Physicians are being told to cut back routine oxygen therapy for hospital patients because the benefit is uncertain and there is clear harm. In particular, oxygen provides no benefit to patients with heart attacks or strokes with normal blood oxygen saturation.
The recommendations come from a panel of international experts led by McMaster University researchers, and they have been published in the prestigious medical journal The BMJ.
"It is commonplace for patients to receive oxygen when they shouldn't, or be given more oxygen than they need," said physician Reed Siemieniuk, one of the panel's co-chairs and a researcher with McMaster's Department of Health Research Methods, Evidence, and Impact.
"As a result, oxygen therapy probably contributes to the deaths of many Canadians."
The advice is part of The BMJ's Rapid Recommendations initiative to produce rapid and trustworthy guidance based on new evidence, to help doctors make better decisions with their patients.
Siemieniuk said oxygen therapy is widely used in hospitals and it is common to give extra oxygen to sick patients, often with relatively little attention paid to when to start and stop it. Until now, clinical practice guidelines for physicians also vary in their advice on when to give oxygen and how much to give.
Oxygen levels are measured by blood saturation (SpO2) - the amount of hemoglobin in the bloodstream that is saturated with oxygen to carry it through the body. Normal oxygen saturation is usually between 96% and 98%, but sick patients are often kept close to 100%.
The new recommendations are a follow-up to a recent review of evidence, also led by McMaster University researchers, which found that giving extra oxygen to hospital patients with normal oxygen levels increases mortality. The authors concluded that oxygen should be given conservatively.
Using the GRADE approach, which is a McMaster-developed system used to assess the quality of evidence, the interprofessional health professional panel made a strong recommendation to stop oxygen therapy in patients with a saturation of 96% or higher.
For most patients, they say a target of 90-94% saturation seems reasonable and is low enough to avoid harm. In all cases, they advise using the minimum amount of oxygen necessary.
The authors point out that while their recommendations apply to most patients, they do not apply to surgical patients, babies, or patients with other uncommon conditions.
Panel co-chair Gordon Guyatt, a McMaster professor of health research methods, evidence and impact, added: "Healthcare workers need to start thinking of oxygen like any other treatment – one that can sometimes be useful but one that can also cause serious harm."