Whether high-flow vs. low-flow nasal oxygen reduces hypoxaemia for sedation during endoscopic retrograde cholangiopancreatography is currently unknown. In this multicentre trial, 132 patients ASA physical status 3 or higher, BMI > 30 kg.m -2 or with known or suspected obstructive sleep apnoea were randomly allocated to high-flow nasal oxygen up to 60 l.min -1 at 100% F I O 2 or low-flow nasal oxygen at 4 l.min -1 . The low-flow nasal oxygen group also received oxygen at 4 l.min -1 through an oxygenating mouthguard, totalling 8 l.min -1 . Primary outcome was hypoxaemia, defined as S p O 2 < 90% regardless of duration. Hypoxaemia occurred in 7.7% (5/65) of patients with high-flow and 9.1% (6/66) with low-flow nasal oxygen (percentage point difference À1.4%, 95%CI À10.9 to 8.0; p = 0.77). Between the groups, there were no significant differences in frequency of hypoxaemic episodes; lowest S p O 2 ; peak transcutaneous carbon dioxide; hypercarbia (transcutaneous carbon dioxide > 2.66 kPa from baseline); requirement of chin lift/jaw thrust; nasopharyngeal airway insertion; bag-mask ventilation; or tracheal intubation. Following adjustment for duration of the procedure, the primary outcome remained non-significant. In high-risk patients undergoing endoscopic retrograde cholangiopancreatography, oxygen therapy with high-flow nasal oxygen did not reduce the rate of hypoxaemia, hypercarbia or the need for airway interventions, compared with combined oral and nasal lowflow oxygen.