Intubation of patients in the intensive care unit (ICU) carries a risk of potentially severe complications, including cardiac arrest. Hypoxemia is common in ICU patients requiring intubation, which must be performed rapidly to avoid aspiration, since the patient is usually not in the fasted state. Studies have assessed interventions designed to improve intubation success rates, such as routine neuromuscular blockade. Care bundles combined with training on simulators have improved the safety of intubation. Nevertheless, intubation in the ICU still carries higher morbidity and mortality rates compared to intubation in the operating room.

Preoxygenation is a cornerstone of safety for intubation in the ICU. Several recent trials have investigated different devices (non-rebreather mask, non-invasive ventilation, high flow nasal cannula, bag valve mask) with conflicting results. A main reason for those results is that efficiency of the preoxygenation period cannot be evaluated in the ICU in opposite to the operating room: gas monitoring are not available in ICU and even if it was the case, high flow demand from the patient, and agitation will make it inefficient. Additionally, desaturation is frequent (from 10% up to 50%) during intubation in ICU and lead to morbidity and mortality; so anticipation of desaturation is a major concern for ICU's physician because it's impacting care (face mask ventilation, early insertion of subglottic device). The oxygen reserve index (ORI) is a new parameter for monitoring oxygen reserve non-invasively.

In this context, the investigators purpose to analyze efficiency of preoxygenation and time allowed by ORI for medical interventions before hypoxemia during intubation in the ICU in a pilot observational study in our medical ICU in a university hospital.

*Severe hypoxaemia is the most common serious adverse event during endotracheal intubation (ETI) in patients admitted to the intensive care unit (ICU), consistently reported around 25%. Preoxygenation is universally recommended as a risk-minimisation measure. While effective in the operating room, preoxygenation is more complex in critically ill patients, most notably those with acute hypoxic respiratory failure, as the low functional residual capacity (FRC) carries a risk of denitrogenation being confined to the healthy lung parenchyma, so that the FeO2 (Fraction of expired oxygen) no longer reflects the partial pressure of oxygen in arterial blood (PaO2). In these patients, 8 minutes of preoxygenation using a bag-valve-mask (BVM) device failed to change PaO2 levels noticeably. Increasing the BVM time was unhelpful, as the increase in PaO2 achieved in some patients was offset by decreases in others.

Non-invasive ventilation (NIV) has theoretical advantages for preoxygenation and has been shown to provide higher PaO2 values in hypoxic patients compared to a non-rebreathing (NRB) mask. However, use of NIV has several contraindications including consciousness alterations and agitation, and drawbacks such as gastric distension.

Apnoeic oxygenation which has been updated is a method consisting in continued oxygenation during laryngoscopy to extend the safe apnoea time. High-flow nasal oxygen (HFNO) therapy ensuring high and stable FiO2 (fraction of inspired oxygen) values by delivering oxygen flows above 60 L/min is another innovation shown to improve oxygenation in hypoxaemic patients. The effect of HFNO for preoxygenation, however, is less clear, as one study suggested benefits vs. NRB mask in patients with moderate hypoxaemia but another found no significant difference vs. BVM in patients with PaO2/FiO2 ratios below 200. No large randomised trial has compared preoxygenation methods. The latest recommendations of the Difficult Airway Society suggest than "In hypoxaemic patients, CPAP (continuous positive airway pressure) and non-invasive ventilation (NIV) may be beneficial."

As preoxygenation is a cornerstone of safety for intubation in the ICU, several recent trials have investigated different devices (non-rebreather mask, non-invasive ventilation, high flow nasal cannula, bag valve mask) with conflicting results. A main reason for those results is that efficiency of the preoxygenation period cannot be evaluated in the ICU in opposite to the operating room : gas monitoring are not available in ICU and even if it was the case, high flow demand from the patient, and agitation will make it inefficient. Additionally, desaturation is frequent (from 10% up to 50%) during intubation in ICU and lead to morbidity and mortality; so anticipation of desaturation is a major concern for ICU's physician because it's impacting care (face mask ventilation, early insertion of subglottic device).

The oxygen reserve index (ORI) (Masimo Corp., Irvine, CA, USA) is a new parameter that indicates real-time oxygenation reserve status in mild hyperoxemia range [partial pressure of oxygen (PaO2) of about 100-200 mmHg). ORI can be measured non-invasively by applying an exclusive sensor on a finger; the ORI is a nondimensional index that changes according to the oxygenation reserve status of the patient. It uses a value range of 0.00-1.00.

In this context, the investigators purpose to analyze efficiency of preoxygenation and time allowed by ORI for medical interventions before hypoxemia during intubation in the ICU in a pilot observational study in our medical ICU in a university hospital.

*Study population : Patients admitted in ICU and planned to be intubated. Recruitment will be done in the Medical ICU of Nantes University Hospital. The study plans to include 50 patients admitted to intensive care and requiring a tracheal intubation procedure, after the provision of the informational letter and collection of non-opposition from the patient or responsible party, or through the emergency inclusion process. In the latter two cases, the non-opposition of the patient will be collected retrospectively after reading the informational letter.

*Design and conduct of the study : In each case, after intubation decision was made and patient consent, a non-invasive sensor capable of measuring ORI (Rainbow® Sensor, R2-25, Revision L, Masimo Corp.) will be applied to the third or fourth finger on the contralateral side of the inflatable cuff for non-invasive blood pressure monitoring. The sensor will be shield to be protected from exposure to light. SpO2 and ORI values, displayed on Root with Radical7 will be recorded each 2 seconds and then will be transferred to a laptop.

Participation in the study will end in case of secondary opposition of the patient.

Patient leaving the protocol will continue the usual follow-up.

*Data management and patient identification : Data collection for each person participating in the research is realized with paper case report form (CRF). Each person responsible for the filling of the CRF (investigator, clinical research assistant ...) will have to be identified in the table of delegations of responsibilities of the center (see investigator's file).

Entering, viewing or modifying data will only be possible via the CRF. The structure of the database will be approved by the person responsible for the research, and will be implemented in conformity with the protocol and the relevant regulations.

Medical data for each patient can only be forwarded to the affiliated organization of the person responsible for research or any person duly authorized by him in conditions ensuring their confidentiality.

The principal investigator and all co-investigators undertake to keep the identities of the persons who participate in the study confidential by assigning them a code.

This code will be used for all the CRF and all the attached documents (reports of imaging exams, biology, etc.). It will be the only information which will make it possible to make the connection with the patient retrospectively. To this end, a correspondence table will be implemented and kept in the service under the responsibility of the investigator who will be the only one to have access to this document.

Sending a person's data for research will hence be possible only provided this code system has been applied; the presentation of the research results must exclude any direct or indirect identification.

No nominative data will be collected during this research. The coding rule is the following: 1st letter of first name + first letter of surname + month and year of birth, Inclusion number.

The documents will be archived under the name of the study in the facilities of the multipurpose ICU of CHU Nantes until the end of the period of practical utility.

These documents are as follows: Protocols and annexes, possible amendments ; Information notices ; Individual data (authenticated copies of raw data) ; Follow-up documents ; Statistical analyses ; Final report of the study.

At the end of the period of practical use, all documents to be archived, as defined in the classification process and document archiving related to CHU Nantes biomedical research, will be transferred to the archive site (CHU Nantes) and will be placed under the responsibility of the institution responsible for research for 15 years following the end of the study in accordance with the institutional practices.

No displacement or destruction can occur without the consent of the institution responsible for the research. After 15 years, the institution responsible for the research will be consulted for destruction. All data, documents and reports may be subject to audit or inspection.

*Statistics : The protocol falls within the scope of the MR003 (reference methodology 003), which conforms to Nantes University Hospital.

Qualitative data will be described in frequency and percentage. Quantitative data will be described using mean, standard deviation, median, and extreme values. All tests will be bilateral, and the level of significance is 0.05.

Stata statistical software (Release 13; StataCorp LP, College Station, TX) will be used.

Statistical justification of the number of inclusions : Duration before drop of SpO2 was approximately 60 seconds +/- 30 seconds (data from Lascarrou JAMA (Journal of the American Medical Association) 2017 and Miguel-Montanes Critical Care Medicine 2015). Assuming that the ORI drop occurs 30 seconds before SpO2 drop, the investigators would like to estimate the average delay between these two drops with 95% confidence interval, with a precision to 10 seconds. Assuming a standard deviation equal to 30 seconds, 35 patients will be necessary. According to Szmuk, only 72% patients will be analyzed so the investigators decide to include 50 patients.

All statistical tests will be performed with a significance level of 5%. No imputation will be realized for this observational study.

*Safety : No adverse event or reaction can be associated with this study insofar as it consists on a non-interventional study with no impact on care.](streamdown:incomplete-link)

Inclusion Criteria:

• ICU admission and need for orotracheal intubation (OTI) to allow mechanical ventilation.
• SpO2/FiO2 ratio upper than 214 (the SpO2/FiO2 ratio will be measured under NIV or HFNC(high flow nasal cannula), and calculated under standard oxygen as follows : FiO2 = 0.21+oxygen flow rate×0.03).

Exclusion Criteria:

Patients will be excluded from the analysis if :

• Failure of computer used to record ORI.
• No ORI values displayed.
• ORI remaining constant at 1 throughout the monitoring.
• SpO2 will stay under 97% during intubation process.

Patients admitted in ICU and planned to be intubated.

Recruitment will be done in the Medical ICU of Nantes University Hospital. The study plans to include 50 patients admitted to intensive care and requiring a tracheal intubation procedure, after the provision of the informational letter and collection of non-opposition from the patient or responsible party, or through the emergency inclusion process. In the latter two cases, the non-opposition of the patient will be collected retrospectively after reading the informational letter.