Characterization of Healthcare-Associated and Community-Associated Clostridioides Difficile Infections Among Adults, Canada, 2015–2019

Tim Du; Kelly B. Choi; Anada Silva; George R. Golding; Linda Pelude; Romeo Hizon; Ghada N. Al-Rawahi; James Brooks; Blanda Chow; Jun C. Collet; Jeannette L. Comeau; Ian Davis; Gerald A. Evans; Charles Frenette; Guanghong Han; Jennie Johnstone; Pamela Kibsey; Kevin C. Katz; Joanne M. Langley; Bonita E. Lee; Yves Longtin; Dominik Mertz; Jessica Minion; Michelle Science; Jocelyn A. Srigley; Paula Stagg; Kathryn N. Suh; Nisha Thampi; Alice Wong; Susy S. Hota


Emerging Infectious Diseases. 2022;28(6):1128-1136. 

In This Article


Using 5 years of CDI surveillance data from acute care hospitals across Canada, we observed a decline in rates of HA and CA CDI that coincided with a marked change in the prevalence of predominant circulating ribotypes. The epidemiologic and molecular characterization of HA and CA CDI revealed differences in patient characteristics and select clinical outcomes, with associations to predominant ribotypes.

The decline in CDI rates in Canada follows a parallel trend observed globally, despite rates being higher in North America.[10,21] We previously reported HA CDI rates ranging from 2.1 to 6.6 cases/10,000 inpatient days during 2011–2016 but showing a decreasing trend over time.[13] We noted an increase in CA CDI rates in that study, but in this study, we found that rates of CA CDI have decreased since 2015. Although the precise reasons for decreased CDI incidence in Canada are unclear, enhanced infection control and antimicrobial stewardship measures combined with improved surveillance methods might have contributed to the overall decline.[22] Furthermore, patients with mild to moderate CA CDI might not be admitted to or tested in a hospital, resulting in underestimation of the true burden of CA CDI.

Although molecular surveillance of CDI in Canada revealed a dynamic and heterogeneous RT population, the predominant circulating types were RT027, RT106, RT020, and RT014. Similar to findings in this study, RT027 has been reported to be decreasing in prevalence in North America, the United Kingdom, and elsewhere; however, RT027 remains a major cause of CDI.[1,23–26] In Canada, the dramatic decrease in RT027 prevalence in HA CDI has continued since its initial reporting.[1] Declining trends observed among HA (−15.2%) and CA (−12.0%) CDI during 2015–2019 in Canada are also consistent with trends in the United States, where HA CDI rates declined from 21% to 15% and CA CDI declined from 17% to 6% during 2012–2017.[25]

Although RT027 prevalence in Canada decreased during 2015–2019, RT106 greatly increased during the same period, from 7.3% to 18.1% in healthcare settings and from 6.5% to 17.6% in community settings. Identified in the United Kingdom in 1999,[27] RT106 is now found worldwide and is one of the most predominant strains in the United States.[28] Studies indicate that RT106 has enhanced spore-producing and biofilm formation capabilities that enable greater persistence in the environment and hospital settings, possibly leading to increased infection rates.[28,29] In addition, studies report that patients infected with RT106 are more likely to experience multiple episodes of relapse.[28,30]

C. difficile RT078 and RT126, which have demonstrated epidemic potential in other countries,[31–33] appear to be uncommon in patients hospitalized with CDI in Canada. Our surveillance shows a small increase in RT078 and RT126 prevalence since a previous report showed rates of 2.4% in HA and 1.9% CA CDI populations.[14]

Similar to previous findings, our study showed that CA CDI patients were more likely to be younger and female.[10,34–36] In addition, this study found that HA CDI is associated with an increased risk for 30-day all-cause mortality compared with CA CDI; however, this association did not remain significant for CDI-attributable deaths. Hospitalized patients with CDI possibly are exposed to other risks and complications during their hospital stay or have underlying conditions that increase their risk for all-cause death. Our findings agree with previously published studies assessing all-cause and CDI-attributable death.[34,37]

We further analyzed the effects of RT027 and RT106, the 2 most prevalent C. difficile strains, on all-cause and CDI-attributable death. We previously showed a significant association between RT027 and attributable mortality.[1] In this study, we concluded that RT027 is a significant predictor of CDI-attributable death even after adjusting for case type (HA or CA CDI). We noted no association between RT106 and all-cause and CDI-attributable deaths.

We found that C. difficile antimicrobial resistance is less common in Canada than in the United States or globally.[38] Stratified by case type, HA and CA CDI isolates revealed no significant difference in resistance, except for moxifloxacin resistance, which was 21.7% for HA and 12.4% for CA CDI, consistent with previously reported findings.[30] In addition, diverse RT populations observed in both HA and CA CDI might be predicative of lower resistance rates observed because RT heterogeneity has been shown to be inversely correlated with antimicrobial resistance as measured by cumulative resistance scores.[12,39]

Our study's first limitations is that hospital participation in HA and CA CDI surveillance varied by year and might have affected trends over time. Furthermore, hospitals self-select whether to participate in both HA and CA CDI surveillance, which might have influenced the comparison of HA and CA CDI patients. To mitigate this limitation, we conducted a sensitivity analysis restricted to hospitals that conducted both HA and CA CDI surveillance. Second, although CDI diagnostic testing methods were collected throughout the study period, data completeness was not consistent from year to year, limiting the inferences we could make regarding the effect of CDI diagnostic testing methods on adult CDI rates over time. Third, for CA CDI surveillance, our study captured data from patients admitted to a CNISP hospital and requiring medical intervention for CDI symptoms or other underlying conditions. The features and outcomes of these patients might not be relevant to patients with CA CDI who do not require hospital care. Finally, although a qualified physician determined the cause of death in patients with CDI, attribution of death is difficult and could be subjective.

In conclusion, rates of HA and CA CDI in Canada declined during 2015–2019, coinciding with a decrease in prevalence of RT027 and increased prevalence of RT106. HA CDI was associated with higher rates of all-cause death than was CA CDI, and RT027 was a major predictor of CDI-attributable death, irrespective of location of acquisition. We noted major decreases in antimicrobial resistance to moxifloxacin in both HA and CA CDI populations, concordant with an overall decrease in prevalence of RT027. Despite declining rates, CDI continues to be a major health burden in Canada. To ensure continued success in combatting this global health threat, robust national surveillance and infection prevention and control programs are integral to clarifying CDI epidemiology, investigation, and control.