OBJECTIVE High-flow nasal oxygen (Optiflow™) is validated in paediatric intensive care but not in adults' patients for severe hypoxemia. The aim of this study was to evaluate this oxygen system delivery in adults' patients for postoperative hypoxemia after cardiac surgery. METHODS Prospective, open study for evaluation of medical practice. METHODS Patients operated upon for cardiac surgery with immediate postoperative hypoxemia characterized by SpO(2) <0.96 with 50% oxygen with a Venturi mask were treated with the high-flow nasal oxygen system (O group) when it was available or with the classical high-flow oxygen face mask (M group). Gas exchanges were measured at the end of the surgery, at the beginning of the treatment and 1 hour, 6 hours after the inclusion and at day 1 and 2 post-treatment. Parameters studied were: duration of hypoxemia, duration of ICU stay, postoperative pneumonia occurrence, requirement of re-intubation, non invasive ventilation and catecholamine. Tolerance was evaluated with measurement of pain (visual scale), satisfaction (visual scale), and dryness of mouth. RESULTS Forty patients were included, 19 in group O, 21 in group M. Patient's characteristics did not differ between the two groups before treatment. There were no significant differences between groups for duration of hypoxemia (3.8±2.2 days in O group versus 4.3±2.3 days in M group), duration of hypoxemia, duration of ICU stay, postoperative pneumonia occurrence, requirement of re-intubation, non invasive ventilation and catecholamine. Pain was not significantly different between groups, satisfaction was better (P<0.001) and mouth drier (P<0.001) in group O than in group M. CONCLUSIONS These results give good arguments for an improvement in gas exchange and better tolerance of high-flow nasal oxygen (Optiflow™) versus classical high-flow oxygen face mask in postoperative cardiac patients. These results must be confirmed by a randomised study with a larger population.