Extended-Spectrum Beta-Lactamases (ESBL)-Producing Enterobacteriaceae (PE) pose a major problem among antimicrobial resistance. The worldwide spread of theses bacteria may be responsible for 10 million death in 2050. Infection with ESBL-PE are associated with a worse prognosis because of delay in the start of adequate antibiotic treatment, especially for severe infections. It has been proposed to identify colonized patients to predict the risk of infection and the risk of nosocomial cross transmission.

This qualitative approach has limit as only 5 to 20% of patients will develop an infection with ESBL-PE. The fecal relative abundance (RA) of ESBL-PE is a ratio of ESBL-PE among enterobacteriaceae that could identify high-risk patients of infection or cross transmission. ESBL-PE RA may be highly variable in patient with antibiotic exposure depending on the molecule received but dynamic data is missing.

The aim of this study is to identify the factor that influence the fecal RA of ESBL-PE in ICU and to evaluate the association between different level of fecal RA and infection or cross transmission with an ESBL-PE.

Antimicrobial resistance is rising since decades with a risk of million of death in the future. Extended-Spectrum Beta-Lactamases (ESBL)-Producing Enterobacteriaceae (PE) have expanded exponentially since 15 years and represent with Carbapenemase-PE one of the major challenges in resistance control. The burden of ESBL-PE infections is major in intensive care units (ICU) because of the delay to identify an effective antibiotic treatment (highly associated with outcome) and because of a higher risk of nosocomial cross transmission.

Identification of digestive carrier of ESBL-PE is based on a qualitative result that categorize the patient as a carrier or as non-carrier. This result makes it impossible to individualize the measures to be taken for an ESBL-PE carrier.

Prevention of cross transmission has no formal guideline. Some practitioners in ICU have stopped to detect for ESBL-PE carriage (specially when prevalence is low) and other prefer to close the ward (specially during outbreak).

Empiric treatment of most infections in ESBL-PE carrier are based on last-resort antibiotic (i.e. carbapenem) until microbiological results of a clinical simple is available.

A quantitative approach based on fecal relative abundance (RA) of ESBL-PE (ratio between ESBL-PE and enterobacteriaceae) has been proposed to individualize the risk of urinary tract infection (UTI) for ambulatory patients. In this setting, a fecal carriage with a very low RA of ESBL-PE safely rule out a risk of infection with ESBL-PE and patients with a high RA had an increased risk of UTI infection with ESBL-PE. Large variations of RA ESBL-PE carriage was observed and prediction of the level of RA was not possible for a patient.

A high variation in fecal RA of ESBL-PE is probable in ICU because of a high proportion of antibiotic exposure. Using the RA in ICU for ESBL-PE carrier could make it possible to identify patients who need for carbapenem in their empiric treatment and those who need continuing contact precautions to prevent cross transmission.

The aim of this study is to identify the factor that influence the fecal RA of ESBL-PE in ICU and to evaluate the association between different level of fecal RA and infection or cross transmission with an ESBL-PE.

Inclusion Criteria:

• Age ≥ 18 years
• Positive sample for ESBL-PE at admission of during ICU stay
• Patient's or relative's consent

Exclusion Criteria:

• Women pregnant, parturient or breast-feeding during the study period
• Patient deprived of liberty by judicial or administrative decision
• Patient undergoing psychiatric care under duress
• Patient subject to a legal protection measure
• Patient with no social security coverage