Asymptomatically colonized patients may account for much of Clostridioides difficile acquisition in leukemia and hematopoietic cell transplant units, with symptomatic infection contributing relatively little to new colonization.
Investigators conducted a prospective cohort study in 2 leukemia and hematopoietic cell transplant units from January 19 to July 22, 2019. Among 421 patients admitted during the study period, 91% (n = 384) consented to testing. Stool samples and rectal swabs were collected at enrollment and weekly until discharge for C difficile culture and nucleic acid amplification testing (NAAT), while clinically indicated toxin enzyme immunoassay testing was used to identify C difficile infection (CDI). The investigators integrated these data with patient movements, room assignments, health care worker staffing patterns, and environmental sampling in a stochastic, individual-based network model of C difficile acquisition and transmission.
The investigators quantified how much new C difficile colonization resulted from importation vs. within-unit transmission and the relative contributions of asymptomatic colonization (AC) and CDI to each pathway. The patients were modeled as being susceptible, AC, or having CDI, with the model accounting for environmental contamination and connections between rooms. An observation model incorporated actual testing schedules and test performance to account for acquisition events that might not have been detected through testing.
Across 1,000 simulations of the calibrated model, 51% of new colonizations were attributed to importation and 49% to within-unit transmission. Among imported colonizations, 87% involved patients admitted as AC. patients who were AC were the source of 92% of transmission events resulting in new colonization, whereas CDI contributed relatively little to new colonizations.
Testing identified a minority of acquisition events. The transmission model estimated 3.9 new colonization events per week, whereas the observation model identified 0.9. The discrepancy highlighted the extent to which acquisition could remain undetected under the testing approach.
The acquisition pattern differed when the analysis was restricted to patients who developed CDI. In that analysis, 87% of new CDI cases were attributed to importation. Among imported cases, 76% involved patients admitted with CDI; the remainder involved patients admitted as AC who subsequently progressed to CDI. When within-unit transmission resulted in CDI, AC patients were the modeled source in 92% of events.
The calibrated model generally reproduced the available surveillance data. Among 26 weekly observations, 19 AC observations and 23 CDI observations fell within the interquartile range of the model simulations, and model-inferred transmission pairs overlapped with the 2 transmission pairs previously identified through whole-genome sequencing. The investigators found that new colonizations were most sensitive to the modeled rates of colonization, environmental contamination, and cleaning.
Source attribution was calibrated using testing data from just 18 CDI cases and 85 AC cases during the 6-month study, resulting in wide ranges and leaving the attribution estimates only weakly constrained by the available data. The study was conducted at a single health care facility, potentially limiting generalizability. The contact network captured whether rooms shared health care workers but not how frequently staff entered rooms, who entered them, or the duration of contact. Transmission from family members and visitors was also not included, which could bias the model toward overestimating importation.
Asymptomatic colonization, particularly colonization present at admission, may play an important role in sustaining C difficile acquisition in these high-risk oncology units. The comparatively small modeled contribution of CDI to new colonizations was consistent with infection-control measures limiting onward transmission from patients with recognized infection. However, targeting AC may be difficult because colonization frequently went undetected, and most patients modeled as AC did not progress to CDI.
“Interventions targeting AC can further reduce CDI by either reducing onward transmission from AC or by reducing the progression from AC to CDI,” wrote lead study author Savannah Curtis, of North Carolina State University, and colleagues.
Full disclosures of the study authors can be found in the study.
