Rectal colonization with Carbapenem-resistant Enterobacterales in intensive care unit patients: prevalence, clinical outcomes, and risk of subsequent infection

Carbapenem-resistant Enterobacterales (CRE) colonization is an important problem in intensive care units (ICUs) and may precede invasive infection. We evaluated the prevalence, clinical course, and outcomes of rectal CRE colonization and investigated risk factors for infection occurring after colonization. This single-center retrospective cohort included adult ICU patients screened by rectal culture at admission and weekly from July 2022 through August 2025. The primary outcome was culture-proven CRE infection occurring after the colonization date; infections detected on the colonization date were evaluated separately. Time zero was the first documented CRE colonization date. Cumulative incidence and risk factors for subsequent infection were evaluated using competing-risk methods, with ICU death and live ICU discharge before infection treated as competing events. CRE colonization was identified in 287 of 3,221 ICU admissions (8.9%). Colonization was present at admission in 88 (30.7%) and acquired during the ICU stay in 199 (69.3%). Sixteen infections were detected on the colonization date and 132 occurred subsequently; therefore, the primary analysis included 271 admissions. The cumulative incidence of subsequent CRE infection was 25.8% at 7 days, 36.2% at 14 days, and 45.0% at 28 days. Mortality was higher among patients with subsequent infection than among those without infection (71.2% vs. 55.4%, p = 0.007). No evaluated factor was independently associated with infection in the Fine-Gray model; prior colistin exposure was borderline but not statistically significant (sHR 1.53, 95% CI 0.99–2.37; p = 0.057). Almost half of evaluable CRE-colonized ICU patients developed infection after colonization, frequently within the first two weeks. These findings support active surveillance and close clinical monitoring while emphasizing the importance of accounting for time at risk when evaluating predictors of infection.