Antimicrobial-resistant Streptococcus pneumoniae in Canadian hospitals : Results from the 2007 CANWARD study

BACKGROUND: The Canadian Ward Surveillance Study (CANWARD 2007) tested isolates collected from January to December 2007 from 12 Canadian hospitals to a range of antimicrobial agents. The present paper focuses on antimicrobial resistance in Streptococcus pneumoniae in Canadian hospitals, with an emphasis on macrolide resistance. METHODS: Minimum inhibitory concentrations of antimicrobial agents were determined using the broth microdilution method and interpreted according to Clinical and Laboratory Standards Institute breakpoints. Macrolide-nonsusceptible strains (clarithromycin minimum inhibitory concentrations 0.5 μg/mL or greater) were analyzed by multiplex polymerase chain reaction for the presence of mefA and ermB genes. RESULTS: S pneumoniae represented 9.0% (706 of 7881) of all isolates collected in CANWARD 2007. Of the 706 S pneumoniae isolates collected, 33.1% (234) were from blood and 66.9% (472) were from respiratory specimens. The overall resistance (resistant and intermediate) rates for S pneumoniae isolated from respiratory and blood specimens, respectively, were: penicillin (23.9%, 14.4%), clarithromycin (22.1%, 12.6%), trimethoprim-sulfamethoxazole (14.7%, 11.5%), doxycycline (7.8%, 5.1%) and clindamycin (7.1%, 3.3%). Multidrug resistance (resistance to penicillin, clarithromycin and trimethoprimsulfamethoxazole) accounted for 2% (n=9) and 0.5% (n=1) of respiratory and blood isolates, respectively. Susceptibility of 95% or greater was found with amoxicillin-clavulanic acid (99.5%, 99.3%), ceftriaxone (99.5%, 100%), cefuroxime (95.0%, 96.8%), ertapenem (99.8%, 100%), meropenem (96.1%, 99.5%) and levofloxacin (99.1%, 100%) for respiratory and blood specimens, respectively. No resistance to vancomycin, tigecycline, cethromycin or telithromycin was found. mefA was present in 53.6% (52 of 97) of respiratory and 59.3% (16 of 27) of blood macrolide-nonsusceptible S pneumoniae, while ermB was present in 38.1% (37 of 97) of respiratory and 37% (10 of 27) of blood isolates. Eight of 97 (8.2%) respiratory and one of 27 (3.7%) blood isolates contained both mefA and ermB genes. CONCLUSIONS: S pneumoniae is a common organism isolated from clinical specimens in Canadian hospitals. Resistance was highest to penicillin and clarithromycin, while ceftriaxone and levofloxacin susceptibility were both greater than 99%. No isolates resistant to vancomycin, tigecycline, linezolid or the ketolides were found. Resistance rates were higher among respiratory tract isolates of S pneumoniae than among blood isolates. Macrolide efflux, mefA, was the predominant mechanism of macrolide resistance among both respiratory and blood clarithromycin-nonsusceptible isolates.

Treatment of CAP is usually empirical, and knowledge of local resistance patterns and predominant mechanisms of resistance are important for successful antimicrobial therapy (5).Macrolides are among the principal therapeutic agents used for the empirical treatment of outpatient CAP, while combinations of beta-lactams and macrolides are used for inpatient CAP; however, this choice is complicated by the increasing prevalence of antimicrobial resistance (4,5).Until the 1960s, resistance to penicillin was uncommon among S pneumoniae isolates, and beta-lactams served as agents of choice for treatment of documented or presumed S pneumoniae infections (3,4).During the 1980s, infections due to penicillinnonsusceptible pneumococci became widespread (3,4).Today, S pneumoniae has acquired resistance to several classes of antimicrobial agents, including penicillins, macrolides, trimethoprimsulfamethoxazole (SXT) and fluoroquinolones, and by variety of mechanisms (1,2,6,9).More alarming is the fact that pneumococcal strains that are not susceptible to penicillin are likely to be resistant to other agents, including macrolides (1,2,6,7).
Clinical outcomes may depend on the level of macrolide resistance (high versus low minimum inhibitory concentration [MIC]) in the pneumococcal strain causing the infection (5,7).There are two common mechanisms of macrolide resistance, which differ in the level of resistance conferred (10,11).The first, most common globally and predominant in Europe, involves modification of the ribosomal macrolide target site by methylases encoded by the ermB gene, and is associated with high-level macrolide resistance (MICs of 64 µg/mL or greater) (11).The second mechanism of resistance, prevalent in North America, involves drug efflux encoded by mefA gene and is associated with low level of macrolide resistance (MICs of 16 µg/mL or less) (11).
The purpose of the present study was to assess the antibiotic resistance in S pneumoniae isolates in hospitals across Canada.Although we reported previously (4) on antimicrobial resistance among S pneumoniae isolates, including in patients in Canadian intensive care units, this is the first national surveillance study assessing the antimicrobial resistance in S pneumoniae blood and respiratory specimens of patients in Canadian hospitals.In addition, because macrolides are among the first-line therapeutics used for treatment of CAP, resistance genotypes of macrolide-nonsusceptible S pneumoniae were determined.

Antibiotic susceptibility
MICs of antibiotics commonly used in empirical treatment of pneumococcal respiratory and blood infections were determined using the Clinical and Laboratory Standards Institute (CLSI) broth microdilution method (12).MICs were interpreted and isolates were categorized as susceptible, intermediate and resistant based on the 2006 CLSI-defined breakpoints (12).Antimicrobial susceptibility testing was performed on 445 respiratory and 219 blood isolates.

Genotyping
Macrolide-nonsusceptible (clarithromycin MIC of 0.5 µg/mL or greater) respiratory (n=97) and bacteremic (n=27) S pneumoniae isolates were analyzed for the presence of ermB and/or mefA genes using a multiplex polymerase chain reaction assay as described by Monaco et al (13) and Pantosti et al (14).mefA, ermB and dual mefA and ermB positive controls as well as ATCC 49619 S pneumoniae (susceptible to macrolides; negative control) were used.

Statistical analysis
All statistical tests were performed using EpiInfo StatCals2 version 6 statistical program (Centers for Disease Control and Prevention, USA).Results were reported as statistically significant when P<0.05.

DISCUSSION
CANWARD 2007 is the first study focusing on pathogens isolated from patients in Canadian hospitals.The goal of the current study was to analyze the antibiotic susceptibility profile of S pneumoniae present among patients attending Canadian hospitals.The least active (based upon MIC only) agents against S pneumoniae collected during CANWARD 2007 study were penicillin (23.9%), clarithromycin (22.1%) and SXT (14.7).Against blood isolates, penicillin (14.4%), clarithromycin (12.6%) and SXT (11.5%) were also least active, although the overall extent of resistance was lower for blood isolates than for respiratory isolates.In comparison, the Canadian Respiratory Organism Susceptibility Study (CROSS) (4,15), which assessed antibiotic resistance in S pneumoniae in both inpatients and outpatients from 1998 to 2006, showed that penicillin resistance (resistant and intermediate) ranged from the lowest of 16.1% to highest of 25.0% during 1998 and 2006.The average penicillin resistance (resistant and intermediate) was 21.6% during the nine years of the study (15).During the first five years of the CROSS study, the increase in penicillin nonsusceptibility was attributed to an increase in penicillin-resistant (MIC 2 µg/mL or greater) isolates and a decrease in penicillinintermediate (MIC 0.12 µg/mL to 1 µg/mL) isolates (4).Penicillin-resistant isolates ranged from 6.4% in 1998 to 13.8% in 2002, while penicillin-intermediate isolates decreased from 14.8% to 10.2% during the same period (4).In 2003, the rate of penicillin-intermediate isolates increased to 14.3% while the rate of penicillin-resistant isolates decreased to 6.9%, and they have remained stable at an average 8.1% penicillinresistant rate and 14.7% penicillin-intermediate rate for the remainder of the study (15).The results from the CANWARD  (8).Compared with our previous studies (4), the resistance rates to all beta-lactams also increased in parallel with increasing resistance to penicillin (4).Among the beta-lactams, ertapenem (99.8%), ceftriaxone (99.5%) and amoxicillin-clavulanate (99.3%) were the most active.Meropenem was less active, with a resistance rate of 3.9%.Macrolide resistance among pneumococci in Canada increased from 8% in 1998 to 18.7% in 2006 as measured by the CROSS study (10,15).During the CANWARD 2007 study, macrolide resistance was 22.1%.These data speak to the continued increase in macrolide resistance in Canada.Effluxmediated resistance has been the most predominant mechanism of macrolide resistance in Canada throughout the CROSS study (10,15), and mefA-positive isolates were the most prevalent among macrolide-resistant respiratory and blood S pneumoniae isolated in CANWARD 2007 as well.Isolates with dual mefA and ermB genes emerged during the CROSS study (10,15) and have been identified among both respiratory and blood S pneumoniae isolates in CANWARD 2007 as well, approaching 10% of macrolide-resistant isolates.The emergence of these dual mefA and ermB isolates is worrisome because these isolates spread by clonal dissemination and are associated with high-level macrolide resistance as well as resistance to multiple antibiotics (10,15).All isolates were susceptible to telithromycin.Cethromycin, an investigational ketolide, demonstrated potent in vitro activity against macrolide-resistant S pneumoniae (10).In addition, cethromycin appears to have slightly greater in vitro activity than telithromycin, particularly against mefA strains, which in turn appears to be affecting the activity of ketolides more than ermB-positive strains (7,10).
Among fluoroquinolones, the resistance to levofloxacin and moxifloxacin was low (less than 1% and less than 2%, respectively), similar to our previous studies (4).No resistance to vancomycin or linezolid was noted.
In light of the recent changes in the penicillin breakpoints for nonmeningitis S pneumoniae, perhaps it is important to note that we applied old breakpoints (meningitis breakpoints: susceptible 0.06 µg/mL or less; intermediate 0.12 µg/mL to 1 µg/mL; resistant 2 µg/mL or greater) (3).In doing so, we are reporting higher than actual penicillin resistance rates among respiratory or blood isolates.The old breakpoints were conservative to ensure accurate interpretation of results for cerebrospinal fluid isolates (3).However, these conservative breakpoints may not have accurately represented lack of response in patients treated with parenteral beta-lactams for pneumococcal pneumonia (3).Application of the new parenteral breakpoints to S pneumoniae will result in a significantly lower rate of penicillin nonsusceptibility in nonmeningeal pneumococcal isolates.With the new breakpoints, a penicillin resistance rate of 2.3% was noted, significantly lower than reported with the old breakpoints.

CONCLUSIONS
S pneumoniae is a common organism isolated from clinical specimens in Canadian hospitals.Resistance was highest to penicillin and clarithromycin, while ceftriaxone and levofloxacin susceptibilities were both greater than 99%.No isolates resistant to vancomycin, tigecycline, linezolid or the ketolides were found.Resistance rates were higher among respiratory tract isolates of S pneumoniae compared with blood isolates.The macrolide efflux gene mefA was the predominant mechanism of macrolide resistance among both respiratory and blood clarithromycin nonsusceptible isolates.

ACKNOWLEDGEMENTS:
Aleksandra Wierzbowski was supported by a PhD research grant from Manitoba Health Research Council (MHRC).This study was supported in part by the Manitoba Institute of Child Health (MICH).Funding for the CANWARD 2007 study was provided in part by the University of Manitoba, Health Sciences Center in Winnipeg, National Microbiology Laboratory-Health Canada, Abbott, Affinium Inc, Astellas, Bayer, Janssen Ortho Inc, Merck, Oryx, Pfizer Canada, TaiGen, Targanta and Wyeth Inc.The authors thank the investigators and laboratory site staff at each medical centre that participated in the CANWARD 2007 study.The investigators of the CANWARD 2007 study are listed in the first paper of the present supplement.CANWARD data are also displayed at www.can-r.ca, the official Web site of the Canadian Antimicrobial Resistance Alliance (CARA).

TAble 2 Activity of various antimicrobial agents against 445 respiratory tract Streptococcus pneumoniae isolates collected by the CANWARD 2007 study
*Tetracycline breakpoints were applied to doxycycline.I Intermediate; MIC Minimum inhibitory concentration; NB No breakpoint defined; R Resistant; S Susceptible; SXT Trimethoprim-sulfamethoxazole

TAble 3 Genotypic and phenotypic data for 97 macrolide- nonsusceptible respiratory tract Streptococcus pneumoniae isolates collected by the CANWARD 2007 study
Susceptiblestudy showed a penicillin-intermediate rate of 17.8% and penicillin-resistant rate of 6.2%.The susceptibility of S pneumoniae to penicillin among patients in CANWARD is similar to patients in CROSS, which suggests that penicillin resistance among S pneumoniae in Canada may have stabilized.The observation that increasing penicillin-intermediate isolates affecting the rate of penicillin nonsusceptibility is not unique and has been published by others