Device-associated infections in Canadian acute-care hospitals from 2009 to 2018 Canadian Nosocomial Infection Surveillance Program

Background: Healthcare-associated infections (HAIs) pose a serious risk to patient safety and quality of care. The Canadian Nosocomial Infection Surveillance Program (CNISP) conducts national surveillance of HAIs at sentinel acute-care hospitals across Canada. This report provides an overview of 10 years of Canadian data on the epidemiology of select deviceassociated HAIs. Methods: Over 40 hospitals submitted data between 2009 and 2018 for hip and knee surgical site infections (SSIs), cerebrospinal fluid shunt SSIs, paediatric cardiac SSIs and/or central lineassociated bloodstream infections (CLABSIs). Counts, rates, patient and hospital characteristics, as well as pathogen distributions and antimicrobial susceptibilities are presented. Results: A total of 4,300 device-associated infections were reported. Central line-associated bloodstream infections were the most common device-associated HAI reported (n=2,973, 69%) and hip and knee arthroplasty infections were the most common SSIs reported (66% of SSIs). Our findings show decreasing CLABSI rates in neonatal intensive care units (4.2 to 1.9 per 1,000 line-days, p<0.0001) and decreasing knee SSI rates (0.69 to 0.30 infections per 100 surgeries, p=0.007). Rates of device-associated HAIs have remained relatively consistent over the 10-year surveillance period. Overall, 4,599 pathogens were identified from device-associated HAI; 70% of these were related to CLABSIs. Coagulase-negative staphylococci (29%) and Staphylococcus aureus (14%) were the most frequently reported pathogens. Gram-positive pathogens represented 68% of identified pathogens, gram-negative pathogens represented 22% and fungi represented 9%. Conclusion: Understanding the national burden of device-associated HAIs is essential for developing and maintaining benchmark rates for informing infection and prevention control and antimicrobial stewardship policies and programs. Affiliation 1 Public Health Agency of Canada, Ottawa, ON *Correspondence: phac.cnisp-pcsin.aspc@canada.ca Suggested citation: Canadian Nosocomial Infection Surveillance Program. Device-associated infections in Canadian acute-care hospitals from 2009 to 2018. Can Commun Dis Rep 2020;46(11/12):387–97. https://doi.org/10.14745/ccdr.v46i1112a05


Introduction
Healthcare-associated infections (HAIs) pose a serious risk to patient safety and quality of care and contribute to prolonged hospital stays, increased antimicrobial resistance, costs to the health system and unnecessary deaths (1). Risk factors for HAIs include the use of invasive devices, surgical procedures and inappropriate antibiotic use (2). In Canada, surgical site infections (SSIs) affect an estimated 26,000 to 65,000 patients annually (3). In a 2017 Canadian point prevalence study at sentinel hospitals, device-associated infections accounted for 35.6% of all HAIs reported. Of the device-associated infections, SSIs associated with a prosthetic implant accounted for 19.4% and central line-associated bloodstream infections (CLABSIs) accounted for 21.2% (4).
Device-associated HAI antimicrobial susceptibility information has important implications for antibiotic resistance (5); impacting length of stay and healthcare costs (6). Cumulative antibiograms are a valuable resource for clinical decision-making while sensitivity results are pending (7). The risk of device-associated HAIs varies among patient populations and hospital types; patients admitted to the intensive care unit (ICU) are at higher risk of developing an HAI (8).
Understanding the trends in device-associated HAIs is essential to effective infection prevention and control. Drawing on a decade of HAI data (2009−2018) from over 40 sentinel acute-care hospitals across Canada participating in the Canadian Nosocomial Infection Surveillance Program (CNISP), this report provides an epidemiological overview of select device-associated HAIs.

Methods Design
Established in 1994, the CNISP, a collaboration between the Public Health Agency of Canada, the Association of Medical Microbiology and Infectious Disease Canada and sentinel hospitals across Canada, conducts national HAI surveillance at sentinel acute-care hospitals across Canada. This report presents data on device-associated HAIs for the following infections: hip and knee SSIs; cerebrospinal fluid shunt SSIs (CSF-shunt-SSIs); paediatric cardiac surgical site infections (paediatric-cardiac-SSIs); and CLABSIs.

Case definitions
Device-associated HAIs were defined according to standardized protocols and expert-reviewed case definitions (Appendix 1). Only CLABSIs identified in ICU settings were included in surveillance. Only complex infections, defined as deep incisional and organ space, were included in hip and knee SSI surveillance.

Data source
Participating hospitals submitted epidemiological data on CSF-shunt-SSIs and CLABSIs occurring between January 1, 2009 and December 31, 2018. Paediatric-cardiac-SSI surveillance started in January 2010. Hip and knee SSI surveillance started in January 2011. Data submission and case identification were supported by annual training sessions and continuous evaluations of data quality.

Statistical analysis
CLABSI rates were calculated by dividing the number of cases by line-day denominators. Hip and knee SSI, CSF-shunt-SSI and paediatric-cardiac-SSI rates were calculated by dividing the number of cases by surgery denominators. Proportions of pathogens were calculated by dividing the number of pathogens by the total number of pathogens identified. Missing and incomplete data were excluded from analyses, therefore denominators may vary. Interquartile ranges (IQR) were calculated. The Mann-Kendall test or negative binomial regression was used to test trends over time. Significance testing was two-tailed and differences were considered significant at p-value ≤0.05. Analyses were conducted using Excel and SAS 9.4.
Overall, 4,599 pathogens were identified from device-associated HAI cases between 2014 and 2018; 69.8% of these were related to CLABSIs. Coagulase-negative staphylococci and Staphylococcus aureus were the most frequently reported pathogens (

Antibiogram
Antimicrobial susceptibility testing results for the most frequently identified gram-positive, gram-negative and fungal pathogens from device-associated HAIs are listed in Table 7

Surgical site infections
Hip and knee-SSIs were the most common SSI reported in our surveillance. Similar to results from the European Centre for Disease Prevention and Control, a decreasing trend in knee SSI was observed among CNISP hospitals, while hip SSI remained stable (9). In addition, a US point prevalence study observed a significant reduction in the prevalence of complex SSIs between 2011 and 2015 (10). Our findings indicate that the most common pathogens identified among hip and knee-SSIs were S. aureus and coagulase-negative staphylococci, consistent with results from other regions (9,11). Frequent identification of S. aureus and coagulase-negative staphylococci may be related to the use of implant devices and contamination from the patient's endogenous skin flora (5). Hip and knee-SSIs affect an older population as joint replacements typically occur among older adults (12). As populations age, hip and knee joint replacements are rising and are linked to a rise in surgical complications (i.e. prosthetic joint infections) (12). High observed rates of readmission and revision surgery highlight the financial and resource burden placed on the healthcare system due to hip and knee-SSI (13).
Our overall rate of CSF-shunt-SSIs (n=3.2/100 surgeries) is on the lower end of what is reported internationally; a 2012 review found that reported rates of infection vary from 3% to 12% of shunt operations (14). Stratification of our CSF-shunt-SSI data by paediatric or adult hospital showed little difference in infection rates and in pathogen distributions between paediatric and adult/mixed settings. However, a previous study among CNISP hospitals, conducted between 2000 and 2002, had identified that CSF-shunt-SSIs were more common in children than in adults (15). In this earlier study, the infection rate among paediatric patients was higher than found in this study ( There was a significant increase in the rate of paediatriccardiac-SSI in 2018 to 7.5/100 surgeries. This increase was limited to two hospital sites, where investigations are ongoing. This increase should be interpreted with caution as rates are calculated from a small number of cases and may be sensitive to random fluctuation at individual hospitals.

Central line-associated bloodstream infections
Central line-associated bloodstream infections were the most commonly reported device-associated HAI (69% of included HAIs); however, it is important to note that the number of hospitals participating in the surveillance of each HAI differs and that the surveillance periods for some HAIs were shorter. In a point prevalence study of HAIs, the frequencies of SSIs (19%) and CLABSIs (21%) were very similar (5).
There were no substantive changes in CLABSI rates among surveyed adult ICUs or PICUs; however, there was a 55% decrease in CLABSI rates among NICUs.

Strengths and limitations
The strength of this study lies in the standardized collection of detailed data from a large network of sentinel hospitals over a decade. While the CNISP network extends across Canada, participating hospitals may not be representative of the general Canadian inpatient population; hospitals participating in CNISP tend to be larger, teaching hospitals in urban centres. The CNISP is currently undergoing a recruitment process to increase representativeness and bed coverage, especially in northern, rural and indigenous populations. The CNISP's data, although standardized, may be sensitive to changes in hospital participation, infection prevention and control practices and the application of surveillance definitions. Differences in surveillance protocols and case definitions limit the ability to compare data from other countries. However, the data presented in this report are routinely used by Canadian hospitals for benchmarking.
For CLABSI surveillance, we do not have data on infections occurring outside of ICU settings; however, in the US, CLABSIs outside of the ICU setting represented 55% of all CLABSIs (19)

Conclusion
This report provides an updated summary of rates, pathogen distributions and antimicrobial resistance among select deviceassociated HAIs and relevant pathogens. Understanding the national burden of device-associated HAIs is essential for developing and maintaining benchmark rates for informing infection and prevention control and antimicrobial stewardship policies and programs.

Funding
This work was supported by the Public Health Agency of Canada.