Community-associated Methicillin-resistant Staphylococcus aureus, Canada

A total of 184 methicillin-resistant Staphylococcus aureus (MRSA) strains were collected from patients who sought treatment primarily for skin and soft tissue infections from January 1, 1999, to March 31, 2002, in east-central Saskatchewan, Canada. Molecular subtyping analysis using pulsed-field gel electrophoresis showed 2 major clusters. Cluster A (n = 55) was composed of a multidrug-resistant MRSA strain associated with a long-term care facility and was similar to the previously reported nosocomial Canadian epidemic strain labeled CMRSA-2. Cluster B (n = 125) was associated with cases identified at community health centers and was indistinguishable from a community-associated (CA)-MRSA strain identified previously in the United States (USA400). Cluster B remained susceptible to a number of classes of antimicrobial agents and harbored the lukF-PV and lukS-PV toxin genes. Over 50% of both clonal groups displayed high-level resistance to mupirocin. This is the first report of the USA400 strain harboring the lukF-PV and lukS-PV toxin genes in Canada.


Setting
The investigation focused on an area in east-central Saskatchewan. This region consisted of a city of ≈7,127 persons that was serviced by a hospital with a central laboratory in which the MRSA strains were identified. This locale also contained a number of community health centers and a long-term care facility housing ≈100 persons. The region also included a number of smaller communities that consisted of small rural First Nations and Metis communities (aboriginal populations) and small towns.

Case Definitions
Surveillance for MRSA was laboratory-based and involved nonrepeat MRSA cases. Case-patients were residents of east-central Saskatchewan with laboratory-confirmed MRSA infections identified from January 1, 1999, to March 31, 2002, by the local hospital or provincial laboratory. All MRSA isolates identified by the hospital laboratory were subsequently confirmed at the provincial laboratory by using standard protocols. Residential status was determined by the location of the patient's treatment facility because case-patient address information was unavailable. General patient demographic information was collected (date of birth, sex, date of sample collection, and invasiveness) for all case-patients. Invasive infections were defined according to the guidelines from the Centers for Disease Control and Prevention Active Bacterial Core Surveillance Program and included obtaining isolates from a normally sterile site, such as blood, cerebrospinal fluid, pleural fluid, peritoneal fluid, pericardial fluid, surgical aspirate, bone, joint fluid, or internal body site (www.cdc.gov/ncidod/dbmd/abcs/meth-case.htm). Antimicrobial susceptibility testing by microbroth dilution was performed according to NCCLS recommendations (13). Breakpoints used for mupirocin were as follows: susceptible, MIC <4 mg/L; low-level resistance, MIC >4 and <256 mg/L; high-level resistance, >256 mg/L (14).

Molecular Characterization of MRSA Strains
mecA and nuc genes from MRSA isolates were coamplified with a multiplex real-time polymerase chain reaction (PCR) assay. Nucleic acid was isolated from 4-5 colony picks by boiling in a 2% (wt/vol) homogeneous suspension of Chelex 100 resin (Bio-Rad Laboratories Ltd, Mississauga, Ontario, Canada). Primer and probe sequences with their reaction concentrations are shown in Table 1. Master mix was composed of the Applied Biosystems TaqMan PCR Core kit (Applied Biosystems, Foster City, CA, USA) that uses reaction concentrations of: 1x PCR buffer A, 4.0 mmol/L MgCl 2 , 200 µmol/L dATP, 200 µmol/L dCTP, 200 µmol/L dGTP, 400 µmol/L dUTP, 1.25 U AmpliTaq Gold, and 0.5 U uracil-DNA-Nglycosylase (UNG) for carryover prevention. Thermal cycling and data collection were performed on an ABI Prism 7700 Sequence Detector using the following conditions: 2 min at 50°C, 10 min at 95°C, followed by 55 cycles of 95°C for 15 s and 60°C for 1 min. Amplification was confirmed for each target by the generation of a sigmoid amplification plot.
lukF-PV and lukS-PV detection was carried out by using PCR with primers and protocols previously described (15). PCR was used to detect the exfoliative toxin genes eta and etb as previously described (16).
Isolates were subtyped by using pulsed-field gel electrophoresis (PFGE) after digestion with SmaI following the Canadian Standardized Protocol as described previously (17). PFGE-generated DNA fingerprints were digitized and analyzed with BioNumerics Ver. 3.5 (Applied Maths, Sint-Martens-Latem, Belgium) by using a position tolerance of 1.0 and an optimization of 1.0. Relatedness was determined following established criteria with major strain clusters (designated by a letter) grouped with banding patterns of <7 band changes (18). Fingerprints were compared to those in the national MRSA fingerprint database, which comprised of >600 unique MRSA fingerprint types. Isolates with specific DNA profiles were grouped into 1 of 6 Canadian epidemic strains of MRSA (CMRSA-1, CMRSA-2, CMRSA-3, etc.) as previously described (19,20). Multilocus sequence typing (MLST) was conducted on a representative isolate from each of the unique PFGE types as previously described (21). Staphylococcal chromosome cassette (SCC) mec typing was conducted as previously described (22).

Epidemiologic Analysis of MRSA
Before 1999, MRSA was rarely isolated in east-central

Methicillin-resistant Staphylococcus aureus, Canada
Saskatchewan. It was identified only 2 other times in persons from this region since 1996. However, in April 1999, a resident of a long-term care facility in the region under study tested positive for MRSA. This patient had recently been hospitalized in Saskatoon, Saskatchewan, and tested positive for MRSA shortly after being transferred back to the long-term care facility. Following this case, an additional 183 nonrepeat MRSA were isolated from infected persons from this region between January 1999 and April 2002. The annual rates of MRSA-related infections in the region under study were 3.1/1,000 persons in 1999, 4.8/1,000 persons in 2000, and 14.6/1,000 persons in 2001. In 1 specific community within the study area with a population of approximately 1,400 persons, MRSA rates were 0/1,000 in 1999, 7.1/1,000 in 2000, and 46/1,000 in 2001. Only a single isolate was considered invasive, having causes a blood infection. Approximately 18% (n = 33) of the cases were identified from a single long-term care facility and made up of >50% of cases identified between April 1999 and June 2000 ( Figure 1A). After June 2000, a larger cluster of cases emerged, which peaked in October 2001 and were primarily identified from patients presenting at local health clinics or nursing stations from surrounding communities ( Figure 1A).

Characterization of Clonal Group A Isolates
Clonal group A isolates were first identified on April 12, 1999, from a patient in a long-term care facility who had been transferred from a hospital in Saskatoon, Saskatchewan. A rapid increase in cases was related to this clonal group from persons from the long-term care facility (n = 33, 60%) as well as community health centers (n = 22, 40%) during the next few months ( Figure 1B). Cases relat-ed to this group peaked June 1999; however, isolates continued to be identified over the course of this study from both patients in long-term care facilities and community health centers (Figure 1). Clonal group A strains were found to cause a large number of noninvasive infections. Only a single isolate was identified as causing an invasive blood infection. A slightly higher number of female patients 58% (n = 32) were identified with clonal group A-related infections. Approximately 64% (n = 35) of these strains were identified from persons >59 years of age, although 29% (n = 16) of cases were identified in persons <10 years of age ( Figure 3). Of the 22 cases reported from the community health centers, 73% (n = 16) were <18 years of age. Comparison of the representative PFGE clonal group A1 fingerprint pattern to the Canadian National Fingerprint Database showed that these strains were similar to the previously identified epidemic MRSA-labeled CMRSA-2 (19,20). A representative strain from each of the unique fingerprint patterns from the A clonal group was typed by using MLST, and all were found to be of the sequence type ST5. All of the unique A subtypes as well as a 25% (n = 12) sample of strains that were of the A1 subtype did not harbor the lukF-PV, lukS-PV, eta, and etb toxin genes. SCCmec typing of 4 A1 subtypes and subtypes A2, A3, A4, A5 showed all contained SCCmec IIIb. Most clonal group A isolates were resistant to ciprofloxacin, erythromycin, gentamicin, tetracycline, mupirocin, and fusidic acid ( Table 2).

Characterization of Clonal Group B Isolates
The first reported case of infection due to clonal group B occurred 16 months after the description of first reported MRSA infection on August 31, 2000, and the cases continued to increase with a peak of occurring in October 2001 ( Figure 1B). Clonal group B correlated with a large number of noninvasive infections reported from community health centers, and only 1 report was made of a longterm care patient with MRSA type 0142 infection. The distribution by sex was similar; 49% of patients were female (n = 62) and 51% were male (n = 64). More than 67% (n = 85) of these strains were identified from persons <20 years of age, although ≈17% (n = 21) of cases were identified in persons >40 years (Figure 3).
The PFGE B1 fingerprint pattern (CDN type 142) was indistinguishable from the USA CA-MRSA strain labeled USA400 ( Figure 2) (21). MLST was conducted on 1 representative strain from each clonal group B subtype, and all were the same sequence type (ST1). In addition, these MRSA strains were shown to contain SCCmec type IV. A selection of ≈25% of the PFGE type 142 strains (n = 33), including all of the unique fingerprint patterns B2 and B3, was shown by PCR to harbor the lukF-PV and lukS-PV genes, and none of the strains tested harbored the eta and etb toxin genes. Clonal group B strains were susceptible to most antimicrobial agents tested; however, 40% and 55% of the isolates displayed resistance to erythromycin and high-level resistance to mupirocin, respectively ( Table 2).

Characterization of Other Molecular Subtypes
Three other unique PFGE subtype groups were identified; these were composed of single isolates labeled C1, D1, and E1, respectively ( Figure 2). None of these isolates carried the lukF-PV and lukS-PV toxin genes or the  exfoliative toxins. Similar to the PFGE B clonal group, these strains remained susceptible to most non-β-lactam drugs tested (data not shown).

Discussion
A dramatic increase of MRSA in the east-central region of Saskatchewan has been documented in this study. Molecular typing analysis has shown that 2 major unrelated strain clusters were in circulation in the communities under study. The emergence of clonal group A was linked to a patient in a long-term care facility who had recently been transferred from a hospital in Saskatoon, and this strain began to spread within the facility. Comparisons to the Canadian National Fingerprint Database showed the A1 pattern is related to the CMRSA-2 epidemic strain cluster observed in Canadian tertiary care hospitals (19,20). Forty percent of infections caused by the A clonal group were identified from community health centers, which suggests that this strain may have spread from the long-term care facility into the community. In fact, CMRSA-2 has been previously shown to be more likely to be associated with community isolations than the other epidemic CMRSA clusters (19). Unlike previously described CA-MRSA strains, this strain does not harbor the PVL or exfoliative toxin genes, and it displays a multidrug resistance phenotype. However, analysis of the community health center cases showed that most (73%) were from persons <18 years of age. Current studies that use comparative genomics are under way in our laboratory to determine whether any genetic differences exist between community, nosocomial, and sporadic MRSA that may explain the association with community isolation (23).
The more predominant clone (cluster B) in this study was, with 1 exception, identified from cases reported from community health centers. Comparison of the B1 finger-print pattern showed that it was indistinguishable to the USA 400 strain, which is responsible for CA-MRSA infections in the midwestern United States (24,25). Although the source of the USA 400 strain into this community cannot be determined, a community-based cluster of MRSA was reported in a rural community in southwestern Manitoba in 1997, and this strain had been identified previously from persons in northwestern Manitoba since 1995 (26). We have demonstrated that this strain is also indistinguishable from the USA400 fingerprint pattern (data not shown). This strain may have spread from northwestern Manitoba to the community in this current study because they are geographically close (J. Wylie, pers. comm.).
When all of the typing information is compared, some observations warrant mention. The single isolate with the molecular subtype designated A6 contains SCCmec IV, which is different from the other A-subtypes that contain SCCmec IIIb. Notably, all A-subtypes are clustered into the MLST group ST5. Although the SCCmec IV region is smaller than that of the SCCmec IIIb, it could not have arisen from a simple deletion event because sequence analysis of these 2 cassettes has shown l significant difference that cannot be explained by a simple deletion event (27). One possible explanation is that a recombinational event occurred, which led to the replacement of SCCmec IIIb with the SCCmec IV. Instability in these regions has been reported, although the events are likely rare (28,29). Since clone B harboring the SCCmec IV is also circulating in this community, a recombination event may have occurred. Furthermore, the MRSA strains with macrorestriction patterns C1 and D1 also contained the SCCmec IV region, which suggests that horizontal transfer of this cassette to additional S. aureus strains may have occurred. Alternatively, the strain with the A6 pattern may have emerged independently from the other strains within the A cluster.
This study documents the finding of high-level mupirocin resistance in >50% of all study strains (clones A and B). Mupirocin can be used to treat superficial skin infections and has been used to decolonize patients. Highlevel mupirocin resistance in MRSA was first described in 1996 among patients in a burn unit in Kuwait (30). Mupirocin resistance is mediated by the mupA gene, which is generally plasmid encoded, although recently the gene has been identified on the chromosome of S. aureus (30,31). Since high-level resistance has been documented on plasmids that vary in size and restriction patterns (1,32), examining the plasmid restriction fragment length polymorphisms may be useful in monitoring monitor plasmid dissemination between and within the clonal types. In a similar manner, we previously described extended-spectrum β-lactamase-containing plasmids in gram-negative organisms (33). Information from this exercise could determine if horizontal transfer of the gene coding for mupirocin resistance occurred between the two clonal groups A and B.
We retrospectively reviewed a sample of charts from patients with skin and soft tissue infections in 2002 in the region of highest CA-MRSA frequency and found that 18% of these infections resulted in a prescription for mupirocin (unpub. data). The use of mupirocin in these communities may be driving resistance. A recent study of patients from a Tennessee medical center documented that decreased usage of mupirocin reduces this form of resistance in MRSA (33). We are currently developing educational programs to decrease the use of this drug, which we hope will decrease mupirocin resistance in this region.
The 2 clonal groups of MRSA described in this study continue to circulate in this area of Saskatchewan (data not shown). We are currently undertaking a case-control study to identify risk factors associated with infections caused by these 2 strains in the community. We call for future studies to include mupirocin in their antimicrobial resistance panels because resistance to this agent may affect treatment outcomes.