Does use of oxygen therapy improve outcomes in patients with suspected myocardial infarction who do not have hypoxemia at baseline?

In patients with suspected myocardial infarction (MI) and normal oxygen levels, giving immediate supplemental oxygen therapy does not improve mortality at one year. Although this study was underpowered because of fewer than expected deaths in the control group, the results were consistent across all subgroups, as well as with findings from other literature. (LOE = 1b)

Hofmann R, James SK, Jernberg T, et al, for the DETO2X–SWEDEHEART Investigators. Oxygen therapy in suspected acute myocardial infarction. N Engl J Med 2017;377(13):1240-1249.  [PMID:28844200]

Randomized controlled trial (nonblinded)

Government

Concealed

Inpatient (any location)

These investigators randomized patients with suspected MI (chest pain or shortness of breath with either electrocardiographic changes of ischemia or elevated troponin levels) and an oxygen saturation of 90% or higher on pulse oximetry to receive either oxygen therapy (n = 3311) or ambient air (n = 3318). Patients were recruited from ambulance services, emergency departments, coronary care units, and catheterization laboratories. The final diagnosis was MI in 76% of the included patients. The remaining 24% were more likely to have unspecified chest pain, other cardiac disease, or angina. In the oxygen group, oxygen was delivered at 6 liters per minute for 6 to 12 hours via an open face mask. The median duration of oxygen therapy was 11.64 hours. Overall, 1.9% patients in the oxygen group and 7.7% patients in the ambient air group received oxygen therapy outside of the protocol because of development of hypoxemia but were included in the intention-to-treat analysis. Furthermore, 9% patients in the oxygen group and 3% patients in the ambient air group did not complete participation through the end of the treatment period and were excluded from the per-protocol analysis. For the primary outcome of all-cause death at one year, there was no significant difference detected between the 2 groups (mortality rate: 5% in oxygen group, 5.1% in ambient air group; P = .80). Results were similar in the per-protocol analysis and across pre-specified subgroups. Additionally, no differences were detected in death at 30 days or in rates of rehospitalization for myocardial infarction at either 30 days or 1 year. Of note, the power calculation for the primary outcome was based on an estimated death rate in the control group of 14%. Since the observed rate was much lower at 5.1% (possibly because 24% of the patients did not actually have MI as a final diagnosis), this study did not have adequate power to detect a small difference in the oxygen group if such a difference existed.