High-flow versus conventional nasal oxygen during EBUS-TBNA under matched nominal FiO ₂ : a multicenter randomized controlled trial

Background Hypoxemia is common during EBUS-TBNA under sedation. Whether HFNC improves oxygenation when compared with conventional nasal cannula (CNC) under matched nominal FiO 2 and standardized sedation remains uncertain. Methods In this prospective multicenter randomized trial across four centers, adults undergoing oral EBUS-TBNA under moderate-to-deep sedation received HFNC (set FiO 2 45%, 60 L·min⁻ 1 ) or CNC (6 L·min⁻ 1 ; nominal FiO 2 ∼45%). The primary endpoint was the proportion of patients without any desaturation, with desaturation defined as SpO 2 <90% for >10 s. Secondary outcomes included severe desaturation, nadir SpO 2 , saturation decline, escalation after airway-opening maneuvers, and exploratory post hoc associations with desaturation. Results Of 126 randomized patients, 120 were analyzed. Any desaturation occurred in 36/63 (57%) CNC and 25/57 (44%) HFNC patients (p=0.20). Severe desaturation (SpO 2 <75% or 75–89% for >60 s) was less frequent with HFNC (3/57 [5%] versus 11/63 [17%]; p=0.048). HFNC was associated with a higher nadir SpO 2 (p=0.01) and a smaller saturation fall (p=0.02). Persistent desaturation after the initial airway-opening maneuver, assessed as an exploratory analysis derived from the prespecified escalation algorithm, was more frequent with CNC (13/63 [21%]) than HFNC (4/57 [7%]; p=0.038). All procedures were completed without advanced airway support. Exploratory post hoc modelling showed associations of higher STOP-Bang score and longer procedure duration with desaturation. Conclusion Under matched nominal FiO 2 and standardized sedation, HFNC did not increase the proportion of patients without any desaturation during EBUS-TBNA, but it was associated with fewer severe and persistent desaturation events and better oxygenation stability.