We in emergency medicine are continually challenged to work in clinical uncertainty and with limited information, so it's nice to be able to fall back on well-established precepts from time to time. Much in the way the Frankenstein's monster understood “Fire bad!” we know that hypoxia is a dangerous condition that often warrants urgent intervention. So doesn't it stand to reason that if low oxygen is bad, more oxygen must be better? The answer, while potentially surprising, is becoming increasingly clear: Too much oxygen can kill.
We've had clues that this was the case for a while now. The esteemed Natalie May, MBCHB, MPHE, wrote about this topic in EMN two years ago, reviewing the evidence showing that oxygen therapy was harmful in STEMI patients and lamenting that the MONA mnemonic needed to go the way of the dodo. (2016;38:1; http://bit.ly/2t79miY.)
That analysis was limited to one particular condition, albeit one of significant clinical import, but we see patients of every stripe in the ED, and we want to know how oxygen affects the acutely ill cohort for whom we care. Enter Derek Chu, MD, et al. and his aptly named IOTA (Improving Oxygen Therapy in Acute illness) systematic review and meta-analysis. (The Lancet 2018;391:1693.)
This review analyzed more than 16,000 patients from a whopping 25 different randomized controlled trials of acutely ill patients receiving either liberal (groups with the higher oxygen target) or conservative oxygen therapy. But what kind of patients were studied? Pretty much all the kinds we worry about the most, including those with sepsis, critical illness, stroke, trauma, myocardial infarction, and cardiac arrest.
So we have a nice cohort of sick patients, and we often tend to throw some oxygen onto their mugs among the million other things we do in the process of assessing, resuscitating, diagnosing and dispositioning them. This happens a lot. One study found that a third of patients arriving by ambulance and a quarter of patients treated in the ED receive oxygen. (Emerg Med J 2008;25:773.) Some of these patients definitely need their hypoxia corrected, but it turns out that we overshoot the mark and cause hyperoxemia quite often, like 43.6 percent of the time in intubated ED patients, to name just one example. (Crit Care 2018;22:9.)
A Lovely Shade of Blue
Much as we'd like to deny it, emergency medicine is chock-full of secret gunners, and a gunner likes nothing more than acing a test. And it is profoundly gratifying to see that blue photoplethysmograph tick up and show us an oxygen saturation of 100%. Unfortunately, all this extra oxygen results in badness for our patients: vasoconstriction, inflammation, and increased oxidative stress on the lungs, the heart, and the brain. If that sounds like I'm placing too much importance on surrogate outcomes, several trials show an association between excess oxygen and increased shock, respiratory failure, arrhythmias, and heart attacks. None of those things are good, and we would of course want to avoid them in our patients whenever possible. The fact that too much oxygen is associated with these kinds of outcomes should give us pause to consider whether this seemingly harmless common practice is really as benign as we've long believed.
Excess shock and heart attacks aren't good, but being dead is even worse. This is the part where you can fairly accuse me of burying the lede because the stunning finding from the IOTA trial is this: Liberal oxygen therapy is associated with increased mortality. This finding was true across all subgroups and for all the time frames analyzed. Liberal oxygen therapy was associated with increased mortality in-hospital, at 30 days, and at longest follow-up, with NNHs for death of 91, 71, and 83, respectively.
The authors also found a dose-dependent response: The relative risk of death in hospital increased by 25 percent and the risk of death at longest follow-up increased by 17 percent for every one percent increase in SpO2 (above the 94-96% range). These are sobering results. In the words of the IOTA authors, “This systematic review and meta-analysis provide high-quality evidence that hyperoxia is life-threatening. This is a distinct viewpoint from the current notion that at worst, liberal oxygen is not beneficial for acute illnesses.”
As important as these findings are and while this meta-analysis was large and of high quality, it was not a perfect work. It didn't include children, pregnant patients, people on ECMO, patients with psychiatric disease, and perhaps most importantly, those with chronic respiratory illness. This limits the applicability of these findings in a cohort of patients likely to present sick and hypoxic. High-quality evidence showed that liberal oxygen therapy did not decrease the risk of hospital-acquired infections, but low-quality evidence from two trials (449 total patients) at high risk of bias showed that liberal oxygen did decrease infections in patients admitted for emergency surgery (RR 0.50, 95% CI, 0.36-0.69).
Quibbles aside, the IOTA analysis should be the last nail in the coffin for indiscriminately applied oxygen therapy. We need to stop putting nasal cannulas and face masks on every sick patient we see and stop shooting for that soothing blue 100% unless we want to be responsible for hammering in a few of those coffin nails ourselves.
Share this article on Twitter and Facebook.
Access the links in EMN by reading this on our website or in our free iPad app, both available at www.EM-News.com.
Comments? Write to us at firstname.lastname@example.org.Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.