Extended-spectrum beta-lactamases-producing (ESBL-E) are disseminating worldwide especially in Intensive Care Units (ICUs) and are responsible for increased health costs and mortality. The aims of this work were to study ESBL-E dissemination in ICU and to assess the impact of ESBL-E fecal carriage on subsequent infections during a non-outbreak situation. We therefore screened every patient at admission then once a week in a medical ICU between January and June 2015. Each ESBL-E isolate was characterized by ESBL genes PCR amplification and the clonal dissemination was assessed by Pulsed-Field Gel Electrophoresis (PFGE). Among the 608 screened patients, 55 (9%) were colonized by ESBL-E. Forty-four isolates were available for further analysis. Most of them (43/44, 98%) contained a ESBL gene from the CTX-M group. Only one case of ESBL-E cross-transmission occurred, even for acquired ESBL-E colonization. Subsequent infection by ESBL-E occurred in 6/55 (11%) patients and infecting ESBL-E strains were the colonizing ones. ESBL-E faecal carriage had a negative predictive value of 100% and a positive predictive value of 40% to predict ESBL-E ventilator associated-pneumonia (VAP). Alternatives to carbapenems consisting in piperacillin-tazobactam, ceftolozane-tazobactam and ceftazidime-avibactam were all active on this panel of ESBL-E. ESBL-E expansion and acquisition in ICU in a non-outbreak situation are not any more fully explained by cross-transmission. Mechanisms underlying ESBL-E dissemination in ICU are still to investigate. Interestingly, as far as we know, our study demonstrates for the first time by PFGE that the colonizing strain is indeed the infecting one in case of subsequent ESBL-E infection. Nevertheless, subsequent ESBL-E infection remains a rare event conferring poor positive predictive value for ESBL-E colonization to predict ESBL-E VAP. Relevance of systematic ESBL-E faecal screening at ICU admission and during ICU stay needs further investigation.< Réduire