Comparison of community-associated and health care-associated methicillin-resistant Staphylococcus aureus in Canada: Results from CANWARD 2007

1Diagnostic Services of Manitoba/Department of Clinical Microbiology, Health Sciences Centre; 2National Microbiology Laboratory, Public Health Agency of Canada; 3Department of Medical Microbiology and Infectious Diseases, Faculty of Medicine, University of Manitoba, Winnipeg, Manitoba Correspondence: Kimberly A Nichol, Department of Clinical Microbiology, Health Sciences Centre, MS673-820 Sherbrook Street, Winnipeg, Manitoba R3A 1R9. Telephone 204-787-4902, fax 204-787-4699, e-mail knichol@hsc.mb.ca CANWARD 2007

Recently, CA-MRSA has begun to replace HA-MRSA in the health care setting. The purpose of the present study was to compare the demographics, antimicrobial susceptibilities and molecular epidemiology of CA-MRSA and HA-MRSA in Canada.

Study isolates
Three hundred eighty-five isolates of MRSA were selected from among 1482 S aureus collected between January and December 2007 as part of the ongoing Canadian Ward Surveillance Study (CANWARD 2007) assessing pathogen prevalence and antibiotic resistance in Canadian hospitals. Isolates were received from 12 sentinel hospital sites that were geographically distributed in a population-based fashion in major cities in seven of the 10 Canadian provinces. Each centre was asked to submit pathogens (consecutive, one per patient per infection site) from blood, respiratory specimens, urine and wound/intravenous (IV) sites. Isolates were collected from Canadian patients affiliated with hospital clinics, emergency rooms, medical/surgical wards and intensive care units. All S aureus were identified at the originating centre using local site criteria. Resistance to methicillin was confirmed at the coordinating laboratory (Health Sciences Centre, Winnipeg, Manitoba) using the Clinical and Laboratory Standards Institute (CLSI)-approved disk diffusion method with cefoxitin, as well as by growth on MRSA Select chromogenic media (Bio-Rad Laboratories, Canada).

Molecular characterization
MRSA status was confirmed by real-time polymerase chain reaction (PCR) of the mecA and nuc genes (12). This triplex PCR assay also included primers for the detection of the lukF-PV and lukS-PV genes encoding the components of the Panton-Valentine leukocidin (PVL) toxin (12). Typing of the staphylococcal cassette chromosome mec (SCCmec) was performed by multiplex PCR as previously described (13).
MRSA strains were typed by pulsed-field gel electrophoresis (PFGE) of SmaI digests following the Canadian standardized protocol (14). PFGE profiles were digitized for analysis with BioNumerics software (v3.5, Applied Maths Inc, USA), and strain relatedness was determined following established criteria (15). PFGE patterns were also compared with the national MRSA fingerprint database and were grouped into one of 10 Canadian epidemic PFGE strain types (CMRSA1 to CMRSA10) (16). For the purpose of the present study, CA-MRSA and HA-MRSA were defined genotypically (ie, on the basis of their PFGE epidemic type) and not epidemiologically as per Centers for Disease Control and Prevention criteria for distinguishing CA-MRSA from HA-MRSA, because epidemiological information was not available. Any MRSA with a CMRSA7 (USA400) or CMRSA10 (USA300) genotype was labelled as CA-MRSA while all other genotypes corresponding to a characterized epidemic type (eg, CMRSA1 [USA600], CMRSA2 [USA100/800], CMRSA3, CMRSA5 [USA500], CMRSA6, CMRSA9, etc) were labelled as HA-MRSA. Isolates that could not be defined as communityor health care-associated based on PFGE patterns were further characterized by staphylococcal protein A (spa) typing as described elsewhere (17). There has previously been shown to be good correlation between PFGE fingerprint clusters and spa types (17), allowing for further classification of strains as either CA-MRSA or HA-MRSA. MRSA with a PFGE pattern or spa type not associated with one of the known Canadian epidemic types were labeled as unique (non-CMRSA).

RESULTS
Based on the genotypic definition of CA-MRSA and HA-MRSA described above, 75 (19.5%) of the 385 MRSA collected during the CANWARD 2007 study were categorized as CA-MRSA while 305 (79.2%) were classified as HA-MRSA. The remaining five (1.3%) MRSA isolates could not be designated as CMRSA1 to CMRSA10 by either PFGE or spa typing and were therefore deemed unique. The patient demographics of the MRSA strains are shown in Table 1 Table 4. CA-MRSA demonstrated lower MIC 50 and MIC 90 (MICs needed to inhibit 50% and

DISCUSSION
Infections caused by CA-MRSA are being reported with increasing frequency worldwide and represent a growing public health concern. These CA-MRSA strains differ from their health care-associated counterparts in their microbiological, epidemiological and molecular characteristics. HA-MRSA typically carry SCCmec types I to III, which contain the mecA gene responsible for resistance to beta-lactam antibiotics, and may also contain multiple determinants for resistance to other classes of antibiotics such as aminoglycosides, fluoroquinolones, macrolides and clindamycin (1). CA-MRSA, by comparison, usually carry unique SCCmec elements (type IV or V) and are often associated with the presence of the PVL toxin encoding a pore-forming protein involved in primary skin infection and pneumonia (1). Most CA-MRSA strains are resistant to beta-lactams but remain relatively susceptible to clindamycin and other non-beta-lactam antibiotics, with variable resistance to macrolides (1). CA-MRSA has also been found to be genotypically distinct from HA-MRSA, with most strains in Canada belonging to two common PFGE types (8,18). In our study, 19.5% and 79.2% of MRSA strains from Canadian hospitals were identified by PFGE as CA-MRSA and HA-MRSA, respectively. CA-MRSA infection was strongly associated with patients in the younger (17 years or less) and the 18 to 64 years age groups, an observation that has been well documented in the literature. Although the majority of CA-MRSA were isolated from wounds and IV sites, CA-MRSA strains were also isolated from bloodstream and respiratory tract specimens, which is likely reflective of the ability of this organism to cause invasive infections including bacteremia and necrotizing pneumonia in addition to skin and soft tissue infections. HA-MRSA carried SCCmec type II or III, were almost exclusively PVL-negative and belonged primarily to the CMRSA2 (USA100/800) genotype. By comparison, CA-MRSA carried only SCCmec type IV, were predominantly PVL-positive and belonged to either the CMRSA7 (USA400) or CMRSA10 (USA300) genotypes. CMRSA7 (USA400) and CMRSA10 (USA300) are the predominant CA-MRSA strains circulating in North America and, although not as prevalent as they are in many centres in the United States, have been increasingly isolated in Canada since 2004 (8,18,19). Presence of the PVL toxin has frequently been linked to CA-MRSA infections and is hypothesized to   play a significant role in increased disease severity (2,20). Recently, however, it has been shown that PVL is not present in all CA-MRSA strains and cannot be used as a definitive marker of CA-MRSA infection (20,21). In this study we found that 5.3% of CA-MRSA were PVL-negative while 0.7% of HA-MRSA were PVL-positive. The isolation of PVL-negative CA-MRSA and PVL-positive HA-MRSA, although rare, indicates that the epidemiology of MRSA in Canadian hospitals continues to change.
As previously described, CA-MRSA strains were more susceptible to beta-lactams, macrolides, clindamycin and fluoroquinolones compared with HA-MRSA. Both CA-MRSA and HA-MRSA displayed low rates of resistance (0.0% to 15.4%) to trimethoprim-sulfamethoxazole. This agent therefore remains a reasonable therapeutic option for empirical treatment of mild to moderate infections caused by CA-MRSA or HA-MRSA. As expected, all MRSA regardless of genotype were susceptible to daptomycin, linezolid, tigecycline and vancomycin. Interestingly, we observed a significant difference between the susceptibility of CMRSA7 (USA400) CA-MRSA and CMRSA10 (USA300) CA-MRSA. With high rates of resistance to clarithromycin and fluoroquinolones, CMRSA10 (USA300) isolates displayed an antimicrobial susceptibility profile intermediate between those of CMRSA7 (USA400) strains and those of HA-MRSA. This raises concerns that some community-associated strains may be able to acquire additional resistance determinants, resulting in an organism capable of causing serious disease that carries the PVL gene and displays a multidrug-resistant phenotype. Similarly, the acquisition of the PVL gene in existing HA-MRSA strains could likewise result in multidrug-resistant organisms with increased virulence, and is of equal concern.
Our study is not without its limitations. The Centers for Disease Control and Prevention criteria for distinguishing CA-MRSA from HA-MRSA state that a CA-MRSA infection can be defined as one that occurs in an individual with no established health care-associated risk factors and who has a positive MRSA culture within 48 h of hospitalization, has no previous history of MRSA infection or colonization, has no history of hospitalization, surgery, dialysis or residence in a long-term care facility within the past year, and has no permanent indwelling catheters or medical devices (22). Because the CANWARD study is a passive surveillance study involving microbiology laboratories, clinical and epidemiological information were not available. Therefore, CA-MRSA and HA-MRSA were defined genotypically based on PFGE epidemic types. Consequently, some CA-MRSA may have been misclassified as HA-MRSA and, conversely, some HA-MRSA may have been incorrectly labelled as CA-MRSA. Thus, the present study may not fully reflect the true proportion of CA-and HA-MRSA in Canadian hospitals. Secondly, due to limited resources, we did not perform SCCmec typing for all HA-MRSA isolates.

CONCLUSIONS
CA-MRSA appears to be an emerging pathogen in Canadian hospitals. As strains that were once thought to be associated only with infections in the community continue to disseminate into the hospital setting and, similarly, as health careassociated strains spread to the community, the distinction between CA-MRSA and HA-MRSA will continue to blur. Ongoing surveillance of the clinical, epidemiological and biological characteristics of CA-MRSA is necessary to increase our understanding of this important pathogen so that effective therapeutic options and infection control measures for combating the spread of CA-MRSA in both the community and the hospital can be established.