Practice Uncertainties

Does routine oxygen supplementation in patients with acute stroke improve outcome?

BMJ 2012; 345 doi: (Published 30 November 2012) Cite this as: BMJ 2012;345:e6976
  1. S J Pountain, study manager of the Stroke Oxygen Study1,
  2. C Roffe, consultant physician 1, professor of stroke medicine2
  1. 1Stoke Stroke Research Group, North Staffordshire Combined Healthcare NHS Trust, Stoke-on-Trent ST4 7LH, UK
  2. 2Institute for Science and Technology in Medicine, Keele University, Keele
  1. Correspondence to: christine.roffe{at}

Stroke is the third most common cause of death and the leading cause of long term disability in developed countries. Specialist care in stroke units is well established as being effective in preventing death and disability after stroke.1 However, which aspects of stroke care are crucial for improving outcome remains unclear. Patients in a stroke unit are more likely than those on a non-specialised general ward to receive oxygen.2 Bravata and colleagues found that treating all episodes of hypoxia with supplemental oxygen was one of three key processes associated with better outcome in acute stroke care.3 Mild hypoxia is common in patients with stroke and may have substantial adverse effects on an ischaemic brain after stroke. Whereas healthy adults with normal cerebral circulation can compensate for mild hypoxia by an increase in cerebral blood flow, this is not possible in patients whose brain is already ischaemic after stroke.4 Hypoxaemia in the first few hours after hospital admission is associated with an increased risk of death.5

Oxygen treatment is not without problems.6 The tubing that connects the patient to the oxygen source impedes early mobilisation and could pose an infection risk. Physiological changes associated with oxygen treatment can include absorption atelectasis; worsening of ventilation-perfusion mismatch; coronary, cerebral, and systemic vasoconstriction; and a reduction in cardiac output.7 Animal models and in vitro studies show that oxygen encourages the formation of toxic free radicals, leading to further damage to the ischaemic brain, especially during reperfusion. Oxidative stress has also been implicated in the activation of cell signalling pathways, which lead to apoptosis and neuronal cell death. Although some research points towards adverse effects of hyperoxia in the ischaemic brain, some evidence supports the notion that eubaric hyperoxia (hyperoxia induced by oxygen treatment at normal atmospheric pressure) may be neuroprotective.8

What is the evidence of the uncertainty?

We searched PubMed, Medline, and Embase databases and the Cochrane Library for randomised controlled trials comparing oxygen treatment with placebo or with no treatment in patients with acute stroke and other relevant literature using the search terms “oxygen” or “hypoxia” or “anoxia” and “stroke” or “cerebrovascular accident”. We included only articles on normobaric oxygen treatment. (Bennett and colleagues conducted a systematic review of hyperbaric oxygen for acute stroke.9)

Should patients with acute stroke be given oxygen routinely?

We found one quasi-randomised and two randomised controlled pilot studies of oxygen supplementation after acute stroke. The quasi-randomised study enrolled 550 consecutive patients within 24 hours of acute stroke and treated them with either oxygen at a rate of 3 L/min via nasal cannulas or no routine oxygen during the first 24 hours. The findings showed no difference in survival at one year or disability at seven months. Subgroup analyses suggested a trend towards better outcomes with oxygen in patients with severe strokes and an adverse trend in those with mild strokes, but the study size was too small to define patients who are likely to derive benefit with certainty.10 In the smaller of the two randomised controlled pilot studies, 16 patients received high flow humidified oxygen at 45 L/min via a simple face mask (using a flow meter that can deliver up to 75 L/min—such as Timeter Classic Series Flowmeter model 0-75, Allied Healthcare Products) or control treatment (room air) within 12 hours of symptom onset, for eight hours. Patients had potentially salvageable tissue in the ischaemic penumbra of the infarct, indicated by a perfusion-diffusion mismatch on magnetic resonance imaging. There was a trend towards better reperfusion on magnetic resonance imaging at 24 hours, but no long term clinical benefit at three months.11 In the larger pilot study (Stroke Oxygen Pilot Study) 301 patients within 24 hours of admission with acute stroke were randomised to receive either oxygen supplementation for 72 hours at a flow rate of 2 L/min or 3 L/min, depending on baseline oxygen saturation, or no routine oxygen supplementation. After one week, their neurological outcome—according to the score on National Institutes for Health stroke scale—was the same in both groups, but the degree of recovery (difference in neurological scores between baseline and week 1) was significantly better in the group receiving oxygen supplementation.12 The researchers found no difference in functional outcome at six months on direct comparison, although they found a trend for better outcome with oxygen after correction for differences in baseline stroke severity and prognostic factors. These results are promising but need confirmation in a much larger study. This is now in progress (see below).

Clinical guidelines on oxygen supplementation after stroke differ across countries and have changed over time without clear justification (box). None of the recommendations are based on evidence from controlled clinical trials. Not surprisingly, doctors treating patients with stroke are uncertain about which treatment approach to take and when to give oxygen. In a survey of British stroke physicians just over half stated they would start oxygen supplementation after stroke at an oxygen saturation below 95%,13 a cut-off that was later included in UK and European guidelines on stroke (box).

Recommendations for oxygen treatment in national and international guidelines

British Thoracic Society Guideline for Emergency Oxygen Use in Adult Patients
  • Supplementary oxygen should only be given to achieve a saturation of 94-98%, or 88-92% in patients at risk of type II respiratory failure (2008)

UK National Clinical Guidelines for Stroke
  • Arterial oxygen concentration should be maintained within normal limits (2004)

  • Give oxygen if saturation <95% (2008)

European Stroke Initiative and European Stroke Organization Recommendations for Stroke Management
  • Give oxygen at a rate of 2-4 L/min when indicated (2003)

  • Give oxygen if saturation <92% (2007)

  • Give oxygen if saturation <95% (2008)

American Stroke Association Guidelines
  • Give oxygen if saturation <95% (2003, 2005)

  • Give oxygen if saturation <93% (2007)

How should oxygen be given in patients with stroke?

A small study (n=46) compared oxygen treatment via a venturi mask delivering 40% oxygen with nasal cannulas at a flow rate of 2 L/min within 48 hours of an acute middle cerebral artery infarct and found non-significant differences in mortality (6% v 24%, P=0.1) and complications.14 When given the choice, patients with stroke prefer nasal cannulas to face masks (73% out of 21 asked).15

When should oxygen supplementation be given?

Hypoxia is common early after stroke, particularly while a patient is being moved from one ward to another or is in the head scanner,5 and at night.16 The mean nocturnal oxygen saturation is about 1% lower than “awake” oxygen saturation, both in patients with stroke and in controls. Twenty three per cent of patients with acute stroke who have normal oxygen saturation during the day spend more than 30 minutes with an oxygen saturation <90% at night.13 Sleep apnoea has been reported in 60-70% of patients early after stroke.17 18 A study in healthy volunteers found that hypoxaemia leads to a compensatory increase in cerebral blood flow during wakefulness but not during sleep and is therefore more likely to result in brain tissue hypoxia at night.19 Oxygen supplementation is therefore probably more important at night than during the day. No studies of oxygen saturation or treatment in patients very early after the stroke (for example, in the ambulance) have been published.

How long should oxygen be given for?

Two studies giving oxygen for eight and 24 hours showed no benefit.8 9 A pilot study of oxygen supplementation for 72 hours showed improved neurological recovery at one week, but no significant difference in functional outcomes at six months.10 We currently do not know whether oxygen supplementation should be given routinely, how early to start, and for how long to continue treatment.

Is ongoing research likely to provide relevant evidence?

A search of the Current Controlled Trials Register (ISRCTN), the metaRegister of Controlled Trials, and the US Government Clinical Trials Register using the terms “stroke” and “oxygen” identified one ongoing, one prematurely terminated, and one completed but unpublished study of oxygen treatment for acute stroke.

The Stroke Oxygen Study (ISRCTN52416964), an ongoing randomised controlled study in 6600 patients with acute ischaemic stroke, is comparing the effects of routine fixed dose supplementation (3 L/min if baseline oxygen saturation is ≤93%, and 2 L/min if baseline oxygen saturation is >93%) for 72 hours; routine fixed dose oxygen supplementation at night only, for three nights; and usual management (oxygen given only when needed) on neurological recovery at one week and functional outcome at three months.

The Normobaric Oxygen in Acute Ischaemic Stroke Trial (NCT00414726), a randomised controlled trial, compared high dose oxygen treatment (30-45 L/min via face mask) or control (room air at the same flow rate) for a period of eight hours on neurological recovery within four and 24 hours in patients within nine hours of acute ischaemic stroke. This trial had aimed to enrol 240 patients but was stopped early because of an imbalance of deaths in favour of the control group (17/43 on oxygen v 7/43 on room air). The results are available on the website but not published in a journal.

A randomised controlled study of the effect of low flow oxygen on capillary blood gases in 40 patients with acute stroke (ISRCTN75718175) is listed as completed but has not yet reported results on the ISRCTN website or in print.

Recommendations for future research

  • How does intensive monitoring compare with routine treatment?

  • How effective is night only supplementation compared with continuous supplementation?

  • Does starting treatment very early (before hospital admission) improve outcome?

  • If oxygen supplementation is effective, what is the optimal treatment dose and duration and the best mode of administration?

  • Are there subgroups of patients who respond better (such as those with ischaemic stroke v haemorrhagic stroke; those receiving thrombolysis; those with a large difference in diffusion-perfusion deficit on magnetic resonance imaging)?

  • Is there a cut-off in oxygen saturation beyond which oxygen supplementation is no longer effective?

What should we do in the light of the uncertainty?

Until evidence shows whether routine oxygen supplementation improves outcome, patients with acute stroke should only be given oxygen to maintain oxygen saturation within the normal range. Patients should be checked regularly for hypoxia, especially at night, during transfers between wards, and in the head scanner. It is important to remember that hypoxia is a symptom of an underlying problem in gas transfer or regulation of respiratory activity, and not a disease in itself. Blind treatment of hypoxia without further investigation risks masking an important warning sign of an underlying life threatening condition and delaying its detection and treatment. Before starting oxygen treatment, basic resuscitative measures, such as checking and clearing the airway and optimising patient positioning, should be taken. It is further important to establish and treat the cause of hypoxia. Secretions in the upper airways, pneumonia, pulmonary embolism, heart failure, and the effects of sedative medications on respiratory drive need to be managed in all stroke patients with low or falling oxygen saturation.


Cite this as: BMJ 2012;345:e6976


  • doi:10.1136/bmj.e6856
  • This is one of a series of occasional articles that highlight areas of practice where management lacks convincing supporting evidence. The series adviser is David Tovey, editor in chief, the Cochrane Library. This paper is based on a research priority identified and commissioned by the National Institute for Health Research’s Health Technology Assessment programme on an important clinical uncertainty. To suggest a topic for this series, please email us at uncertainties{at}

  • The views and opinions expressed in this article are those of the authors and do not necessarily reflect those of the National Institute for Health Research, the NHS, or the Department of Health.

  • Contributors: Both authors contributed to the planning, drafting, revising, and final approval of the article, and both are guarantors.

  • Competing interests: Both authors have completed the ICMJE uniform disclosure form at (available on request from the corresponding author) and declare: both authors had support for the submitted work from North Staffordshire Combined Healthcare NHS Trust and are collaborators on the Stroke Oxygen Study with grant funding by the National Institute for Health Research (NIHR) Health Technology Assessment programme, previously funded by the NIHR Response for Patient Benefit programme; no financial relationships with any organisations that might have an interest in the submitted work in the previous three years; no other relationships or activities that could appear to have influenced the submitted work.

  • Provenance and peer review: Commissioned; externally peer reviewed.


View Abstract