View commentaries on this research

This is a plain English summary of an original research article

There was no benefit to routinely giving oxygen to people who have had a stroke. Oxygen given continuously, or just overnight, did not reduce disability or death and it did not improve people’s ability to do everyday tasks or live independently. There were no oxygen-related adverse events reported.

Strokes occur when the blood supply to the brain is disrupted by either a blocked or burst blood vessel. They can lead to death or disability as parts of the brain are deprived of blood. Therefore, giving oxygen to reduce the potential damage may appear to make sense.

Guidelines from NICE and the British Thoracic Society recommend that people are not routinely given oxygen after a stroke unless their oxygen levels drop. This large NIHR funded trial provides evidence to support these recommendations and reinforces the need to monitor oxygen levels to guide the appropriate use of oxygen therapy on an individual basis. Given the size and quality of this UK-based trial, it is unlikely that future research would change these recommendations.

Why was this study needed?

A stroke is when the brain is deprived of blood, either due to a blocked blood vessel (the commonest form) or a bleed (less common). There are over 100,000 strokes annually in the UK.

The reduced flow of blood and oxygen can lead to areas of the brain being damaged or dying. Stroke treatment is improving, and twice as many people survive strokes now than they did in 1990. However, around two-thirds of stroke survivors are left with a long-term disability.

Giving oxygen to people who have had a stroke could plausibly help to prevent or reduce brain damage. However, high levels of oxygen can also be harmful – causing constriction of the blood vessels, reduced blood flow to the brain, damage to the lungs and restricting people’s mobility.

This trial aimed to provide further clarity about whether routinely offering people low-dose oxygen after a stroke affected a range of outcomes important to patients.

What did this study do?

The UK-based SO2S randomised controlled trial equally allocated 8,003 adults admitted to hospital after stroke to groups who either received continuous oxygen, night-time oxygen or no routine oxygen (control) within 24 hours. Participants’ oxygen levels were monitored four times a day.

The average age of participants was 72 years, 55% were men, and 92% were living independently before their stroke. Oxygen was administered before hospital admission in 20% of people, and the average oxygen saturation was 96.6% at randomisation.

Diagnoses were an ischaemic stroke (82%), haemorrhagic stroke (7%), stroke of unknown type (4%), transient ischaemic attack (2%), non-stroke diagnosis (4%) and missing data (1%).

Around 82% of people managed to take the oxygen in the way intended. The remainder did not, and this was mainly due to confusion and restlessness. Monitoring at 6 am and midnight were introduced halfway through the trial to check this “adherence”.

What did it find?

  • There was no difference in people’s level of disability after 90 days whether they received oxygen or not (adjusted odds ratio [aOR] 0.97, 95% confidence interval [CI] 0.89 to 1.06), and no difference between people who received continuous oxygen or night-time oxygen (aOR 1.01, 95%CI 0.92 to 1.12).
  • The number of people who died was similar in all people who received oxygen compared with those who did not (hazard ratio [HR] 0.97, 99% CI 0.78 to 1.21). There was also no difference between continuous and night-time only oxygen (HR 1.15, 99% CI 0.90 to 1.48).
  • There was no difference between the groups in the number of people at 90 days who were living independently or living in their own home, their ability to perform basic or extended everyday activities, or quality of life. Neurological outcomes after one week were also similar.
  • The number of serious adverse events was similar in all groups, and no oxygen-related adverse events were reported.

What does current guidance say on this issue?

Guidelines by NICE (2008), the Royal College of Physicians (2016) and the British Thoracic Society (2017) recommend giving oxygen to people who have had a stroke if their oxygen levels drop. The accepted threshold at which oxygen is advised is around 95% oxygen saturation. Routine oxygen administration is not recommended.

Oxygen monitoring is advised every four hours day and night. Oxygen should be administered via a tube in the nose, using the lowest concentration possible to restore oxygen levels. Keeping people with existing heart or respiratory conditions sitting upright can help to maintain oxygen levels.

What are the implications?

This trial provides robust UK data that reinforces guideline recommendations to not routinely give additional oxygen to people after they have had a stroke. Rather, people’s oxygen levels should be monitored so that supplemental oxygen can be used when medically necessary.

These findings highlight the need for regular monitoring of oxygen levels for all people who have had a stroke and monitoring of oxygen use to ensure adherence when it is used.

Not using oxygen routinely is likely to save the NHS money and nurse time for use on other things.

Citation and Funding

Roffe C, Nevatte T, Sim J, et al; Stroke Oxygen Study Investigators and the Stroke Oxygen Study Collaborative Group. Effect of Routine Low-Dose Oxygen Supplementation on Death and Disability in Adults With Acute Stroke: The Stroke Oxygen Study Randomized Clinical Trial. JAMA. 2017;318(12):1125-35.

This project was funded by the National Institute for Health Research Health Technology Assessment Programme (project number 09/104/21) and the Research for Patient Benefit Programme.



British Thoracic Society Emergency Oxygen Guideline Development Group. BTS guideline for oxygen use in adults in healthcare and emergency settings. Thorax. 2017;72(suppl 1):256-73.

Intercollegiate stroke working party. National clinical guideline for stroke. London: The Royal College of Physicians; 2016.

NHS website. Stroke. London: Department of Health and Social Care; updated 2016.

NICE. Stroke and transient ischaemic attack in over 16s: diagnosis and initial management. CG68. London: National Institute for Health and Care Excellence; 2008.

Stroke Association. State of the nation: stroke statistics. London: Stroke Association; 2017.

Produced by the University of Southampton and Bazian on behalf of NIHR through the NIHR Dissemination Centre


  • Share via:
  • Print article


Expert commentary

It is easy to see the logic for giving oxygen in acute stroke if neurons are dying for lack of it, and that alone has dictated widespread practice before now.So this study addresses a real issue in acute stroke care and, as a large well-conducted trial, we can have substantial confidence in the findings, especially as participants were pretty typical of stroke patients admitted to hospitals in the UK.The trial gives us an unambiguous message: ‘if you're not hypoxic, you don't need oxygen', which will help clinicians, patients and their families in the crisis of acute stroke.Dr Martin James, Consultant Stroke Physician/Honorary Associate Professor, Royal Devon & Exeter Hospital/University of Exeter Medical School

Expert commentary

Strokes occur when blood flow to part of the brain is cut off, starving cells of oxygen, so it is not surprising that clinicians have wondered whether giving extra oxygen might aid recovery.In this study, people with acute stroke and normal blood oxygen levels received no additional oxygen, continuous oxygen via a mask, or oxygen at night only for three days following stroke.There was no difference in recovery between the groups, meaning that we can now be confident that, provided blood oxygen is normal, there is no benefit in giving extra oxygen after a stroke.Professor Philippa Tyrrell, Chair in Stroke Medicine, Greater Manchester CLAHRC

Expert commentary

The results of the trial influence practice by showing that administering prophylactic oxygen does not reduce death or disability at three months in patients who are not classed as hypoxic.This further supports The Royal College of Physicians National Clinical Guidelines for Stroke, which states “patients with acute stroke should only receive supplemental oxygen if their oxygen saturation is below 95%”.Ensuring these standards of care are adhered to enhances safety and outcomes for our patients.Maria Fitzpatrick, Lead Consultant Nurse for Stroke in South East London, Friends Stroke Unit, Kings College Hospital & Governance Lead Nurse Representative Pan-London
Back to top