PupillOmetry for Prediction of DelirIUM

Delirium in intensive care unit (ICU) is a serious event. It is associated with short-term complications (agitation, self-extubation, accidental removal of catheters, prolonged length of stay and ventilation), excess mortality, functional and cognitive impairment. It is particularly frequent in patients requiring mechanical ventilation but diagnosis is not easy. There are screening scales, but it is insufficiently used in clinical practice: Confusion Assessment Method for the Intensive Care Unit (CAM-ICU) or Intensive Care Delirium Screening Checklist (ICDSC). These scales are time consuming and require trained personnel. Automated pupillometry (AP) is a new device to objectively, rapidly, and reproducibly identify acute brain dysfunction. Recent data suggest that AP could be used to predict delirium in the ICU. This would need to be validated for routine use in the ICU. Evaluate AP parameters on day 3 of invasive mechanical ventilation as a predictive tool for CAM-ICU diagnosed delirium during the first 14 days of ICU stay. Study design: Prospective, multicenter, non-interventional cohort Measurement of the AP parameters at day 3 after ICU admission and their predictive performance for delirium: pupillary diameter, variation of the pupillary diameter, pupillary constriction speed, pupillary dilatation speed, photomotor reflex latency, NPi and symmetry of pupillary responses.

Inclusion criteria : * Adults ≥ 18 years * mechanical ventilation since 48h * Information and no opposition of patient or close person if he is present at the time of inclusion, otherwise urgent inclusion Exclusion criteria : * Patient diagnosed confused by CAM-ICU on D0 * Ophthalmological pathology modifying the photomotor reflex * Acute or chronic neurological pathology: Cerebro-injured patients (head trauma, stroke, cardiopulmonary arrest, hypoglycemic coma, meningitis / encephalitis / brain abscess), dementia with MMSE \ 72h or duration of ventilation\> 72h * Readmission in intensive care * Moribund patient * Patient under guardianship or curatorship * No affiliation to social security (beneficiary or assignee)