Abstract
High-flow oxygen therapy (HFOT) has become the most common oxygen-supplementation modality in patients with acute respiratory failure, based chiefly on data from tightly selected patients included in randomized trials. Our objective was to obtain real-life data on HFOT failure. This prospective observational multicenter French study done in 13 intensive care units in 2019–2020 was designed to determine the proportion of patients with failed HFOT, defined as intubation, noninvasive ventilation, standard oxygen therapy with an estimated FiO(2)>50%. We included 257 patients before prematurely ending the study due to the COVID-19 pandemic. HFOT failed in 79 patients (32%), including 42 (17%) who required intubation. Mean HFOT duration was 2.4 ± 2.2 days. By multivariate analysis, a low ROX index (ratio of pulse-oximeter saturation (SpO(2)) over FiO(2), over the respiratory rate (RR) was associated with HFOT failure (adjusted hazard ratio [aHR], 0.83; 95%CI, 0.77–0.90, P < 0.0001), as were elevated mean arterial pressure (aHR, 1.03; 95%CI, 1.02–1.05), vasopressor therapy (aHR, 3.35; 95%CI, 1.44–7.79), and worse Glasgow Coma Scale score (aHR, 0.69; 95%CI, 0.57–0.80, P < 0.001). The identification of factors independently associated with HFOT failure should help to determine which patients are most likely to benefit from HFOT, thereby preventing delayed intubation. Additional studies are needed to further assess these risk factors and to determine the optimal HFOT weaning modalities. Trial registration: The study was registered on the ClinicalTrials.gov registry on February 11, 2019 (NCT04141956). SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1038/s41598-026-39969-4.