Saturday, November 29, 2014

Oxygen: Savior or Devil in a Green Dress?

Prepared by Michael. C Berrier, NRP, FP-C, C-NPT, AAS
Reviewed by: Yen Chow, MD 

If you ask any small child what doctors do or what happens in a hospital, you’ll probably get some variation of “they make sick people better.” Were you to ask the same question of one of those doctors or someone working inside the hospital, you’d likely get a more complicated, wordy and educated version of the same answer. We all tend to believe that our goal is to improve our patient’s health and, truly, make them better. However, that isn’t true at all. The human body is an impressive machine which needs little improvement from the likes of us in order to function. Our goal as end users is to maintain our own body with diet and exercise, but its original design is pretty fantastic. 

This version of a human being works best by following several basic physiological rules:

*We eat our food (enteral nutrition)
*We circulate our blood through normal vessels (Heart beating, stroke volume,     drug free)
*We breathe our air (negative pressure ventilation, 21% oxygen)
*We do all of this without tubes, lines and wires. (No ET tubes, no trach, no         foley, no NG)

The human body is, however, fragile and prone to breakdown, usually through neglect and misuse (see diet and exercise again). This is where medicine comes in, not to make people better but to help return them to the best working version of themselves.

Modern medicine is often about adding and taking away to achieve this goal. We add drugs, we take away excess fluid. We add positive pressure, we take away afterload. This is the kind of physiological arithmetic that a well body will do for itself, but the unwell machine cannot or will not. And while the interventions by which we do this medical math are lifesaving in the short term, in the long run they are at worst harmful and at best contrary to our stated goal of returning to the model of humankind. If we stop eating, out gut will stop working. Too many drugs (recreational or otherwise) will cause permanent damage and remodeling to our heart and vessels. Positive pressure ventilation for as little as 18 hour will cause diaphragmatic atrophy, and tubes left in too long can lead to things like infection, stenosis and tracheomalacia.

But, probably the most commonly used medical interventions by providers from first aid to anesthesia is supplemental oxygen. One of the very first things I learned in basic EMT school was how to operate an oxygen system to provide oxygen to…everybody. No patient was spared the healing wonder of compressed gases. After all, if they’re sick enough to need an ambulance, they must be sick enough to need the green skinned breath of life carried within. Even as I progressed through EMT-Intermediate and paramedic schools, the mantra for nearly every patient began with “IV, O2, Monitor”. 

And it is true, for a very specific patient population, supplemental oxygen can be the difference between life and death. Unfortunately, that population is a good bit smaller than we have been led to believe. Even the patients who will benefit from oxygen concentrations above atmospheric would benefit from another treatment even more, such as CPAP or inhaled bronchodilators or suction. The truth is that, unless your patient hails from a planet where the atmosphere is made up of more than 21% oxygen, applying more than that does not fix a problem, it only treats a symptom. Based on the evidence, in the absence of hypoxia, supplemental oxygen seems to do no good and may in fact do harm to patients with a variety of acute illnesses and injuries.

Science & History: The presence of oxygen in the atmosphere has been studied and well understood for centuries, and the ancient Greeks had an understanding that air was vital to life. However, the role of oxygen in body metabolism has been well documented and understood for only a little over a century. The first account of continuous oxygen therapy for a specific illness was published in 1890, describing a woman with pneumonia who received about “200 gallons in twenty four hours” and improved.   However, Even the early pioneers of oxygen therapy strongly cautioned against large doses of the panacea “on account of the dangerous degree of stimulation to the system and the increased combustion of tissue” (1) In fact, almost every scientist and physician responsible for a significant advancement in the medical applications of supplemental oxygen added statements warning against overzealous and excessive oxygen usage. Yet somehow a “more is better” approach in acute care medicine and especially EMS has survived to the present day. Oxygen is often applied routinely as part of a protocol for conditions like chest pain, dyspnea and severe illness without an accompanying assessment to determine if it is actually indicated.  

So, why is supraatmospheric oxygen bad? It has been known for quite some time that newborns, especially those born prematurely, were are risk for damage to their eyes and other organs from increased oxygen exposure. Further research into children born with certain congenital heart defects (CHD) found that they were dependent on a patent ductus arteriosus (PDA) to circulate oxygenated blood to the rest of their body. Since increased oxygen levels in the blood (PO2) lead to closure of the PDA, these patients are often given limited oxygen at atmospheric or sometimes even subatmospheric concentrations to maintain PDA patency until they can get a surgical repair (2).

For adults, the dangers of supplemental oxygen are a little more nuanced with effects that are usually observed in the long term rather than acutely as noted in neonates. As oxygen surpasses adequate levels in the blood, arterial receptors send signals to reduce the vessel size, demonstrating oxygen as a powerful and fast vasoconstrictor(3). This mechanism is useful in the minutes after birth as a neonate converts from fetal to newborn circulation, but for adults this can cause problems that will be outlined in the Evidence section.

Incompletely metabolized or excess oxygen is also prone to increase production of oxygen free radicals, or reactive oxygen species (ROS). These molecules are necessary for certain function inside of cells, but an abundance of ROS can lead to cellular damage which is thought to exacerbate ischemia during heart attack and stroke.(4)

The Evidence: It is important to separate recent evidence supporting hyperoxygenation in preparation for rapid sequence induction and intubation (RSII) from research that addresses ongoing supplemental oxygen therapy. NO DESAT and other methods for oxygen administration prior to RSII are specifically designed for a patient with an imminent predicted period of extended apnea as a means to create a reservoir of oxygen in the lungs and in the plasma (5). Since active respiration will cease as a matter of course, it is imperative that we provide this extra inventory of oxygen in compensation for the patient’s failed ability to provide it for themselves. However, many patients who are candidates for RSII do not have respiratory pathology as the reason for their intubation and thus will likely require little or no supplemental oxygen after airway management is complete. Rather, they will only need positive pressure ventilation, and most can tolerate that at room air concentrations or slightly higher.

For patients who are not intubated and/or will not be intubated, using supplemental oxygen is very much like using any other drug. In fact, I’d say it’s exactly like using any other drug, and the evidence bears this out. The most recent development on this front is the AVOID trial, which studied the outcome of patients with myocardial infarction who received high flow oxygen versus patients who did not. An important feature of the trial was that they first separated the patients into groups that were hypoxic and those who were not, then randomized the latter group to either get oxygen at 8 lpm by mask or none at all. Long story short, the group that got indiscriminate oxygen had higher cardiac enzymes, bigger infarcts and more recurrent infarcts during their hospital stay (6). Similar studies in recent years have shown similar results, with one demonstrating a threefold increase in mortality in a group that received high flow oxygen versus compressed air(7). The main reason for this outcome is thought to be increased coronary microvascular constriction, and other studies have demonstrated this under observation in the cath lab.

The current version of ACLS makes reference to this in print an during the acute coronary syndromes video when it shows “Hugh” suffering his big, fat STEMI without receiving a single drop of oxygen.

Acute coronary syndrome and STEMI are not our only areas of concern when it comes to oxygen; however, though they seem to get the most attention. This is likely due to the anoxic nature of tissue death related to coronary artery occlusion during myocardial infarction. A greater population in transport for whom oxygen administration is a factor is the mechanically ventilated. People are intubated and put on a ventilator for many reasons that have nothing to do with their ability or inability to use and metabolize oxygen. As noted above, these patients would likely die without the mechanics of positive pressure ventilation but many do not need supplemental oxygen.

Ventilated patients with acute serious illness such as sepsis and ARDS are routinely hypoxic and as such often receive substantial amounts of supplemental oxygen prior to and after the initiation of mechanical ventilation. While this seems intuitive, mounting evidence suggests that this strategy may cause more harm than good in the long term. Conversely, these patients are much better served by increasing their mean airway pressure (MAP) with the application of PEEP to improve oxygenation.  It is not uncommon in the ICU and transport setting to see PEEP set as high as 15-20 cmH20 while trying to maintain as low a FiO2 as possible. Two studies found that increased FiO2 worsened oxygenation index over 48 hours and lead to increased days on ventilator and ICU days (8). There has been theoretical belief that oxygen induced vasoconstriction might be helpful to patients in either septic or hemorrhagic shock, but no study has found this to be true while other evidence suggest that deleterious lung effects would far outweigh any perceived vascular benefit. 

For other patients who are not intubated, specifically those with stroke or COPD, evidence is compelling that supranormal oxygen concentrations lead to poor outcomes. It has long been taught (this was pounded into my head during EMT school) that COPD patients would surely die if given anything more than two liters per minute of oxygen by nasal cannula thanks to the ominous “hypoxic drive” While this is indeed a thing, the number of people who actually live and die by hypoxic drive is very small. So, rest easy tonight knowing that you probably haven't killed anyone with a nasal cannula. However, multiple studies demonstrate poor outcomes for patients who receive indiscriminate high flow oxygen versus titrated oxygen based on clinical presentation. For patients with stroke, the microvascular constriction attributed to oxygen leads to decreased perfusion to the ischemic penumbra and thus has been blamed for worse outcomes for non-hypoxic patients who received high flow oxygen versus those who did not (3).


Oxygen. It’s easy, it’s intuitive, and it often makes people feel better, including the caregivers. The most likely reason it’s been left alone in most EMS, acute care and CCT protocols is the attitude that if it doesn’t do any good, at least it probably doesn’t do any harm. Now we know, with the available research, that this simply is not true. It can hurt, it probably does hurt and the frivolous application of oxygen “because the book says so” needs to stop.


Oxygen is a drug, and like any other drug it needs to be used judiciously. We wouldn’t dream of blindly starting a patient on a vasopressor like levophed because his blood pressure may fall without doing an assessment to first find out if he is actually hypotensive. Oxygen administration should be initiated the same way after the same assessment. Just as vasopressor therapy is for the hypotensive, oxygen therapy is for the hypoxic, not just the sick. 

We tend to take a “cover all bases” approach to caring and preparing a patient for transport to make sure we don’t miss anything. For example, patients who suffer traumatic injury, especially head injuries, are particularly sensitive to hypoxia and hypoperfusion. A single episode of either is associated with significantly increased mortality and morbidity. So, we often turn up the oxygen on the ventilator to unequivocally remove one of those hazards right from the beginning. However, both of these problems can be managed at the cellular level in the same way: ensure adequate mean arterial pressure to perfuse the injured brain with appropriately oxygenated blood. Theoretically, excessive oxygen will provide little benefit and may do harm by causing vasoconstriction upstream from the injury. And, if an attempt to hyperoxygenate a head injured patient is erroneously attempted with hyperventilation, the lack of CO2 will almost assuredly result in poor blood supply to the injured area.

It is also helpful, and important, to remember that there are four different types of hypoxia, only one of which has anything to do with a lack of available oxygen (aptly named hypoxic hypoxia"). Hypoxic hypoxia is usually suffered by someone in a place devoid of oxygen, like space. If your patient isn't in space, or in a room full of smoke or carbon monoxide, high flows of oxygen are temporizing at best. The other types of hypoxia: stagnant, hypemic and histotoxic are products of the body's inability to either transport, absorb or metabolize oxygen. To truly heal these conditions, you must isolate and eliminate the cause. Providing supplemental oxygen will only poorly manage one symptom without going any distance to fixing the problem. 

We should also think about how our care and reporting affects the initial care of those who take patients from us at the end of the transport. Reporting oxygen demand to receiving staff is as important as vasopressors and sedation; it lets them know where to begin. If you spent the entire transport at 100% FiO2 and the oxygen saturation was 100%, no one has any real idea what the oxygen demand is. And since we know that many of these patients do not and will not require that much oxygen, and that excessive oxygen can do harm, there is every reason to adopt a stepwise process for titrating oxygen to a goal of < 50% to keep SpO2 between 90%-94%, depending on the pathology(8). 

The bottom line is, no study shows the benefit of hyperoxia while plenty at least suggest that it causes harm. This should lead us to think of oxygen as a drug available to those who need it (blue people) rather than a shotgun blast of “appropriate care” or part of some silly mnemonic to help us remember the most basic tenets of medicine. In doing so, we tend to ignore the most important one. Do no harm.

Now, turn off the tap.

Thanks for reading. As always, feel free to comment here or on twitter, @Crit_Care_Excel.

1.       Grainge C, Breath of life: the evolution of oxygen therapy. J R Soc Med. 2004; 97: 489-493
2.       Reviews CT. Studyguide for Egan's Fundamentals of Respiratory Care by Robert M. Kacmarek, ISBN 9780323082037. Academic Internet Publishers; 2013.
3.       Cornet A, Kooter A, Peters M. The Potential harm of oxygen therapy in medical emergencies.  Critical Care 2013, 17:313. Available at
4.       Chabot F, Mitchell JA, Gutteridge JM, Evans TW. Reactive oxygen species in acute lung injury. Eur Respir J. 1998;11(3):745-57.
5.       Weingart S, Levitan R. Preoxygenation and prevention of desaturation during emergency airway management. Annals of Emergency Medicine. 2012; 59: 165-75
6.       Avoid the oxygen. Available at: Accessed November 27, 2014.
7.       Wijesinghe M, Perrin K, Ranchord A, Simmonds M, Weatherall M, Beasley R. Routine use of oxygen in the treatment of myocardial infarction: systematic review. Heart. 2009;95(3):198-202.
8.       Rachmale S, Li G, Wilson G, Malinchoc M, Gajic O. Practice of excessive F(IO(2)) and effect on pulmonary outcomes in mechanically ventilated patients with acute lung injury. Respir Care. 2012;57(11):1887-93.

No comments:

Post a Comment