ROX Index for the Timing of Intubation in Nasal High Flow

This is a pilot study to assess whether the use of an algorithm incorporating the ROX index to traditional criteria for intubation in patients with AHRF supported with NHF is:(i) feasible; (ii) associated with a decreased time to intubation. The ROX-1 trial is an international, multicenter, parallel, open-label randomized controlled two-arm trial that will be performed in patients with AHRF, supported by nasal high flow, who are admitted to the intensive care unit (ICU). Ethics approval will be sought from each participating institution before starting enrolmentand consent will be obtained for each patient. The coordinating center will be Hospital Universitari Vall d'Hebron in Barcelona, Spain. All consecutive patients older than 18 years with acute hypoxemic respiratory failure who need to be supported with nasal high flow (NHF) will be enrolled if they meet all the following criteria: 1. Respiratory rate \> 25 breaths/min 2. A pulse oximetry (SpO2)\<92% while receiving standard oxygen administered through a facemask at 10L/min or more. 3. Pulmonary infiltrate on chest X-ray. Patients already treated with NHF for acute respiratory failure prior to ICU admission will be enrolled if duration of NHF prior to randomization does not exceed one hour. Non-inclusion criteria are: Patients younger than 18 years old, Patients with indication for immediate intubation, patients with do not intubate order, patients electively intubated for diagnostic or therapeutic procedure (fibrobronchoscopy, surgery), patient with post-extubation AHRF, awake ECMO, pregnancy and refusal to participate or participation in another interventional study with the same primary outcome. Reasons for non-inclusion in screend patients will be registered. Patients admitted to the ICU with acute hypoxemic respiratory failure will be screened, included after having received study information, and signed informed consent. They will be randomized after stratification according to PaO2/FiO2, through an online 4 automatic centralized and computerized system to one of the study groups (1:1 ratio): traditional intubation criteria or traditional intubation criteria + ROX criteria. Due to the nature of the intervention, blinding will not be feasible. Randomization will be stratified to PaO2/FiO2 in order to analyse separately those patients with a PaO2/FiO2 ratio ≤ 200 and \> 200. Intubation criteria for each study arm HFNC failure will be defined as the subsequent need for invasive MV. After randomization, clinicians will set NHF to the maximum tolerated flow (up to 60L/min) and FIO2 will be titrated targeting a SpO2 between 94-98% (88-92% in case of underlying chronic lung disease). In both groups, the presence of intubation criteria will be assessed every two hours during the first 12 hours after randomization and after any clinical deterioration. Between 12 and 24 hours of NHF therapy, the presence of intubation criteria will be assessed every 4 hours. Beyond the first 24 hours after randomization, the presence of intubation criteria will be assessed once a day and after any clinical deterioration. In the intervention arm, patients will be intubated according to both the traditional intubation criteria and the ROX index, which ever are met first. Moreover, during the first two hours of NHF supportive therapy after randomization, the ROX index will be measured every 30 minutes. If the patient has a ROX index \<2.85, the NHF support will be increased to the maximum tolerated flow (up to 60L/min) and FIO2 of 1 and subsequently titrated with the target SpO2. Then, the ROX index will be recalculated in 30 minutes. If the patient is already treated with to 60L/min and FIO2 of 1 and no further increase could be done, the ROX index will be recalculated after 30 minutes of full NHF support. Then: 1) if the ΔROX is \<0 the patient will be intubated; 2) if the ΔROX is 0-0.5, the ΔROX will be assessed in 30 minutes; and 3) if the ΔROX is \>0.5 the patient will not be intubated, NHF will be managed as protocolized and respiratory condition will be reassessed every two hours or at any new clinical deterioration. Between the 2 and 6 hours after randomization, the ROX index will be measured every 60 minutes. If the ROX index is \<3.47, the NHF support will be also increased to the maximum tolerated flow (up to 60L/min) and FIO2 of 1 and subsequently titrated with the target SpO2. If the patient is already treated with to 60L/min and FIO2 of 1 and no further increase could be done and the ROX index will be recalculated after 30 minutes of full NHF support. As described before, the ROX index will be recalculated in 30 minutes: 1) if the ΔROX is \<0 the patient will be intubated; 2) if the ΔROX is 0-0.5, the ΔROX will be reassessed in 20 minutes; and 3) if the ΔROX is \>0.5 the patient will not be intubated, NHF will be managed as protocolized and respiratory condition will be reassessed hourly or at any new clinical deterioration. Finally, between the 6 and 12 hours after randomization, the ROX index will be measured every 60 minutes. If the ROX index is \<3.85, the NHF support will be increased to the maximum tolerated flow (up to 60L/min) and FIO2 of 1 and subsequently titrated with the target SpO2. If the patient is already treated with to 60L/min) and FIO2 of 1 and no further increase could be done and the ROX index will be recalculated after 30 minutes of full NHF support. Then, the ROX index will be calculated in 30 minutes and decision of intubation will be taken according to the previously described ΔROX. In the ROX group, if there is any clinical deterioration beyond the first 12 hours of treatment , the algorithm should be restarted from the beginning, considering time 0 as the time when new clinical deterioration started. End of follow-up. Enrolled patients will be followed for one month or until ICU discharge or death, whichever occurs first.