Background

When is liberal high flow oxygen of benefit in the ED?¹

• Spontaneous pneumothorax: high flow oxygen has been shown to increase the resorption of air from a pneumothorax four-fold in animal studies.
• Cluster headaches: high flow oxygen has been shown to produce relief in 56-85% of patients.
• Carbon monoxide and cyanide poisoning: BTS Guidelines recommend the use of high flow oxygen regardless of pulse oximeter readings.
• Critical illness: BTS guidelines state that many CPR guidelines recommend high flow oxygen. However, not many of those are evidence based, mainly due to the ethical considerations of any RCT investigating this. There is some evidence that hyperoxia causes a paradoxical decrease in whole body consumption in critical illness and hyperoxia leads to impaired oxygen delivery in septic patients.

When shouldn’t you use liberal high flow oxygen in the ED?

COPD: The effect of high flow oxygen causing hypercapnia in COPD patients is well documented.¹ A retrospective review of case notes by Roberts et al. found that mortality increased from 7.2% to 11.1% in those given greater than 35% oxygen.²

This can be challenging in the pre-hospital setting, where patient history may be limited. There is limited data looking at the effect of high flow oxygen in the pre-hospital setting. The only study described in a recent Cochrane review³ is an RCT by Austin et al⁴ comparing high flow oxygen with titrated oxygen in patients with COPD in the pre-hospital setting. This showed 9% deaths (21 patients) in the high flow arm compared with 4% (7 patients) in the titrated arm. The BTS guidelines therefore recommend aiming for a target saturation of 88-92% in anybody suspected of COPD prior to blood gas measurements.

There are other conditions which put patients at risk of hypercapnic respiratory failure. BTS guidelines advise treating these as for COPD. These groups include:

• Cystic fibrosis
• Non-cystic fibrosis bronchiectasis
• Severe kyphoscoliosis or severe ankylosing spondylitis
• Severe lung scarring from old TB,
• Morbid obesity (BMI >40 kg/m2)
• Musculoskeletal disorders with respiratory muscle weakness
• Overdose of respiratory depressant drugs e.g. opioids and benzodiazepines

Parquet poisoning and bleomycin lung injury: BTS guidelines states that oxygen should only be given if saturations drop below 85%, as oxygen is hazardous in these conditions.

Post return of spontaneous circulation: A recent meta-analysis of RCTs in Resuscitation⁵ concluded that conservative oxygen therapy was associated with a statistically significant reduction in mortality at follow-up compared to liberal oxygen therapy. However, the certainty of available evidence was low or very low due to bias, imprecision, and indirectness. Similar findings were found by a meta-analysis of observational studies by Wang et al⁶. Based on current evidence from the Resuscitation Council and BTS guidelines, it is recommended to aim for target saturations of 94–98% post ROSC.¹

Myocardial infarction: Current BTS guidelines advise targets saturations of 94–98% in patients with myocardial infarction. There have been no studies demonstrating the benefit of oxygen administration in non-hypoxaemic patients.¹ The AVOID study comparing provision of routine oxygen to no-supplemental oxygen showed infarct sizes were 55% larger in the oxygen group at 6 month-MRI scan.7 DETOX2AMI study showed that supplemental oxygen in those without hypoxaemia did not reduce 1 year mortality in patients with MI.⁸

Stroke: There have been four randomised trials of oxygen therapy in patients with stroke with normal oxygen saturation, none of which showed any benefit. A small randomised trial of high concentration therapy 10 L/min showed no benefit and a study in 2009 looking at very high flow oxygen 45 L/min had to be terminated due to excess mortality. One observational study reported increased mortality associated with hyperoxaemia in patients with stroke in ITU. Therefore, the current recommendation in stroke is to aim for target saturations of 94-98%¹.