The use of antibiotic therapy is common in intensive care units and primarily involves beta-lactams. Its optimal implementation is made difficult by the pharmacokinetic changes inherent in critically ill patients.

Despite the current recommendations from the French Society of Anesthesiology and Intensive Care (SFAR) and the French Society of Pharmacology and Therapeutics (SFPT), there are no recommendations on prescription modalities for patients under veno-arterial extracorporeal membrane oxygenation (VA-ECMO). The use of antibiotic therapy is common in VA-ECMO patients and their pharmacokinetic variability factors are then exacerbated.

We aim to conduct a prospective, multicenter, interventional study designed to identify predictive factors for failure to achieve therapeutic target circulating concentrations of beta-lactams in patients under VA-ECMO treated with one of the studied beta-lactams

The use of antibiotic therapy is common in intensive care units and primarily involves beta-lactams. Its optimal implementation is made difficult by the pharmacokinetic changes inherent in critically ill patients (fluid resuscitation, use of catecholamines, shock state, organ dysfunctions, especially those involved in drug elimination, implementation of extracorporeal circulations...). It has been demonstrated that delayed initiation of appropriate antibiotic therapy in severe infections is an independent predictor of poor prognosis.

The current joint recommendations from the French Society of Anesthesiology and Intensive Care (SFAR) and the French Society of Pharmacology and Therapeutics (SFPT) formulated in 2018 have codified the implementation of beta-lactam antibiotic therapy in critically ill patients, specifying administration modalities, initial dosages, as well as pharmacological therapeutic monitoring methods (plasma monitoring of the administered molecule after 24-48 hours of administration in all patients for whom pharmacokinetic variability is expected) and target plasma concentrations. However, there are no recommendations on prescription modalities for patients under veno-arterial extracorporeal membrane oxygenation (VA-ECMO).

VA-ECMO is a salvage therapy for cardio-pulmonary support via extracorporeal circulation, used in severe cardiac or respiratory failures. The use of antibiotic therapy is common in these patients. Pharmacokinetic variability factors common to intensive care patients are then exacerbated (massive fluid resuscitation, haemodilution by extracorporeal circulation, adsorption of xenobiotics on the exchange membrane and circuit, intensity of organ failures) while the severity of the patient necessitates immediate optimal antibiotic therapy administration.

The literature on antibiotic pharmacokinetics in patients under VA-ECMO is scarce in the adult population. Two exploratory approaches with distinct objectives can be found:

A "fundamental" approach, aiming to define the pharmacokinetic parameters (apparent volume of distribution, elimination half-life...) specific to each molecule within this population and to compare them with those observed in a reference population.

A "clinical" approach, focusing on the frequency of concentrations within the therapeutic range within the studied population (proportion of patients for whom the measured plasma concentration in practice conforms to the targets formulated in the recommendations).

While this second approach seems to correspond more closely to our clinical practice, there are few studies in this area. Bouglé et al. conducted a prospective monocentric study in 2019, including all patients under ECMO-VA receiving antibiotic therapy. The authors described the frequencies of plasma concentrations exceeding the lower limit of the therapeutic range. Thus, within this population of critically ill patients, where the failure to implement optimal antibiotic therapy is likely a factor of poor prognosis, 50 to 90% of patients did not reach the recommended therapeutic targets for the different beta-lactams studied. Moreover, the study considered overdoses as pharmacokinetic successes, a highly debatable point given the complications inherent in beta-lactam overdoses. However, this pilot study did not focus on factors that could predict failure to achieve target concentrations.

We aim to conduct a prospective, multicenter, interventional study designed to identify predictive factors for failure to achieve therapeutic target circulating concentrations of beta-lactams in patients under VA-ECMO treated with one of the studied beta-lactams

Inclusion Criteria:

• Patients undergoing VA-ECMO
• Patients requiering initiation of a novel antibiotic therapy with one of the studied beta-lactam (Piperacillin-Tazobactam, Cefepime, or Meropenem) while under VA-ECMO
• With an expected survival exceeding 24 hours
• Patient's or their trusted person's consent

Exclusion Criteria:

• Subject under administrative or judicial surveillance
• Non-affiliation to social insurance
• Pregnant or lactating patient
• Antibiotic therapy of interest already initiated before VA-ECMO implantation
• Contraindication to the use of beta-lactam
• Administration modalities of beta-lactam not compliant with current recommendations
• Subject in exclusion period from another study