Longitudinal acquisition and infection risk associated with multidrug-resistant organisms and Candida colonization in intensive care unit patients

Intensive care unit (ICU) patients are at high risk for colonization with multidrug-resistant organisms (MDROs) and Candida species, which frequently precedes invasive infections. However, the longitudinal acquisition dynamics and time-dependent relationship between colonization and subsequent clinical infection remain incompletely defined. We conducted a prospective longitudinal cohort study including 290 adult ICU patients. Patients were followed throughout their ICU stay with serial weekly surveillance cultures. Active surveillance screened for four major organism groups: carbapenem-resistant or extended-spectrum β-lactamase (CR/ESBL)-producing Gram-negative bacteria (GNB), methicillin-resistant Staphylococcus aureus (MRSA), vancomycin-resistant enterococci (VRE), and Candida species. Prior colonization was modeled as a time-varying exposure. The association between colonization and subsequent infection was evaluated using time-dependent Cox proportional hazards models with a one-week lag period. The mean ICU length of stay was 23.4 ± 19.1 days. Overall cumulative colonization rates were 37.6% (n = 109) for CR/ESBL-producing GNB, 25.5% (n = 74) for VRE, 16.6% (n = 48) for Candida species, and 16.2% (n = 47) for MRSA. Although VRE was the second most common colonizer, zero clinical VRE infection events occurred during ICU follow-up, precluding its inclusion in multivariable infection risk modeling. In time-dependent Cox regression analyses, prior colonization significantly increased the hazard of subsequent clinical infection across all evaluable pathogen groups: CR/ESBL-producing GNB (adjusted hazard ratio [aHR] 2.89, 95% confidence interval [CI] 1.82–4.59, P < 0.001), MRSA (aHR 4.02, 95% CI 1.37–11.79, P = 0.011), and Candida species (aHR 3.99, 95% CI 1.78–8.92, P = 0.001). Prolonged ICU stay and pre-acquisition broad-spectrum antimicrobial exposure were consistent independent predictors of colonization acquisition. Prolonged ICU stay is associated with progressive cumulative acquisition of MDRO and Candida colonization, which was independently associated with an increased hazard of subsequent infection. Although VRE colonization was relatively frequent, no transition to clinical infection was observed during ICU follow-up. Given the observational design of our study, this finding should be interpreted cautiously as a reflection of sample size constraints rather than a measure of infection-control efficacy.