Characterization of methicillin-resistant Staphylococcus aureus , vancomycin-resistant enterococci and extended-spectrum beta-lactamase-producing Escherichia coli in intensive care units in Canada : Results of the Canadian National Intensive Care Unit ( CAN-ICU ) study ( 2005 – 2006 )

243 1Department of Medical Microbiology, Faculty of Medicine, University of Manitoba; Departments of 2Medicine; 3Clinical Microbiology, Health Sciences Centre; 4Nosocomial Infections Branch, National Microbiology Laboratory, Winnipeg, Manitoba Correspondence: Dr George G Zhanel, Clinical Microbiology, Health Sciences Centre, MS673-820 Sherbrook Street, Winnipeg, Manitoba R3A 1R9. Telephone 204-787-4902, fax 204-787-4699, e-mail ggzhanel@pcs.mb.ca Received and accepted for publication February 12, 2008

The purpose of the present study was to genotypically and phenotypically characterize the isolates of MRSA, VRE and ESBL-producing E coli collected from patients in Canadian intensive care units (ICUs) between 2005 and 2006.

Bacterial isolates
Study isolates were obtained from the Canadian National Intensive Care Unit (CAN-ICU) study <www.can-r.ca>(16).The CAN-ICU study included 19 medical centres across Canada (western Canada -British Columbia, Alberta, Saskatchewan and Manitoba; central Canada -Ontario and Quebec; and eastern Canada -Maritime provinces).The ICUs selected represented tertiary care medical centres from all regions of the country.Between September 2005 and June 2006, each centre collected a maximum of 300 consecutive, clinically significant isolates obtained from blood, urine, tissue or wound, and respiratory specimens (one pathogen per cultured site per patient) of ICU patients.Surveillance swabs; duplicate isolates; and eye, ear, nose and throat swabs were excluded, as were anaerobic bacteria and fungi.Isolates were shipped to the reference laboratory (Health Sciences Centre, Winnipeg, Manitoba) on Amies charcoal swabs, subcultured onto appropriate media and stocked in skim milk at -80°C until further testing was conducted.

Antimicrobial susceptibility testing
Following two subcultures from frozen stock, the in vitro activities of cefazolin, ceftriaxone, cefepime, ciprofloxacin, clarithromycin, clindamycin, dalbavancin, daptomycin, doxycycline, ertapenem, gentamicin, levofloxacin, linezolid, meropenem, moxifloxacin, piperacillin-tazobactam, tigecycline, trimethoprim-sulfamethoxazole and vancomycin were determined by microbroth dilution in accordance with the 2006 Clinical Laboratory Standards Institute (CLSI) guidelines (M7-A7 and M100-S16) (16).Antimicrobial agents were obtained as laboratory-grade powders from their respective manufacturers.Stock solutions were prepared and dilutions were made as described by the CLSI guidelines (M7-A7).The minimum inhibitory concentrations (MICs) of the antimicrobial agents for the isolates were determined using 96-well custom-designed microtitre plates.These plates contained doubling antimicrobial dilutions in 100 μL/well of cationadjusted Mueller-Hinton broth and were inoculated to achieve a final concentration of approximately 5×10 5 colony-forming units/mL.They were then incubated in ambient air for 24 h before reading.Colony counts were performed periodically to confirm inocula.Quality control was performed using American Type Culture Collection Quality Control organisms -Streptococcus pneumoniae 49619, S aureus 29213, Enterococcus faecalis 29212, E coli 25922 and Pseudomonas aeruginosa 27853.
For all antimicrobials tested, MIC interpretive standards were defined according to the CLSI break points (M100-S16).For tigecycline, the following susceptible, intermediate and resistant interpretive breakpoints (Food and Drug Administration) were used -S aureus, less than or equal to 0.5 μg/mL (susceptible); Enterococcus species, less than or equal to 0.25 μg/mL (susceptible); and Enterobacteriaceae family, less than or equal to 2 μg/mL (susceptible), less than or equal to 4 μg/mL (intermediate), and greater than or equal to 8 μg/mL (resistant).

Characterization of MRSA, ESBL-producing
Enterobacteriaceae family and VRE MRSA: Potential MRSA isolates were confirmed using the CLSI disk-diffusion method and mecA polymerase chain reaction (PCR).All isolates of MRSA were tested for Panton-Valentine leukocidin and typed using pulsed-field gel electrophoresis following the Canadian standardized protocol to assess whether the isolates were CA-MRSA or HA-MRSA (10,11,17,18).Pulsed-field gel electrophoresis fingerprints were analyzed with BioNumerics version 3.5 (Applied Maths, Belgium) using a position tolerance of 1.0 and an optimization of 1.0.Strain relatedness was determined as previously described (19).Fingerprints were compared with the national MRSA fingerprint database and were grouped into one of 10 Canadian epidemic MRSA strains (CMRSA-1, CMRSA-2, etc) as previously described (11).In the present study, CA-MRSA and HA-MRSA were defined genotypically and not epidemiologically.Any MRSA with a CMRSA-7 (USA400/MW2) or CMRSA-10 (USA300) genotype was labelled as CA-MRSA, while all other genotypes (eg, CMRSA-1 [USA600], CMRSA-2 [USA100] and CMRSA-4 [USA200]) were labelled as HA-MRSA.ESBL testing: Any E coli or Klebsiella species with a ceftriaxone MIC of greater than or equal to 1 μg/mL was identified as a potential ESBL producer.ESBL producers were confirmed using the CLSI double-disk diffusion method.PCR and DNA sequence analyses were used to identify bla SHV , bla TEM and bla CTX-M genes among isolates, as previously described (9,12,13).VRE: Potential VRE isolates were confirmed using the CLSI vancomycin disk-diffusion testing; they underwent vanA and vanB PCR, as well as DNA fingerprinting to assess genetic similarity, as previously described (14,20).

DISCUSSION
The CAN-ICU study was the first national prospective surveillance study assessing antimicrobial resistance in Canadian ICUs.It determined that more than one-half of all infections in the ICUs were respiratory in origin, irrespective of patient age and sex (16).Bloodstream, wound or intravenous origin, and urinary tract infections were less common than respiratory infections in ICU patients, as has been previously documented (21).
The CAN-ICU study documented that MSSA and MRSA are important pathogens causing respiratory tract infections, bacteremia, and wound or intravenous infections in ICUs in Canada.MRSA accounted for 21.9% of all S aureus, and 9.3% of all MRSA causing infections in the ICU were CA-MRSA genotypes.This has not been previously documented in Canada and shows the infiltration of CA-MRSA into Canadian ICUs.All 18 of the CA-MRSA isolates were either USA400 (CMRSA-7) or USA300 (CMRSA-10) genotypes.These two genotypes are the two primary CA-MRSA genotypes reported across North America (11,18,(22)(23)(24).It appears that infections in Canadian ICUs caused by CA-MRSA are primarily occurring in western Canada, with USA400 (CMRSA-7) the predominant genotype in Saskatchewan or Manitoba, and USA300 (CMRSA-10) the predominant genotype in British Columbia (Table 2).
The present study is the first to document that ESBLproducing E coli are becoming more common than ESBLproducing Klebsiella species in Canadian ICUs (3.7% of E coli were ESBL-producing E coli, and 1.8% of Klebsiella species were ESBL-producing Klebsiella species [data not shown]).The 18 ESBL-producing E coli (89.5%) were primarily obtained in central and eastern Canada from urine, blood, the respiratory tract and wounds.All 18 ESBL-producing E coli that were isolated displayed an MDR phenotype, with 77.8% demonstrating concomitant resistance to fluoroquinolones and 55.6% demonstrating resistance to trimethoprim-sulfamethoxazole.The study showed that CTX-M with bla CTX-M-15 was the predominant genotype (72%) of ESBL-producing E coli in Canada.Other studies (8,9,12,13) assessing ESBL-producing E coli have shown that the CTX-M genotype is spreading rapidly in both community and hospital settings.Pitout et al (13)  The CAN-ICU study showed that VRE represented 6.7% of all enterococci tested, with the vanA genotype (mostly E faecium) making up 88.2% of VRE.This relatively low level of VRE across Canada has been previously documented and likely reflects the active surveillance programs in Canadian hospitals (14).Such programs have been reported to prevent VRE colonization and bacteremia (25).Previous data (5,14) have suggested that the E faecium carrying vanA is the predominant genotype in North America.
Resistance rates of MRSA were high with fluoroquinolones and macrolides, such as clarithromycin and clindamycin (range 76.1% to 91.8%); they were lower, at 11.7%, with trimethoprimsulfamethoxazole.Thus, trimethoprim-sulfamethoxazole still represents a reasonable empirical treatment for mild to moderate infections caused by CA-MRSA or HA-MRSA.It should be noted that the study found a significant difference between the susceptibilities of CA-MRSA and HA-MRSA.Like others, we report that CA-MRSA was more susceptible to beta-lactams, trimethoprim-sulfamethoxazole, macrolides, clindamycin and fluoroquinolones than HA-MRSA (11).All CA-MRSA and HA-MRSA isolates were susceptible to vancomycin, linezolid, tigecycline and daptomycin.Likewise, all VRE in the study proved to be susceptible to linezolid, tigecycline and daptomycin.MDR ESBL-producing E coli isolates were all susceptible to the carbapenems, ertapenem, meropenem and tigecycline.Because nosocomial infections in the ICU are frequently MDR (frequently associated with prior antimicrobial use [6]), some have suggested that involvement of an infectious diseases specialist may help to improve, cure and minimize further resistance development (2).
A total of 4133 organisms from 2580 patients (or 1.6 isolates per patient) were collected from ICUs across Canada in 2005 and 2006; they were taken during any time of the patient's ICU admission.59.3% (2451 of 4133) of isolates were collected from male patients, while 40.7% (1682 of 4133) were collected from female patients.Patient age breakdown was 17 years of age or younger, 13.7%; between 18 and 64 years of age, 46.7%; and 65 years of age or older, 39.6%.54.8% of organisms were obtained from respiratory sites -17.7% from blood, 13.9% from wounds and intravenous sites, and 13.6% from urine.Most common organisms isolated from ICUs Table 1 describes the 20 most common organisms isolated from ICUs across Canada in 2005 and 2006.The most common Gram-positive cocci included methicillin-susceptible S aureus (MSSA), coagulase-negative staphylococci/Staphylococcus epidermidis, Enterococcus species, S pneumoniae and MRSA, which represented 39.5% of all pathogens.The most common Gramnegative bacilli included E coli, Pseudomonas aeruginosa, Klebsiella pneumoniae, Enterobacter cloacae, Stenotrophomonas maltophilia and Serratia marcescens, which made up 45% of all pathogens in ICUs.All MRSA, VRE and ESBL E coli underwent genotypic characterization and antimicrobial susceptibility studies.

TABLE 2
Characteristics of 18 community-associated methicillin-resistant Staphylococcus aureus (MRSA) isolates from patients in Canadian intensive care units in 2005 and 2006

TABLE 3 Antimicrobial
susceptibilities of community-associated methicillin-resistant Staphylococcus aureus (CA-MRSA) and health care-associated MRSA (HA-MRSA) isolated from patients in Canadian intensive care units in 2005 and 2006 investigated the molecular epidemiology of ESBLproducing E coli collected between 2000 and 2005 in the Calgary Health Region in Alberta.These investigators reported Results of the 2006 CAN-ICU study Can J Infect Dis Med Microbiol Vol 19 No 3 May/June 2008 247 that 64% (354 of 552) of ESBL-producing E coli were PCRpositive for bla CTX-M genes, with CTX-M-14 (59.6%) and CTX-M-15 (36.2%) reported most commonly.This study highlights the rapid spread of MDR ESBL CTX-M-15 E coli in Canadian ICUs.This genotype is likely spreading rapidly due to the extensive use of third-generation cephalosporins and fluoroquinolones.