National MRSA Reference Laboratory, St. James s Hospital, James s St., Dublin 8, Ireland 1 ; Department of Clinical Microbiology, University of Dublin, Trinity College, St. James s Hospital, James s St., Dublin 8, Ireland 2 ; and Microbiology Research Unit, Division of Oral Biosciences, Dublin Dental School and Hospital, University of Dublin, Trinity College, Dublin 2, Ireland 3 Received 31 January 2007/Returned for modification 29 March 2007/Accepted 12 June 2007 Community-acquired methicillin-resistant Staphylococcus aureus (CA-MRSA) carrying pvl is an emerging problem worldwide. CA-MRSA tends to harbor staphylococcal cassette chromosome mec type IV (SCCmec IV), to be non-multiantibiotic resistant, and to have different genotypes from the local hospital-acquired MRSA (HA-MRSA). However, in Ireland, 80% of HA-MRSA isolates have the non-multiantibiotic-resistant genotype ST22-MRSA-IV. This study investigated MRSA isolates from Ireland (CA-MRSA, health care-associated MRSA, and HA-MRSA) for the carriage of pvl and determined the genotypic characteristics of all pvl-positive isolates identified. All 1,389 MRSA isolates were investigated by antibiogram-resistogram typing and SmaI DNA macrorestriction analysis. pvl-positive isolates were further characterized by multilocus sequence typing and SCCmec, agr, and toxin gene typing. Twenty-five (1.8%) MRSA isolates belonging to six genotypes (ST30, ST8, ST22, ST80, ST5, and ST154) harbored pvl. Nineteen of these (76%) were CA-MRSA isolates, but a prospective study of MRSA isolates from 401 patients showed that only 6.7% (2/30) of patients with CA-MRSA yielded pvl-positive isolates. Thus, pvl cannot be used as a sole marker for CA-MRSA. Fifty-two percent of pvl-positive MRSA isolates were recovered from patients with skin and soft tissue infections; thirty-six percent were from patients of non-irish ethnic origin, reflecting the increasing heterogeneity of the Irish population due to immigration. All 25 pvl-positive isolates carried SCCmec IV; 14 (56%) harbored SCCmec IV.1 or IV.3, and the remaining 11 isolates could not be subtyped. This study demonstrates that pvl is not a reliable marker for CA-MRSA in Ireland and reveals the emergence and importation of diverse genotypes of pvl-positive MRSA in Ireland. * Corresponding author. Mailing address: Microbiology Research Unit, Division of Oral Biosciences, Dublin Dental School and Hospital, University of Dublin, Trinity College, Dublin 2, Ireland. Phone: Fax: dcoleman@dental.tcd.ie. Published ahead of print on 20 June Methicillin-resistant Staphylococcus aureus (MRSA) is a predominantly nosocomial pathogen, but in recent years it has been seen with increasing frequency in the community (50). Although there were reports of community-acquired MRSA (CA-MRSA) in 1982 and in the early 1990s, it was the deaths from necrotizing pneumonia of four young children without underlying health care-associated (HCA) risk factors that brought CA-MRSA to worldwide prominence in the late 1990s (42). Clinically, CA-MRSA usually causes skin and soft tissue infection, often with abscess and furuncle formation. However, it can cause serious life-threatening conditions, which in addition to necrotizing pneumonia include necrotizing fasciitis, bloodstream infection, and septic shock (10, 20). Unlike patients with infections caused by hospital-acquired MRSA (HA- MRSA), patients infected with CA-MRSA lack HCA risk factors, tend to be younger, and are often young children. Outbreaks of CA-MRSA infection among Australian aborigines, Native Americans, inmates of jails, participants in contact sports, and children attending crèches have been reported. Other risk groups include men who have sex with men and intravenous drug users (21). Some reports have suggested that certain strains of CA- MRSA may be more virulent than HA-MRSA (4, 42). The expression of Panton-Valentine leukocidin (PVL), a two-component, pore-forming, cytolytic toxin that targets mononuclear and polymorphonuclear cells and causes cell death by necrosis or apoptosis, has been strongly associated with CA-MRSA (4). The PVL toxin consists of two synergistic proteins, LukS-PVL and LukF-PVL, encoded by the pvl genes lukf and luks, which are carried on a temperate bacteriophage (41, 50). Methicillin resistance in MRSA is coded for by the meca gene, which is carried on a mobile genetic element termed the staphylococcal cassette chromosome mec (SCCmec) (7). Six types of SCCmec have been recognized to date (31). The two smallest SCCmec types, SCCmec IV and SCCmec V, have been associated with CA-MRSA (5, 50). It has been suggested that the small size of SCCmec IV incurs a lower cost on fitness when acquired by S. aureus, which favors its acquisition and retention (29). 2554

2 VOL. 45, 2007 DIVERSE GENOTYPES OF pvl-positive CA-MRSA IN IRELAND 2555 Study Isolation period a MRSA group TABLE 1. MRSA isolates investigated for carriage of pvl genes Hospital type or laboratory source (no. of hospitals) Specimen site No. of isolates pvl (%) No. of isolates pvl CA-MRSA (%) Retrospective 2003, Q1 Q4 I EARSS b (28) Blood (0.5) 2 (100) 2003, Q2 II Tertiary (1) Various (0.3) 1 (100) 2003, Q1 Q4 III Maternity/pediatric (2) Various 53 1 (1.9) 1 (100) IV NMRSARL (21 c ) Various (12.2) 12 (75) V EARSS (28) Blood 44 2 (4.6) 1 (50) Prospective 2004, Q4 VI Pediatric (2) Various 76 2 d (2.6) 2 (100) 2004, Q4 VII Tertiary (1) Various d (0.3) 0 (0) Total 1, (1.8) 19 e (76) a Q, quarter. b Irish hospitals that participated in the European Antimicrobial Resistance Surveillance System. c Isolates submitted to the NMRSARL with phenotypic characteristics and/or patient clinical details suggestive of CA-MRSA. d In addition to these three pvl-positive MRSA isolates, three pvl-positive methicillin-susceptible S. aureus isolates (one from a pediatric patient and two from patients in the adult hospital) were detected among isolates submitted as MRSA. These isolates are undergoing further investigation but were excluded from the present study. e Five isolates were health care-associated MRSA; one isolate was hospital acquired. Multilocus sequence typing (MLST) has shown that CA- MRSA isolates are associated with a number of different genetic backgrounds. In a study of 117 CA-MRSA isolates from three continents, Vandenesch et al. reported that different sequence types (STs) were associated with different geographic areas (ST80 is associated with Europe, ST1 and ST8 with the United States, and ST30 with Oceania) and that within each geographic area, CA-MRSA isolates were unlike local HA- MRSA isolates (50). In that study, CA-MRSA isolates tended to harbor the accessory gene regulator (agr) group 3, to not be multiantibiotic resistant, and to have lower oxacillin MICs (50). Definitions of CA-MRSA often use time-based criteria where the recovery of MRSA isolates within 48 or 72 h of hospital admission is considered indicative of CA-MRSA. However, time-based criteria do not consider patients with MRSA following recent health care exposure. Hence, a category of HCA-MRSA or CA-MRSA with risk factors has been suggested for patients who have had risk factors for the acquisition of nosocomial MRSA during the previous 12 months (42). Reported prevalence rates of CA-MRSA vary among different studies, partly reflecting problems with definitions of CA- MRSA, but also because rates may vary in different settings. A meta-analysis of studies reported a pooled prevalence rate of 30.2% among hospitalized patients in 27 retrospective studies. Among community members without HCA risk factors, the rate was 0.2% (21). Reported prevalence rates vary from 1% (among children in New York and adults in Chicago) to 42% (in some rural communities in Western Australia). Another study from Chicago reported that CA-MRSA accounted for 22% of all MRSA isolates recovered from children. Nosocomial outbreaks following hospital admission of patients with CA-MRSA have also been reported (28). Successive studies have suggested that the lack of a multiantibiotic resistance phenotype and the carriage of SCCmec IV are useful characteristics to aid in the recognition of CA- MRSA (29, 50). However, this distinction is unhelpful in Ireland, where 80% of MRSA isolates recovered from blood in 2003 did not exhibit a multiantibiotic-resistant phenotype and carried SCCmec IV. The genotype of these isolates is ST22- MRSA-IV, similar to that of the United Kingdom epidemic strain EMRSA-15 (40). Other studies suggest that CA-MRSA may be resistant to kanamycin, show decreased susceptibility to fusidic acid, and be susceptible to tobramycin and fluoroquinolones (28). Many studies have also suggested that carriage of pvl could be used as a marker for CA-MRSA (4, 17, 50). MRSA has been a widespread problem in Irish hospitals, where strains have been well characterized (35, 40, 45) for many years, but there are very few published data regarding CA-MRSA in Ireland (36). In order to increase our understanding of CA-MRSA so that effective infection control measures for combating the spread of CA-MRSA in both the community and the hospital can be established, it is vital that the clinical, epidemiological, and biological characteristics of CA-MRSA are determined. The aims of the present study were to determine whether carriage of pvl could be used as a surrogate marker for CA- MRSA in Ireland, to carry out comprehensive phenotypic and genotypic analyses of pvl-positive MRSA isolates recovered in Ireland, and to determine whether any additional phenotypic or genotypic characteristic could be used to aid in the recognition of CA-MRSA. MATERIALS AND METHODS Hospitals. MRSA isolates investigated for the presence of pvl genes were recovered from patients in the following hospitals: any of the 28 Irish hospitals from geographically diverse areas that participated in the European Antimicrobial Resistance Surveillance System (EARSS) during 2003 and reported patients with MRSA-positive blood cultures (n 18) (this level of participation provided an estimated population coverage of 89%; the 28 hospitals consisted of 7 tertiary referral university teaching hospitals, 2 regional hospitals, 12 general hospitals, 2 private hospitals, 2 maternity hospitals, 2 pediatric hospitals, and 1 long-stay care hospital) (27, 40); 1 maternity hospital (with 188 beds); 2 pediatric university teaching hospitals (1 with 236 and 1 with 150 beds); and 1 tertiary referral adult university teaching hospital (with 936 beds). Bacterial isolates. The sources and numbers of isolates investigated are summarized in Table 1. The study was performed in two phases. Initially, retrospective studies were undertaken to determine the proportion of pvl-positive isolates of MRSA recovered from the following sources: group I, patients with bloodstream infections in any of the 28 Irish EARSS participant hospitals during 2003 (n 430); group II, any site from patients in an adult tertiary referral hospital

3 2556 ROSSNEY ET AL. J. CLIN. MICROBIOL. with a widespread MRSA problem during the second quarter of 2003 (n 330); group III, any site from patients attending 1 maternity hospital and 1 pediatric hospital throughout 2003 (n 53); group IV, MRSA isolates referred to the Irish National MRSA Reference Laboratory (NMRSARL) between January 1999 and December 2005 with clinical or epidemiological details suggestive of possible community acquisition or where phenotypic characteristics (such as decreased susceptibility to fusidic acid and susceptibility to ciprofloxacin) (17, 28) suggested possible CA-MRSA (n 131); and group V, MRSA isolates with the phenotypic characteristics described above from blood culture specimens submitted to the NMRSARL for investigation under the EARSS project during 2004 and 2005 (n 44). The second phase of the investigation was a prospective study to determine the proportion of pvl-positive MRSA isolates among MRSA isolates recovered during the last quarter of 2004 from all patients attending the two pediatric hospitals (group VI; n 76) and the tertiary referral hospital (group VII; n 325). Epidemiological data were obtained to determine whether isolates were considered to be CA-MRSA, HA-MRSA, or HCA-MRSA according to the criteria published by the Centers for Disease Control and Prevention, Atlanta, GA ( MRSA strains were considered to be CA-MRSA strains if isolates were recovered from patients in outpatient settings or from inpatients within 48 h of hospital admission who were not known to have ever had MRSA, were not hospitalized or residents in long-stay care facilities, had not undergone surgery or dialysis during the previous 12 months, and did not have permanent indwelling catheters in situ (see the URL mentioned above). Isolates from outpatients suffering from cystic fibrosis and from children less than 12 months of age were considered to be HCA-MRSA. One isolate per patient was investigated unless a patient was found to carry more than one strain of MRSA. Control reference strains of S. aureus used in the present study are listed in Table 2. On receipt in the NMRSARL, all isolates were confirmed as MRSA as described below, inoculated onto Protect beads (Technical Service Consultants Limited, Heywood, United Kingdom), and stored at 70 C prior to subsequent detailed analysis. Chemicals, enzymes, and oligonucleotides. All chemicals used were of analytical grade or molecular biology grade and were purchased from Sigma-Aldrich Chemical Co. (Tallaght, Dublin, Republic of Ireland), BDH (Poole, Dorset, United Kingdom), Oxoid Ltd. (Basingstoke, Hampshire, United Kingdom), Merck (Nottingham, United Kingdom), Bio-Rad (Marnes La Coquette, France), Roche Diagnostics Ltd. (Lewes, East Sussex, United Kingdom), or Abtek Biologicals Ltd. (Liverpool, United Kingdom). Enzymes for molecular biology procedures were purchased from Promega Corporation (Madison, WI), Roche Diagnostics Ltd., or New England Biolabs, Inc. (Beverley, MA) and were used according to the manufacturers instructions. DNA molecular weight markers were purchased from Promega or New England Biolabs. Oligonucleotide primers were custom synthesized by Sigma-Genosys Biotechnologies Ltd. (Cambridge, United Kingdom). Identification and confirmation of methicillin resistance. All isolates were identified by staphylocoagulase production (39). Isolates yielding equivocal results were tested for the presence of clumping factor by using the Pastorex Staph-Plus kit (Bio-Rad), thermostable DNase, and if necessary, API Staph identification strips (biomérieux SA, Marcy-l Etoile, France) as described previously (37). Methicillin resistance was detected by using 10- g methicillin disks on Columbia blood agar (CBA; International Diagnostics Group PLC, Bury, United Kingdom) containing 7% (vol/vol) defibrinated horse blood as described previously (15). Isolates were also tested against 30- g cefoxitin disks on BBL Mueller- Hinton II agar (Becton Dickinson and Company, Sparks, MD) by using Clinical and Laboratory Standards Institute (CLSI) methodology (8). Isolates appearing susceptible to either antimicrobial were further tested against oxacillin (1- g and 5- g disks) on CBA at 30 C. Isolates were also tested for PBP2a production by using the Mastalex MRSA kit (Mast Diagnostics, Bootle, United Kingdom), and oxacillin MICs were determined by using the Etest system (AB Biodisk, Solna, Sweden). The presence of meca was investigated in any isolate exhibiting an oxacillin MIC of 4 mg/liter by PCR using the primers described previously by Oliveira and de Lencastre (30). AR typing. All isolates were typed by antibiogram-resistogram (AR) typing using disk diffusion as described previously, with the modification that susceptibility was determined by using the method recommended by the CLSI (8, 37, 38, 40). The pattern of susceptibility was determined with the following antimicrobials (figures in parentheses are the concentrations [ g/disk] used and breakpoints for antimicrobials not included in CLSI documentation or that differ from CLSI breakpoints [breakpoints are expressed in millimeters using the following TABLE 2. S. aureus reference strains used in the study Purpose of reference strain S. aureus reference strain Susceptibility testing ATCC Low-level resistance control for methicillin susceptibility testing DNA extraction and endonuclease digestion in all PFGE experiments ATCC NCTC Source or reference agr typing agr group 1 NCTC agr group 2 ATCC (Mu3) agr group 3 HT J. Etienne a agr group 4 RN pvl detection pvl-positive control HT J. Etienne SCCmec typing SCCmec I.1 (formerly I) Phenotype II 45 (43.2) SCCmec I (formerly IA) AR22 (0308) 45 SCCmec II.1 (formerly II) AR07.4 (0237) 45 SCCmec III.1 (formerly III) AR01 (0.1109) 45 SCCmec IVA JCSC SCCmec IV.1 (formerly IVa) CA05 24 SCCmec IV.2 (formerly IVb) 8/63-P 24 SCCmec IV.3 (formerly IVc) JCSC4788 T. Ito b SCCmec IV.4 (formerly IVd) JCSC Toxin gene detection of: seb NY19 46 seh FRI sea 42CR3-L 6 sea, sec, see, and tst FRI sed, seg, sei, and sej FRI eta and etb RN etd TY a Jerome Etienne, Centre National de Référence des Staphylocoques, Lyon, France. b Teruyo Ito, Department of Bacteriology, School of Medicine, Juntendo University, Tokyo, Japan. format: resistant {R}, intermediate {I}, and susceptible {S}]): amikacin (30) (R, 14; I, 15 to 19; and S, 20), ampicillin (10), cadmium acetate (130) (R, 10; I, 11 to 15; and S, 16), chloramphenicol (30), ciprofloxacin (5), erythromycin (15), ethidium bromide (60) (R, 13; I, 14; and S, 15), fusidic acid (10) (R, 23; I, 24 to 26; and S, 27), gentamicin (10), kanamycin (30), lincomycin (2) (R, 14; I, 15 to 16; and S, 17), mercuric chloride (10) (R, 13; I, 14; and S, 15), mupirocin (5) (R, 12; I, 13 to 9; and S, 20), mupirocin (200) (R, 15; I, 16 to 29; and S, 30), neomycin (30) (R, 15; I, 16 to 17; and S, 18), phenyl mercuric acetate (10) (R, 24; I, 25 to 28; and S, 29), rifampin (5), spectinomycin (500) (R, 13; I, 14 to 19; and S, 20), streptomycin (25) (R, 13; I, 14 to 15; and S, 16), sulfonamide (300), tetracycline (30), tobramycin (10), trimethoprim (5), and vancomycin (30). Isolates were assigned an AR type number based on the susceptibility pattern obtained as described previously (40). Biotyping. All isolates were characterized by a biotyping method which investigated hydrolysis of urea, hydrolysis of Tween 80, and pigment production as described previously (40). DNA macrorestriction digestion analysis. DNA macrorestriction digestion analysis was performed on all isolates. Chromosomal DNA was extracted from isolates as described previously (26). DNA was digested with the restriction endonuclease SmaI (Promega), fragments were separated by pulsed-field gel electrophoresis (PFGE), and analysis of banding patterns was performed using the GelCompar software package (version 4.1; Applied Maths, Belgium) (40). Visual inspection was used for final analysis, pattern differences were interpreted

4 VOL. 45, 2007 DIVERSE GENOTYPES OF pvl-positive CA-MRSA IN IRELAND 2557 as recommended by Tenover et al., and PFGE patterns were assigned alphanumeric designations as described previously (48). PFGE patterns were also assigned 5-digit pulsed-field types. These pulsed-field types were abbreviated to 2-digit pulsed-field groups (PFGs) and combined with AR typing results to yield AR-PFG types (40). DNA extraction. DNA was extracted from S. aureus isolates grown on either CBA or Trypticase soy agar (Oxoid) for 18 h at 35 C. Genomic DNA for use in MLST, SCCmec typing, and toxin gene detection was extracted using the QIAGEN DNeasy kit (QIAGEN, Crawley, United Kingdom) according to the manufacturer s instructions. DNA for pvl detection was extracted by the suspension of three to four colonies in 30 l sterile Tris-EDTA, ph 7.5, incubation at 100 C for 15 min, and placement on ice prior to immediate use. DNA for agr typing was extracted by suspending two to three colonies in 500 l sterile water and centrifuging at 12,000 g for 1 min. Cells were resuspended in 600 l of lysostaphin (80 g/ml), incubated at 37 C for 30 min, and centrifuged at 12,000 g for 2 min. Thereafter, DNA was extracted using InstaGene matrix solution (Bio-Rad) according to the manufacturer s instructions. Detection of pvl. All isolates were investigated for the carriage of pvl genes by PCR amplification as described previously (22). The identity of the PCR product was confirmed by sequencing using Big Dye terminator chemistry on an ABI 3730 genetic sequencer at the Genomics Core Facility, Queen s University, Belfast, Northern Ireland. Molecular characterization of pvl-positive MRSA isolates. In addition to AR typing, biotyping, and DNA macrorestriction analysis, all pvl-positive isolates were investigated by agr typing, MLST, SCCmec typing, and toxin gene detection. agr typing. All pvl-positive isolates were agr typed in a multiplex PCR using previously described primers specific for agr groups 1 to 4 (14). MLST. MLST was performed by PCR amplification, purification, and sequencing of an internal fragment of seven unlinked housekeeping genes by using a method and primer pairs described previously (12). SCCmec typing. The SCCmec type of all pvl-positive isolates was investigated using both a simplex technique and a multiplex SCCmec typing method because an earlier study of MRSA from Ireland found that the combination of both methods was required to identify previously described SCCmec types and local SCCmec variants (45). A previously described simplex SCCmec typing method was used to determine the ccr type and mec gene complex class (ccr-mec genes) harbored by each isolate (34). A multiplex SCCmec typing method utilizing nine primer pairs was used as described previously (30). The primer pairs in the multiplex assay were designed to amplify the partial nucleotide sequences of SCCmec types I to IV, including SCCmec subtypes I (formerly IA), III.1 (formerly IIIA), and III (formerly IIIB) (according to nomenclature suggested previously by Chongtrakool et al. [7]) and IVA (as described previously by Oliveira and de Lencastre [30]). Isolates found to carry SCCmec IV were subtyped using previously described primer pairs specific for the J1 region of SCCmec IV subtypes IV.1 (formerly IVa), IV.2 (formerly IVb), IV.3 (formerly IVc), and IV.4 (formerly IVd) (according to the nomenclature suggested previously by Chongtrakool et al. [7]). The primer pairs 4a1/4a2 (29), IVb-F/IVb-R (52), 4c1/4c2 (16), and IVd-F5/ IVd-R6 (52) were used. Detection of toxin genes. All pvl-positive isolates were tested for the presence of 13 staphylococcal toxin genes using a previously described method utilizing a series of multiplex PCRs (17). The targeted genes were the staphylococcal enterotoxin genes sea, seb, sec, sed, see, seg, seh, sei, and sej; the exfoliative toxin genes eta, etb, and etd; and the toxic shock toxin (TST) gene tst. Phenotypic detection of the toxic shock syndrome toxin TSST-1 was performed using a reverse passive latex agglutination kit (TST-RPLA test kit; Oxoid) as described previously (18). RESULTS AR typing, biotyping, and DNA macrorestriction analysis. In total, 1,389 MRSA isolates recovered between 1999 and 2005 were investigated for the carriage of pvl. The most frequently occurring AR-PFG type among all isolates was AR- PFG 06-01, a non-multiantibiotic-resistant phenotype with the ST22-MRSA-IV genotype (40). AR-PFG isolates exhibit resistance to -lactam antibiotics only or to -lactams and one or more of the following clinically useful antimicrobials: ciprofloxacin, erythromycin, fusidic acid, and trimethoprim. In contrast, isolates exhibiting multiantibiotic resistance phenotypes, in addition to resistance to -lactam antibiotics, exhibit resistance to a range of antimicrobials varying from 7 to 15 of the antimicrobials in the AR typing panel. The proportion of AR-PFG isolates varied from 80 to 85% (among isolates recovered from blood culture specimens and from the adult hospital) to 69 to 70% (among isolates from the pediatric hospitals). Among the former group of isolates, 1.5% exhibited a variety of non-multiantibiotic-resistant AR patterns, were urease positive, and yielded PFGE patterns assigned to a PFGE pattern group, PFG 99, reserved for sporadically occurring PFGE patterns as described previously (40). The proportion of such isolates was 21% among isolates from pediatric patients. Carriage of pvl. Twenty-five MRSA isolates (1.8%; 25/1,389) were pvl positive. Table 1 summarizes the proportions of pvlpositive isolates observed among the different groups of MRSA isolates investigated. In the retrospective studies, the lowest proportions of pvl-positive isolates were found among the predominantly HA-MRSA blood culture isolates in group I (0.5%; 2/430) and among isolates from patients in the adult tertiary referral hospital in group II (0.3%; 1/330) (Table 1). Among MRSA isolates recovered from pediatric and maternity patients (group III), the proportion of pvl-positive isolates was 1.9% but overall numbers were low (1/53). Subsequent epidemiological investigation showed that all four pvl-positive isolates detected during these retrospective studies (groups I to III) were CA-MRSA isolates (Table 1). Among isolates submitted to the NMRSARL where the clinical details of patients or phenotypic characteristics of isolates suggested possible CA-MRSA (group IV), 12.2% of isolates (16/131) were pvl positive and 75% (12/16) of pvl-positive isolates were CA-MRSA strains (Table 1). Epidemiological data and/or clinical details indicated that 48.9% (64/131) of all group IV isolates were HCA (either HA-MRSA or HCA- MRSA), and 41.2% (54/131) were CA-MRSA; data were unavailable for 9.9% (13/131) of isolates. Of the 54 CA-MRSA isolates, 55.6% (30/54) were from patients attending accident and emergency departments or outpatient departments or from external sources (such as general practitioners); 33.3% (10/30) of these were pvl positive. There was a statistically significant difference between the rates of carriage of pvl among CA-MRSA and HCA-MRSA (P ), but the difference in rates between isolates from adult and pediatric patients did not reach statistical significance (P ). The majority of CA-MRSA isolates (77.8%; 42/54) were pvl negative. Two of 44 isolates (4.6%) with phenotypic characteristics suggestive of pvl-positive MRSA recovered from blood culture specimens investigated during 2004 and 2005 (group V) carried pvl (Table 1). This figure represents a proportion of 0.2% (2/983) of all patients from whom blood culture isolates were submitted for investigation during this period. In the prospective study, overall 0.8% of isolates (3/401) were pvl positive but the proportion varied from 0.3% among isolates from adults (group VI) to 2.6% from isolates from pediatric patients (group VII); this difference was not significant (P ), but the number of pvl-positive isolates was small. The proportions of HA-MRSA, HCA-MRSA, and CA- MRSA strains among the 401 isolates studied were 86% (280/ 325), 8% (26/325), and 6% (19/325), respectively, in the adult

5 2558 ROSSNEY ET AL. J. CLIN. MICROBIOL. TABLE 3. Phenotypic characteristics of 25 pvl-positive MRSA isolates recovered from patients in Ireland between 1999 and 2005 a Isolate (AR type) Year of isolation Hospital source Ethnicity of patient or place of origin Age (yr) Sex Clinical details Acquisition category Urease production Oxacillin MIC (mg/liter) Resistance pattern d E1401 (AR06) 2003 EARSS Irish 69 M Bacteremia CA 32 AMP, ERY E1760 (NT) 2004 EARSS Irish 28 M Bacteremia HA 256 AMP, CAD, CIP, ERY, FUS E1394 (NT) 2003 EARSS Asian 34 M Buttock abscess; bacteremia CA 256 AMP, CAD M04/0069 (NT) 2004 NMRSARL African 27 F Skin abscess CA 32 AMP, CAD M05/0099 (NT) 2005 NMRSARL African 28 F Staff screening during outbreak HCA 16 AMP, CAD M04/0244 (NT) 2004 Pediatric Irish 6 F Dermatology b CA 128 AMP, CAD, FUS M04/0071 (NT) 2004 NMRSARL Irish 5 M Buttock abscess c CA 16 AMP, CAD, FUS M04/0072 (NT) 2004 NMRSARL Irish 3 M Colonized c CA 32 AMP, CAD, FUS M04/0073 (NT) 2004 NMRSARL Irish 35 M Colonized c CA 8 AMP, CAD, FUS M04/0074 (NT) 2004 NMRSARL Irish 33 F Colonized c CA 32 AMP, CAD, FUS M05/0076 (NT) 2005 NMRSARL Irish 1 M Ear discharge HCA 16 AMP, CAD, FUS MF181 (NT) 2004 Adult Irish 39 F Surgical site infection HCA 16 AMP, CAD, FUS M05/0134 (NT) 2005 NMRSARL Irish 40 M Eyebrow folliculitis CA 32 AMP, CAD, CIP, FUS M05/0100 (NT) 2005 NMRSARL Irish 46 F Staff screening during outbreak HCA 2.0 AMP, CIP, ERY M05/0028 (Unf-1) 2005 NMRSARL African 39 M Buttock abscess CA 256 AMI, e AMP, ERY, KAN, NEO M05/0259 (Unf-1) 2005 NMRSARL Unknown 36 M Face infection CA 64 AMP, ERY, KAN, NEO M04/0266 (Unf-1) 2005 Pediatric USA 1 M Inguinal lymphadenitis CA 4.0 AMI, e AMP, ERY, KAN, NEO M05/0060 (Unf-1a) 2005 NMRSARL Irish 26 M Skin scalp abscess CA 64 AMI, e AMP, CHL, ERY, KAN, LIN, NEO M05/0199 (Unf-1b) 2005 NMRSARL Irish 23 M Pneumonia (travel to the United States) CA 16 AMP, CIP, ERY, KAN, NEO ML224 (Unf-1c) 2003 Adult Irish 21 F Skin abscess CA 16 AMP, CAD, CIP, ERY, KAN, NEO M05/0267 (Unf-1d) 2005 NMRSARL Irish NK M Perianal abscess CA 128 AMI, e AMP, CIP, ERY, KAN, NEO, TET M04/0101 (Unf-2) 2003 Maternity Asian 1 F Colonized CA 2.0 AMP, CAD, FUS, KAN, NEO, STR, TET M02/0088 (Unf-2a) 2002 NMRSARL Middle Eastern 25 F Episiotomy wound infection HCA 8.0 AMP, CAD, CHL, FUS, KAN, NEO, STR, TET E1758 (NT) 2004 EARSS Asian 26 M SSTI; osteomyelitis; bacteremia CA 256 AMP, CAD, CIP, SUL, TET M04/0068 (NT) 2004 NMRSARL African 21 F Skin abscess CA 16 AMP, CAD a National MRSA Reference Laboratory (NMRSARL), isolates were submitted for investigation from various Irish hospitals. NT, no AR type was assigned to these isolates because, although the patterns suggested AR type AR06, the isolates were urease positive and previous studies had shown that similar isolates required confirmation by molecular typing (40); Unf, unfamiliar AR pattern; SSTI, staphylococcal skin and soft tissue infection; NK, not known; M, male; F, female. b This patient was a sibling of a patient with an MRSA infection prior to the start of the study period; the isolate from the first patient was unavailable for investigation. c Four isolates were recovered from four members of one family. One member of the family was infected (MRSA buttock abscess), and the other three were colonized. d Antimicrobials tested: AMI, amikacin; AMP, ampicillin; CAD, cadmium acetate; CHL, chloramphenicol; CIP, ciprofloxacin; ERY, erythromycin; ethidium bromide; FUS, fusidic acid; GEN, gentamicin; KAN, kanamycin; LIN, lincomycin; mercuric chloride; mupirocin; NEO, neomycin; phenyl mercuric acetate; rifampin; spectinomycin; STR, streptomycin; SUL, sulfonamide; TET, tetracycline; tobramycin; trimethoprim; and vancomycin. e Intermediately susceptible to amikacin. tertiary referral group and 5% (4/76), 80% (61/76), and 15% (11/76), respectively, in the adult and pediatric hospitals. The proportion of CA-MRSA strains among the pvl-positive isolates was 67% (2/3). The two pvl-positive isolates from pediatric patients were CA-MRSA strains, while the one isolate from an adult was an HCA strain. Three additional pvl-positive S. aureus isolates were detected but were found to be susceptible to methicillin by disk diffusion and MIC determination and lacked the meca gene. These isolates were not investigated further in the present study. There was a statistically significant difference between the proportions of pvl-positive MRSA strains among CA-MRSA isolates and those among HCA- MRSA isolates (including both HA and HCA isolates) (P ). Nevertheless, only 6.7% (2/30) of CA-MRSA isolates carried pvl. When data from isolates in group IV (NMRSARL isolates) were combined with data from isolates in the prospective study (groups VI and VII), the statistical difference between CA- MRSA and HCA-MRSA isolates was extremely significant (P ), but only 16.7% (14/84) of CA-MRSA isolates were pvl positive. Clinical details of patients with pvl-positive MRSA. Clinical details of the patients with pvl-positive MRSA are shown in Table 3. The mean age was 25.5 years (range, 1 to 69 years). The age of one patient was unknown. Fifteen patients were male, and 10 were female. Skin infection, in particular abscess formation, accounted for 52% (13/25) of clinical presentations; one patient presented with pneumonia, and four presented with bloodstream infections. Six patients, including two health care workers with isolates detected during screening for

6 VOL. 45, 2007 DIVERSE GENOTYPES OF pvl-positive CA-MRSA IN IRELAND 2559 TABLE 4. Molecular characteristics of 25 pvl-positive MRSA isolates recovered from patients in Ireland between 1999 and 2005 Isolate PFGE pattern a agr group MLST CC ST Allelic profile Simplex (ccr-mec) SCCmec typing PCR Multiplex SCCmec IV subtype b Toxin gene profile c E1401 A (01003) B IV IV.1 seg sei E1760 A1 (01018) 1/3 d B IV None sec seg sei E1394 C (02033) B IV IV.3 seg sei M04/0069 C (02033) B IV IV.3 seg sei M05/0099 C (02033) B IV IV.3 seg sei M04/0244 C1 (02034) B IV None seg sei M04/0071 C1 (02034) B IV None seg sei tst M04/0072 C1 (02034) B IV None seg sei tst M04/0073 C1 (02034) B IV None seg sei tst M04/0074 C1 (02034) B IV None seg sei tst M05/0076 C1 (02034) B IV None seg sei MF181 C2 (02040) B IV None seg sei tst M05/0134 C1 (02034) B IV None seg sei tst M05/0100 D (99023) B IV None Negative M05/0028 D (99023) B IV IV.1 Negative M05/0259 D (99023) B IV IV.1 Negative M04/0266 D1 (99017) B IV IV.1 Negative M05/0060 D2 (99024) B IV IV.1 Negative M05/0199 D3 (99025) B IV IV.1 seg sei ML224 D (99023) B IV IV.1 Negative M05/0267 D4 (99028) B IV IV.1 Negative M04/0101 F (99019) B IV IV.3 etd M02/0088 F1 (99020) B IV IV.3 etd E1758 E (99021) 3 Unknown e B IV None Negative M04/0068 B (03050) B IV IV.1 sea seg sei a PFGE patterns are assigned an alphanumeric pattern as described previously (48); MRSA isolates are assigned 5-digit numbers, where epidemiological relatedness is inferred among isolates exhibiting PFGE patterns within the same series of 1,000 numbers except for series 99000, which is reserved for sporadically occurring patterns (40). b Each pvl-positive S. aureus isolate that demonstrated SCCmec IV type 2 ccr-class B mec amplimers by the simplex PCR method was subtyped using primers specific for SCCmec IV subtypes IV.1, IV.2, IV.3, and IV.4. Isolates that could not be subtyped using these primers are indicated as having none of the SCCmec IV subtypes investigated. c The toxin genes, other than pvl, investigated by PCR were the staphylococcal enterotoxin genes (sea, seb, sec, sed, see, seg, seh, sei, and sej), the exfoliative toxin genes (eta, etb, and etd), and the toxic shock toxin gene (tst). d This isolate yielded an equivocal result, suggesting agr group 1 and/or group 3. e Unknown, unknown clonal complex. MRSA, were asymptomatic. Acquisition was deemed CA for 19 patients (76%), HCA for 5 patients (20%), and HA for 1 patient (4%). In total, 9/25 (36%) pvl-positive isolates were from patients of non-irish ethnic origin; the ethnicity of one patient was unknown. One patient of Irish ethnic origin was diagnosed following travel to the United States. Phenotypic characterization of pvl-positive MRSA isolates. The antimicrobial resistance patterns of the 25 pvl-positive MRSA isolates are shown in Table 3. Sixty percent of isolates (15/25) exhibited non-multiantibiotic-resistant AR patterns and were urease positive. Most of the remaining isolates (36%; 9/25) exhibited one of two unfamiliar patterns and were resistant to a larger number of antimicrobials, including a range of aminoglycosides (Table 3). Overall, 96% of MRSA isolates (24/25) were urease positive. Rates of resistance to fusidic acid, ciprofloxacin, and erythromycin were 44% (11/25), 28% (7/25), and 40% (10/25), respectively. Only 16% of isolates (4/25) exhibited oxacillin MICs of more than 256 mg/liter; the MIC range of the remaining isolates was 2.0 to 28 mg/liter (Table 3). The two isolates with oxacillin MICs of 2 mg/liter were meca and PBP2a positive and were resistant by disk diffusion. Molecular characterization. Details of PFGE type, agr group, MLST, SCCmec type, and toxin gene profiles of the 25 pvl-positive isolates are shown in Table 4. PFGE analysis. As shown in Table 4, pvl-positive isolates exhibited six apparently unrelated PFGE patterns (A to F) according to the criteria of Tenover et al. (48). Almost half the isolates (44%; 11/25) yielded pattern C, while 32% (8/25) yielded pattern D. agr typing. Fifty-six percent of isolates (14/25) belonged to agr group 3, while 36% (9/25) were assigned to agr group 1 (Table 4). One isolate belonged to agr group 2, and one yielded an equivocal result, indicating that it belonged to agr group 1 and/or group 3. MLST. Six MLST STs were identified among the 25 pvlpositive isolates investigated (Table 4). ST30 was most prevalent (44%; 11/25), followed by ST8 (32%; 8/25), ST22 (8%; 2/25), ST80 (8%; 2/25), ST5 (4%; 1/25), and ST154 (4%; 1/25). The pvl-positive MRSA isolates from patients of Irish nationality (n 15) belonged to three genotypes (ST22, ST30, and ST8), whereas isolates from patients of non-irish ethnic origin (n 9) exhibited five genotypes (ST30, ST8, ST80, ST154, and ST5); the ethnic origin of one patient was unknown (the genotype of this patient s isolate was ST8) (Tables 3 and 4). Five of the six STs identified clustered into five clonal complexes (CCs) (CC30, CC22, CC8, CC5, and CC80) (Table 4). ST154 belongs to an as yet unnamed CC that differs from

7 2560 ROSSNEY ET AL. J. CLIN. MICROBIOL. any previously described CC and includes only ST154 and its single locus variant, ST96. Correlation between PFGE, agr, and MLST typing. There was 100% correlation between the assignment of isolates to PFGE types and CCs (Table 4). agr grouping showed that agr group 1 was common to CC8 and CC22 isolates, while agr group 3 was shared by CC30 and CC80 isolates. The single agr group 2 isolate belonged to CC5 (Table 4). SCCmec typing. All isolates harbored SCCmec IV (Table 4). SCCmec IV subtyping showed that 36% of isolates (9/25) harbored SCCmec IV.1, 20% (5/25) harbored SCCmec IV.3, and 44% (11/25) failed to subtype using primers specific for SCCmec IV subtypes IV.1 to IV.4. No isolate harbored SCCmec IV.2 or IV.4 (Table 4). Genotypes. Six genotypes (ST-SCCmec type) were identified among the pvl-positive MRSA isolates, the predominant being ST30-MRSA-IV (subtype IV.3 or nonsubtypeable) (44%; 11/ 25), followed by ST8-MRSA-IV (subtype IV.1 or nonsubtypeable) (32%; 8/25), ST22-MRSA-IV (subtype IV.1 or nonsubtypeable) (8%; 2/25), ST80-MRSA-IV.3 (8%), ST5-MRSA- IV.1 (4%), and ST154-MRSA-IV (nonsubtypeable) (4%) (Table 4). Detection of toxin genes. Six staphylococcal toxin genes were detected among the 25 pvl-positive isolates investigated (Table 4). The most common were the enterotoxin genes seg (60%; 15/25) and sei (60%; 15/25) and the toxic shock toxin gene tst (24%; 6/25). The six isolates carrying the tst gene all expressed the TST toxin TSST-1 phenotypically. Other toxin genes detected included the enterotoxin genes sea (4%; 1/25) and sec (4%; 1/25) and the exfoliative toxin etd (8%; 2/25). All ST30 and ST22 isolates harbored seg and sei, but six ST30 isolates also carried tst (Table 4). One ST22 isolate harbored sec in addition to seg and sei. The single ST5 isolate investigated had the toxin profile sea seg sei (Table 4). The two ST80 isolates were the only isolates harboring etd. With the exception of pvl, none of the toxin genes investigated were detected in seven of the eight ST8 isolates (Table 4). The remaining ST8 isolate harbored seg and sei. The enterotoxin genes seb, sed, see, seh, and sej and the exfoliative toxin genes eta and etb were not detected among any of the isolates tested. DISCUSSION The results of this study show that 76% of pvl-positive MRSA isolates recovered in Ireland as part of the present investigation were CA, confirming an earlier preliminary observation that CA-MRSA is an emerging problem in Ireland (36). However, although the overall proportion of MRSA isolates carrying pvl was 1.8% among the 1,389 isolates investigated, during the prospective study where 7.5% (30/401) of isolates were CA-MRSA strains, only 6.7% of isolates (2/30) from patients with an epidemiological history of CA-MRSA carried pvl genes and the carriage of pvl was not restricted to CA-MRSA. Similarly, 78% (42/54) of CA-MRSA isolates referred to a central reference facility for investigation of the carriage of pvl were negative whereas 25% (4/16) of pvl-positive isolates in this group were HCA. These findings agree with recent reports that carriage of pvl cannot be used as a sole marker for CA-MRSA (9, 41). Overall, 24% of pvl-positive isolates were from patients with HCA risk factors, but distinguishing between CA-MRSA, HA- MRSA, and HCA-MRSA using epidemiological data was difficult in some cases. Using the recommendation from the Centers for Disease Control and Prevention that isolates from patients with hospital admissions within the previous 12 months should be considered HCA-MRSA strains may result in a number of CA-MRSA isolates, especially those from very young children, being erroneously considered HCA-MRSA strains ( For example, neither pediatric hospital in the present study has a widespread MRSA problem; hence, categorizing MRSA isolates from some patients aged less than 1 year as HCA may be inaccurate. Furthermore, many of the pediatric patients suffered from cystic fibrosis and had frequent health care contact, and their isolates were considered to be HCA-MRSA strains, but some of these isolates may have been CA. In addition, time-based criteria are problematic and it is now known that CA-MRSA can give rise to outbreaks in hospitals (43). Descriptions of MRSA isolates recovered from patients in nursing homes as CA-MRSA strains further complicate definitions. In Ireland, there is an additional problem because MRSA isolates similar to the predominant HA-MRSA strain (ST22-MRSA-IV) have been recovered in veterinary practice (32). For these reasons, further studies are required to determine the true prevalence of CA-MRSA in Ireland. The clinical presentations of patients from whom pvl-positive MRSA isolates were recovered were generally similar to those reported in other studies in that the predominant presentation was skin and soft tissue infection (4). Two patients presented with surgical site infection by HCA-MRSA, and two isolates were recovered from asymptomatic health care workers. The isolates from the two staff members were detected during an investigation of a putative nosocomial outbreak, and although each carried a different strain (ST30 or ST8), the carriage of pvl-positive MRSA by health care workers increases the risk that nosocomial spread may occur. Surgical site infection with pvl-positive HCA-MRSA also provides evidence for the potential of nosocomial transmission. Phenotypic and genotypic analyses showed that the pvl-positive MRSA isolates identified were generally unlike the current predominant MRSA population in Irish hospitals (i.e., they were not AR-PFG type 06-01, genotype ST22-MRSA-IV) as only two of the pvl-positive MRSA isolates had this profile. The detection of pvl in this prevalent nosocomial epidemic strain, which is similar to the United Kingdom epidemic strain EMRSA-15, is a cause of serious concern because of the potential increased virulence associated with pvl-positive strains (4, 42). An investigation of pvl carriage among S. aureus in the United Kingdom has shown that methicillin-susceptible S. aureus isolates with the ST22 genotype were also pvl positive (17). Unlike earlier reports that specific individual genotypes are associated with particular geographic areas, six distinct genotypes (ST30, ST8, ST22, ST80, ST5, and ST154) were identified among pvl-positive MRSA isolates in Ireland (50). This relatively diverse range of genotypes from a small country may reflect recent significant changes in the characteristics of the Irish population, which has changed from being very homogeneous up until the 1990s to an increasingly heterogeneous population, where 5.8% of respondents to the 2002 national population census reported their nationality as non-irish,

8 VOL. 45, 2007 DIVERSE GENOTYPES OF pvl-positive CA-MRSA IN IRELAND 2561 with 0.5, 0.5, and 0.4% coming from Africa, Asia, and the United States, respectively ( /personsclassbyplaceofbirth2002.htm). Reflecting this change in population, of the pvl-positive MRSA isolates investigated in the present study, 36% (9/25) were recovered from patients of non- Irish nationality, with 16 (4/25), 12 (3/25), and 4% (1/25) coming from Africa, Asia, and the United States, respectively. Furthermore, of the six genotypes identified, only three were found among the 15 MRSA isolates from patients of Irish nationality (ST22 [n 2]; ST30 [n 8]; ST8 [n 5]), whereas five genotypes were identified among the nine MRSA isolates from patients of non-irish ethnic origin (ST30 [n 3]; ST8 [n 2]; ST80 [n 2]; ST154 [n 1]; ST5 [n 1]; the ethnic origin of one patient was unknown). One ST80 isolate was recovered from a patient from the Middle East, and one ST8 MRSA was isolated from an Irish patient following travel to the United States. This suggests that one reason for the greater genotypic diversity of pvl-positive MRSA from Ireland than that in studies from other centers may be that strains are being imported; the proportion of patients of non-irish ethnic origin among patients with pvl-positive MRSA was sixfold higher than the proportion among the general population. Interestingly, the single pvl-positive ST154-MRSA-IV isolate identified in this study was recovered from a patient from central Asia. The only previous report of the ST154 genotype comes from a study of pvl-positive MRSA isolates from Mongolian hospitals, where three of six isolates investigated had the genotype ST154-MRSA-IV.3 and had characteristics suggestive of CA-MRSA (33). These authors also reported that a CA-MRSA strain with the ST154 genotype was recovered in Sweden from a child of Mongolian parents (33). Genotypes ST8, ST1, ST59, and more recently, ST30 have been associated with CA-MRSA in the United States (9, 11, 50). Isolates with the ST8-MRSA-IV genotype include the pvl-positive strain ST8USA300 and the pvl-negative strain ST8USA500 (9, 49). ST8USA500 isolates tend to cause sporadic disease but are being displaced by ST8USA300, which is epidemic in some communities in at least 16 states in the United States (9). ST8USA300, which tends to be highly transmissible and virulent, was recently sequenced and shown to carry a novel mobile genetic element, termed the arginine catabolic mobile element, encoding an arginine deiminase pathway and an oligopeptide permease system that may act as an additional virulence factor. ST8USA300 also carries genes encoding staphylococcal enterotoxins Q and K (9, 10). In the present study, pvl-positive isolates were not investigated for the carriage of sek and seq but the detection of these enterotoxins, together with the recently described characteristic signature AT repeat sequence within the conserved hypothetical gene SACOL0058 located within SCCmec, would confirm whether the ST8-MRSA-IV.1 isolates from Ireland belong to the ST8USA300 clone (3). In Europe, the carriage of pvl genes has been associated predominantly with ST80 (11, 13, 17, 50). In contrast, in the present study, only two isolates exhibited the genotype ST80- MRSA-IV, whereas the genetic background of the majority of isolates was from genotypes ST30 (44%) and ST8 (32%). CA- MRSA isolates with the ST30 genotype were originally reported from Australia, New Zealand, and Samoa, and the genotype consequently became known as the Southwest Pacific clone (50). However, the potential of the ST30 clone to spread worldwide has subsequently been reported (47). Four of the ST30 isolates identified in the present study represent an outbreak in a family where a 5-year-old boy was infected and a sibling and both parents were colonized. One ST8 isolate was recovered from a patient in a family where 1 year later, a sibling presented with an indistinguishable isolate, while one of the two ST22-MRSA-IV isolates was recovered from a bloodstream infection in a patient with diabetes who 2 years later presented with a second bloodstream infection from which an indistinguishable ST22-MRSA-IV isolate was recovered. Previous studies reported that pvl-positive MRSA isolates were associated with the agr group 3 background (50). Although 56% of isolates in the present study belonged to agr group 3, our results reflected the fact that the agr 3 background was common to isolates in clonal complexes CC30 and CC80, whereas agr group 1 was common to both CC8 and CC22. Only one isolate exhibited agr group 2, and this was the only isolate with the genetic background ST5-MRSA-IV. All pvl-positive MRSA isolates in the present study harbored SCCmec IV, but only 56% (14/25) could be subtyped. Isolates exhibiting SCCmec IV subtypes were found to harbor either subtype IV.1 (36% of isolates) or subtype IV.3 (20% of isolates); no isolates harbored SCCmec IV.2 or IV.4. The prevalence of subtypes IV.1 and IV.3 among pvl-positive MRSA isolates in the present study was not surprising as these are the most frequently identified subtypes among pvl-positive CA- MRSA isolates (2, 17, 49). Other studies also report SCCmec IV elements which fail to subtype (2, 23). It has been suggested that SCCmec IV may be highly recombinant, especially in relation to the J1 region, and that this results in the emergence of novel subtypes. Isolates that failed to subtype in the present study belonged to the following STs, ST30, ST8, ST22, and ST154, but other isolates with ST30, ST8, and ST22 STs did subtype (ST22-MRSA-IV.1 [n 1], ST8-MRSA-IV.1 [n 7], and ST30-MRSA-IV.3 [n 3]). The pvl-positive isolates in the present study carried genes encoding an array of toxins, but no specific toxin gene profile was associated with pvl-positive CA-MRSA (n 19). Although 58% (11/19) of isolates harbored the enterotoxin genes seg and sei either alone or in combination with sea or tst, 11% (2/19) harbored only etd and 37% (7/19) had none of the toxin genes investigated. The two etd-positive isolates had the ST80 genetic background, confirming previous reports of the association between etd, pvl, and the ST80 genotype (17). The number of pvl-positive MRSA isolates (n 25) carrying the tst gene and expressing TSST-1 (24%; 6/25) was unexpectedly high. Five of the six isolates were CA-MRSA strains, one was an HCA- MRSA strain, and all had the ST30 genotype. In contrast, Diep et al. found that none of 29 pvl-positive ST30 MRSA isolates harbored tst (9). Interestingly, in that study, all 16 ST36- MRSA-II (USA200) isolates, which like ST30 isolates belonged to CC30 but were pvl negative, carried tst. ST36- MRSA-II (USA200) isolates are similar to United Kingdom strain EMRSA-16 (ST36-MRSA-II), which also carries tst (1), and to Irish strain AR-PFG (ST36-MRSA-II) (40). The emergence of a CA-MRSA strain in Ireland carrying tst in addition to pvl is alarming, especially in a CC that includes a major epidemic nosocomial strain, as the toxins encoded by