Oxygen of No Benefit, Even in MI Patients With Low-Normal Saturation

Supplemental oxygen in patients with confirmed myocardial infarction has no clinical benefit irrespective of oxygen saturation levels at baseline, a new analysis of the Determination of the Role of Oxygen in Suspected Acute Myocardial Infarction (DETO2X-AMI) study shows.

Even in patients with low-normal saturation levels, which was defined as saturation between 90% to 94%, supplemental oxygen did not lower the risk of all-cause mortality or rehospitalization for MI or heart failure when compared with patients who received ambient air alone, report Stefan James, MD, PhD (Uppsala University, Sweden), and colleagues in their paper published online December 11, 2019, in JACC: Cardiovascular Interventions. 

DETO2X-AMI included 6,629 patients with normal oxygen saturation and suspected acute MI who were randomized to oxygen at 6 L/min for 6-12 hours or to ambient air. The new analysis focused on 5,010 patients with confirmed MI, stratifying them into two groups based on oxygen saturation at baseline (low-normal of 90-94% and high-normal of 95-100%).

“In the main trial, there was a mixed population, all-comers, with suspected MI,” senior investigator Robin Hofmann, MD, PhD (Karolinska Institutet, Stockholm, Sweden), told TCTMD. “We were looking at patients with normal saturation, that being between 90% and 100%. In the broad population, we couldn’t see any difference. So the question, which I think is a reasonable one, is how does it look in patients with low-normal oxygen saturation at baseline, where we get closer to the level of hypoxemia where oxygen use is recommended? It’s reasonable to test if giving oxygen in those patients might be worthwhile.”

Like the main trial, which was presented and published in 2017, there was no benefit in the 836 patients with low-normal oxygen levels. The primary endpoint of all-cause mortality or rehospitalization for MI or heart failure occurred in 18.0% of patients with low-normal saturation levels who received oxygen and 16.7% in patients who received ambient air, a nonsignificant difference. Similarly, the primary composite endpoint occurred in 9.2% of high-normal patients treated with oxygen and 9.8% of patients who received ambient air, also nonsignificant.

The overall event rates for the primary endpoint differed significantly between those with low-normal and high-normal oxygen saturation (17.3% vs 9.5%; P < 0.01), even after adjusting for relevant variables. The difference was largely driven by hospitalizations for heart failure (6.7% vs 2.6%; P < 0.01). Patients in the low-normal oxygen cohort were older, more likely to be female, had more comorbid conditions and cardiovascular disease, and were being treated with more intensive medical therapy than those with high-normal oxygen saturation levels.

“There’s a big difference in outcomes in these [low-normal versus high-normal] patients regardless of oxygen therapy,” said Hofmann. “Oxygen doesn’t change the outcome. The outcomes are driven not by hypoxemia itself, but other factors. Basically, the [low-normal] patients are sicker. This fact remains even after we adjust for all baseline differences. . . . The fact that we apply oxygen in this group does not change their risk.”

Oxygen Provided vs Oxygen Utilized

European Society of Cardiology (ESC) 2017 guidelines for the treatment of STEMI no longer recommend routine oxygen in MI patients when oxygen saturation is 90% or greater, nor do guidelines from the Canadian Cardiovascular Society/Canadian Association of Interventional Cardiology. In the United States, the American College of Cardiology/American Heart Association only recommend routine use of oxygen in patients with clinically significant hypoxemia.

If you still believe it’s reasonable to give oxygen to a normoxemic individual, do the trial. Of all that is known so far, oxygen is not beneficial in this group. Robin Hofmann

In an editorial, Joseph Fisher, MD (Toronto General Hospital, Canada), and Duminda Wijeysundera, MD, PhD (St. Michael’s Hospital, Toronto), question whether the oxygen intervention resulted in a “physiologic change sufficiently large to have clinical effects,” which could have been assessed by monitoring end-tidal PO2 to calculate alveolar PO2 or arterial blood gases in a subset of patients. The inspired oxygen in the trial was uncontrolled and unmeasured, they state.

“In other words, we cannot make any assumptions as to the physiological effects—and by extension, clinical effects—of oxygen blowing into a mask,” write Fisher and Wijeysundera.

To TCTMD, Hofmann said the editorialists raise a valid point, but that this was not the point of DETO2X-AMI. The researchers’ goal, which Fisher and Wijeysundera acknowledge, was to test the way oxygen was administered in clinical practice.

“In routine practice, you supply oxygen with a face mask or with a nasal cannula and test oxygenation with pulse oximetry, which isn’t the best technique available, but this is how it is,” he said. “The points the editorialists make are valid that that’s not what we do. If you want to check the efficacy of utilized oxygen, then they’re right: you have to use arterial blood gases, change the flow rate according to the individual. But that wouldn’t happen in routine clinical practice.”

Likening oxygen to medical therapy, Hofmann said it shouldn’t be used in normoxemic STEMI patients given the absence of benefit.

“If you want to supply this drug, then you have to do a trial that shows efficacy,” he said. “It’s what we always do. If you still believe it’s reasonable to give oxygen to a normoxemic individual, do the trial. Of all that is known so far, oxygen is not beneficial in this group.”