Avoid wrong wayWe’ve been waiting for the AVOID study, since we mentioned it a few years ago in another post on the harm of excessive oxygen. AVOID (Air Versus Oxygen in Myocardial Infarction). Now, it’s out. As expected, it shows that unnecessary oxygen supplement worsens outcome. The surprise is just how big a difference it makes! In this study, too much oxygen increased recurrent MI fivefold!

Avoiding hyperoxemia isn’t new. Normoxemia has been a trend, but lacked hard evidence in form of an RCT, and the reflex-O2-mask in ED has been hard to fight. Right now, the full AVOID article has yet to be released, but the results have just been presented at AHA’s congress in Chicago last week. And AHA has posted a video interview with Dr. Stub, one of the investigators of the AVOID trial, on the results, as well as posted his presentation slides here.

Prospective, multicenter trial, encompassing both pre-hospital and in-hospital treatment of the patients. Included STEMI patients with an initial Sat >94% and STEMI symptoms less than 12 hours.

Here’s the flow chart for patients in AVOID. Pre-hospital is above the stapled line, in-hospital below.

Study endpoints
The endpoint was infarct size measured by TnT and CK-MB, as well as judging infarct size by MRI scans 6 months post MI. There was a (just) significant increase in cardiac enzymes and borderline significant increase in infarct size by MRI in the oxygen group.

Important clinical endpoints
Of clinical interest was that the oxygen arm had highly significant increase of recurrent MI during admission – the risk was 5x as high! The no-oxygen group had a recurrent MI rate of 0.9%, but it was 5.5% in the oxygen group! Also, they found an increase of significant arrythmias in the oxygen group. These are based on rather small numbers, so the big effect seen here might not be replicated. There were no reported positive effects of oxygen supplement, except for the happy bleep of the monitor measuring 100% sats.

This is a quite small study, but falls nicely into place with other evidence on oxygen toxicity, and how we’ve moved away from supranormal levels of oxygenation. There’s currently a larger RCT underway in Sweden that will measure more hard endpoints, the DETO2X study. Read about it on deto2x.se. It’ll be interesting to see those results too. Also, go read the old evidence on oxygen toxicity stretching back to the 50s in this post: Oxygen. Enough, already.

We’ll be back with a more detailed report when the full AVOID article is out. But for now, turn off that oxygen for patients with sats >94% on room air! It’s been an infatuation for way too long, it’s time to break up with MONA.

AVOID trial abstract

Oxygen is commonly administered to patients with ST-elevation myocardial infarction (STEMI) despite previous studies suggesting a possible increase in myocardial injury due to coronary vasoconstriction and heightened oxidative stress.
We conducted a multicenter, prospective, randomized, controlled trial comparing oxygen (8 L/min) with no supplemental oxygen in patients with STEMI diagnosed on paramedic 12-lead electrocardiogram. Of 638 patients randomized, 441 were confirmed STEMI patients who underwent primary endpoint analysis. The primary endpoint was myocardial infarct size as assessed by cardiac enzymes, troponin (cTnI) and creatine kinase (CK). Secondary endpoints included recurrent myocardial infarction, cardiac arrhythmia and myocardial infarct size assessed by cardiac magnetic resonance (CMR) imaging at 6 months.
There was a significant increase in mean peak CK in the oxygen group compared to the no oxygen group (1948 U/Lvs.1543 U/L; means ratio, 1.27; 95% CI, 1.04 to 1.52; P= 0.01). Mean peak troponin was similar in the oxygen and no oxygen groups (57.4 mcg/L vs. 48.0 mcg/L; ratio, 1.20; 95% confidence interval [CI], 0.92 to 1.56;P=0.18).There was an increase in the rate of recurrent myocardial infarction in the oxygen group compared to the no oxygen group (5.5%vs.0.9%, P=0.006) and an increase in frequency of cardiac arrhythmia (40.4%vs.31.4%; P=0.05). At 6-months the oxygen group had an increase in myocardial infarct size on CMR n=139;20.3grams vs. 13.1grams; P=0.04).
Supplemental oxygen therapy in patients with STEMI but without hypoxia increased early myocardial injury and was associated with larger myocardial infarct size assessed at six months.

Full article now out:
Air Versus Oxygen in ST-Segment-Elevation Myocardial Infarction, Circulation May 2015.

Earlier ScanCrit post on oxygen: OXYGEN. ENOUGH, ALREADY.

This entry was posted in Cardiology, Emergency Medicine, Intensive Care. Bookmark the permalink.


  1. Brian O'Byrne says:

    I am in agreement with the mentality behind
    The findings. The problem occurs when there are no tools to measure
    The oxygen levels or training not given in the use of these items. On any occasions oxygen is given to alleviate chest discomfort
    Or anxiety, would it not be better to train people first
    Then reduce the oxygen less given to a therapeutic amount
    So as not to hyper oxyGe ate stemi patients

  2. tonvanderwijst says:

    What to do with patients in shock who develop an MI caused by hypoperfusion with normal Spo2 in the prehospital EMS setting?

    • Thomas D says:

      What this study looks at is pretty stable patients with MI. Your patient sounds sick. So this study doesn’t really apply. Usually, in these sick pts sats aren’t normal either. Anyway, titrating oxygen doesn’t take priority here. In this case, it’s about doing the basics right. I’d crank up the oxygen so I could manage the patient and get him quickly to hospital.

      In shock, or any other setting, delivery of oxygen (DO) to the tissues is expressed by the DO equation, which can be simplified to these three factors:

      DO2= CO (cardiac output) x Hb x Sat%

      Those are the three factors you can manipulate. If hypoperfusion is due to hypovolemia, CO is low, but will increase with fluids. In shock due to sepsis the same will often apply. Hb can be checked and transfused, but probably not pre-hosp. And sat%/SpO2 still doesn’t go beyond 100%. So theoretically, there’s not much good to come from extra O2 if sats are already maxed out. But, in this setting, I wouldn’t worry about details. Fine tuning can be done in hospital when you got an overview of the patient and the resources to care about details. That’s my take.

  3. Rachana says:

    I have always been convinced with the fact mentioned above. , if sats are >94% then. Oxygen therapy is not adding to any useful effect on patients conditions, N have been proving otherwise. Tho I guess our cardiology fraternity should take the Cal.

  4. Pingback: Oxygène et infarctus | thoracotomie

  5. Thulani Nkomo says:


  6. is there any written protocols on this subject
    If so please email a copy to me

  7. rachel says:

    8lpm is a lot. I wonder what the result would have been with 3 or 4lpm

    • Thomas D says:

      True, but in the places I have worked, it’s not unusual to find pts with no oxygen problem, but with a non-rebreather cranked up to 10lpm – just to be on the safe side.

      The DETO2X-AMI study uses 6lpm. We’re looking forward to their conclusions.

      Still, the lower you go, the less of an effect – and side effect – you’ll get. So having a study with, say, air vs O2 on nasal cannula 2lpm, doesn’t make any sense. It’s highly unlikely you’ll find any difference. Still, the low flow nasal cannula is often annoying and tying pts to their beds – and often unnecessary. We’ve all seen them.

    • Charles Dusha says:

      Referencing practices of Respiratory Therapy, 6-8 lpm is the minimum that any regular face mask (except a venturi mask, which is a special case) should flow. The basic reason for this is that the mask itself forms a reservoir which traps exhaled CO2, and unless these flow rates are met, the CO2 trapped in the dead space of the mask is then rebreathed, negating the advantages of the oxygen enhancement. Similarly, a nasal cannula above 6 lpm does not increase the FiO2, but causes significant difficulty with drying of the mucous membranes and irritation because of the flow rate.

  8. Greg says:

    The of Etco2 capnography is much better tool to use in lieu of SPO2 in theEMS clinical environment. There are many variables that influence the measurements of Spirometery. Oxygen delivery ultimately depends on what is allowable by local protocol.

    • Rob says:

      PetCO2 has nothing to do with oxygenation, it shows only ventilation.
      SPO2 = Oxygenation (or what’s binding with hemoglobin)
      PetCO2 = Ventilation

      Certainly PetCO2 is very useful (I use it often), and should be used along with SPO2, but they provide entirely different information.

  9. I’m not supporting oxygen use but this study’s primary end point was a disease oriented end point in which it doesn’t mean a lot to me as a clinical physician. I hope it was a patient oriented outcome (MACE for instance).
    Great Review buddy.

    Abdul from Saudi Arabia

    • Thomas D says:

      Thanks for reading and commenting! I partly agree, but I must say it does mean something to me if my pts get re-MI due to my (unneccesary) treatment. Mortality isn’t the only endpoint that matters.

      When that’s said, the DETO2X study will be looking at “the hardest endpoint” – mortality.

      And we’re supporting oxygen use, as long as there’s an indication and it’s not just a reflex.

  10. Paul Ransley says:

    Thanks for this review – a long debated subject.

  11. Simon says:

    This is something that as paramedics we practice following resuscitation and JRCALC guidelines from about 5 years ago. We titrate spO2 levels to 94% to avoid vasal constriction as common practice. It does however cover the use of Entonox (nitrous oxide and oxygen 50:50), which does effect spo2, but is a very good tool in controlling anxiety, pain and relieving tension. I will always deliver to the ACQI, but if it isnt indicated, dont use it! I will strive to administer Morphine ( for cardiac preload and analgesia), entonox, GTN and aspirin. Good package of care to deliver to a cath lab for the tremendous skills they have!

  12. Auday Alkhunaizi says:

    Well O2 is harmful in many other small studies.. Big Q is: what was their sofa/apache score before randomization? Were they equivalent? Were management given to both arms similar? Plus the definition of damage based on cardiac enzyme is a bit vague and doesnt contribute outcome.. mRI was equivalent which is ultrasensitive…what was pre cath echo EF? Did they have similar MI? Any Cardiogenic shock?alot of confounding factors and agree we need harder end points to clarify although to show mortality diff is almost to impossible..

    • Thomas D says:

      I agree with all those comments, and that’s why I’m looking forward to reading the full article. And also to the DETO2X study. In the meantime, as you say, the AVOID study seems to add the the many small studies pointing to harm from excessive use of O2. So, as with anything else: Use in moderation.

  13. Pingback: LITFL Review 158

  14. dr muneer km says:

    Can anybody explain,

    • Thomas D says:

      DO2 = CO x Hb x Sat%
      DO2 (I misspelled it as DO above) is the delivery of oxygen to the tissues. It is determined by the three factors on the other side of the equation: CO (cardiac output), Hb and Sat% (or SpO2).

      The ‘real’ equation is DO2 = [1.39 x Hb x SaO2 + (0.003 x PaO2)] x Q

      Where Q is cardiac output (which I name CO). And SaO2, which I often call Sat%). So, This above equation is the correct one, but it can be simplified to give a better and easier understanding of what determines DO2, in my opinion.

      The part with (0.003 x PaO2) is how much the freely dissolved oxygen in blood adds to the delivery of oxygen to the tissues. As seen from the 0.003 factor, it is negligable – so I just strike it from the equation. The other number I take out, is the conversion factor: 1.39. It is only of interest when trying to compute the actual DO2. For understanding how CO, Hb and Sat contribute to the delivery of oxygen, the 1.39 factor is irrelevant.

      Now, the DO2, delivery of oxygen, is determined by how many oxygen molecules attached to hemoglobin that gets pumped around every minute. That’s what this equation says. DO2 = CO x Hb x Sat%

      The oxygen train
      I was once told to look at DO2 as a train. This is simple, but made me understand DO2. Oxygen steps on the oxygen train at the main staition in the lungs, and is delivered/get off at the various stations in the tissues.

      So the oxygen train looks like this: Hb is the number of carriages on the train, Sat% is how full each carriage is. Cardiac Output is the locomotive/engine, and determines how quickly the train moves through the body and returns to the lungs to get more oxygen.

      So, to deliver more oxygen, you can either have more carriages (hb), fill up your carriages better (sat%), or make the whole train go faster (CO).

  15. Great review!

    Less O2, less fluids, etc…

    Slowly we seem to be learning how to to follow something we learned about a long time ago: Primum Non Nocere…


  16. Vidar Magnusson (@HEMSonICE) says:

    I agree that the results are consistent with the recent studies on oxygen but find the trend toward increased mortality worrying, even though it doesn’t reach statistical significance. Will be good to see larger study.

  17. Pingback: מתן שגרתי של חמצן לחולי STEMI מזיק - המלר''ד

  18. Pingback: Oxygen in MI | Clinical Focus

  19. Pingback: July 2015 REBELCast - R.E.B.E.L. EM - Emergency Medicine Blog

  20. Pingback: THE OXYGEN-ICU TRIAL |

Leave a Reply

Your email address will not be published. Required fields are marked *