Beta-lactams agents are time-dependent antibiotics. Their bactericidal effects are associated with the time spent above the minimal inhibitory concentration (MIC) of the bacteria. In ICU patients, many pathophysiological changes may occur, which significantly alter beta-lactam pharmacokinetics (increasing in distribution volume, hypoalbuminemia or change in glomerular filtration rate…), leading to antibiotic underexposure, which may explain some treatment failures and risk of emergence of a multi-resistant bacteria.

Many guidelines advocate to reach a fixed target for the steady-state free plasma concentration of beta-lactam between 4 and 8 times of the MIC. However, clinical evidences supporting that septic patients in ICU would benefit from such optimization remain low and contradictory, and use of surrogate MICs based on critical breakpoints instead of actual MICs are significant limitations in the description of beta-lactam exposure among ICU patients with sepsis.

In this study, the investigators aim to indentify profiles of intensive care unit patients which are more likely to be associated with pharmacological failure (< 4 time the MIC of the bacteria). The investigators also aim to identify association between failure and poor clinical outcome.

The duration of patient participation extends from the inclusion visit (D1), to discharge from intensive care; if the participant is still in intensive care at D90, the last visit with data collection will be at D90. The maximum duration of participation in the study will therefore be 3 months.

Taking into account 3 months of set-up, 1 year of recruitment, 3 months of participant follow-up, 3 months of data processing and 3 months of interpretation, 2 years will be sufficient to carry out this project.

Inclusion Criteria:

• Participant over 18 years of age.
• Hospitalized in intensive care unit.
• Bacterial infection documented by culture.
• Treatment with beta-lactam for less than 48 hours.
• Person not objecting to participation in the study. If the particpant is unable to receive the information, it will be given to a relative and his or her non-opposition will be collected; the patient's non opposition will be sought and collected as soon as his or her condition permits.

Exclusion Criteria:

• Measurement of MIC of the bacterium in question impossible.
• Patients allergic to beta-lactam antibiotics.
• Patients with an estimated hospital stay of less than 3 days.
• Participants with "complex" infectious sites: central nervous system infection, osteoarticular infection, etc.
• Patients already included in the present study on the occasion of a previous beta-lactam administration.
• Adult patients under legal protection (safeguard of justice, curatorship, guardianship), persons deprived of liberty.

Patients hospitalized in intensive care with sepsis.