Several Studies Point Out Harm in Hyperoxia



David Anderson, MStJ, BSc, MB, ChB | From the July 2011 Issue | Friday, July 1, 2011

Learning Objectives
>> Review current research on supplemental oxygen use in the prehospital field.
>> Evaluate the use of oxygen for myocardial infarction, COPD, stroke and neonatal patients.

Key Terms
Cochrane review: A group of people who prepare reports so that healthcare professionals can make informed decisions about the care they provide.
Cofactor: A substance that acts with another substance to bring about certain effects.
Hyperoxia: A bodily condition characterized by a greater oxygen content of the tissues and organs than normally exists at sea level.
Meta-analysis: A statistical technique for combining the results of several smaller studies into one large study.
Metalloprotein: A conjugated protein in which the prosthetic group is a metal.
Mitochondria: A component of a cell that generates energy for the cell to use.
Odds ratio: A statistical measure of the probability or chance of an event occurring.
Partial pressure: The pressure generated by a gas in a solution; used as a measure of how much gas is present.
Relative risk: A statistical measure of the risk of an event occurring in one group compared with the risk in another group.

Oxygen is a ubiquitous treatment in prehospital care. Most of us remember the excitement of being trained to use it, because it’s typically the first drug that EMS providers are taught to administer. As a core EMS skill, all providers should be highly proficient with oxygen delivery equipment. Less well taught, however, is the fact that oxygen is a drug, which, like all others, has indications, contraindications, adverse effects, a dose and an intended duration of administration.

I remember being taught as a young volunteer EMT to administer oxygen for everything from nausea and vomiting to lower limb fractures. (I was even told at one point that it had analgesic properties.)

This article highlights some current research on supplemental oxygen use in prehospital care and discusses controversies that are arising from this research.

Oxygen is used in every cell of the body—except, ironically, red blood cells—to convert chemical potential energy from food into potential energy in the form of adenosine triphosphate (ATP). The mitochondria are the powerhouses of the cells, containing enzymes that gradually break down fats, sugars and proteins into single carbon compounds that release a small amount of energy each time.

At the end of this chain of carbon breakdown, an enzyme called cytochrome oxidase uses oxygen as a cofactor to phosphorylate (i.e., add a phosphorous atom) to adenosine diphosphate (ADP) to produce ATP. ATP is an energy store, allowing the body to use energy when it needs it.

Getting the oxygen to the cells is a bit trickier now than it was when single-celled organisms started using oxygen billions of years ago. A complex circulatory system has evolved to get the oxygen from the alveoli to the cells, and a complex respiratory system has evolved to get oxygen into the blood.

The problem is that oxygen isn’t particularly soluble in blood. So the solution is the hemoglobin contained within the red blood cell. The red blood cells aren’t really cells, because they don’t contain mitochondria, nuclei, Golgi apparatus or all the other things that make cells cells. They’re better thought of as little bags of hemoglobin.

Hemoglobin is a metalloprotein made up of four subunits, each containing a protein (globin) and an iron-containing compound (heme). It’s the iron that’s the key. As we all know, iron and oxygen happily combine to form iron oxide—rust. Rust is red. Oxygenated blood is red. Each time you take a breath in, you cause some of your blood to rust. The beauty of hemoglobin is that it can also “un-rust,” allowing it to offload oxygen to body tissues and be ready to pick more up in the lungs.

The four molecules of hemoglobin work together in a system known as cooperative or allosteric binding. This means that when there isn’t much oxygen around (e.g., in the tissues) hemoglobin doesn’t bind to oxygen very well, so the oxygen dissociates and goes into the tissues. However, when there’s lots of oxygen around (e.g., in the lungs), the hemoglobin avidly binds to oxygen, allowing greater uptake into the blood.

The affinity of hemoglobin for oxygen can also be altered by the conditions of the tissue the blood is flowing through. So in hot, acidotic tissues with high carbon­ dioxide (CO2) levels (e.g., exercising muscle), the oxygen affinity of hemoglobin is poor, allowing more oxygen to be offloaded to the tissues. However, in cold, alkalotic blood with low CO2 levels, the affinity for oxygen is much higher, allowing increased loading in the lungs. This occurs as part of the response to high altitude.

Oxygen directly affects the tissues it travels through by interacting with blood vessels. In most, oxygen acts as a vasoconstrictor, like norepinephrine. It’s thought that this response is part of the cardiovascular system’s function known as autoregulation, whereby organs alter their own blood supply in response to such factors as oxygen and CO2 levels, acidity, potassium and lactate levels in the blood. The exception to this rule is in the lungs, in which a phenomenon known as hypoxic pulmonary vasoconstriction occurs, allowing lung units that are poorly ventilated to autoregulate their blood supply and decrease ventilation perfusion mismatch.

Myocardial Ischemia
Oxygen has been referenced as an integral component of the treatment of myocardial ischemia since 1900. In EMS, it’s considered a primary treatment; indeed the only treatment that can be administered by EMTs or emergency medical responders in many jurisdictions. Hearing that giving supplemental oxygen to patients with cardiac chest pain may not be beneficial—and may even be harmful—may shock many EMS professionals. But the theory isn’t that new.

In 1976, a randomized controlled trial was published in the British Medical Journal in which 157 patients with uncomplicated myocardial infarction (MI) were randomized to receive either supplemental oxygen or air for the first 24 hours following onset of symptoms.1 There was no significant difference in the primary outcome of death or the secondary outcomes of ventricular dysrhythmia or pain requiring analgesia.

There was, however, a non-significant trend toward increased mortality in the group that received oxygen. A caveat of this study is that it took place in the era before revascularization therapy, so care must be taken in comparing it with modern studies of MI management.

Despite this small study, healthcare providers continued to give oxygen for many years. Further, small, randomized studies in the later part of this past century also indicated there may be limited benefit from oxygen use. But it wasn’t until the past year that a systematic review and then a Cochrane review gave a strong indication that oxygen may be harmful in uncomplicated MI.2

The Cochrane review, which analyzed the 1976 paper and two other studies from 1997 and 2004, concluded that the limited evidence available showed no benefit and potential harm from oxygen use. The review indicated that a large, randomized controlled trial should be conducted.

On the basis of the Cochrane review, the American Heart Association (AHA) recommended in the 2010 CPR Guidelines that supplemental oxygen no longer be administered to patients with uncomplicated cardiac chest pain who have an oxygen saturation greater than 94%.3

Cardiac Arrest
A very important paper published in the past year gave a strong indication that hyperoxia may actually be more harmful than hypoxia in adults during and immediately following out-of-hospital cardiac arrest.

A paper in the Journal of the American Medical Association studied the partial pressure of oxygen (PaO2) in arterial blood of patients brought to the emergency department (ED) following successful resuscitation by EMS.4 The patients were stratified based on an arterial blood gas taken within 24 hours of hospital admission into those with hypoxia (PaO2 less than 60 mmHg), those with normoxia (PaO2 61–299 mmHg) and those with hyperoxia (PaO2 greater than 300 mmHg).

The study was multicenter and quite large, enrolling 6,326 patients over five years. The primary outcome was survival to hospital discharge. It found that hyperoxia was a significant independent risk factor for in-hospital mortality. The odds ratio for death was 1.8 (95% CI 1.5–2.2) in the hyperoxia group compared with the normoxia group. This was even higher than the odds ratio for death in the hypoxia group.

The authors of the study correctly note that correlation doesn’t equal causation. However, there’s a plausible mechanism by which hyperoxia could cause increased mortality in patients following cardiac arrest, and the chance of causation is great enough to consider limiting the amount of oxygen given to patients following cardiac arrest to a sufficient amount to keep arterial oxygen saturation 94–96%. The one time 100% oxygen is always indicated (via non-rebreather mask or BVM) for a patient is prior to an intubation, to effectively “wash out” the nitrogen and prevent rapid onset of hypoxia during an intubation attempt.

Acute exacerbations of chronic obstructive pulmonary disease (COPD) are a common reason to call an ambulance, and supplemental oxygen (often high-flow) is often administered by EMS professionals to these patients. However, a recent randomized controlled trial conducted in the prehospital environment has produced dramatic results indicating that providing high-flow oxygen to patients with an exacerbation of COPD—and possibly even patients with undifferentiated shortness of breath who don’t have a firm diagnosis of COPD—can significantly increase mortality. 5

The study was carried out in Hobart, Australia. Instead of randomizing patients, paramedics were randomized to either provide high-flow oxygen (8–10 LPM via a non-rebreather mask) or titrated oxygen (variable flow via nasal cannula titrated to keep arterial oxygen saturation at 88–92%).5

When nebulized drugs were required, they were delivered by oxygen in the control arm and by compressed air in the intervention arm. The study was extended over a one-year period and randomized 226 patients into the control arm and 179 patients into the intervention arm.

A large number of patients were excluded from both groups after being retrospectively identified as not having a diagnosis of COPD, leaving 117 patients in the control arm and 97 patients in the intervention arm.

Analysis was by intention to treat, and a large number of patients in the intervention arm received high-flow oxygen but were still included in the analysis as though they had received titrated oxygen, making the result of the study all the more remarkable.

The study showed a relative risk of death in the intervention arm of 0.42 (0.02–0.89, p= 0.02) in all patients and a relative risk of death in the intervention arm of 0.22 (0.05–0.91, p=0.04) in those with confirmed COPD. So even in patients without COPD, high-flow oxygen was associated with a doubling of mortality. When the results were analyzed as treatment received rather than intention to treat, the same trends in mortality were identified; however, due to the smaller numbers, the results were no longer statistically significant.

Neonatal Resuscitation
It’s been known for some time that prolonged administration of high oxygen concentrations to premature neonates suffering from apneas is associated with retinopathy of prematurity, which causes blindness. A more recent finding, however, is that resuscitation of neonates with air rather than 100% oxygen results in a decrease in mortality. A 2004 systematic review and meta-analysis pooled the results of five studies involving a total of 1,302 neonates.6

These were all randomized trials of oxygen versus air in the resuscitation of neonates. Three were blinded, and two were unblinded. Although none of the studies showed a significant decrease in mortality, the pooled results showed a significant decrease in mortality with air compared with oxygen. The relative risk of death in the air group was 0.71 (0.54–0.94), with a number needed to harm of 20. (Twenty babies would have to be treated with 100% oxygen to cause one death). The deaths occurred during follow up over a 24-month period. Interestingly, there was no significant increase in such neurological complications as cerebral palsy.

This review has caused a significant change in international protocols for neonatal resuscitation, with the latest edition of the AHA resuscitation Guidelines recommending initial resuscitation of neonates with air, switching only to oxygen when there’s clear evidence of significant hypoxaemia following resuscitation with air.7

Modern emergency management of acute stroke has evolved along similar lines to that of myocardial ischemia, with the potential for reperfusion therapy leading to a significant decrease in morbidity and mortality. Supplemental oxygen has long been a standard component of emergency care for patients with acute stroke. However, a study conducted more than a decade ago in Scandinavia cast some doubt on this treatment.8

The study randomized 310 patients with acute ischemic stroke (hemorrhagic stroke and subarachnoid hemorrhage were excluded) to receive either oxygen (3 LPM via nasal cannula) or air for the first 24 hours following hospital admission. Outcomes were survival, measured at one year and stroke severity (using the Scandinavian Stroke Scale) and disability (using the Barthel index)—both measured at 7 months.

The study showed no significant difference in survival (69% in the oxygen group, 73% in the air group, p= 0.3). The authors did, however, comment on a non-significant trend toward decreased survival in the oxygen group.

Analysis of stroke severity and disability index for patients also showed no significant difference, but when a subgroup analysis of patients with severe stroke at presentation was carried out, there was a statistically significant decrease in survival in the oxygen group (82% versus 91%, p=0.023; OR 0.045, 95% CI 0.23–0.9.)

Although these results should not be viewed as conclusive, they suggest oxygen perhaps has no benefit and may be harmful to some patients suffering acute ischemic stroke. On the basis of this study, some EMS services have removed oxygen from the protocols for managing adult patients with uncomplicated acute stroke.

Harm from Supplemental O2
Many theories exist to explain why supplemental oxygen may be harmful. It’s well accepted that prolonged, supranormal blood and tissue oxygen tension is detrimental. Breathing 100% oxygen for longer than three days causes certain death in many species (thankfully, not humans). Breathing 100% oxygen causes thickening of alveolar membranes and restriction of lung expansion. In hyperbaric situations, it can cause seizures.

Hyperoxia also causes free-radical damage. Free radicals are oxygen atoms with a charge due to an unequal number of protons and electrons. These radicals are known to cause intracellular damage and cell destruction (much like pouring hydrogen peroxide into an open would). It’s known that ischemic tissues are particularly sensitive to free-radical damage. This may be a major factor in the pathogenesis of hyperoxia.

Another possibility is that increased oxygen tension causes widespread vasoconstriction. This is known to occur in cerebral arteries, but it may also occur in other vessels, including coronary arteries. This may explain why administration of oxygen following MI may not be beneficial.

Hypoxia—Not So Bad?
In May 2007, a group of intensivists climbed Mt. Everest.9 When they got to the top, they removed their oxygen masks, allowed their bodies time to equilibrate and took arterial blood gas (ABG) samples from each other. The samples were rushed back down to base camp by Sherpa guides and analyzed. The results were startling.

The intent of the study was to see how low the PaO2 could get in the blood of fit and healthy volunteers and then to use this data to guide the management of critically ill patients in intensive care with respiratory failure. The results showed that the average PaO2 was 24.6 mmHg, and the lowest was 19.1 mmHg. The normal PaO2 range for an adult at sea level is 80–100 mmHg. Many clinicians would view the levels achieved in this study as being incompatible with life. The average oxygen saturation (SaO2) was 54.0%, and the lowest SaO2 was 34.4%. These numbers would horrify a paramedic. However, none of these subjects were gravely ill. Some felt giddy and slurred their words slightly. Importantly, no long-term harm came to them. The authors of the paper hope to use their data to further study the limits of hypoxaemia in the critically ill.

Benefits of O2
Having reviewed a few conditions where hyperoxia may be harmful, we must bear in mind that profound and prolonged hypoxia is universally fatal. The timely administration of oxygen can be life saving in many cases. Supplemental oxygen should still be administered to patients who have an oxygen saturation of less than 94% for any reason. (In many jurisdictions, however, it will be acceptable to target an oxygen saturation of greater than 88% in patients with COPD).

Oxygen should also be administered to any patients with actual or potential airway compromise (e.g., epiglottis or airway burns) and drug assists and/or rapid sequence intubations.

The rationale for this treatment isn’t to increase tissue oxygenation but to increase oxygen reserves in the lungs so that if the airway is lost for any reason, the patient will take longer to desaturate before a critical level of hypoxia is reached, which allows providers time to rescue the airway with a supraglottic or surgical airway.

Providers should continue to administer oxygen to patients who have evidence of globally poor tissue perfusion (e.g., hypovolemic or septic shock patients). The rationale for this treatment is to ensure the hemoglobin molecules are fully saturated with oxygen and that there’s as much oxygen as possible dissolved in the blood, increasing the driving pressure to aid internal respiration.

Oxygen is a drug and should be administered to appropriate patients in an appropriate dose. A recent editorial coined the term “Goldilocks effect” to highlight that we should strive to deliver an amount of oxygen that’s not too little, not too much but just right.10 In the 2010 Guidelines, the AHA target an oxygen saturation should be 94%, following successful resuscitation.11

Although we have traditionally erred on the side of caution in prehospital care by providing liberal amounts of supplemental oxygen, we may actually have been erring on the side of harm. EMS providers are skilled professionals who are used to learning the ins and outs of a new drug before administering it appropriately.

Perhaps it’s time we treated oxygen like a new, exciting drug and begin to administer it with the knowledge that, like all other drugs, it has indications, contraindications and, in particular, adverse effects. Until further studies clarify practice and provide a stronger evidence base for guidelines, EMS professionals should at least think before they pick up the mask.

1. Rawles J, Kenmure A. Controlled trial of oxygen in uncomplicated myocardial infarction. Br Med J. 1976;1(6018):1121–1123.
2. Cabello J, Burls A, Emparanza J, et al. Oxygen therapy for acute myocardial infarction. Cochrane Database of Syst Rev. 2010;(6):CD007160.
3. O’Connor R, Brady W, Brooks S, et al. Part 10: acute coronary syndromes: 2010 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation. 2010;122(18 Suppl 3):S787–817.
4. Kilgannon J, Jones A, Shapiro N, et al. Association between arterial hyperoxia following resuscitation from cardiac arrest and in-hospital mortality. JAMA. 2010;303(21):2165–2171.
5. Austin M, Wills K, Blizzard L, et al. Effect of high flow oxygen on mortality in chronic obstructive pulmonary disease patients in prehospital setting: randomised controlled trial. Br Med J. 2010;341:c5462.
6. Davis P, Tan A, O’Donnell C, et al. Resuscitation of newborn infants with 100% oxygen or air: a systematic review and meta-analysis. Lancet. 2004;364(9442):1329–1333.
7. Kattwinkel J, Perlman J, Aziz K, et al. Part 15: Neonatal Resuscitation: 2010 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation. 2010;122(18 Suppl 3):S909–919.
8. Ronning O, Guldvog B. Should stroke victims routinely receive supplemental oxygen: A quasi-randomized controlled trial. Stroke. 1999;30(10):2033–2037.
9. Grocott M, Martin D, Levett D, et al. Arterial blood gases and oxygen content in climbers on Mount Everest. N Engl J Med. 2009;360:140–149.
10. Hommers C. Oxygen therapy post-cardiac arrest: The ‘Goldilocks’ principle? Resuscitation. 2010 (12):1605–1606.
11. Peberdy M, Callaway C, Neumar R, et al. Part 9: post-cardiac arrest care: 2010 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation. 2010;122(18 Suppl 3):S768–786.

This article originally appeared in July 2011 JEMS as “Behind the Mask: Is oxygen harming your patient?

Connect: Have a thought or feedback about this? Add your comment now
Related Topics: Patient Care, Airway and Respiratory, supplemental oxygen, oxygen, normoxia, hypoxia, hyperoxia, Jems Features


David Anderson, MStJ, BSc, MB, ChB

David Anderson, MStJ, BSc, MB, ChB, is a senior resident in critical care medicine from Sydney, Australia. Before qualifying as a physician, he was a paramedic in Auckland, New Zealand, and has significant experience in both frontline prehospital care and clinical education.


What's Your Take? Comment Now ...

Featured Careers & Jobs in EMS



Get JEMS in Your Inbox


Fire EMS Blogs

Blogger Browser

Today's Featured Posts


EMS Airway Clinic

Innovation & Progress

Follow in the footsteps of these inspirational leaders of EMS.
More >

Multimedia Thumb

Colorado Bill Makes EMS Violence a Felony

Mandatory jail time for people convicted of assaulting an EMT or paramedic.
Watch It >

Multimedia Thumb

Michigan Ambulance Crash Causes Building Explosion

MMR ring crashed into a gas line at their company quarters.
Watch It >

Multimedia Thumb

Two Burned in Phoenix Tanker Fire

40,000-gallon rail car tanker explodes in industrial area.
Watch It >

Multimedia Thumb

Maryland Ambulance Company Temporarily Suspended

Company failed to meet Frederick County code standards.
Watch It >

Multimedia Thumb

South Korea Ferry Disaster

Death toll expected to rise dramatically.
More >

Multimedia Thumb

Boston Bombing Heroes

A look at four who became well known in the tragic bombing.
More >

Multimedia Thumb

Fatal, Fiery California Bus Crash

Seven students and two drivers killed in crash.
More >

Multimedia Thumb

Braun Ambulances' EZ Door Forward

Helps to create a safer ambulance module.
Watch It >

Multimedia Thumb

The AmbuBus®, Bus Stretcher Conversion Kit - EMS Today 2013

AmbuBus®, Bus Stretcher all-hazards preparedness & response tool
Watch It >

Multimedia Thumb

LMA MAD Nasal™

Needle-free intranasal drug delivery.
Watch It >

More Product Videos >