Intended for healthcare professionals


Avoidance of high concentration oxygen in chronic obstructive pulmonary disease

BMJ 2010; 341 doi: (Published 18 October 2010) Cite this as: BMJ 2010;341:c5549

This article has a correction. Please see:

  1. B Ronan O’Driscoll, consultant respiratory physician1,
  2. Richard Beasley, professor2
  1. 1Salford Royal University Hospital, Salford M6 8HD, UK
  2. 2Medical Research Institute of New Zealand, Wellington Hospital, Wellington 6021, New Zealand
  1. ronan.o.driscoll{at}

Routine use may also be harmful in several other medical emergencies

Oxygen was used to treat several diseases as early as the 1790s, and the belief that “purified air” was better than ordinary air persisted among doctors and the general public for the next two centuries.1 In the early 1960s, Moran Campbell recognised the danger of hypercapnic respiratory failure (carbon dioxide retention) caused by high concentrations of oxygen in the treatment of acute exacerbations of chronic obstructive pulmonary disease.2 Subsequent research has consistently reported adverse outcomes after the use of uncontrolled oxygen treatment in this condition.3 4 Evidence on the benefits and harms of the use of oxygen in acute exacerbations of chronic obstructive pulmonary disease has led all specialties in the United Kingdom to recommend controlled oxygen treatment with a target saturation range (usually 88-92%) in this condition.5 However, this recommendation was limited by the lack of level 1 evidence, because no randomised controlled trials were available.6

The linked cluster randomised trial by Austin and colleagues (doi:10.1136/bmj.c5462) finally fills this gap, and it provides robust evidence that the routine administration of high concentration oxygen in acute exacerbations of chronic obstructive pulmonary disease is associated with increased mortality.7 The trial compared high concentration oxygen treatment with titrated oxygen treatment in the pre-hospital setting in 405 patients with a presumed acute exacerbation of chronic obstructive pulmonary disease. Mortality was significantly lower in patients receiving titrated oxygen rather than high concentration oxygen (relative risk, 0.42, 95% confidence interval 0.20 to 0.89). In the subgroup of patients with confirmed chronic obstructive pulmonary disease (n=214) mortality was reduced even further (0.22, 0.05 to 0.91). In patients who had arterial blood gas measurements within 30 minutes of presenting to hospital, those who received titrated oxygen were significantly less likely to have hypercapnia (mean difference in alveolar carbon dioxide tension −34 mm Hg) or respiratory acidosis (mean difference in pH 0.12) than were those who received high concentration oxygen treatment. These physiological effects probably contributed to the increased risk of death, especially if high concentration oxygen treatment was continued within the hospital setting or if management of respiratory acidosis required exposure to the hazards of invasive ventilation.

Two other less well recognised mechanisms may also have contributed to the increased risk of death. Firstly, hyperoxaemia causes coronary artery vasoconstriction and reduced coronary artery blood flow.8 Cardiac troponin concentrations are raised in about 25% of patients with acute exacerbations of chronic obstructive pulmonary disease and positively correlate with the degree of hypercapnia and acidosis.9 Many patients with acute exacerbations of chronic obstructive pulmonary disease have coexisting ischaemic heart disease, and high oxygen concentrations may increase the risk of death by causing myocardial damage as a result of reduced coronary blood flow.

Secondly, if supplemental oxygen is withdrawn abruptly from patients in whom high concentration oxygen treatment has caused hypercapnia and a marked increase in alveolar carbon dioxide tension, the oxygen tension in the alveoli will fall rapidly to below that seen before the start of oxygen treatment, and this may result in profound hypoxaemia.5 This phenomenon, known as rebound hypoxaemia, can occur if oxygen treatment is suddenly stopped—for example, during a meal or in response to a blood gas sample showing hyperoxaemia in association with severe hypercapnia and acidosis.

The 9% mortality in patients given high concentration oxygen treatment in Austin and colleagues’ study is comparable to the 7.4% mortality reported in a 2003 UK national audit of acute exacerbations of chronic obstructive pulmonary disease.10 However, the mortality of only 2% in patients randomised to oxygen titrated to achieve an oxygen saturation of 88-92% sets a new “gold standard” for management of this condition. The priority for future randomised controlled trials will be to define the ideal target oxygen saturation levels; further trials of high concentration oxygen in this condition would not be ethical.

It is important to recognise that concerns about the safety of the routine use of high concentration oxygen treatment extend beyond chronic obstructive pulmonary disease. Evidence suggests that this therapeutic approach may worsen outcomes in a wide range of other clinical situations including, but not limited to, myocardial infarction, stroke, neonatal resuscitation, and postcardiac arrest resuscitation in adults.5 11 12 Sufficient evidence exists to recommend avoidance of routine administration of high concentration oxygen treatment in the emergency setting. The UK National Patient Safety Agency has expressed concerns about inadequate oxygen prescription, administration and monitoring.13 The use of oxygen should be limited to patients with hypoxaemia, and it should be titrated to relieve hypoxaemia and avoid hyperoxaemia. After more than 200 years of haphazard use it should be recognised that oxygen should be prescribed for defined indications in which its benefits outweigh its risks and that the patient’s response must be monitored.


Cite this as: BMJ 2010;341:c5549


  • Research, doi:10.1136/bmj.c5462
  • Competing interests: All authors have completed the Unified Competing Interest form at (available on request from the corresponding author) and declare: no support from any organisation for the submitted work; no financial relationships with any organisations that might have an interest in the submitted work in the previous three years; BRO’D had travel and accommodation funded by Fiser Paykell as an invited speaker at the CARE conference in 2010.

  • Provenance and peer review: Commissioned; not externally peer reviewed.


View Abstract