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


Our analysis included a total of 18,455 adult inpatient cases of primary CDI from participating hospitals during 2015–2019. HA CDI accounted for 74.4% (n = 13,735) of cases and CA for 25.6% (n = 4,720). The number of hospitals participating in HA CDI surveillance each year ranged from 58–64, and 46–54 hospitals participated in CA CDI surveillance (Appendix Table 1). We also completed a sensitivity analysis to restrict hospitals that conducted both HA and CA CDI surveillance but observed no stastically significant differences in results (data not shown).

During 2015–2019, HA CDI rates decreased by 23.8%, from 4.74 to 3.61 cases/10,000 patient-days (p<0.02), and CA CDI rates decreased by 18.8%, from 1.33 to 1.08 cases/1,000 admissions (p<0.33) (Figure 1). Regionally, HA CDI rates decreased significantly in the central (p<0.02) and western (p<0.02) regions of Canada, but rates fluctuated in the eastern region (p = 0.62), peaking at 4.06 cases/10,000 patient-days in 2019. Despite a decline, CA CDI infection rates remained highest in the central region, at 1.53 cases/1,000 admissions in 2019. Of the 64 hospitals for which data were available for adult CDI surveillance, 58 (91%) reported data for the entire 5-year period of surveillance. After restricting our analyses to these 58 hospitals, interpretation of our results did not change. Incidence rates for HA CDI decreased by 22.8%, CA CDI incidence decreased by 18.0%, and rates were consistent with those reported and generated with data from 64 hospitals.

Figure 1.

National and regional healthcare-associated (A) and community-associated (B) Clostridioides difficile infection rates among adults, Canada, 2015–2019. Western region is British Columbia, Alberta, Saskatchewan, and Manitoba; Central region is Ontario and Quebec; Eastern region is Nova Scotia, New Brunswick, Prince Edward Island, and Newfoundland and Labrador.

We aggregated patient characteristics and outcomes by case type (Table 1). Compared with HA CDI patients, CA CDI patients were younger (median age 67.0 vs. 70.0 years; p<0.01), and more CA CDI patients were female (56.0% vs. 49.1% male; p<0.01).

Clinical Manifestations

Of the 18,455 cases, 3,084 had clinical and outcome data available; these data are collected during a 2-month targeted surveillance period (March–April) each year. Overall, 10.4% (316/3,033) of patients with CDI died, and 2.9% (86/3,019) of deaths were CDI-attributable (Table 1). Of 316 deaths among patients with CDI, 27.2% (86/316) were CDI-attributable. Patients with HA CDI had significantly higher 30-day all-cause mortality than patients with CA CDI (11.4% vs. 7.3%; p<0.01). Of 3,073 patients with CDI, 207 (6.8%) required ICU admission, 27.5% (57/207) of whom were admitted because of CDI complications, and 1.9% (57/3,073) all patients with CDI were admitted to the ICU because of CDI complications. We observed no statistically significant differences in ICU admission by acquisition type.

During 2015–2019, ICU admission data were available for 2,340 HA CDI patients (433–507 patients annually). ICU admissions decreased significantly among HA CDI cases, from 9.1% (46/507) in 2015 to 5.9% (26/442) in 2019 (p<0.02). We saw no statistically significant trends for age, sex, or 30-day outcomes for all-cause or CDI-attributable deaths (Appendix Table 2).

Ribotyping Analysis

Of the 18,455 cases, a total of 3,189 stool samples were received for C. difficile isolation at the National Microbiology Laboratory (Winnipeg, Manitoba, Canada), and 2,506 samples met inclusion criteria. Of samples tested, 1,887 (75.3%) were HA CDI and 619 (24.7%) were CA CDI. We performed capillary gel electrophoresis ribotyping and antimicrobial susceptibility testing to further characterize isolates.

Among 1,887 HA CDI isolates characterized during the study period, we noted 170 unique PCR RTs (Figure 2). The most common RTs among HA CDI were RT027 (16.0%), RT106 (11.5%), RT014 (8.6%), RT020 (6.4%), and RT002 (5.7%). The 15 most prevalent RTs accounted for 69.6% of isolates tested (Appendix Table 3). The prevalence of RT027 in HA CDI cases decreased from 24.6% in 2015 to 9.4% in 2019 (p<0.01), but the incidence of RT106 increased from 7.3% in 2015 to 18.1% in 2019 (p<0.01).

Figure 2.

Prevalence of Clostridioides difficile ribotypes detected each year from healthcare-associated (A) and community-associated (B) infections among adults, Canada, 2015–2019.

Of 619 CA CDI isolates, we noted 115 unique RTs, of which RT106 (12.3%), RT020 (8.4%), RT014 (8.1%), RT027 (7.9%), and RT056 (5.0%) were the most prevalent. For CA CDI, the 15 most prevalent RTs accounted for 66.1% of isolates tested (Appendix Table 3). As for HA CDI, we noted a decrease in prevalence of RT027, from 14.8% in 2015 to 2.8% in 2019 (p<0.01) and RT106 increased from 6.5% in 2015 to 17.6% in 2019 (p<0.01). Despite a steady decline in prevalence over the study period, RT027 remained the most commonly isolated strain type with an overall combined prevalence of 14.0% (351/2,506 isolates).

RT078 and RT126 are livestock-associated strains that correlate with increased virulence and disease severity and have been identified in human CDIs. RT078 and RT126 prevalence among HA CDI cases averaged 2.4% (range 2.0%–3.2%), but for CA CDI, RT078 and RT126 prevalence averaged 1.9% (range 0.8%–3.2%) (Appendix Table 4).

All-cause and CDI-attributable Deaths

Among patients with reported 30-day all-cause mortality (n = 316) and 30-day CDI-attributable mortality (n = 86), most were HA CDI: 80.2% (263/316) of all-cause and 83.7% (69/86) of CDI-attributable deaths. In addition, more deaths occurred among female patients, who made up 55.4% (175/316) of all-cause and 57.0% of CDI-attributable (49/86) deaths, and more patients ≥65 years of age, who comprised 79.8% (252/316) of all-cause and 83.7% of CDI-attributable deaths (72/86).

After multivariable analysis, patient characteristics significantly associated with 30-day all-cause mortality and 30-day CDI-attributable mortality were age ≥65 years and severe CDI (Table 2). The adjusted odds ratio of 30-day all-cause mortality among patients with HA CDI was 1.83 (95% CI 1.23–2.72) times more than for patients with CA CDI (p<0.01). Similarly, the adjusted odds ratio of 30-day CDI-attributable mortality was 1.25 (95% CI 0.67–2.35) times higher among HA CDI than CA CDI, but this difference was not statistically significant.

Analysis of RT027 and RT106 Outcomes

Among 2,320 case-patients with available data on 30-day all-cause mortality, 316 (13.6%) were reported to have died (Appendix Table 5). Of 235 deaths among patients with associated ribotyping data, 44 (18.7%) deaths were associated with RT027 and 30 (12.8%) deaths with RT106. Among RT027 cases, a significantly higher proportion of 30-day all-cause mortality was associated with HA CDI cases than with CA CDI cases (p = 0.01). We saw no statistically significant difference in 30-day all cause mortality between HA and CA CDI cases associated with RT106. We also saw no statistically significant difference in CDI-attributable deaths when stratified by HA and CA CDI cases for RT027 and RT106.

Of 162 cases with severe outcomes for which ribotype analysis was available in the HA CDI population, 33 (11.7%) were associated with RT027 and 10 (4.8%) were associated with RT106 (p<0.01). We also noted a small number of severe outcomes associated with RT027 (n = 2) and RT106 (n = 3) in CA CDI cases; however, we noted no statistically significant differences between HA and CA CDI.

Multivariate analysis found RT027 was significantly associated with 30-day CDI-attributable mortality (adjusted odds ratio [aOR] 2.85, 95% CI 1.64–5.00) compared with non-RT027 cases (p<0.01). However, the association of RT027 with the outcome of 30-day all-cause mortality did not remain statistically significant compared with non-RT027 cases when controlling for other factors within the multivariate model (aOR 1.10, 95% CI 0.74–1.63). When compared with non-RT106 cases, RT106 was not significantly associated with either 30-day all-cause (p = 0.68) or CDI attributable (p = 0.88) mortality in the univariate model.

Antimicrobial Susceptibility

We conducted antimicrobial resistance testing for HA and CA CDI isolates collected during 2015–2019 (Figure 3; Appendix Tables 6, 7). During the study years, HA CDI resistance was 21.7% to moxifloxaxin, 31.0% to clindamycin, and 1.9% to rifampin and CA CDI resistance was 12.4% to moxifloxacin, 33.6% to clindamycin, and 1.5% to rifampin. Of note, HA CDI resistance to moxifloxacin decreased from 34.3% in 2015 to 13.5% in 2019. Similarly, CA CDI resistance to moxifloxacin declined from 18.7% in 2015 to 11.1% in 2019. Resistance to clindamycin was more variable in both study populations, overall resistance was 32.3% (range 19%–54%) (Figure 3).

Figure 3.

Antimicrobial resistance rates for HA and CA Clostridioides difficile infections among adults, Canada, 2015–2019. CA, community-associated; HA, healthcare-associated.

RT027 accounted for 60.2% (293/487) of identified moxifloxacin-resistant isolates. Of note, 83.5% (293/351) of all RT027 isolates examined were moxifloxacin-resistant, of which 97.3% (285/293) had MICs ≥32 μg/mL. Among RT027 isolates, resistance was higher in HA CDI (85.4%; 258/302) than CA CDI (71.4%; 35/49) cases. In contrast, RT106, the second most prevelant strain type (11.7%), accounted for 6.0% of all moxifloxacin-resistant isolates. Fluoroquinolone resistance in RT106 isolates was much lower (9.9%; 29/293), and resistance values were similar for both HA (10.6%) and CA settings (7.9%).

RT027 strains also were more likely to be associated with resistance to ≥1 antimicrobial drug. Of 172 isolates resistant to both moxifloxacin and clindamycin, 79 (45.9%) were RT027. Of 22 isolates found to be resistant to moxifloxacin, clindamycin, and rifampin, 68.2% (15/22) were RT027; of these, 12 were from HA CDI cases and 3 were from CA CDI cases. No other RT strain exhibited resistance to ≥1 drug with a prevalence >5%.

We did not observe resistance for metronidazole, vancomycin, or tigecycline for any study year in either HA or CA CDI populations. One adult patient with HA CDI in 2019 had intermediate susceptibility to vancomycin (MIC 6 μg/mL) but sensitivity to all other drugs tested.