Weaning failure is associated with prolonged duration of mechanical ventilation that itself can worsen patient's prognosis. Therefore, identification of the causes and mechanisms leading to weaning failure is important in daily practice. While diaphragm dysfunction and pulmonary edema are two main causes of weaning failure, there are currently no data that quantify their respective contribution neither their potential co-existence. During weaning from mechanical ventilation, heart and lung interaction plays a major role. As a key factor, diaphragm function is therefore certainly involved in the occurrence of weaning-induced pulmonary edema. However, both phenomenons diaphragm dysfunction and weaning induced pulmonary edema have not been simultaneously investigated so far. This current project aims at exploring the relationship between diaphragm dysfunction and pulmonary in patients experiencing weaning failure. The objectives are 1) to quantify the respective contribution and co-existence of diaphragm dysfunction and weaning induced pulmonary edema and 2) to investigate the role of diaphragm function in the occurrence of weaning induced pulmonary edema.

The study will take place in the Medical ICU of Pitié-Salpêtrière Hospital in Paris. The ICU have an established and successful clinical research program and infrastructure.

ICU patients will be enrolled if eligible (based on inclusion and non inclusion criteria) after their first spontaneous breathing trial failure.

Diaphragm function will be assessed by its capacity to generate a pressure with the phrenic nerves stimulation technique (reference method) and by ultrasonography (thickening fraction). Weaning induced pulmonary edema will be diagnosed with cardiac echo and markers of plasma contraction (plasma protein concentration).

Phrenic nerves stimulation, diaphragm and cardiac ultrasound will be repeated as part of the research before the spontaneous breathing trial.

After obtaining consent and before starting the second spontaneous breathing trial, clinical and biological will be collected. EKG and maximal inspiratory pressure will be done. Then, a diaphragm ultrasound and cardiac echo will be done and the phrenic nerves stimulation method will be undertaken. After these measurements, the spontaneous breathing trial will start. At the end of the spontaneous breathing trial (30 minutes with pressure support 7 cmH2O and 0 positive end expiratory pressure) or sooner in case of intolerance (respiratory distress), a new cardiac echo and diaphragm ultrasound will be performed as part of standard care before resuming the initial ventilator settings. As recommended by the usual medical practice, a second phrenic nerves stimulation will be eventually done after resuming the initial ventilator settings.

Diaphragm capacity to generate pressure :

Twitch airway pressure (Ptr, stim) will be measured during phrenic nerve stimulation. Phrenic nerve stimulation will be performed by bilateral anterior magnetic stimulation.

Cardiac echo :

Cardiac echo will be performed with a Philips Sparq ultrasound with a dedicated probe. The following variables will be obtained: left ventricular ejection fraction, E wave, A wave, E' wave, TAPSE

Diaphragm ultrasound :

Diaphragm thickening fraction will be measured by placing 13 MHz transducer on the right chest wall in the ninth intercostal space mid-way between the anterior and mid-axillary line and held perpendicular to the chest wall. Thickening fraction will be computed as the percentage change in thickness between end-expiration (i.e. minimum muscle thickness) and peak inspiration (i.e. maximal muscle thickness) visualized in M-mode.

Inclusion Criteria:

1. Patients adult (age ≥ 18 years)
2. Intubation and invasive mechanical ventilation longer than 48 hours;
3. Failure to the first spontaneous breathing trial;
4. Presence of weaning readiness criteria
5. Oral informed consent
6. Patient with affiliation to the french healthcare system

Exclusion Criteria:

1. Pregnant women
2. Patient opposition
3. Contraindication to perform magnetic phrenic nerves stimulation technique (chest tube, cardiac pacemaker or implanted defibrillator, cervical implants, chest drains,...)
4. Impossibility to consider ventilator weaning (total ventilator dependence : preexisting neuromuscular disorders, cervical spine injury…)