Several factors can influence oxygenation monitoring and respiratory support through oxygen therapy: the type of oximeter used, skin pigmentation, and the oxygenation target. It is essential to have an accurate measurement of SpO2 in order to optimize the flow of oxygen or FiO2 administered to patients. The question arises as to the impact of these confounding factors on the FiO2 set during high-flow nasal oxygen therapy. The aim of the study is to evaluate the impact of the oxygenation target and the oximeter used on FiO2 in patients receiving high-flow nasal oxygen therapy.

The oximeters evaluated will be the Nonin (Plymouth, MN) and the Philips (Eindhoven, Netherlands). Two SpO2 targets will be evaluated: 90% and 94%.

A total of four 10-minute periods will be performed in randomized order: Nonin 90, Nonin 94, Philips 90, and Philips 94.

At the end of each 10-minute period, arterial gas will be collected through the arterial catheter.

Inclusion Criteria:

• ≥ 18 years of age
• Admiitted in intensive care unit
• Presence of an arterial catheter
• Ongoing on high flow nasal oxygen therapy with SpO2 between 88 and 100% with a FiO2 between 40 and 80% with the usual oximeter .

Exclusion Criteria:

• No SpO2 signal with oximeter in use
• False nails or nail polish
• Methemoglobinemia >0.015 on last available arterial gas
• Patient in isolation (multi-resistant bacteria, C-Difficile, SARS-CoV-2 ...)
• Expected to use another respiratory support within two hour of inclusion (NIV or mechanical ventilation)