Antimicrobial-resistant pathogens in Canadian ICUs: results of the CANWARD 2007 to 2016 study Free

Abstract

Objectives

To describe the microbiology and antimicrobial resistance patterns of cultured samples acquired from Canadian ICUs.

Methods

From 2007 to 2016, tertiary care centres from across Canada submitted 42938 bacterial/fungal isolates as part of the CANWARD surveillance study. Of these, 8130 (18.9%) were from patients on ICUs. Susceptibility testing guidelines and MIC interpretive criteria were defined by CLSI.

Results

Of the 8130 pathogens collected in this study, 58.2%, 36.3%, 3.1% and 2.4% were from respiratory, blood, wound and urine specimens, respectively. The top five organisms collected from Canadian ICUs accounted for 55.4% of all isolates and included Staphylococcus aureus (21.5%), Pseudomonas aeruginosa (10.6%), Escherichia coli (10.4%), Streptococcus pneumoniae (6.5%) and Klebsiella pneumoniae (6.4%). MRSA accounted for 20.7% of S. aureus collected, with community-associated (CA) MRSA genotypes increasing in prevalence over time (P < 0.001). The highest susceptibility rates among MRSA were 100% for vancomycin, 100% for ceftobiprole, 100% for linezolid, 99.7% for ceftaroline, 99.7% for daptomycin and 99.7% for tigecycline. The highest susceptibility rates among E. coli were 100% for tigecycline, 99.9% for meropenem, 99.7% for colistin and 94.2% for piperacillin/tazobactam. MDR was identified in 26.3% of E. coli isolates, with 10.1% producing an ESBL. The highest susceptibility rates among P. aeruginosa were 97.5% for ceftolozane/tazobactam, 96.1% for amikacin, 94.7% for colistin and 93.3% for tobramycin.

Conclusions

The most active agents against Gram-negative bacilli were the carbapenems, tigecycline and piperacillin/tazobactam. Against Gram-positive cocci, the most active agents were vancomycin, daptomycin and linezolid. The prevalence of CA-MRSA genotypes and ESBL-producing E. coli collected from ICUs increased significantly over time.

Introduction

Antimicrobial resistance represents a critical threat to human health. Antimicrobial utilization and over-utilization in both hospitals and the community is a strong impetus for antimicrobial-resistant pathogens such as MRSA, ESBL-producing Enterobacteriaceae, carbapenem-resistant Enterobacteriaceae (CRE) and MDR Pseudomonas aeruginosa. Over the past decade, in conjunction with continued surveillance and careful infection control, antimicrobial stewardship programmes have become the standard of care.¹ Despite these measures, adequate control of the epidemic of antimicrobial resistance has not been achieved, as resistance rates continue to climb globally.² Clinicians have and will continue to have increasing difficulty identifying antimicrobial options for their patients, a problem compounded by a scarcity of new treatment alternatives. Infection with an antimicrobial-resistant organism has broad implications, including longer hospital stays, increased morbidity and increased mortality when compared with infection with antimicrobial-susceptible strains.³ Concerns regarding antimicrobial resistance and utilization are particularly important in ICUs, as resistance rates are highest and antimicrobial use is most pervasive in comparison with other areas of medicine; it is estimated that 70% of ICU patients are receiving antimicrobials at any one time.^4–6 Furthermore, the high burden of illness in ICU patients contributes to altered immune function and a corresponding increased susceptibility to infection.⁷ Such patients rely on prompt identification of infection and the administration of effective antimicrobial therapy, a premise that is increasingly difficult in the era of MDR and XDR organisms.^8,9

The purpose of this prospective surveillance study was to describe the microbiology and antimicrobial resistance patterns of cultured samples acquired from Canadian ICUs between 2007 and 2016.

Materials and methods

Bacterial isolates

Bacterial and fungal isolates were collected as part of the ongoing CANWARD national surveillance study from 2007 to 2016 as previously described by Zhanel et al.¹⁰ Briefly, tertiary care centres (listed in the Acknowledgements section) representing 8 of the 10 Canadian provinces (12 centres in 2007, 10 in 2008, 15 in 2009, 14 in 2010, 15 in 2011, 12 in 2012, 15 in 2013, 13 in 2014, 13 in 2015 and 13 in 2016) were asked to submit clinically significant, consecutive isolates from patients with respiratory, urine, wound and bloodstream infections. Demographic data provided by the submitting centre included patient gender, age and specimen source.

Ethics

The CANWARD study receives annual approval by the University of Manitoba Research Ethics Board (H2009:059).

Antimicrobial susceptibility testing

Antimicrobial susceptibility testing was performed using broth microdilution in accordance with CLSI guidelines.¹¹ Following two subcultures from frozen stock, the MIC of the antimicrobials tested was determined using in-house, custom-designed 96-well microtitre panels, as previously described.¹⁰ MIC interpretive standards were defined by CLSI breakpoints,¹² unless otherwise noted. MDR and XDR were defined as resistance to three or more and five or more different antimicrobial classes, respectively. Specific antimicrobial classes used in this determination can be found outlined by Magiorakos et al.¹³ MDR and XDR criteria for Streptococcus pneumoniae were defined by Golden et al.¹⁴

Characterization of MRSA, ESBL-producing Enterobacteriaceae and VRE

Potential MRSA isolates were confirmed by mecA PCR. Isolates were further characterized by staphylococcal protein A (spa) typing and Canadian epidemic PFGE strain types were inferred by spa repeat pattern analysis as previously described.¹⁵ Community-associated (CA) MRSA and healthcare-associated (HA) MRSA genotypes were assigned based on the inferred epidemic strain type, as previously described by Nichol et al.¹⁵ Potential ESBL-producing Enterobacteriaceae were identified as any Escherichia coli or Klebsiella pneumoniae with a ceftriaxone and/or ceftazidime MIC of ≥1 mg/L.¹² Potential ESBL-producing isolates were confirmed by the CLSI confirmatory disc test and molecular methods as previously described.¹⁶ Potential VRE isolates were confirmed by vanA and vanB PCR, as previously described.^17,18

Statistical analysis

Statistical significance was calculated by the χ² test or Fisher’s exact test in the case of small sample sizes using the SPSS statistics (Version 23) program (IBM Corporation). Temporal analysis was performed using the Cochran–Armitage test of trend (XLSTAT). Statistical significance in this study was defined as a P value ≤0.05; P values >0.05 have been denoted not significant throughout this article.

Results

Patient demographics and specimen types

From 2007 to 2016 the CANWARD study collected a total of 42938 isolates, of which 8130 (18.9%) were from ICU patients. There were 5137 (63.2%) isolates obtained from male patients while 2992 (36.8%) were from females (1 isolate from a patient of unknown gender). According to specimen type, there were 4730 (58.2%) isolates obtained from respiratory specimens, 2951 (36.3%) from blood specimens, 252 (3.1%) from wound specimens and 197 (2.4%) from urine specimens. By age group, 1451 (17.8%), 3666 (45.1%) and 3013 (37.1%) isolates were received from patients 0–17, 18–64 and ≥65 years of age, respectively. The only demographic parameter demonstrating statistically significant variation over time was the proportion of isolates collected from respiratory specimens, which increased significantly (P < 0.001) from 45.3% in 2007 to 63.0% in 2016. The geographical locations of the submitting centres were as follows: 2659 (32.7%) from western provinces (British Columbia, Alberta, Saskatchewan and Manitoba), 2664 (32.8%) from Ontario, 1497 (18.4%) from Quebec and 1310 (16.1%) from maritime provinces (New Brunswick and Nova Scotia).

Most common pathogens

Among the 8130 isolates collected, the 20 most common organisms represented 7385 (90.8%) of the total (Table 1). The 20 most common organisms isolated from ICUs across Canada were 3177 (43.0%) Gram-positive organisms (methicillin-susceptible and -resistant Staphylococcus aureus, S. pneumoniae, CoNS/Staphylococcus epidermidis, Enterococcus faecalis, Enterococcus faecium, Streptococcus agalactiae and Enterococcus spp.) as well as 4034 (54.6%) Gram-negative organisms (P. aeruginosa, E. coli, K. pneumoniae, Haemophilus influenzae, Enterobacter cloacae, Stenotrophomonas maltophilia, Serratia marcescens, Klebsiella oxytoca, Moraxella catarrhalis, Klebsiella aerogenes, Proteus mirabilis and Acinetobacter baumannii) (Table 1).

Antimicrobial susceptibility

The activity of antimicrobials tested against the most common Gram-positive pathogens isolated from Canadian ICUs is summarized in Table 2. More than 95% of MSSA isolates were susceptible to all agents tested, with the exception of the fluoroquinolones (90%–93% susceptible) and clarithromycin (78.9%). In contrast to MSSA, a significantly (P < 0.001) lower proportion of MRSA isolates were susceptible to trimethoprim/sulfamethoxazole, gentamicin, clindamycin, levofloxacin, ciprofloxacin and clarithromycin (Table 2); however, all MRSA isolates tested were susceptible to ceftobiprole, linezolid and vancomycin, and only five (1.4%) isolates demonstrated a vancomycin MIC of 2 mg/L. S. pneumoniae isolates were >90% susceptible to all antimicrobials tested except clarithromycin, doxycycline, penicillin and trimethoprim/sulfamethoxazole. E. faecalis and S. epidermidis were ≥97.9% and 100% susceptible, respectively, to daptomycin, tigecycline and vancomycin.

The activity of antimicrobials tested against the most common Gram-negative pathogens isolated from Canadian ICUs is summarized in Table 3. Antimicrobials with the greatest activity against P. aeruginosa included ceftolozane/tazobactam, amikacin, colistin, tobramycin and gentamicin, while reduced susceptibilities were noted for ceftazidime (75.4% susceptible), ciprofloxacin and levofloxacin (76.0 and 59.3%, respectively) and meropenem (73.7%). Susceptibility of E. coli was >90% for amikacin, cefepime, ceftolozane/tazobactam, colistin, meropenem, piperacillin/tazobactam and tigecycline, with ceftazidime and ceftriaxone maintaining activity against 87.9% and 85.4% of isolates, respectively. Resistance rates >20% were noted for cefazolin, ciprofloxacin, levofloxacin and trimethoprim/sulfamethoxazole. K. pneumoniae was ≥91.1% susceptible to all antimicrobials tested except for cefazolin. Similarly, amoxicillin/clavulanate, ceftaroline, ceftriaxone, cefuroxime, fluoroquinolones, meropenem and piperacillin/tazobactam maintained activity against ≥97.2% of H. influenzae. However, H. influenzae demonstrated decreased susceptibility to clarithromycin (89.8%), trimethoprim/sulfamethoxazole (82.5%) and ampicillin (79.2%). Both E. cloacae and S. marcescens were >95% susceptible to amikacin, cefepime, gentamicin, levofloxacin, meropenem and tobramycin. S. maltophilia was 25.8%, 68.0% and 98.7% susceptible to ceftazidime, levofloxacin and trimethoprim/sulfamethoxazole, respectively. The MIC₅₀ and MIC₉₀ of moxifloxacin and tigecycline for S. maltophilia were 0.5/4 and 1/4 mg/L, respectively. For both Gram-positive and Gram-negative organisms, susceptibility values for numerous antimicrobials varied by region (data not shown).

MDR among select Gram-positive and Gram-negative pathogens

The presence of MDR and XDR was assessed in the 10 most common organisms where clear criteria have been defined. Overall, 19.0% and 2.3% of S. aureus collected were found to be MDR and XDR, respectively. The majority of MDR and XDR S. aureus isolates were MRSA, with only 3.3% and 0.4% of MSSA isolates demonstrating MDR and XDR phenotypes, respectively. All MDR and XDR S. aureus remained susceptible to vancomycin and linezolid. Among S. pneumoniae isolates collected, 5.1% and 1.9% were MDR and XDR, respectively. All MDR and XDR S. pneumoniae remained susceptible to fluoroquinolones.

Fifteen percent of P. aeruginosa isolates were MDR, while only 2.0% were found to be XDR. The majority (82.4%) of XDR P. aeruginosa were resistant to five classes of antimicrobials, while one isolate was resistant to six antimicrobial classes and two were resistant to seven. No single antimicrobial used in the determination of MDR/XDR¹³ maintained 100% activity against P. aeruginosa.

In total, 26.3% and 14.9% of E. coli were MDR and XDR, respectively, with both phenotypes demonstrating a significant increasing trend over the study period (P < 0.0001). The proportion of MDR E. coli increased >2-fold from 13.4% in 2007 to 30.4% in 2016 and the proportion of XDR E. coli increased >5-fold from 3.2% in 2007 to 18.8% in 2016. Tigecycline remained the only antimicrobial with activity against all MDR/XDR E. coli. The proportion of both MDR and XDR E. cloacae and K. pneumoniae also increased significantly across this study; MDR/XDR E. cloacae increased from 4.3%/2.2% in 2007 to 22.2%/11.1% in 2016 (P ≤ 0.002), while MDR/XDR K. pneumoniae increased from 2.5%/0% in 2007 to 13.6%/12.8% in 2016 (P ≤ 0.02). Across all study years, the proportion of MDR/XDR E. cloacae and K. pneumoniae was 18.2%/5.3% and 8.6%/5.4%, respectively.

Characteristics of MRSA

Of the 1746 S. aureus collected, 361 (20.7%) were MRSA, including 230 (63.7%) HA-MRSA and 118 (32.7%) CA-MRSA genotypes. The proportion of MRSA isolates by study year is presented in Table 4. The national prevalence of MRSA dropped significantly from 2007 to 2009 (P = 0.001), but remained constant subsequently. Among HA-MRSA genotypes, the most common epidemic type was CMRSA-2, while CMRSA-10 was most common among CA-MRSA genotypes (Table 4). The proportion of CA-MRSA genotypes compared with HA-MRSA increased significantly across the study (P < 0.001), from 15.3% in 2007 to 76.2% in 2016.

Characteristics of VRE

In total, 36 VRE isolates were collected across all study years, of which 34 (94.4%) and 2 (5.6%) were vanA and vanB genotypes, respectively. All VRE isolates were identified as E. faecium and VRE constituted 28.3% of the 127 E. faecium collected. The proportion of VRE among E. faecium collected by study year was as follows: 16.7% in 2007, 25.0% in 2008, 28.6% in 2009, 27.3% in 2010, 26.3% in 2011, 46.2% in 2012, 14.3% in 2013, 62.5% in 2014, 25.0% in 2015 and 40.0% in 2016 (P = not significant).

Characteristics of ESBL-producing E. coli and K. pneumoniae

Of the 848 E. coli and 521 K. pneumoniae collected in this study, 86 (10.1%) and 26 (5.0%) isolates, respectively, were identified as ESBL producers. The prevalence of ESBL-producing E. coli increased significantly across this study (P < 0.001); rates of ESBL-producing E. coli by study year and genotype are presented in Table 5. The majority (96.5%) of ESBL-E. coli produced a CTX-M-type ESBL, where CTX-M-15 was the most common variant. The proportion of ESBL-producing K. pneumoniae demonstrated a borderline significant increasing trend across the study period (P = 0.057). The proportions of ESBL-producing K. pneumoniae by study year were as follows: 1.3% in 2007, 1.6% in 2008, 5.6% in 2009, 10.9% in 2010, 2.0% in 2011, 2.8% in 2012, 7.5% in 2013, 6.3% in 2014, 2.6% in 2015 and 11.4% in 2016. Genotypes were more diverse among the limited number of ESBL-producing K. pneumoniae, with CTX-M-15 and SHV-type ESBLs being most common.

Discussion

Our group has previously reported on antimicrobial-resistant pathogens in Canadian ICUs from 2005 to 2006 (CAN-ICU).¹⁹ Since this publication 10 years ago, there has been a paucity of large-scale surveillance data in the literature specific to antimicrobial resistance in the ICU. The present study seeks to update this work in the hope of aiding clinicians in the difficult task of treating infections in critically ill patients.

The CANWARD study collected 8130 bacterial and fungal isolates from patients in ICUs from 2007 to 2016. Notably, an increasing number of specimens in this study were found to be respiratory in origin; by 2016 approximately two-thirds of all specimens received were collected from the lung. Interestingly, Quan et al.²⁰ have reported a 180% increase in community-acquired pneumonia (CAP) admissions in the UK from 1997 to 2014, a trend that increased dramatically from 4.2%/year in 2007/2008 to 8.8%/year post-2008. Further to this, Woodhead et al.²¹ have reported a 128% increase in CAP admissions to UK ICUs over a time when ICU admissions overall increased by only 24%, suggesting a growing proportion of CAP among ICU patients.

The top five bacterial pathogens collected in this study were S. aureus, P. aeruginosa, E. coli, S. pneumoniae and K. pneumoniae, constituting 55.4% of the overall total. Similarly, both Vincent⁵ and Streit et al.²² have reported S. aureus, P. aeruginosa and E. coli to be the top three organisms isolated from European and North American ICUs, respectively. Despite S. aureus being the most commonly isolated pathogen, we report that Gram-negative organisms represent a slight majority of the bacterial isolates recovered from Canadian ICU specimens. This finding is similar to that of the Extended Prevalence of Infection in ICU (EPIC II) study, where 47% and 62% of isolates were Gram-positive and Gram-negative, respectively.⁵ In contrast to this, the Canadian Sepsis Treatment and Response Registry has shown Gram-positive organisms to constitute the majority of isolates collected from 12 Canadian ICUs, though this result is specific to patients with severe sepsis.²³

The antimicrobials with the greatest activity against the most common Gram-positive isolates in this study included vancomycin, ceftaroline, ceftobiprole, linezolid, tigecycline and daptomycin. Against S. aureus, our group has previously reported similar susceptibility rates from the overall CANWARD study and from CANWARD blood culture isolates, suggesting that S. aureus collected from the ICU is no more resistant to antimicrobials in vitro when compared with non-ICU isolates in our data set.^10,24 This runs contrary to conventional wisdom that antimicrobial resistance tends to be higher in the ICU when compared with other areas of the hospital, as well as several published studies assessing unit-specific antibiograms among MSSA and MRSA isolates.^25–28 Specific to MRSA, we report a statistically significant decline in MRSA rates from 2007 to 2009, as well as a trend to increasing representation of CA-MRSA genotypes within our cohort. Our group, as well as others, have published previously on declining MRSA rates in a variety of hospital settings.^29–33 Such declines are multifactorial and are attributed to a variety of interventions, including altered prescribing practices and antimicrobial stewardship, improved hand hygiene, screening and decolonization, contact precautions and strategies to decrease hospital-acquired infections.^30–33 The increasing number of CA-MRSA genotypes in this study appears to be driven by an expansion of the CMRSA-10 (USA300) clone, which is consistent with other reports.^34,35 Woods et al.³⁶ have commented on high rates (5.1%) of MRSA isolates demonstrating vancomycin MICs of 2 mg/L and increased mortality in patients admitted to a single ICU in Washington, DC. Fortunately, among MRSA in this study only five (1.4%) isolates possessed a vancomycin MIC of 2 mg/L. Furthermore, these isolates appeared sporadically across the duration of the study.

The most active agents against Gram-negative bacilli were the carbapenems, tigecycline and piperacillin/tazobactam, while for P. aeruginosa only amikacin, ceftolozane/tazobactam, colistin and tobramycin demonstrated >90% susceptibility. Unlike S. aureus, Enterobacteriaceae and P. aeruginosa isolated from Canadian ICUs demonstrated a tendency to decreased susceptibility in comparison with the overall CANWARD cohort.¹⁰ The largest differences were noted for the extended-spectrum cephalosporins and piperacillin/tazobactam, as well as meropenem in the case of P. aeruginosa, likely reflecting the extensive use of these agents in the critically ill.¹⁰ Though comparable data are limited, Sader et al.²⁵ have published their findings from the SENTRY Antimicrobial Surveillance Program, which collected 6848 Gram-negative isolates from US and European ICUs between 2009 and 2011. Compared with what is reported here, Sader et al.²⁵ demonstrated similar susceptibility among E. coli from both US and European ICUs. Two exceptions were identified: notably, fluoroquinolone susceptibility was greater among Canadian isolates (6.4% and 10.7% greater for ciprofloxacin and levofloxacin, respectively) when compared with the USA; and piperacillin/tazobactam susceptibility was greater in Canada when compared with both the USA and Europe (3.1% and 9.6% greater when compared with the USA and Europe, respectively).²⁵ For K. pneumoniae isolated from CANWARD, we report increased susceptibility to extended-spectrum cephalosporins, fluoroquinolones, gentamicin, meropenem and piperacillin/tazobactam in comparison with both US and European ICUs, where susceptibility rates were on average 7.2% and 17.1% lower, respectively.^25,P. aeruginosa notably demonstrated higher susceptibility to meropenem, piperacillin/tazobactam and tobramycin when compared with what is reported by SENTRY.²⁵ We report that 26.3% and 14.9% of E. coli were MDR and XDR, respectively. In addition, E. coli, E. cloacae and K. pneumoniae in this study demonstrated a statistically significant trend to increasing MDR. Inter-study comparisons of MDR are limited due to a lack of consistent criteria throughout the literature and differing patient populations. Lockhart et al.³⁷ have reported that MDR E. coli increased from 0.0% in 1993 to 2.0% in 2004 in US ICUs. Additionally, Sievert et al.³⁸ showed small increases in MDR among E. coli and E. cloacae isolated from ICU patients with hospital-acquired infections collected between 2007 and 2010; however, these increases were not statistically significant. The CAN-ICU study reported low levels of MDR in Canadian ICUs, especially among Enterobacteriaceae, while the current study highlights growing rates of MDR within Canada as well as the emergence of XDR pathogens.¹⁹

We have published previously that ESBL-producing E. coli increased significantly in Canadian hospitals from 2007 to 2011.¹⁶ To our knowledge, this study is the first to report a statistically significant increase in ESBL-producing E. coli in Canadian ICUs; the rate of ESBL-producing E. coli in Canada has long trailed behind that in the USA and Europe.³⁹ According to SENTRY, 13.7% and 16.6% of E. coli collected from the ICUs in the USA and Europe, respectively, demonstrated an ESBL phenotype between 2009 and 2011.²⁵ Such a trend as that reported here must continue to be monitored. As the rate of ESBL-producing Enterobacteriaceae increases in Canadian hospitals, increased use of carbapenems will create selection pressure for the expansion of CRE. These data show a low prevalence of CRE in Canadian ICUs, where only 0.5% (11 of 2328) of the top 20 Enterobacteriaceae were found to be meropenem non-susceptible; however, further characterization is required to determine the true rate of carbapenemase-producing Enterobacteriaceae (CPE). Mataseje et al.⁴⁰ have previously published their finding that nationally 0.1% of Enterobacteriaceae collected between 2009 and 2010 demonstrated reduced susceptibility to carbapenems. Beyond this, however, two smaller provincial studies, as well as multiple case reports, suggest CRE and CPE do appear in noteworthy and potentially growing numbers.^41–46

The CANWARD study suffers from a number of limitations. Firstly, we acknowledge that the CANWARD study receives ICU isolates from a small number (10–15) of Canadian tertiary care centres. However, we feel that it is representative, as CANWARD is population-based and encompasses many of the largest centres in each region. Additionally, we are unable to guarantee all isolates submitted represent true infection. The participating centres have been asked to submit only clinically relevant isolates as defined by their individual laboratory criteria; however, we are unable to guarantee that these criteria are followed universally at all times. Secondly, CANWARD is limited in that the demographic data received are minimal, including only simple identifying and descriptive information. This limitation impacts our ability to conduct a more thorough analysis of each isolate received within the clinical context from which it was collected. Finally, this study utilizes in vitro susceptibility testing data in an attempt to predict the activity of antimicrobials in vivo. While pharmacodynamics (PD) are an important component of antimicrobial efficacy, pharmacokinetics (PK) also represent a significant determinant. Altered PK in critically ill patients have been documented consistently in the literature, where failure of multiple organ systems can greatly impact antimicrobial concentration at the site of infection.⁴⁷ As a result, PK/PD target attainment has been shown to be inadequate in large numbers of critically ill patients, a finding that contributes to failure of therapy and poor outcomes.⁴⁸ While such differences are worthy of mention, we believe the analysis as outlined in this article is still valid. However, as antimicrobial activity is multifactorial, the results described must be interpreted within this context of a larger picture.

In summary, this study reports on the prevalence of infectious organisms, including resistant pathogens and antimicrobial resistance patterns, in Canadian ICUs from 2007 to 2016. Key pathogens such as CA-MRSA and ESBL-producing organisms have increased in prevalence in ICUs over time; in addition, increasing rates of MDR were identified in the 10 most common ICU organisms. This study highlights the complexity of selecting empirical antimicrobial coverage for critically ill patients, a task that is further confounded by the potential for decreased susceptibility in ICU patients in comparison with non-ICU areas of the hospital, as well as variable and unpredictable PK/PD in such patients. Additionally, due to the variability of regional susceptibility rates, local antibiograms remain a crucial guide to selecting appropriate empirical therapy.

Acknowledgements

These data were previously presented in part at the ASM Microbe Meeting, Atlanta, GA, USA, 2018 (Poster 407). CANWARD data can also be found at www.can-r.ca, the official website of the Canadian Antimicrobial Resistance Alliance (CARA).

 We would like to thank the centres mentioned below for their participation in the CANWARD 2007–16 study.

Member laboratories of the Canadian Antimicrobial Resistance Alliance (CARA)

Vancouver Hospital (Vancouver, British Columbia), University of Alberta Hospital (Edmonton, Alberta), Royal University Hospital (Saskatoon, Saskatchewan), Health Sciences Centre (Winnipeg, Manitoba), London Health Sciences Centre (London, Ontario), St Joseph’s Hospital (Hamilton, Ontario), Mount Sinai Hospital (Toronto, Ontario), St Michael’s Hospital (Toronto), Children’s Hospital of Eastern Ontario (Ottawa, Ontario), The Ottawa Hospital (Ottawa), Royal Victoria Hospital (Montreal, Quebec), Montreal General Hospital (Montreal), Hôpital Maisonneuve-Rosemont (Montreal), Centre Hospitalier Universitaire Sherbrooke (Sherbrooke, Quebec), CHRTR Pavillon Ste Marie (Trois-Rivières, Quebec), Hôpital de la Cité-de-la-Santé (Laval, Quebec), L’Hôtel-Dieu de Quebec (Quebec City, Quebec), South East Regional Health Authority (Moncton, New Brunswick) and Queen Elizabeth II HSC (Halifax, Nova Scotia).

Funding

Funding for CANWARD was provided in part by the University of Manitoba, Health Sciences Centre (Winnipeg, Manitoba, Canada), Astellas, Merck, Pfizer, Sunovion, The Medicines Company, Abbott, Achaogen, Cubist, Paladin Labs, Bayer, Janssen Ortho/Ortho McNeil, Affinium, Basilea, AstraZeneca, Paratek, Tetraphase, Theravance, Sanofi-Aventis and Zoetis.

Transparency declarations

D. J. H. and G. G. Z. have received research funding from Astellas, Merck, Pfizer, Sunovion, The Medicines Company, Abbott, Achaogen, Cubist, Paladin Labs, Bayer, Janssen Ortho/Ortho McNeil, Affinium, Basilea, AstraZeneca, Paratek, Tetraphase, Theravance, Sanofi-Aventis and Zoetis. All other authors: none to declare.

References