Hypoxemia is usually observed during ICU stay. Nevertheless, the prevalence of hypoxemia among patients admitted in ICU is not well known, in partly due to a lack of a consensual definition. Some of these patients meet the criteria of Acute Respiratory Distress Syndrome whom the definition has been recently modified in 2011. Many uncertainties are remaining in considering the patients with ARDS: prevalence of the category of "mild ARDS", the proportion of patients with invasive mechanical ventilation, the therapeutic modalities and the prognosis, in particular, the move towards a more severe ARDS category. Finally, many patients with hypoxemia do not meet ARDS definition. The prevalence of this population in ICU is unknown, likewise its prognosis. In this way, it would be interesting to compare, for each levels of hypoxemia (mild, moderate and severe) the prognosis of the patients with and without ARDS.

The main objective of the SPECTRUM study is to assess the prevalence of hypoxemia in French-speaking Intensive Care Unit in 2016 in using a single-day point-prevalence study design. Two periods of inclusion have been, actually, planned: from the 29th Mars to the 31st Mars and from the 5th to the 7th April. In each participating center, the day of the study will be chosen among these periods.

Inclusion Criteria:

• Age 18 and older
• Hospitalized in ICU during the 24 hours of the study
• Information provided to the patient or proxy

Exclusion Criteria:

• Pregnant or breast-feeding woman
• Patient under guardianship or curators
• Lack of social insurance

Critically ill patients admitted in Intensive Care