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

Disclosures

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

In This Article

Methods

Case Definition

We used previously described case definitions for primary CDI[14] (Appendix, https://wwnc.cdc.gov/EID/article/28/6/21-2262-App1.pdf). A case of HA CDI was defined on the basis of laboratory confirmation of CDI and a compatible clinical syndrome developing ≥72 hours after admission, or <72 hours after admission if the patient had a previous admission to the hospital and was discharged within the previous 4 weeks. CA CDI was defined as clinical manifestation of CDI symptoms ≤72 hours before admission with no history of hospitalization or healthcare exposure, including outpatient healthcare exposures, within the previous 12 weeks.

Severe outcomes were defined as CDI-attributable admission to an intensive care unit (ICU), colectomy, or death ≤30 days after admission. All deaths were reviewed by an infectious disease physician or medical microbiologist by using clinical judgement to determine whether deaths were CDI-attributable.

Data Sources and Collection

CNISP has conducted prospective surveillance for HA CDI in hospitalized patients in Canada since 2007, and CA CDI surveillance was added in 2015. By 2019, CNISP included a network of 76 acute care hospitals across 10 provinces and 1 territory (Appendix Table 1). We analyzed data collected during 2015–2019 from adult and mixed (adult and pediatric) hospitals. The Canadian Network for Public Health Intelligence collected and verified clinical and laboratory surveillance data to ensure accuracy, as previously described.[14]

Bacterial Culture and Molecular Characterization

We performed C. difficile isolation by using an ethanol shock treatment method, then selected for C. difficile on Clostridium difficile Moxalactam Norfloxacin agar (Oxoid, https://www.oxoid.com), as previously described.[15,16] We prepared DNA for PCR analysis and ribotyping by using InstaGene Matrix (Bio-Rad, https://www.bio-rad.com), as previously described.[17] We performed multiplex PCR targeting toxin A (tcdA), toxin B (tcdB), binary toxin (cdtB), negative regulator of toxin production (tcdC), and triose phosphate isomerase (tpi) housekeeping gene, as previously described,[15,18,19] with slight modifications. We substituted an in-house A3B primer (5'-ACCATCAATCTCGAAAAGTCCAC-3') for the tcd-R reverse primer for detecting tcdA (420 bp amplicon) and the detection of tcdA deletion variants (147 bp amplicon).

PCR Ribotyping

We performed capillary gel electrophoresis–based ribotyping targeting the 16S-23S intergenic spacer region, as previously described.[17] We assigned RTs by comparing query profiles to those of a reference set of RTs used in a previous multicenter international study.[17]

Antimicrobial Susceptibility Testing

We used Etest strips (bioMérieux, https://www.biomerieux.com) to perform susceptibility testing for metronidazole, clindamycin, vancomycin, rifampin, moxifloxacin, and tigecycline, as previously described.[16,20] We interpreted antimicrobial resistance in accordance with Clinical and Laboratory Standards Institute guidelines.[20]

Statistical Analysis

We calculated HA CDI incidence rates as number of cases per 10,000 patient-days and CA CDI incidence rates as number of cases per 1,000 patient admissions. We used the Cochran-Armitage test for categorical variables and the Mann-Kendall test for continuous variables to assess statistically significant trends over time for patient characteristics and outcome results. To compare characteristics of patients with HA CDI versus CA CDI, we used the χ 2 test for categorical variables and the Student t test or Wilcoxon rank sum test for continuous variables.

We used multivariable logistic regression to model the association between RTs and outcomes (i.e., 30-day all-cause and 30-day CDI-attributable mortality) and adjusted for a priori–selected confounders of age, sex, severe CDI cases (albumin level <30 g/L, leukocyte count ≥15 ×109/L, or both), and CDI case types (i.e., HA vs. CA CDI). We used 2-tailed statistical tests and considered p≤0.05 statistically significant. We performed all analyses in SAS version 9.4 (SAS Institute Inc., https://www.sas.com).

processing....