The investigators propose to study the value of non invasive continuous transcutaneous PtC02 monitoring for ventilatory withdrawal guidance in neuro-injured patients and to predict the risk of extubation failure in this category of patients, particularly at risk of re-intubation.

Extubation is part of the daily routine in intensive care and intensive care units. If the procedures for weaning from mechanical ventilation are well supervised by the establishment, in particular, of standardized treatment protocols, they nevertheless remain technical procedures at risk and are associated with significant morbidity and mortality.

Extubation, defined by the withdrawal of the intubation probe, is to be distinguished from the period known as ventilatory "weaning" or "ventilation" which is represented by the separation of the patient from the invasive mechanical ventilation device. These 2 successive phases (deventilation then extubation) must be anticipated as best as possible by the intensivist in order to detect the risk factors and complications inherent in one or the other of the stages. A patient can thus be ventilable but not extubable and vice versa.

Extubation failure can be defined as the need for early reintubation within 48-72 hours of scheduled intensive care extubation. This definition takes into account the contribution of non-invasive ventilation (NIV) that can be applied in the direct consequences of extubation.

The risk factors specific to failure of extubation are well identified in the literature and differ from those associated with failure of ventilatory weaning. These risk factors are multiple but are represented in particular by obstruction of the airways, ineffectiveness of cough or bronchial congestion, swallowing disturbances and disturbances of consciousness among others.

Monitoring of transcutaneous pressure in CO2 (PtCO2) has been developed over the past twenty years as a reflection of arterial pressure in CO2 (PaCO2). To date, this type of monitoring has been developed mainly in neonatal care due to its non-invasive nature. This continuous monitoring could be an interesting tool in the evaluation of the gas exchange of the patient under mechanical ventilation, in particular in a medical or surgical resuscitation environment. This tool could thus make it possible to participate in the prediction of extubation failure in medical intensive care or general surgery during ventilatory weaning tests. There is indeed a good correlation between PaC02 and PtC02 in the intensive care patient population for PaC02 values <60 mmHg.

We therefore propose to study the benefit of using transcutaneous PtC02 monitoring for guiding ventilatory weaning in neuro-injured patients and predicting the risk of extubation failure in this category of patients, particularly at risk of re-intubation.

The main objective of the study carried out is therefore to assess the predictive value of the observed variation in PtCO2 during a spontaneous ventilation weaning test in a population of neuro-resuscitation patients at high risk of reintubation.

Inclusion Criteria:

• major patients (>18 years)
• admitted in ICU for an acute brain injury defined by a Glasgow Coma Scale < 8 and the presence of a cerebral lesion identified on a referent Tomodensitometry (TDM) or Magnetic Resonance Imaging (MRI)
• requiring mechanical ventilation during >48 hours
• eligible for a ventilatory weaning test

Exclusion Criteria:

• preexisting neurologic evolutive or degenerative affection or preexisting diaphragmatic dysfunction
• preexisting decision of therapeutic limitation
• impossibility to collect an informed consent
• patients aged <18 years
• pregnancy
• mental illness
• impossibility for the subject to have a good comprehension of the study
• lack of health insurance coverage

The measured values of continuous PtC02 will not be disclosed to care provider or participant during the ventilatory weaning test