Midazolam and propofol are the most used intravenous (IV) sedative agents, but their use is associated with well-known adverse effects such as accumulation, myotoxicity, tachyphylaxis, and unpredictable wake-up time.

For benzodiazepines, an increased tolerance, possible accumulation after long-term use, and an increased risk of acute withdrawal syndrome are reported. In patients on extracorporeal membrane oxygenation (ECMO) for cardiogenic shock, the negative hemodynamic effects of these drugs are a particular matter of concern. Besides the extracorporeal circuit itself may affect the pharmacokinetics of these IV sedatives. Indeed, drug sequestration in ECMO circuits is a well-known phenomenon influenced by drug chemo-physical properties. Given the large surface area of tubing and membrane, considerable quantities of drugs used in ECMO patients may be sequestered over a period, resulting in a significant increase in their volume of distribution. Similarly, frequent hemodilution and organ dysfunction would also contribute to an increase in the volume of distribution.

Propofol, which is lipophilic is significantly sequestrated in the circuit. Consequently, it is commonly observed that patients receiving ECMO have substantially higher sedative and analgesic drug requirements than patients without ECMO.

To date, there is no ideal concept for analgesia and sedation of patients on ECMO in the ICU.

A drug that sedates effectively but with minimal residual sedation after the end of the administration and without the aforementioned drawbacks of the current agents would be valuable.

Interestingly, a recent randomized controlled non-inferiority trial that randomized 338 patients showed that, compared with propofol, sedation with inhaled anaesthetics was non-inferior. Sedation with inhaled anaesthetics resulted in a higher rate of spontaneous breathing and a shorter wake-up time after 48h of sedation. Indeed, inhaled sedation, which has been associated with reduced opioid consumption and less delirium in ICU patients, is a promising alternative to IV sedation. Moreover, inhaled anaesthetics might be associated with less myocardial injury and lower doses of inotropic support in patients undergoing cardiac surgery. However, to date, the experience with volatile agents remains limited in patients on ECMO.

We hypothesized that the use of inhaled isoflurane with the Sedaconda anaesthetics conserving device (ACD) in cardiogenic shock patients on ECMO will reduce the mortality and increase the number of ventilation-free days at day 28 following ECMO onset compared to usual IV sedation by propofol and/or midazolam.

Inclusion Criteria:

1. Cardiogenic shock on VA ECMO support for less than 24 hours
2. Patients on invasive mechanical ventilation receiving propofol and/or midazolam at the time of randomization
3. Invasive mechanical ventilation for less than 48 hours
4. Expected invasive ventilation and sedation for at least 24h, with a prescribed Richmond agitation scale target within the range of -1 to - 4
5. Social security registration (AME excluded)

Exclusion Criteria:

1. Age <18 and >75
2. Pregnancy or breastfeeding
3. Initiation of ECMO >24 hours
4. Initiation of mechanical ventilation >48 hours
5. Cardiopulmonary Resuscitation >20 minutes before randomization
6. Patient moribund on the day of randomization, SAPS II >90
7. Suspected or proven intracranial hypertension
8. Corrected QT interval > 450ms or with a known or suspected genetic predisposition to malignant hyperthermia
9. Chronic liver disease defined as a Child-Pugh score of 12-15
10. Patients ventilated with a tidal volume < 4ml/kg predicted body weight
11. Participation in another interventional study or being in the exclusion period at the end of a previous study.
12. Contraindication or allergies to isoflurane, propofol, midazolam or other halogenated anaesthetics

The procedures carried out with the French data privacy authority (CNIL, Commission nationale de l'informatique et des libertés) do not provide for the transmission of the database, nor do the information and consent documents signed by the patients.

Consultation by the editorial board or interested researchers of individual participant data that underlie the results reported in the article after deidentification may nevertheless be considered, subject to prior determination of the terms and conditions of such consultation and in respect for compliance with the applicable regulations.

Beginning 3 months and ending 3 years following article publication. Requests out of these time frame can also be submitted to the sponsor

Researchers who provide a methodologically sound proposal