BACKGROUND: Stroke is a major cause of death and disability worldwide. Hypoxia is common after stroke and is associated with worse outcomes. Oxygen supplementation could prevent hypoxia and secondary brain damage. OBJECTIVES: (1) To assess whether or not routine low-dose oxygen supplementation in patients with acute stroke improves outcome compared with no oxygen; and (2) to assess whether or not oxygen given at night only, when oxygen saturation is most likely to be low, is more effective than continuous supplementation. DESIGN: Multicentre, prospective, randomised, open, blinded-end point trial. SETTING: Secondary care hospitals with acute stroke wards. PARTICIPANTS: Adult stroke patients within 24 hours of hospital admission and 48 hours of stroke onset, without definite indications for or contraindications to oxygen or a life-threatening condition other than stroke. INTERVENTIONS: Allocated by web-based minimised randomisation to: (1) continuous oxygen: oxygen via nasal cannula continuously (day and night) for 72 hours after randomisation at a flow rate of 3 l/minute if baseline oxygen saturation was ≤ 93% or 2 l/minute if > 93%; (2) nocturnal oxygen: oxygen via nasal cannula overnight (21:00-07:00) for three consecutive nights. The flow rate was the same as the continuous oxygen group; and (3) control: no routine oxygen supplementation unless required for reasons other than stroke. MAIN OUTCOME MEASURES: Primary outcome: disability assessed by the modified Rankin Scale (mRS) at 3 months by postal questionnaire (participant aware, assessor blinded). Secondary outcomes at 7 days: neurological improvement, National Institutes of Health Stroke Scale (NIHSS), mortality, and the highest and lowest oxygen saturations within the first 72 hours. Secondary outcomes at 3, 6, and 12 months: mortality, independence, current living arrangements, Barthel Index, quality of life (European Quality of Life-5 Dimensions, three levels) and Nottingham Extended Activities of Daily Living scale by postal questionnaire. RESULTS: In total, 8003 patients were recruited between 24 April 2008 and 17 June 2013 from 136 hospitals in the UK [continuous, n = 2668; nocturnal, n = 2667; control, n = 2668; mean age 72 years (standard deviation 13 years); 4398 (55%) males]. All prognostic factors and baseline characteristics were well matched across the groups. Eighty-two per cent had ischaemic strokes. At baseline the median Glasgow Coma Scale score was 15 (interquartile range 15-15) and the mean and median NIHSS scores were 7 and 5 (range 0-34), respectively. The mean oxygen saturation at randomisation was 96.6% in the continuous and nocturnal oxygen groups and 96.7% in the control group. Primary outcome: oxygen supplementation did not reduce disability in either the continuous or the nocturnal oxygen groups. The unadjusted odds ratio for a better outcome (lower mRS) was 0.97 [95% confidence interval (CI) 0.89 to 1.05; p = 0.5] for the combined oxygen groups (both continuous and nocturnal together) (n = 5152) versus the control (n = 2567) and 1.03 (95% CI 0.93 to 1.13; p = 0.6) for continuous versus nocturnal oxygen. Secondary outcomes: oxygen supplementation significantly increased oxygen saturation, but did not affect any of the other secondary outcomes. LIMITATIONS: Severely hypoxic patients were not included. CONCLUSIONS: Routine low-dose oxygen supplementation in stroke patients who are not severely hypoxic is safe, but does not improve outcome after stroke. FUTURE WORK: To investigate the causes of hypoxia and develop methods of prevention. TRIAL REGISTRATION: Current Controlled Trials ISRCTN52416964 and European Union Drug Regulating Authorities Clinical Trials (EudraCT) number 2006-003479-11. FUNDING DETAILS: This project was funded by the National Institute for Health Research (NIHR) Research for Patient Benefit and Health Technology Assessment programmes and will be published in full in Health Technology Assessment; Vol. 22, No. 14. See the NIHR Journals Library website for further project information.

Original publication




Journal article


Health Technol Assess

Publication Date





1 - 88