AVOID: Excessive oxygen in MI

Excessive O2 in MI can cause harm

The routine use of supplemental oxygen in MI is of no benefit and may actually be of harm. In fact, data from 1976 (!) suggested that the use of high concentration oxygen therapy in uncomplicated MI results in a significant increase in infarct size.

This is supported more recently with 2 reviews in 2009 (Farquar et al, Wijesinghe et al) that routine supplemental oxygen may lead to increased coronary artery tone and reduction in coronary artery blood flow, reductions in cardiac output, increased systemic vascular resistance and increased infarction size.  In a 2013 Cochrane review of 4 trials (n=430), analysis revealed an increased mortality rate in the group who received supplemental oxygen, although the conclusions of this are limited by the small sample.

Current guidelines suggest that supplemental O2 be reserved for patients with hypoxaemia evidence by low SaO2, pO2 or other evidence of respiratory compromise. The evolving recommendations for oxygen in STEMI have shifted away from use of supplemental oxygen:

  • 1999: Class I - oxygen for SaO2 <90%; Class IIa - routine oxygen for first 2-3 hours of uncomplicated MI
  • 2004: Class I - oxygen for >6 hours in hypoxaemia
  • 2013: Limited data; no recommendation

AVOID: Air vs Oxygen


This multicentre, pragmatic clinical trial comes from Melbourne, with 638 patients identified with STEMI in the prehospital arena, of whom 470 were included and continued through to PCI; 441 had confirmed STEMI. Patients with SaO2 <94%, required oxygen for other reasons or had already received oxygen were excluded. The study calculated a sample size of 600 prehospital enrollments with 490 eligible on hospital arrival to detect a 20% difference in myocardial injury in a two-sided test.

Patients were randomised to either 8L/min supplemental oxygen or no oxygen (although supplemental oxygen was administered if the SaO2 fell below 94%). Oxygen provision was not titrated to SaO2—this has been previously demonstrated to be difficult to achieve in the prehospital setting (the HOT or NOT trial).


The primary endpoint was the extent of myocardial injury using peak cardiac enzymes as a surrogate marker. Secondary endpoints at hospital discharge and 6 months were mortality, major adverse cardiac events, MI size and ST segment resolution.

There was no benefit noted in the oxygen group across the primary endpoints. The oxygen-receiving group where noted to have a larger infarct size as measured using peak CK (but not toponin I), as well as on 6-month cardiac MRI in a subgroup (n=139, 32%).

They conclude:

Our study does not demonstrate any significant benefit of routine oxygen therapy for reducing myocardial infarct size, improving patient hemodynamics, or alleviating symptoms. Instead, we identified some evidence for increased myocardial injury when oxygen was administered during uncomplicated AMI.

The Upshot

This article, despite its limitations, adds further evidence to not routinely using supplemental oxygen in uncomplicated MI, but reserving if for patients with hypoxaemia. There is a good accompanying editorial that also explores a number of these issues. There is an ongoing study in Sweden that is powered for mortality and morbidity with the use of supplemental oxygen in MI that may provide more definitive answers.


Share your thoughts