Objective: The proportion of intensive care unit (ICU)-acquired infections that are a consequence of nosocomial cross-transmission between patients in tertiary ICUs is unknown. Such information would be useful for the implementation of appropriate infection control measures.
Design: A prospective cohort study during 18 months.
Setting: Five ICUs from two university hospitals.
Patients: All patients admitted for >/=48 hrs.
Measurement: ICU-acquired infections were ascertained during daily bedside patient and chart reviews. Episodes of potential cross-transmission were identified by highly discriminating genetic typing of all clinical and surveillance isolates of the ten bacterial species most frequently associated with nosocomial infections in ICUs. Isolation of indistinguishable isolates in two or more patients defined potential transmission episodes.
Main results: During 28,498 patient days, 431 ICU-acquired infections and 141 episodes of nosocomial transmissions were identified. A total of 278 infections were caused by the ten species that were genotyped, and 41 of these (14.5%) could be associated with transmissions between patients.
Conclusion: Infections acquired during treatment in modern tertiary ICUs are common, but a causative role of direct patient-to-patient transmission can only be ascertained for a minority of these infections on the basis of routine microbiological investigations.